Machsupport Forum
General CNC Chat => Show"N"Tell ( Your Machines) => Topic started by: usfwalden on August 17, 2007, 11:42:28 PM
-
I just got an Ex-Cell-O Spindle Wizard mill to retrofit. It's a large machine with 40" x 20" travels on the table and 5" on the quill. The table slides on tubular ways/linear guides. It has 5 tpi precision ground ballscrews all around.
The original control was a bandit hooked up to siemens drives and motors with tachometers and resolvers. I'm looking for new motors with encoders but the motors are Nema 42 and the places I know of like homeshopcnc, etc don't seem to have anything larger than Nema 34. I've found some Nema 42 steppers but I much prefer the idea of getting servo motors. Does anybody know of a source for Nema 42 servos fitting a home cnc retrofit budget?
I'm also considering picking up ac servo motors and drives off of ebay. It seems like there are more comercial systems that use ac servos for motors of this size.
The overal plan is: mach on dell's latest pc model on sale -> grex -> new drives -> new motors
I figure with the grex I'll be able to hook up all of the limit and home switches as well as the air controled devices like the mist coolant, spindle brake, spindle lock, back gear, and speed control.
-
If the servos and drives are functional you might want to consider keeping them and getting a Step,Direction to analogue converter such as the pixie board. Would certainly be a lot cheaper and would save a lot of hassle.
Hood
-
Thanks. That sounds like an idea. I haven't even put power to the machine since it's three phase. I haven't decided what I want to do about that yet.
If the servos and drives are functional you might want to consider keeping them and getting a Step,Direction to analogue converter such as the pixie board. Would certainly be a lot cheaper and would save a lot of hassle.
Hood
-
Also the bellows on my table ballscrews seem to be flaking on the inside. Anybody know a good place to order rod bellows?
-
is there a defacto standard way to remove the spindle motor on these things. I want to pull this motor off so I can measure the shaft and bolt circle and find a single phase motor to replace it with. I'm new to this and would rather not dissassmble more of the head than I have to out of ignorance.
-
I'm guessing I remove the front plate of the head that says Ex-Cell-O? It has motors that drive chains to operate whatever is in there to mechanicall shift gears and control spindle speed. Unfortunately I don't know what's on the inside of that face plate until I take it off.
(http://www.ecutune.com/mill/mill1.jpg)
(http://www.ecutune.com/mill/mill2.jpg)
-
I've done some cleaning up and probing with the idea of keeping the motors and drives that are on it. It turns out that a lot of the documentation I got for it is incorrect. I don't know what the motors are but the drives are neither the siemens drives or the standard drives described in the bandit documentation I have. I'll post pictures. Maybe somebody is famliar with the bandit contro/drives and can share some helpful information.
I checked the power supply and it is functioning. It is a 60V DC power supply. I'd like to check the drives and if they are functional retain them and use pixie boards with them. My hurdle to this is I don't have any documentation for the drives to tell me the input pins and they way they are wired doesn't give any hints.
The following drawing of the motors seems to be correct.
(http://www.ecutune.com/mill/mill7.jpg)
In the back of the Summit Dana Bandit box are 3 drive cards which connect to the resolver phase logic card/motor control card with a ribon cable.
(http://www.ecutune.com/mill/mill3.jpg)
Up in the card rack is the resolver phase logic card. The resolver cables plug into that card and then it outputs its signals to the drives through a 16 pin ribon cable.
(http://www.ecutune.com/mill/mill4.jpg)
The ribon cable connects to all 3 drives. I pulled the ribon cable from each socket and wrote down which pins are present on each connector in order to identify which pins are common to all drives and which pins are unique. There are 4 pins that are common to all the drives and 4 pins unique to each drive. The pin numbers I wrote down isn't an actual pin number it was just an easy system to use to draw a connector without making mistakes.
(http://www.ecutune.com/mill/mill5.jpg)
(http://www.ecutune.com/mill/mill8.jpg)
If anyone knows the pinouts of these sockets or knows where I can find the pinout for these sockets or how I can figure out the pin out of these sockets, or has any idea at all to get me past this hurdle PLEASE share.
-
Anybody know a good place/way to find out what pins are what on these drives?
Anybody have any advice or suggestions on where to ask about the best way to free the spindle motor of the belt and even more importantly propperly tension the belt?
-
Ok, I'm finally getting to this.
I've decided to rip out and keep only the power supply from the control. I'm keeping my DC servo motors and replacing the tachometers and resolvers with encoders.
The set up I'm going with is:
US Digital H5MD-500 encoders http://www.usdigital.com/products/h5/
US Digital EA-R10-W5 differential receivers http://www.usdigital.com/products/ea/
GeckoDrive G320 DC servo drives http://www.geckodrive.com/product.cfm?pid=13
GeckoDrive Grex G100 http://www.geckodrive.com/product.cfm?pid=19
I'm using the G100 so I have enough I/O for all the relays and limit and home switches and to give me acurate timing--I'm thinking with the G100 handling the timing I may run the mach mill software on a laptop and not have a computer dedicated to the control of the mill.
-
I also decided what to do about the 3 phase spindle motor. I ordered a varispeed drive from WEG that takes single phase input. This both takes care of my 3 phase problem and makes it easy for me to control my spindle speed with mach. It has a pwm input.
http://www.weglibrary.com/pdf/cfw10.01.07.pdf
I still haven't taken the top of this thing apart. If anyone is familiar with them and has any insights on what to take apart first to remove the belt and get into the mechanical mechanism it has for speed control now please do share.
-
Are the servos direct driving the screws? I don't think the Geckos will push those motors fast enough to your liking you need some bigger drives like the Vipers from Larken Automation.
Are you using the g100 mostly for extra inputs? I think of these will be much better: http://cubloc.com/product/01_01.php
very easy to use through serial ModBus in Mach3.
-
The motors for the x and y axiis are directly coupled to the ballscrews. The motor for the spindle z axis belt drives the spindle ballscrew--the pulleys are still 1:1. The ballscrews are 5 tpi. The motors are 110v 5.9A 1800 rpm. The power supply in the bandit that the mill originally came with is only 60V though. So ya, I suppose if I changed the power supply too I could close to double my max feed rate. I'm just thinking there must have been a reason spindle wizzard matched the 60v power supply with the motors in first place. Opinions anyone?
-
On the note of the g100, I was just doing some poking around and found the latest install guide here on the forum and it says probing isn't working as of sept 2007--has that changed? Is that going to change soon? I definitely do want to get a contact probe.
I also want to run on a laptop over a usb cable without any worries of timing difficulties
-
I think the only reason they put 60v if it is actually 60v is to keep the feed speed controllable, by my estimate at 60v the motors will give you ~190ipm rapid.
It this case you will need a higher count encoders at least 1000 cpr. Because the 500 cpr will give you .0004 between counts.
Have you read this thread about the g100:
http://www.machsupport.com/forum/index.php/topic,5506.0.html
-
Because they are quadrature encoders 500 cpr gives me a .0001 resolution, right?
-
well I see the plugin for the g100 is a beta and that link shared above would indicate that there may be firmware changes need to get it fully functional with mach and gecko isn't going to make them...I guess I should assume the g100 and mach won't be fully functioning together any time soon?
I see the plugin for the ncpod is a full release but I can't find the ncpod for sale anywhere except as a beta still at oemtech. Is that correct? Was the timing error ever fixed? Is it fully functional now or not?
I'm going to guess not many are going to wander into my thread but I'm posting my questions here anyway because this is what I'm going through to get the mill up and running and I'm just not finding the answers in my searches.
-
Hello usfwalden,
Man....You're doin' a bunch of "Showin' n' Tellin'....and I'm "Watchin' n' Listenin'.
That is a nice project you have there. I wish I knew enough about it to help you out.
I wish you good luck with it and hope you find the help that you need.
I'll be keeping an eye out and if I find anything I'll respond.
RC
-
Thanks,
Ya i wish i were hearing more feed back (particularly on the plugin situation) but oh well. I've also started considering using a galil controler and drives through ethernet. It occurs to me that will end up working out much better for the 4th axis. With a large rotary table I think I'm going to need a much higher count encoder to get any resolution on high diameter parts and an index in order to home it. Of course the galil plugin is listed as prototype.......
I'll post in the galil plugin development thread. So far no response in the g100 or ncpod plugin threads but then it's only been a day.
-
Thanks, that really doesn't look encouraging. What do you think of the ncpod over usb with the gecko drives vs a galil over ethernet with galil drives?
I think the only reason they put 60v if it is actually 60v is to keep the feed speed controllable, by my estimate at 60v the motors will give you ~190ipm rapid.
It this case you will need a higher count encoders at least 1000 cpr. Because the 500 cpr will give you .0004 between counts.
Have you read this thread about the g100:
http://www.machsupport.com/forum/index.php/topic,5506.0.html
-
Ok.....
Here's the new, new plan:
Galil DMC-2160 stand alone control which i purchased used connected to mach on my laptop over ethernet
Galil AMP-19540 amp which I am purchasing new from galil
using my 60v power supply from the bandit that I'm scrapping
H5MD-1000I 1000 line encoders with indexes from US Digital on my xyz axiis--I'll figure out the 4th axis motor and encoder after I am running
-
I sure hope the new plan is the final plan...since I already purchased the galil contol. I had already purchased the 500 line non indexed encoders from US Digital and dropped them along with my motors and motor mount parts off at a machine shop to have the mounting plates modified to mount the encoders but the 1000 line indexed encoders are externally, physically identical so I'll be able to swap them.
Hopefully the admins don't take a dissliking to me. After not getting any response to my queries on plugin status in any of the applicable threads (I posted in all of them) I emailed and posted the suggestion that they update the status info and make it a policy to keep it updated.
So anyway...now I'll be using a galil control and am charging ahead with the project.
-
I've received my galil 2160 control and the weg variable speed drive for the 3 phase spindle today. I returned the 500 line encoders without indexes and ordered the 1000 line encoders with indexes having been able to machine the motor mount plates to fit the new encoders in place of the resolvers since externally the 500 and 1000 line encoders are identical.
-
Does anybody know what pull studs these use? There's a tool holder in the spindle now so when I get it out I'll be able to measure the pull stud on it but if anybody knows I'd appreciate the info.
-
hmmm nmtb 40 tool holders don't use pull studs? The drawbar screws into the tool holder? It's just a guess at this point but from what I can find of pictures of power drawbars I'm thinking the unit on this mill is a kurt 101-02. The parts look a lot like kurt parts and the kurt 101-02 parts list says it uses an electric impact wrench as the motor which is what the unit on this machine has.
-
The amp-19540 and cables hav been ordered.
-
the US digial h5md-1000-I differential 1000 cpr with index encoders and cables have come. Everything has a very high quality feel about it.
-
Hello, I've been following along.....wish I had something to offer in the way of help or advice.
I Googled some of your components out of curiosity to get an idea of what is generally involved in a retro.
I may tackle one myself one day.
Some of these parts are right pricey. But you should end up with a real nice machine.
Keep it up,
RC
-
Thanks,
Yes, going the Galil stand alone control with amplifiers route over the "smart drives" with all control functions handled by the pc route certainly added to the cost of the retrofit. On the other hand the cost is at least a little bit offset by not having to buy a pc to dedicate to controling the mill. More important to me, I don't have to install a pc and monitor to the mill. In my particular case that is a big plus. If you look at the pictures of the mill you can see why. The thing has a 60" table not counting the servo on the end. Acounting for the table needing the clearance to slide all the way one way and then all the way the other way the thing takes up an entire wall length of space. I can't give up an entire wall of space. Since the machine is also quite deep, however, I can put large work benches on the wall behind and to the sides of it. Getting rid of that big contol box and not replacing it with something just as big to accomodate a pc and monitor will make that right hand work bench much more comfortable to use.
Hello, I've been following along.....wish I had something to offer in the way of help or advice.
I Googled some of your components out of curiosity to get an idea of what is generally involved in a retro.
I may tackle one myself one day.
Some of these parts are right pricey. But you should end up with a real nice machine.
Keep it up,
RC
-
I'm where the machine is now and starting the actual retrofit. I've been going back and forth between schematics and spindle wizard engineering diagrams and the actual machine and have developed a much better understanding of what exactly I've got here. As I understand it now the machine casting and spindle drive are Ex-Cell-O's standard 602 contribution. From there the knee, huge tubular ways, ballscrews, and "Position Wizard" super upgrade table which give it the huge and acurate travels that attracted me to the machine are Spindle Wizard upgrades. Spindle Wizard also put on the quil motion, power draw bar, coolant, and their own control for things like gear changes, spindle speed, tool changing, and coolant. One of the buttons on the 11 button front panel is an auto/manual button. With that button set to manual the other buttons can be used to control all those functions. With it set to auto the spindle wizard control takes commands from a cnc control (originally a bandit) to control those functions.
(http://www.ecutune.com/mill/mill1.jpg)
The way this Spindle Wizard control is set up a custom "M card" in the Bandit control would send m codes to Spindle Wizard control. The Spindle Wizard control has its own power supply supplying 5 volts for the ttl logic and 24 volts for the signals to operate relays, a 'logic board" that translates the binary input of the m code into a single pin output to the "power board," a "power board" that amplifies and distributes those signals to the designated relay or solenoid, and a "jog board" that supplies the signal voltage to shift between high and low gears and the spindle jog timing to engage the gears after they have shifted.
The Spindle Wizard power supply, logic board, and power board housed in a cabinet on the right side of the machine. I have the chasis with holds the power supply and is the card rack unscrewed here but you don't actually have to unscrew it to pull the cards out. Each wire from the harnesses plugs individually into securely mounted card edge plugs at the bottom of the card file.
(http://www.ecutune.com/mill/DSC_5536.jpg)
The logic card pulled out of the card file
(http://www.ecutune.com/mill/DSC_5537.jpg)
Fortunately I have the logic diagram for the logic card. Looking at this, it looks like I can easily interface with the control the same way the bandit did through the same wires the bandit did. It's using 4 wires ( card edge pins L, M , K, 19 ) as binary input of the decimals 0 through 9 and 3 wires ( card edge pins P=10, R=40, S=20 ) for the decimals 10, 20, and 40. That's only 7 wires I need to hook up to the Galil control outputs to be able to send the m codes to the Spindle Wizard control. Then card edge pin Y is the input for emergency stop. That's 7 I/O pins. Since I have a 6 axis model Galil I have 16 uncommitted I/O pins.
(http://www.ecutune.com/mill/DSC_5534.jpg)
(http://www.ecutune.com/mill/DSC_5533.jpg)
The power card
(http://www.ecutune.com/mill/DSC_5539.jpg)
The relay box. The jog car is in the box behind all the relays.
(http://www.ecutune.com/mill/DSC_5553.jpg)
-
I found a few of the auxilliary actuators on the microswitches on the spindle (limit and home switches) were bent and broken. I suspect both by me and someone previously taking the front panel off of the mill. The nice thing is that since they are auxilliary actuators, as in not part of the microswitches themselves, and since they are actual micro switch brand micro switches the auxilliary actuators (JE-5) are still available. I put them on order along with a 20 amp tyco line filter to install with the vfd.
(http://www.ecutune.com/mill/DSC_5560.jpg)
On that note, I was also able to order a replacement for the broken air pressure regulator that's been stopping me from putting air to machine to get the backgear, spindle brake, power draw bar, and mist coolant working. The exact Watts Fluidair part number (163-2) isn't around still but it looks like their current model ( r364-02c) is a direct replacement.
I mounted the new US digital encoders in place of the no longer wanted resolvers. They aren't mounted in exactly the same place since they aren't exactly the shame shape and size but I was able to fit them in there anyway. I had the pulley bored out to fit the 1/4" shaft and the mounting plate drilled and tapped for the encoder mounting screws. I retained the tachometers just because they hold the other pulley allowing me to keep the same belt routing. The encoder blocks off where a pole used to be in the center of the electrical box to screw into to retain the electrical box cover so I had 4 holes drilled and tapped into the cormers of the elctrical box area so I can replace the one pole in the middle with one in each corner. The pocket in the second motor mount plate which sandwiches with this one and covers the pulleys and belt needs to be milled out for a little more clearance around the encoder pulley.
(http://www.ecutune.com/mill/DSC_5563.jpg)
(http://www.ecutune.com/mill/DSC_5566.jpg)
-
With the motor mount plates that cover the encoder pulleys back from having the pockets milled out for clearance for the new encoder pulley mounting I found I needed to drill out some clearance for the end of the encoder shafts too. I did that by hand and now the motor mounts are done.
(http://ecutune.com/mill/DSC_5582.jpg)
While replacing the broken auxilliary actuators on the spindle limit and home switches I figured out that someone had messed with them before and switched locations between the home and plus over-limit switch. I put them back where they are supposed to be and adjusted them according to the documetation. The over-limit switch cuts the drive power supplies at 0.200" before hitting the bumper. The limit switch activates 0.030" before the over-limit switch and the home switch activates 0.020" before the limit switch. I'm waiting till I mount the quill servo motor (still waiting for the belt) to adjust the lower limit switches.
(http://ecutune.com/mill/DSC_5632.jpg)
With the limit switches corrected and working some of the front panel push buttons started working like they should and this sent me off examining the wiring and relays to find out why the rest of them weren't working yet. In that process I found that many of the bulbs in the series 3 micro switch push buttons were burnt out or missing all together. Honeywell lists them as discontinued with no replacement but emailed me some old documentation on the series which told me that the bulbs were j1 bulbs. With that information and a bulb and a set of calipers in hand I was able to determine that the bulbs are j1 with sub-micro flange bottoms which are available from chicago miniature. I continued into the relay panel to find that my relay lay out didn't quite match the logic diagram I thankfully was close enough that with a bunch of backtracking of circuits from what I did find that matched I was able to figure out what was different. The logic diagram I have has a note at the top about the addition of the cr18 relay which is included in the diagram. My mill does not have the cr18 relay. My relays go in order and match the logic diagram for 1 through 16. Then where you would expect to find relay cr17 there is a blank receptacle used for wiring posts in the back which is in fact what is called p17 in the logic diagram-- then the next relay is cr 3 which is where it is drawn in the paperwork, then relay cr-17 is where cr-18 is drawn in paperwork. I labeled them all with a paint marker to avoid future confusion.
(http://ecutune.com/mill/DSC_5605.jpg)
While I had the relay panel off doing all my tracing of circuits I pulled the "jog board" which handles zogging the spindle while switching the back gear and examined it. It had an electrolytic capictor on it that had done what old electrolytic capacitors do and sprayed it's contents causing a little bit of corrosion on some adjacent parts but nothing bad. I replaced the capacitor and reinstalled the board. While tracing circuits I also verified the pinouts of the amphenol mil spec connectors to the harnesses for everything and labled the connectors with a paint marker. This is where I found my problem. When I removed the spindle motor starters and 3 phase wiring and connected the panel to 115 single phase I got a wire wrong. With that corrected the rest of the front panel controls work and the drawbar panel has power when it should though itself is not working yet.
(http://ecutune.com/mill/DSC_5597.jpg)
(http://ecutune.com/mill/DSC_5591.jpg)
-
Since the internet seems completely devoid of information illustrating how a varispeed spindle drive works I'm going to post some pictures and information here on it now that I've taken mine apart.
Not having any idea what was on the inside I took the rpm indicator off of the front plate. It turned out that wasn't necessary. The rpm indicator dial is driven by a worm gear connected to the varispeed adjustment shaft--both mounted to the front plate. Turning the input shaft turns the indicator dial and moves a fork mechanism on the inside of the plate up and down. It all pulls off in one piece.
(http://ecutune.com/mill/DSC_5624.jpg)
(http://ecutune.com/mill/DSC_5627.jpg)
That fork pulls up or pushes down a throw out bearing which moves the top half of the spindle pulley up and down thus adjusting the effective diameter of the spindle pulley. The varispeed belt is basically a sturdy wide v belt. When the two halves of the pulley are moved closer together the belt rides further out on the pulley; when the two halves of the pulley are moved further apart the belt rides closer in towards the center. The pulley on the motor is also two seperate halves--idential to the spindle pulley parts except that a spring works to try to compress the two halves together so that it can react in oposition to the adjustment made to the spinlde pulley. As the spindle pulley halves are moved closer together making that pulley effectively larger it pulls on the belt spreading the two halves of the motor pulley making it effectively smaller. As the spindle pulley halves are moved further apart making that pulley effectively smaller tension on the belt is reduced alowing the spring on the motor to compress the two halves of the motor pulley making it effectively larger. Thus the spring on the motor pulley maintains belt tension at a constant and the effective size of the motor pulley adjusts in the opposite direction to changes in the spindle pulley size in order to maintain that tension.
(http://ecutune.com/mill/DSC_5616.jpg)
(http://ecutune.com/mill/DSC_5623.jpg)
(http://ecutune.com/mill/DSC_5619.jpg)
I found the following documents explaining the numbering of varispeed belts and how measurement of "pitch length"
http://www.jdv-belts.com/imagesjd/VARIABLE%20SPEED%20V-BELTS.pdf
http://www.gates.com/facts/documents/Gf000204.pdf
The belt I removed is very worn. It was a 1922v332. 19 = the width of the v groove at the outside of the pulley at it's furthest closed position is 19 1/16ths inches wide. I measured my pulleys and verified that was correct. 22 = the v groove angle. I measured my pulleys with an angle finder and verified that was correct as well. v = varispeed type belt. 332 = 33.2" length I measured the centers of the pulleys to be 8.25" apart so double that for a contribution of 16.5" to the length of the belt. That leaves 16.7" of belt length contributed by the circumfernce of the pulleys which means a 5.3" diameter at the pitch line of the belt. My pulleys are 7.5" in diameter. I'm estimating the pitch line on the belt to be .3" in from the top of the belt so I figure the diameter of the fully open pulley at the pitch line to be 7.2" and the diameter of the fully closed pulley at the pitch line to be 3.4" Those two numbers averaged together = 5.3" So the belt is the correct belt--probably the original 30 year old belt.
-
I finally got the power draw bar working and I got the spindle going so it's nice to know it isn't in the condition the power draw bar was, lol.
The spindle is a 3 phase multi-voltage motor. I'm running it on 3 phase 230V provided by a variable frequency drive. I used a 4 wire cord to bring line 1, line 2, neutral, and ground to the machine. That goes to a 30 amp breaker I put in the relay cabinet. I haven't moved the 115 stuff over to that yet but I'm going to; that's why I used the 4 wire cord and brought neutral over in it. From there I used a 3 wire cord to bring line 1, line 2, and ground up to a tyco corcom 20 amp emc filter which I mounted to the back of the aluminum bracket the vfd is mounted on. I mounted that bracket to the bracket that the old reversing switch was mounted to. From the EMC filter line 1 and line 2 provide single phase 230 to the vfd. The vfd grounds through the back plate. The vfd provides 3 phase 230 to the spindle motor. Since it's being run with 230V the spindle motor is wired phase 3 to ( 1&7 ), phase 2 to ( 2&8 ), phase 1 to ( 3&9 ), and ( 4,5,&6 ) tied together.
(http://ecutune.com/mill/DSC_5634.jpg)
(http://ecutune.com/mill/DSC_5636.jpg)
The power drawbar while kind of cool was a mess electronically. It uses an electric impact wrench for the turning and a pneumatic cyllinder to bring the impact wrench and drawbar socket down onto the drawbar. It uses a 115v air solenoid to actuate the cyllinder to bring it down onto the drawbar. The coil on the solenoid was bad. It uses a relay to switch poles for forward and reverse and has some electronics to enable an adjustable delay between when the solenoid is activated bringing the socket down onto the nut and when the impact wrench starts going as well as adjusting the speed difference in the impact wrench between tool in and tool out. This makes it work beautifully when it works and suitable for automatic tool changing but it needed a little diagnosis to get it going. I started by replacing the electrolytic caps just because they are 30 years old and are electrolytic caps. I noticed a bad transistor and replaced that as well. I plugged it in to test it out and as soon as it got power it ran wild (and began blowing the new transistor). At that point I had to figure out how the thing worked. Since it's part of the summitt dana bandit gear and not actually manufactured by spindle wizard I don't have the luxury of the schematics for this part but it wasn't impossibly complex. I found the diode going from the transistor that was melting had 20 ohms resistance instead of no continutity and needed to be replaced. Then I found the scr had less than an ohm resistance instead of no contiunity and needed to be replaced. Now it works beautifully.
(http://ecutune.com/mill/DSC_5639.jpg)
(http://ecutune.com/mill/DSC_5642.jpg)
Able to reinstall a tool holder and ready to run the spindle I did just that and did some testing and adjusting of the varispeed so that the indicator which I removed when I took the varispeed housing apart was set back to the correct place.
-
And now I've diagnosed the last broken thing I know of--the gear changer. It's operated by a couple MAC Valves air valves--one for high gear and one for low gear. The coil on the one for high gear is bad. Interestingly while trying to look up a replacement coil in the MAC Valves catalog online I found out they say they have a lifetime waranty on coils on every valve in the catalog. I don't know if they have a lifetime waranty on the coils on valves they made 30 years ago or not but I'll sure be asking come Monday.
(http://ecutune.com/mill/DSC_5645.jpg)
-
I went ahead and tuned the vfd a bit. The motor is rated at 5.8 amps for 230v and a max of 8 amps.
I set the motor overload current to 6.5 amps. The vfd manufacturer recomends a setting 10 to 20 % higher than the motor is rated for this setting. This is the setting where the vfd starts doing whatever it does to bring the amperage down short of cutting the frequency.
I set the maximum output current to 7.5 amps. This is the setting where the vfd starts cutting the frequency to bring the current down.
I found the auto torque boost produced better results than the manual torque boost so I set the manual torque boost to 0 and the auto torque boost to 30%. That gave me a pretty flat amperage across frequencies without coming close to hitting overcurrent during acceleration.
I set the acceleration time to 3 seconds. I could run it a little faster without hitting over current but it didn't really seem to get the spindle up to speed faster so why stress it.
I set the deceleration time to 3 seconds which gives some comfort from hitting over voltage for now. I expect to turn this down to almost nothing when I'm actually running but I didn't have the spindle brake operating for this tuning session because I've got the airvalves taken apart for the back gear.
-
Well I've run into a nasty bump here. I have the Galil AMP-19540 which I purchased new along with all the cables and modules to hook up to the DMC-2160 I also installed and had purchased used. I wasn't able to connect at first because my laptop has vista and unbeknownst to me since it isn't listed in the documentation there are some special settings you need to make to get the galil software to work in vista. I called there tech support and after we went back and forth about a bunch of stuff that wasn't causing the word xp came out of his mouth and I said "hey I'm using vista" and he said oh well you have to change some settings for vista. Anyway after I got the file that told me what settings to change and changed them I was able to connect to the controller no problem and sends some simple motion commands to make things move just to check everything. This is when I found out it was only displaying 2 of the 6 axiis the 2160 is supposed to have and that the controller is displaying as a 2120 in the galil software although the serial number on the board checks out and it has all the chips on it for 6 axiis.
I tried calling Galil tech support to work out this last issue before buying the software to tune the servos and the secretary informed me that she had been told the tech support department would no longer talk to me because my controller is used. I talked to the tech dept manager and told him I'd purchased everything from them excep the control but he didn't care. His response was "Upgrade to a new 2260 and I'll shower you with support." Asside from it just not making business since to only support products for original purchasers I don't consider it good service and don't want to stay in bed with such a company. I'm returning the Galil stuff.
The DSPMC/IP from Vital Systems looks like an interesting control. I haven't spoken with anyone there yet but the isntructions are certainly friendlier and it looks like it integrates well with mach. Anybody have any experience or opinions on it?
-
I got four 30A8 Advanced Motion Controls amplifiers. Comparing these to the Galil amp-19540 amplifier I returned I have the impression that Galil documentation is intentially vague and confused to keep their customers from realizing that they have limited options under software control for the same settings other company's products have been setting much more accurately and easily with potentiometers for the last 30 years. The 30 year old bandit control I'm replacing was easily tunable with 3 or 4 potentiometers. The Advanced Motion Controls amplifiers are easily tunable with the same few potentiometers. Since I'll be using the amplifiers in current mode and closeing the loop in the control I'll have the pid parameters to configure there--I suspect I won't find them that mystical to configure.
After my experience with Galil I'm being much more cautious about giving my money out. I seem to be in this kind of limbo period where there are about to be products that do everything I want but they just aren't there yet. I really want to get my machine up already but then I'm not about to do business with companies I consider as unscrupulous as I consider Galil and I'm not going to jump into any of these other products that have a lack of response or slow response about what features work and don't work and what's being done to make them work. Right now I'm holding out a lot of hope for the dspmc/ip though. The feature set/capabilities seel bang on for running a machine like mine--they just need to finish it up. I'm hoping to see the index pulse added to homing for the dspmc/ip control very soon and expect to go ahead and pull the trigger on that control then.
-
Usfwalden,
I am working on a similar machine a bridgeport boss machine. (just look at my thread) I am interested in the 30A8 advance Motion control amplifiers..... are those servo drives? My tool changer is manually operated by air (pushbutton) the speed control (vari drive) was controlled by air motor with electrical relays. I plan to use a stock manual crank in place of it and control the speed with a vfd within limits. if I need slower or faster speed window I can then crank it up and down manually. the head has backgear as well. It looks like the Smooth Stepper and cnc4pc dedicated breakout board are the trick system as of now. there seems to be a couple of bugs still but that is the way I am going. it also has a tac feed so we may be able to comfortably tap with a floating tap holder. as it stands right now I intend to use the larkin Viper servo drives with outboard mounted heatsinks and fans blowing on the inside components. I am not sure that is necessary but that is the way I am going with it.
I have enjoyed your thread. I am surprised I had not stumbled upon it before.
Happy Hunting
archie =) =) =)
-
Thanks Archie,
I'm very happy with the 30A8 servo amplifiers. I would definitely recomend Advanced Motion and Control amplifiers to anyone wanting to run big DC servo motors. The 30a8's are actually bigger than necessary for my motors but they fit where I was installing them and that's what I found in the quantity I needed. a-m-c has both larger and smaller amps.
The setup is extremely easy and they let you test the system without having a control. In the 30a8 doc there was a note to check the engineering notes for additional installation pointers which I did. I'm using them in current mode so all I had to look at was page 13 of the engineering notes where there is a step by step for hooking them up and testing the system. http://www.a-m-c.com/download/document/support/general/instnotes.pdf The two things I would point out is that with these drives you are setting them to run your motors at the current the motor wants to run which makes them perform extremely well and that you can test the motion of the axis with the amplifier before you have a control.
The current adjustment is described on page 5 of the 30a8 doc http://www.a-m-c.com/download/datasheet/30a8.pdf The 30A8 amps are 30 amps peak 15 amps continuous. My motors are 5.8 amps continuous. The current adjustment potentiometer is 12 turns + 1 inactive turn on each side and linear. So all I had to do was take 5.8 amps divide by 15 amps multiply by 12 turns +add one turn for the inactive turn = turn the potentiometer 5.64 turns from the ccw click to perfectly tune each amp for my motors.
I'm still holding off on the control but the motion from turning the test pot on the amps is very nice.
Usfwalden,
I am working on a similar machine a bridgeport boss machine. (just look at my thread) I am interested in the 30A8 advance Motion control amplifiers..... are those servo drives? My tool changer is manually operated by air (pushbutton) the speed control (vari drive) was controlled by air motor with electrical relays. I plan to use a stock manual crank in place of it and control the speed with a vfd within limits. if I need slower or faster speed window I can then crank it up and down manually. the head has backgear as well. It looks like the Smooth Stepper and cnc4pc dedicated breakout board are the trick system as of now. there seems to be a couple of bugs still but that is the way I am going. it also has a tac feed so we may be able to comfortably tap with a floating tap holder. as it stands right now I intend to use the larkin Viper servo drives with outboard mounted heatsinks and fans blowing on the inside components. I am not sure that is necessary but that is the way I am going with it.
I have enjoyed your thread. I am surprised I had not stumbled upon it before.
Happy Hunting
archie =) =) =)
-
The Belt size should be the standard one for the converted mill They would not have changed it for the conversion. Trust me they are a hard job to put back in >:( Did it twice on the mill I used.
-
I'm not following you. This mill wasn't converted. It is a spindle wizard.
The Belt size should be the standard one for the converted mill They would not have changed it for the conversion. Trust me they are a hard job to put back in >:( Did it twice on the mill I used.
-
I ordered the DSPMC/IP, three 7711 break out boards, and a 7535 digital IO board today.
-
My dspmc/ip and breakout boards have arrived. I'm sidelining the install until i have the new ac installed in the shop in 3-5 days. It's at the top of the list after that.
-
I read the mach3 with vista thread and think I'll just install xp on my laptop to prepare for running mach.
-
I read the mach3 with vista thread and think I'll just install xp on my laptop to prepare for running mach.
probably a wise choice ;)
Hood
-
I ordered the DSPMC/IP, three 7711 break out boards, and a 7535 digital IO board today.
can you provide a link to this.....
are you saying that you ar going to run this machine over IP like th Grex ?
archie =) =) =)
-
Yes, the dspmc/ip is an ethernet control with a plugin for mach3. It's made specifically for mach3 and it has the features to let you run differential encoders and analog +/-10v input servo amplifiers that industrial size equipment uses. They haven't sold many of them yet but it seems like THE setup for running servos with mach and they actually want your business. I'm going to be giving feedback and video of my machine running it to try to help them prove themselves in the market (assuming it works out well).
They are listed on the mach plugins page: http://www.machsupport.com/plugins.php
Their website is: http://www.vitalsystem.com/
Thieir section in the 3rd party support part of the mach forums is: http://www.machsupport.com/forum/index.php/board,66.0.html
-
I installed the dspmc/ip, the break out boards for the analog io and encoders and hooked up my encoders. I figured out that the 7535 digital io board has sinking outputs and I need sourcing outputs. They are making another digital io board with sourcing outputs so you have a choice but it will be a few weeks before it's done. I went ahead and got an opto22 board and modules.
G4PB16T io module rack with terminal strip connections for both field and control hook ups
http://www.opto22.com/site/pr_details.aspx?item=G4PB16T&qs=100310111003,,,9,12&
8 g4odc5 ouput modules to send the +24 volt signals for my M functions
http://www.opto22.com/site/pr_details.aspx?item=G4ODC5&qs=100310071003,,,31,45&
8 g4idc5 input modules to accept the +24 volt signals from my limit switches.
http://www.opto22.com/site/pr_details.aspx?item=G4IDC5&qs=100310071003,,,22,45&
My machine has overlimit switches which cut the servo power outside of the limit switches but there's no reason to conect those to the control. That's the whole point behind them. There's a home switch on the z which light's the light on the front panel and enables the tool changer. It does it's own thing and there's no reason to connect it to the control either. I'm connecting the + and - minus limit switches to the control and using the -x, -y, and +z limit switches as the home signals. There will be a home switch on the 4th axis so that's a 7th input. That leaves me one open input which I will connect to what was the "tape start" button which is just a button to send a signal to the control. Maybe I'll use that as a feed hold button.
-
I also looked at the grayhill io options. Grayhill and Opto22 seem to be pretty much equivalent and the standards in the industry from what I can tell. They had equivalent offerings to match my needs. They had the 70GRCK16T rack with terminal strips for both field and control wiring and 70godc5 output modules and 70gidc5 input modules--in fact I think they are interchangeable with the opto22 parts. I went with the opto22 stuff because I found new surplus on ebay for all the parts. There are a lot of grayhill modules on ebay and a lot of grayhill racks but none of the racks at the time i was looking had the terminal strip connections for the control wiring.
-
I've installed the opto22 board into the control box, if you call drilling the holes and screwing it in installing it. I haven't wired up any of the io yet. I did get the connectors to hook up the amps to the control though so I've tuned my amps and started tuning the control.
I covered the initial current limit setting of the amps earlier. That's the starting point with the amc amps no matter what mode you are running them in. I was going to run the amps in current mode so that the controller's pid loops would be the only feedback loop with no chances of the amps and control fighting eachother but while continuing to set up the amps I realized/decided that was almost "silly." I'll itemize my amp tuning in a minute but for now--while adjusting the input gain I realized that in current mode the highest current the amps would ever send the motors would be the continuous current. In voltage mode the amps will give the motors whatever current they need up to the peak current limit for 2 seconds to attain the commanded voltage. It seems to me, being that the amps are electrically in the best situation to quickly make the adjustments wanted to get the commanded voltage and advanced motion control amps are good amps so I trust them to do it and hey being able to run twice as much current when needed to accelerate they have to be able to perform better. It felt like running them in current mode I wouldn't be using them anywhere near their full potential so I switched to voltage mode. Here's how I tuned the amps:
-switch to and tune voltage mode:
-switch 1 3 5 on and all others off
**********************************************
-disconnect motor leads
-use current limit harness with pin p1-11 swapped into p-4 set 0v reference
-adjust offset pot to 0v across motor leads
**********************************************
-use current limit harness with pin p1-11 pulled out and p1-2 reconnected to p1-4 for +10v reference
-adjust ref gain pot to full buss voltage across motor leads
-verify voltage across p1-9 and p1-2 indicates correct continuous current limit and adjust current limit pot if necessary
**********************************************
-disconnect test harness
-reconnect motor leads
-remove motor from axis
-turn loop gain pot clockwise until motor oscillates
-turn loop gain pot back 1 turn
-
Controller's PID tuning:
At first I tried a bunch of different tuning methods where you adjust one parameter at a time and wasn't having much luck. Then I tried a method where you set both p and d at the same time in a ratio determined by the pid loop sample rate. The DSPMC/IP has a 5kHz sample rate which means the D should be 25 times the P--then if your amps are in current mode you probably have to turn up d but as you will see I didn't have to. I have my scale set to 100. This will allow me to have a small integer I--I will probably barely need any I at all.
First iteration I tried
P=100
D=2500
(http://www.ecutune.com/mill/1st.jpg)
There was no motion on the first try to I multiplied by 10
Second iteration I tried
P=1000
D=25000
.....oops, I forgot to save that screenshot
There was very sluggish motion so I multiplied by 2
3rd iteration I tried
p=2000
d=50000
(http://www.ecutune.com/mill/3rd.jpg)
The motion was much better but still sluggish so I multiplied by 2 again
4th iteration I tried
p=4000
d=100,000
(http://www.ecutune.com/mill/4th.jpg)
motion was excellent so I tested larger moves
(http://www.ecutune.com/mill/1000.jpg)
(http://www.ecutune.com/mill/10000.jpg)
Vital systems is still working on the axisworks software. That's an alpha I'm working on now. I asked Rufi/Abdul to add check boxes allow traces to be added for position error, commanded velocity, and commanded acceleration and to remove the commanded position and actual position. When he has that done it will be easy to Dial in the feed forwards which I want to do before even bothering with throwing an I in the pid loop.
-
Oh--so nobody has to go back and search for what encoders I am running--they are 1000 line encoders so 4,000 counts per turn and I have 5 turn per inch ballscrews so that's 20,000 counts per inch so I was tuning with a velocity of 150 inches per minute. My max is about 180 inches per minute so this was aggressive without any worries of clipping by either the amps or DACs.
-
I just bought a 10" troyke 4th axis to add to the mix.
-
I installed the one opto 22 g4pb16t board and wired up my io so now I have limits and I'm pretty much ready to run mach. I left space for a second g4pb16t board in my control box but I don't forsee needing it. I've got 8 ouput modules in the board for my 8 m codes (wired up and functioning) and I've got the first 8 of the 16 inputs from the 1st db25 io connector on the dspmc wired to it but I'm only using 6 of them so far--for my + and - limits. I decided not to use the spindle wizard m boards. There is no reason to output binary codes of m functions to m boards to be deciphered and then amplified and sent to the relays. Each M function is activated by a single 24v signal that was connected on the terminal strips in the relay panel. Since I can output 24v from my opto22 modules I pulled the m boards and wired straight through to the terminal strips so now I'm using a discrete ouput pin for each m code.
The amc amps have a "jumper" inside which is actually a soldered in surface mount jumper not a pull it off by hand type of jumper to switch the inhibits into enables (you can order them configured like this to being with). I removed the jumpers so now the amps are inhibited until they are enabled by the control. To enable the amps now the inhibit, + inhibit, and - inhibit pins must be grounded. I wired these straight to the 2nd db25 io connector on the dspmc. This prevents the servos from lurching when the control is powered down.
-
The control installation is done so here are some pictures. The advanced motion control amps are installed where the bandit amps used to be and the control and io boards are installed where the bandit card file used to be. The 4th axis stuff is all there but not plugged in until the 4th axis actually gets here. I used pieces of one of my old tachometer cables for the +-10v analog control signals to the amps and a piece of one of my old resolver cables for the amp enable signals to the amps. My US digital encoder cables were barely long enough to make it into the control section of the box since they had been cut to length for the galil amp back when I was going to use the galil stuff. I also used the original io wires. They go from the control box to spindle wizard side control box where the m boards used to be. There I soldered and heat shrinked the wires from the control box to the wires to the relay box. It's all 100% tinned, extremely well shielded, probably much better than I can buy today even for a small fortune cable. I could make things much prettier with some ties but that would just be making things less serviceable so I'm not going to.
(http://www.ecutune.com/mill/DSC_5908.jpg)
(http://www.ecutune.com/mill/DSC_5917.jpg)
-
Comeing right along. Looking good. This has been an interesting thread. Looking forward to a video of some chip slinging. :)
Brett
-
With my wiring done I didn't want to wait for the next version of axis works to finish my tuning and start making chips so I didn't. With a little trial and error I settled on quadrupling my gains and throwing in some I. I dropped the scale down to 1 so all my gains have 2 less zeros than before. This time I tuned at my full rapid speed of 180 ipm. I tuned all the way down to doing 2 count moves (0.0001") and never have any instability or significant following error even with ridiculous accelerations for the size moves.
Here's a ten count move at 3000, 30000, and 300000 count accelerations.
(http://www.ecutune.com/mill/10-3000.jpg)
(http://www.ecutune.com/mill/10-30000.jpg)
(http://www.ecutune.com/mill/10-300000.jpg)
Here's a 100 count move at 3000, 30000, and 300000 count accelerations.
(http://www.ecutune.com/mill/100-3000.jpg)
(http://www.ecutune.com/mill/100-30000.jpg)
(http://www.ecutune.com/mill/100-300000.jpg)
Here's a thousand count move at 30000 and I seem to have a 100 count move over the one with 300000 acceleration and axis works doesn't like doing step moves that take too long so that's it for this one
(http://www.ecutune.com/mill/1000-30000.jpg)
And here's a ten thousand count move at 300,000 count acceleration
(http://www.ecutune.com/mill/10000-300000.jpg)
I think I'm set now and should be able to push my mill around to my satisfaction.
-
I'm in complete "awe". What a great project, thanks for sharing - :)
-
Thanks guys,
I'm in the process of setting up Mach now. I got my input and output pins mapped out easily enough and have decided I don't like the way mach handles +/- limits. I'm going to leave them wired up just the way they are except I'm not going to enable the limit switches in mach and I'm going to wire the outputs from the opto 22 modules for them to the +/- enables on the a-m-c amps as well as the input pins on the dspmc. The amps will only inhibit motion in the direction towards the triggered limit so you can still move away from the limit without having to dissable it.
-
where do you set the pin for m5? I see you can define the pins for m3 and m4 when you enable relay spindle control but it doesn't say anything about m5 for stopping the spindle....
-
I think M5 just DeActivates M3 or M4. It's in the VB code, no pin assignment.
Just guessing, not sure.
RC
-
Thanks,
sounds like I'll have to do some learning to customize mach before giving it spindle control. My m3, m4, and m5's are discrete momentary inputs. Actually all my m codes except for quill lock are momentary outputs. I don't know where m codes are defined yet but maybe i can just de-define them and write a brain that activates the appropriate output for the appropriate time when an m code is called so I can do them all in one place?
-
Anybody have any suggestions for adding an index to my spindle? I'm thinking a little magnet/bump mounted on the side of my spindle pulley and some sort of hall effect sensor/proximity switch to pick up when it passes. Does anybody make a kit for this?
-
I bought this indexer and it work great
http://www.cnc4pc.com/Store/osc/product_info.php?cPath=25&products_id=129
-
Thanks,
That's kind of cool but to mount that (at least the way they show it)I would have to pull my spindle belt off to get a ring on the bottom half of my varispeed pulley (the top half moves up and down for the varispeed). I'm hoping to do something where I can screw a screw into the side of the bottom half of my spindle pulley so I can do it without pulling it apart. An optical sensor like that would be ok but it would have to count breaks in the beam instead of connections (maybe it counts the same if the beam is connected for 359 degrees?).
Why do you need the board and not just the interupt module? http://www.cnc4pc.com/Tech_Docs/nte3100.pdf I'll ask Rufi what the minimul pulse width the dspmc will pick up is. I'd think I could just use the interupt module wired straight to the dspmc and it should pick up breaks from a set screw passing through it.
I bought this indexer and it work great
http://www.cnc4pc.com/Store/osc/product_info.php?cPath=25&products_id=129
-
You could also use one of these with a piece of reflective tape on the shaft, this work like an indexer.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=270267755646&ssPageName=STRK:MEWA:IT&ih=017
-
how good are those at only picking up reflective tape and not picking up splotches of not rusty vs rusty. Actually I guess there aren't any not rusty spots on my pulley. http://ecutune.com/mill/DSC_5616.jpg
You could also use one of these with a piece of reflective tape on the shaft, this work like an indexer.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=270267755646&ssPageName=STRK:MEWA:IT&ih=017
-
i talked to rufi at vital systems and he said since we aren't using the parallel port driver he thinks the dspmc needs to do the counting and to hook it up to one of my spare encoder inputs. So what I'm going to do is get the photocoupler interupter dealy--I guess that nte part linked to above....and a us digital EA-D-L-10-W5 differential encoder adapter driver to convert the signal into a differential encoder signal...and a us digital CA-FC10-W8-NC-1 cable to connect the driver to the dspmc break out board for encoder connections. I went ahead and ordered the us digital parts while I had them on the phone telling me what parts would work.
-
What kind of backlash do most precision ground ballscrews have? I checked and I have .0010" backlash on the y and .0008" backlash on the x. I have the instructions for adjusting the backlash on the table and it's supposed to adjust down to .0002" but I wasn't able to reduce the backlash on the y. I suspect the spring in the adjusting mechanism has lost a lot of it's tension over the last 30 years and that's why it's beyond its adjustment range. I need to replace the bellows for the ballscrews too so I will look into replacing the springs in the backlash adjusting mechanism when I'm doing that in case I accidentally dissengage part of the mechanism that requires removing the table to get to.
-
A word of warning for others who might buy surplus opto22 stuff on ebay. I bought eight g4 idc5 input modules that were supposed to be new old stock on ebay. I don't know if they were actually qc rejects or what but 3 of the 8 intermittently go out--which if it's on a limit switch is pretty annoying. I spoke with opto 22 tech support and reviewed my set up and all my wiring with them and verified that they are just bad modules and not anything I'm doing wrong or some other problem. The date codes on them are from 96 so they are 12 years old. Anyway I almost ordered new modules while I was on the phone with them but then decided to order grayhill modules through sager. The grayhill modules through a distributor are cheaper than the retail opto22 modules and they have a lifetime warranty.
-
Just for good measure I hooked my analog oscilloscope up to the 24v from my limit switches to see if it's clean and it triggered at 24.8V and displayed a clean, smooth, triangle wave with about an 8.25mS period (122Hz) and 2.2V from peak to trough.
-
Anybody know a good place to get springs? How about to guesstimate what the appropriate spring rate is to preload this screw/nut?
-
Just reporting that there is nothing to report. As much as I'd like to have this machine up and running I had to change my focus to other projects. In a couple or few weeks hopefully I can get back to it. In the mean time if anyone has any info on my questions about springs for the backlash removing mechanism on the ballscrews I'd appreciate hearing it.
-
Are the springs Belleville washers? Basically a cupped spring steel washer.
They should be available fairly easy with a quick search of bearing etc suppliers in your area.
Wayne....
-
No they are regular coiled springs that keep pressure on pushrod mechanisms.
-
Do you have a pic of the mechanism?
Wayne...
-
Wayne, nope. I have the spindle wizard drawing of the position wizard table with the parts break out but I haven't removed the table to get to the mechanism beyond the adjustment points which can be reached without removing the table. I figure anyone who would know the information that would help me would already understand what I am talking about.
-
I haven't taken the table off to replace the ballscrew bellows and fix/adjust the backlash adjusting mechanism but the machine does move so I've given into the temptation to try to start using it already.
I started by trying to surface a big piece of steel and figured out both that I need to get some decent tooling and I needed to square up the head. I had thought just setting the head to zero according to the indicators would get the head pretty square with the table but apparantly not. When I tried surfacing the steel I got a terrible surface finish and could feel channels in it from the end mill not being square with the table.
I found that my nsk dial indicator set mounted well to an empty 3/4" tool holder via sticking the magnetic base to the end of the tool holder--go figure, it was actually quite rigid and a good set up. I set up the indicator so that rotating the spindle would spin the indicator around in as large a diameter circle as possible without hitting the t-slots--about 5.4".
(http://www.ecutune.com/mill/DSC_6304.jpg)
At start the difference between the high reading and low reading was huge--something like 70 thousandths. I first adjusted whatever you call the side to side rotation of the head which was the most off. Then I adjusted whatever you call the front the back rotation of the head. I went through several itterations of that with my nsk .001 indicator which has a nice, easy, smooth movement. When I got it down to one thousandth run out (difference between the highest and lowest reading) I switched to my starret .0001 indicator. At this point I had the bolts which lock down the head half tightened(otherwise the readings change when you lock down the adjustment). I was getting a fine enough adjustment now that I would overshoot and probably went through another dozen itterations before I got it dialed in but I did in the end get it dialed in. In the end I had 2 tens run out (.0002") that was variation in the table surface not alignment of the head.
After squaring the head I tried surfacing the piece of steel again and was getting a good surface; however, I destroyed my chineese end mill about 1/4 through the job. I did a bad job "touching" and got my z offset wrong--that much I know. I probably had my speed and feed wrong too. I need to get myself a handbook since I don't know what I'm doing. I also went straight to the internet and ebay to try to find a good yet affordable tool for surfacing. I ended up getting a 2" diameter 90 degree Iscar face mill that uses tangimill inserts and a pack of new ic328 lnat15 inserts for it at a steel on ebay. I would have prefered IC908 inserts as I'll be wanting to surface aluminum too but oh well--the budget is very tight and I still have to order a nmtb40 arbor for it.
-
One milestone reached....making chips.
A good sign.
Congrats.,
RC
-
I got a speed and feed calculator off of ebay: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&ssPageName=STRK:MEWNX:IT&item=230294668188
I'm hopefull with this I won't be destroying any/many more tools.
-
I took the table top off so I could get to the inner workings of the ballscrew preloading mechanism. It wasn't all that hard--just very heavy. Basically had to remove the motor assembly which is on one side of the x ballscrew, undo the bellow from the table top, unbolt the bearing block at the other end of the ballscrew from the table top, and unbolt the table top from the linear guides, then lift it off.
Here it is with the table top off, one of the linear guides out, and the cover off of the saddle to expose the ballscrew nuts.
(http://www.ecutune.com/mill/DSC_6343.jpg)
Here you can see the ballscrew nuts. The screw on the top with the nut on the left is the x axis. The rod preloading the x axis is on the bottom and you can see the spring thats pushing the rod to apply the preload down there. The screw on the bottom is the y and you can see the rod applying the preload on it extending in from the right. You can't see the spring because the rod is much longer.
(http://www.ecutune.com/mill/DSC_6363.jpg)
Here you can see the rod, and the spring, and the adjusting mechanism after I pulled them out. The spring pushes between the adjustment mechanism and the lock nut on the rod to apply pressure through the rod to the ballscrew nuts. When I took the picture I had already backed the lock nut out a quarter inch or so to make up for some shortening of the spring over 30 years.
(http://www.ecutune.com/mill/DSC_6375.jpg)
While I had the table top off I greased the linear ball bushing bearings for the x axis. Here's a picture of one of them before I greased it. They have 5 channels of recirculating balls that the round linear guides slide on. I found some info on the Thomson website recomending ep2 grease so that's what I used.
(http://www.ecutune.com/mill/DSC_6357.jpg)
-
After having moved that lock nut out to increase the spring pressure and putting the table back together I quickly able to set the adjustment to get zero backlash with minimal preload. I set up a dial indicator with a magnetic base on the table and used a flat point for getting a good indication against the spindle. I used a dial indicator graduated in .0001" I moved the table so that the indicator touched the column just enough to move then moved it .001" more then back then increased the spring tension and repeated. I only had to do this a few times and it quickly jumped from over .001" backlash to zero backlash. Then I tightened down the pin on the end to set the spec'ed .015 air gap from the end of the rod and tightened down the nut locking in the adjusment.
(http://www.ecutune.com/mill/DSC_6391.jpg)
(http://www.ecutune.com/mill/DSC_6400.jpg)
My facemill, arbor, and speed and feed calculator had come so I returned to "making chips." I got a sandvik coromant arbor (actually I got both a 3/4" and a 1" but this one is the 1") and 3" diameter iscar tangimill cutter. I liked the quality of both. The inserts that came in the facemill were pretty badly beaten up but all had 1 good cutting edge left. My finish quality was at first less than I expected so I checked the squareness of the spindle again and what do you know it showed a couple thousandths run out. I adjusted it again and still didn't get the finish quality I was hoping for. I checked it again and what do you know it showed run out. At that point I'm confident my adjustments aren't coming loose or anything. Now I started moving the table back and forth with the indicator on it to see if it's flat and what do you know the center of the table is 5 thousandths lower than the left and right ends. I found a section dead in the center of the table that gave the same indication over a length longer than needed and reset the head to actually be square, finally. I then put the facemill and the piece of steel back into the mill and used a feeler guage to check for run out by measureing the distance between insert cutting edges and the steel. I marked a point on the steel and turned the spindle to measure the gap between each cutting edge and that mark. Doing that I found that there was a couple thousandths variation. With some cleaning off broken edges on the inserts and shuffling of the inserts I was able to get them all at least close enough that the .008 feeler gage would slip through but the .009 wouldn't. I then moved the table to put that mark at 4 opposing sides of the cutter to check for run out and it was good. I'm sure I'll do better with new inserts but I wanted to go ahead and give it a go with these. It came out pretty well (the bottom right quadrant is the one i did this time). It's smooth to the touch now. With new inserts there may not be any visible mill lines at all. I did one pass with a .010" cut depth and then another with a .0015" cut depth both with a feed rate of 25 ipm and spindle speed of 530 rpms.
(http://www.ecutune.com/mill/DSC_6483.jpg)
(http://www.ecutune.com/mill/DSC_6495.jpg)
-
Excellent info here and great pics as always.
I feel like I'm learning al little bit more each time you post - :)
Thanks for sharing all of this-
Dave
-
Thanks, It's my hope the info on what I go through may save others some head scratching.
Excellent info here and great pics as always.
I feel like I'm learning al little bit more each time you post - :)
Thanks for sharing all of this-
Dave
-
I machined my first non scrap piece of metal today. I machined what is called a "step deck" into a flat 6 block. Most engine builders machine the deck of a block flat and the cylinder head mating surface flat in order to get even pressure accross the entire head gasket. On this particular type of engine I use a step deck which is where you machine the rest of the deck a slight step lower than the cylinder liners. This increases the pressure on the head gasket around the cyllinders where it has to hold in combustrion pressures and decreases the pressure on the head gasket around the coolant and oil passages where there is really very little pressure it needs to hold in. Doing this allows me to hold in higher cylinder pressures with less torque on the head studs allowing for more thermal expansion of the block without stretching the studs past yield. :)
Running the machine from a laptop proved to be a huge benefit for this. I was able to walk around and peak around all the tight corners as I jogged the machine from the laptop.
(http://www.ecutune.com/posts/DSC_6504.jpg)
(http://www.ecutune.com/posts/DSC_6536.jpg)
(http://www.ecutune.com/posts/DSC_6542.jpg)
(http://www.ecutune.com/posts/DSC_6551.jpg)
-
Great Job usfwalden!
Great idea and use of your laptop too!
Thanks for sharing, Glenn :)
-
Thanks,
I've got my vice mounted now. If anybody wants me to make plates to mount their motors or something like that I wouldn't mind some paying work to fund licensing and tooling.
-
Rufi at vital systems, the maker of the dspmc, just emailed me the new files. He's added in index homing. This is going to be great for me. In the end here I'm going to be manufacturing high end fishing reels. It's a small market and as we all know too well materials are expensive these days. Homing to the index pulses will let me do small runs without being eaten up by set up time. I'm going to have 3 stations on my table: my vice, a scroll chuck mounted facing upwards from the table, and a scroll chuck mounted on the 4th axis. As long as I don't move the stations with index homing I can just throw a piece of stock in a chuck and run my saved programs with my saved offsets. The .00005" resultion of the machine is better than I would accomplish setting up myself.
I know this thread gets pretty good reading from new visitors who are where I was when I got into this...starting from scratch trying to figure out up front what they need to retrofit a machine to do what they want to do. For what it's worth I can recomend the dspmc. Vital systems is a small business that actually wants your business and Rufi actually wants to know what you need and will do his best to make make it happen. I think the more people who purchase from him the more time he'll be able to devote to that--and my personal assessment is that it's already the most functional solution out.
-
I haven't really put this machine to work yet but it has worked very nicely on the odd job I've put to it as the need arose. I ran the road runner g-code that comes with mach with a pen in a tool holder and a couple facing programs from some of the free wizards during the testing phase but since then it's all been setting a % feed and jogging keys. Last week I finally licensed Mach and the nfs wizards. I've also been collecting tooling and am just about to be tooled up to the point of being fully capable.
I have a piece of 14" inch long 4x4 t6 sitting next to it now that is going to be the prototype piece for the first part I'm going to be producing on this machine. It has some 3d profiling in it. While I could machine it in the vise I may go ahead and set it up in the 4th axis to avoid having to flip the part over in the vise. Either way I'll videotape it and put it on youtube to showcase what a commercial machine with the dsmpc and mach can do.
-
I used one of the NewFangled wizards for the first time today. I build racing transmissions for Subarus and there's a drive gear/awd transfer clutch drum assembly that tends to break on cars that have a lot of power. The drum sheers off of the gear where they are welded so I cross drill them and through bolt them.
First I used German Bravo's circle center wizard to find the center of the gear and set my work offset--thank you German Bravo.
Then I used the new fangled wizard's circular hole wizard to drill the bolt circle. I'm using a 4mm solid carbide drill bit. I set it as 4mm in the wizard and found the values it was giving me for chipload and feedrate were crazy. I'm guessing they went to metric too but I didn't want to rack my brain or risk anything so I just entered the bit as a .1575" diameter and got values that made me happy. 3031 rpms, .0019" chip load, feedrate was I think 11.5 ipm if I remember correctly--anyway it worked very well and drilled the hardened gear like it was butter.
(http://www.ecutune.com/mill/DSC_7653.jpg)
(http://www.ecutune.com/mill/DSC_7661.jpg)
(http://www.ecutune.com/mill/DSC_7666.jpg)
-
It's also worth mentioning for those who look at this thread for feedback on the parts I've used like the dspmc/ip and the a-m-c amplifiers...this machine is really purring now. If you are considering going this route you can expect good things.
-
Thats good to hear that the DSPMC is working well. Its analogue capability is something that has been needed for a while in Mach :)
Hood
-
Here's a multi-purpose picture. You can see a number plate for a yacht I just machined. I both pocketed out around the letters/numbers and profiled around the letters/numbers in 5 successively deeper passes at 57 ipm. Each pass matched the last perfectly so I think that's pretty good proof the dspmc is up to the task. I also only have to zero the z for any job; Since I use the index pulses to home the machine it homes acurate to half a ten which I guarantee you is better than i would do setting up each job. I just keep the corner of the vise saved as my main fixture and when I throw something into it all I have to do is zero the z with feeler gages. I have a square atlas 4 jaw scroll chuck I used to have mounted on the table and saved as my 2nd fixture but it was in the way of machining this number plate so I will have to set it up again next time I do something I want to work from the center on.
You can also see the troyke nc10a 4th axis I just broke down, refinished, and have half way reassembled. I already installed the amc amp and wiring to the dspmc in the control box when I set up the other 3 axes. Now I need to finish putting it together, machine a mounting plate, and set the thing up.
(http://www.ecutune.com/mill/DSC_7941.jpg)
-
great thread. thank you so much for sharing this journey. good read. i work on the opposite side of the spectrum to you with most my jobs happening inside of one cubic inch so its always great to see other people using their tools.
Regards
Chris
-
Thanks Chris,
Speeking of wanting to machine big stuff, I've stalled my spindle a few times and have decided I need more power. I had gotten the electronics to count spindle rpm a while back but never got around to installing them. Now I won't have to. I just bought a fanuc 20s servo to replace my spindle motor. It's a 4.7hp servo with a 2500 pulse coder on it.
Pico makes a converter to pull fanuc's proprietary commutation signals out of their pulse coder signals to give standard hall and encoder signals. I'm getting one of these: http://pico-systems.com/oscrc4/catalog/product_info.php?cPath=4&products_id=29
I also ordered an a-m-c b30a40ac amplifier. It takes AC input so I'm not limited by the dc power supply I have running my positioning axes. I'll be able to scale the gain to get the 2000 rpm max motor speed at 146v and run 15 amps continuous 30 amps peak. The motor's stall current is 20 amps.
This aught to let me take much deaper cuts with big face mills, easily and acurately control spindle start up and rpm with the normal gcode, perform semi-rigid tapping, and since it has spindle orientation I can implement a tool changer if I decide to.
-
AMC have just launched a open forum for support questions.
www.a-m-c.com/forum.html
Nosmo.
-
Very nice setup! well done.
just a word of warning about expanding-sheave drives like that, we took an XYZ drive apart a few years ago and the spring leapt out and smashed a big chunk out of our workshop wall. Best to be careful ;) Obviously you've got away with it so fair enough.
-
It's also worth mentioning for those who look at this thread for feedback on the parts I've used like the dspmc/ip and the a-m-c amplifiers...this machine is really purring now. If you are considering going this route you can expect good things.
can/would you provide links to the exact parts you used. minus the spindle.
Thanks
archie =) =) =)
-
I've been seeing a lot of a-m-c sinusoidal drives for sale. Does anyone know of an inexpesive/reasonably priced means of generating the sinusoidal control signal those drives require? Rufi said the dspMC can't do it so I'm wondering if there is something I could put in between to do the job.
-
The problem is the plain S series require two sinusoidal command signals with 120° phase shift between them.
Probably a sine wave output oscillator that uses the digital step frequency to synchronize it would work.
It should be fairly simple circuit, but I know of nothing out there off the shelf.
N.
-
I think it has to actually use the commutation signals from the motor to correctly generate the sinusoidal command signals the s series amps require. I don't really know and am not ready to invest too much time into it. I've just been seering a lot more big s series than b series amps for sale and figure their might be a cool doodad out there to run them that you have to ask to find out about...like the pico fanuc encoder converter I got for the motor.
-
Bump
-
CUI has come out with a programmable BLDC commutation version that is cheaper than the Pico, the pole count, phase direction and encoder resolution is programmable. The commutation to stator phase relationship is simply done by energizing A & B phase with a DC voltage to orient the rotor, and then the info is stored in flash memory.
The technology is the same as used by Mitotoyo in their digital calipers.
Nosmo
-
Wow,
This gives me flashbacks, I ran one of those for 4 years! The Excello is an excellent machine, like Bridgeport, only better built. The belt speed change is more robust than a Bridgeport too. I pulled the motor off and replaced the belt too. As I recall the motor is bolted to a top cover and you just remove four bolts and lift it off. I wouldn't toss the variable speed drive, even with a VFD. There is NO substitute for gearing to get lots of low speed torque. That little gear motor that drives the speeds up an down in steps set by the setscrews actually works quite well. You could make it give you just two ranges if you wanted, high and low, and you get to pick what those ranges are.
The table drive is not ball screw. It has an anti-backlash nut with a spring loaded mechanism to tighten a second nut against the first. It seems to work okay, but a fair amount of friction so it will never move real fast.
I hope you got the NMTB 40 spindle, because it is way better than the R8. I would also recommend you put a drive on the knee immediately. That sucker is heavy, and with a five inch quill travel you WILL spend lots of time cranking that sucker up and down for tool changes! If you have NMTB start looking for short integrated chucks, and screw machine length drills. I found a really nice NMTB short chuck, and went to buy more, they had discontinued making them, I got he last one!
Good luck with your project.
Gary H. Lucas