Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Jeff_Birt on May 08, 2007, 04:21:23 PM
-
Well, it is alive but not tuned.
We had an old Dyna 2400 table top CNC mill in the lab with its wonderful 900 line Conversational control. A few years ago I wrote a drip feed program and modified a MasterCAM post processor enough that it would work (most of the time) feeding the program line by line from the PC. Enter Mach III, I whipped up a homebrew 4 axis opto-isolation board with a charge-pump safety circuit (still have to add spindle control). It is fully opto-isolated and will run on 3.3 or 5V LPT and it steals 5V from a USB port (for PC side of circuit), the existing machines +5 and +24V power supplies were retained. As the machine was already stepper based and the OEM drivers were a standard Step/DIR control not much else was needed.
So, now I have to do is motor tuning and get homing set up and we should be good to cut some parts!
-
Jeff,
The last one that I did i had to use Sherline mode to make it run right :)
Just wanted you to know that before you started!
BTW Nice mill 8)
Later
Brian
-
Thanks Brian, that's good to know. I know the optos on my isolation board are fast enough, but I have no idea about the drivers themselves. I guess maybe that is where the problem might lie.
Thanks again, I'll get her tuned up tomorrow.
-
Hi everybody ,I have found one Dyna 2400 two years ago on the e-buy and I did full retrofit by renewing old Dyna control, power supply and driver. Dyna has made originally at the power limit not for stabile long-term run and low quality components. Made not to last, at least my show that marks. We have made new driver and now run in the half step mode. Motors are tuned up and now we have 30% power more in full step and some 15-20 % more. This is the largest problem of Dyna. 27 oz power for this cast iron machine is to small but it can run. I have made 5 axis control but it was run only in 4 axis simultaneously. Still work on construction to make precise ficsation of one sherline rotary table positioned normally to another. First of all I needed to change rolling bearings on the spindle and my advice is to use p4 standard of bearings and SKF LGLT2 grease for lubrification (permanent) .Spindle will run smoothly and with less noise. We did make new control for the spindle and now we have achieved 180 rpm minimal and regular 10000 rpm for the top speed, but we have made that we can reverse direction and have vector control . That means we can cut the screw (sorry for my English) I don't know proper word for all technical process but I hope you will understood. this means that we can control speed from the Mach 3 but Dyna's head is the weaker part of the macine, beacouse is made from Dural and vibration that pass thru make small problems. I have found Precise SC52 spindle with 60000 rpm and it is now on service but after I believe that we will doo much bather job. Precision of old head is to make around 0.01 mm in production and much precise in positioning. I wanted to have control of positioning corrected from linear scale but Mach3 can't support that and I hope that new version will do that .So all yellow cabinet was put in the waste and new electronic has made. My advice is don't try to make old electronics to work, I have spend 6-7 months try to figure out what is wrong. If you have question send me a message. Best regards Raca Sasa
-
Well, tuning went so-so today. I can't get eth velocity above 7.5 Inch/min without missing steps. I suspect that even though my laptop appears to pass the driver test, that it may be the source of the trouble. And yes, Brian Sherline mode did help a bunch.
-
I hooked up the new PC today, did a fresh and lean XP build on it yesterday, installed Mach III, ran driver test, copied over the profile from my laptop and finidhed tuning. Got the velocity up to 20 IPM for jogging, 25 IPM seemed OK 30 IPM was too much. I did find that the infamous qt.exe was running on my laptop which could have been causing the problem.
I got the homing mostly figure out BUT...the home switches on this machine are at the far positive end of each axis, I got homing set up so it homes and sets the machine coords with the proper offsets. But the local coords are not affected, I would like it to reset the local coords to the machine coords on a home.
-
it should set the Machine coordinates to zero when you home (Unless you changed the settings under the homing and limits)
Check you settings under Config/Homing and Limits
Thanks
Brian
-
Brain, since the limits are at the positive extremes of all three axis I have added in offsets to that the machine coords are correct. For simplicty I just want it to also set the local coords to the same values so the machine is forced into working in machine coords unless/until you reference to a workpiece/fixture. This machine is used for a few classes I teach where I only have one semester to give an introduction to CNC, PLC's, robotics, MMI's, etc; so I'm thinking that having the local-machine after a home (in a known state after a 'reset') will take some of the guess work out of it for the students. I think I found a post about this yesterday referring to adding some addtional code to the 'Ref All' button.
-
After squashing a self induced hardware bug today (cold solder joint), I moved my computer cart around to the front of the machine and thought I would try running the first part program. I was greatly confused at first as all the X and Y axis were moving backwords (but homing OK). It was then that I realized that when I had the computer cart on the backside of the machine (so I could run Mach and probe stuff withteh Oscope) that I set up the X&Y backwords! So, after removing my head from my butt, I changes things around in 'Ports and Pins' and 'Homing' and now all is well. All the home switches are infact at '0', so I had to remove my offsets in homing. I now have the machine moving the table to about mid position after homing. I tried to force the soft limits and machine coords on (and LED's) but my script will only toggle them (below). Any idea what I may be foing wrong?
Thanks
DoButton( 24 )
DoButton( 23 )
DoButton( 22 )
DoButton( 25 )
'' Wait until the axis are done moving and then
' Set to machine coords and move table to front
While IsMoving()
Wend
Code "G53"
Code "G0 X3 Y2"
'' Wait until the axis are done moving and then
' set the proper modes
While IsMoving()
Wend
'' Make DRO and Screen reflect Machine coord mode
If GetUserLed(16) = 0 Then
DoOEMButton(107)
End If
'' <ake sure that we enter into soft limit mode
If GetUserLed(23) = 0 Then
DoOEMButton(119)
End If
-
Try GetOEMLED(23) = false it is not a user led :)
That should do it
Nice job!
Brian
-
Try GetOEMLED(23) = false it is not a user led
Doh! Thanks, I'll give that a go in the AM. I spent a good part of the day ripping apart the X and Y axis and cleaning things up. Plan to get Z done tomorrow. I spent some time today looking for a refernece design (other than a RC network) for an PWM to voltage converter. I always like to see what others have tried before designing my own. Anyone know of one on teh web?
Thanks!
-
look at the Digispeed here
http://www.homanndesigns.com/DigiSpeedDeal.html
-
Thanks Brian! Changed teh code as you suggested and now all is fine. The Digispeed looks great, a very flexible product indeed, I like how they keep the manual pot as a spped override of sorts. I guess first I will have to see what voltage reference range the built in spindle driver board on the Dyna uses and go from there.
-
I got the old Dyna tuned in pretty well but I had to back the rapid down to 15 IPM to avoid lost steps. I think that's as good as the stock drivers/steppers will do, though I have a hunch that new drivers would work wonders. The backlash in both X and Y stinks at 0.002" but any more tweaking on the anti-backlash nut stalls the axis. I ran a 30,000+ line program today (took 2+ hours) and all went well other than me futzing up the tool change (wrong Z offset) so the finish pass left a lot to be desired. But, heck is was just wax after all.
Now that it is running good I would like to get some new tool holders; called Dyna today and they can only get the 1/4" shank (guess I could bore it out) for $85 and the 'quick change nut' is $70. Does anyone know whay type of tool holder these machiens use? There is a measured drawing in the manual but it does not state if it is a standard taper or not.
Thanks,
Jeff
-
Can you post a scan of the tool dimensions, maybe QC30 when you mention "quick change nut"
Hood
-
I guess that would have made sense ::)
-
That is a tiny taper, sorry no idea what it is :(
Hood
-
I see now how it works, the collets go direct into the spindle. I just did a google for ESX16 and heres one place http://www.reneswiss.com/Catalog1/esxser.htm
-
Looks like ESX16 and ER16 might be the same, have a look here for more tooling http://www.floydautomatic.co.uk/arbors.htm
-
Thanks Hood, that last link nailed it I think. I feel silly now for asking as I guess I should have assumed that maybe the ESX16 meant something. Anyhow, thanks again, have to see how to get these in the US now.
-
I did a more extensive search today for quick change tooling for an ESX16 taper and it looks like Floyd Automatic is about it. I have emailed them to see if they have something that will work with the Dyna 2400. The good news is that the ESX16 collet seem to be somewhat popular in the jewlery making industry so at least collets should be easy to get. I'll post details when I get them.
On another note, I just noticed that the stock Dyna 2400 stepper drivers have a programming jumper to select the axis and a voltage reference to set for current adjustment. The Z-driver has seperate adjustments for UP and DOWN, with DOWN being only about 85% of up. The UP current limit is also 20% or so hiegher than X or Y (according to the chart on control box cover). My X and Y drivers are currently switched, so I guess I will switch them back (I think direction is the only thing diffrent but I dont know) and check the current limits to see if I can then get a little more out of them.
Here is a cell phone picture of the first thing we cut with Mach III. I screwed up the tool change the finsh pass didn't do much, but it proved that Mach and the machine work.
-
If the ESX16 and ER16 are the same then you will get collets almost any place.
Looks like your mill is running well.
Hood
-
Got a quick reply from Floyd Automatic: Thank you for the enquiry, unfortunately our quick change system is only available with ESX20 or ESX25 taper so it is too big for your spindle.
So that's ano go for my mill at least. Did some more web surfing this mornign and found this: http://www.hightechsystemsllc.com/index_files/Page2045.htm. Which is pretty interesting and low cost. It is made to thread directly to a Sherline spindle but they do have an adapter that you install in the exisisting chuck, http://www.hightechsystemsllc.com/index_files/Page993.htm, and then the quick change threads onto the adapter. And it uses the same collets the my machine surrently does. There are some videos on their website. I called them this morning and found out that the quick change comes with three tool holders. Anyhow, it looks it might work for our purposes.
I also adjusted the voltage reference for the X and Y drivers (OEM) as they were a bit low. Now I am happily traversing at 21 IPM! A pretty good increase from 15 IPM.
I'll report back after I recieve teh quick change listed above adn let you guys know how it works out.
-
You may finde at Schaublin or Iscar!
-
I got an email the other day after somebody read this thread. Just a quick update:
I recently re-updated this machine to a G540. WOW! what a huge difference. The motors and power supply are stock, but I ripped out the old stepper drives and a G540 dropped right in the same location. It is like a whole new machine, running 40IPM rapids even with 101,600 steps/inch (thinks to the SmoothStepper putting out such a nice pulse train.)
I'll post some pictures soon.
-
It's good to hear somebody that continues work on Dyna. I have finish mine, and results is: First I have to remove all electronics form Dyna 2400, second I have removed all transmission in all 3 axes and changed to THK ground ball screw witch I have obtain from "http://www.ebay.com/sch/jak8858/m.html?_trksid=p4340.l2562" this guy. No need to buy new one, this are in excellent condition. Look for C1 quality, and 14mm diameter with 4mm lead. You need to do some more adaptation to connect this to machine. I have used R+W couplers for connecting this ground ball screw to motors. This is my recommendation, and for third I have changed motors for servo Sanyo Denky from this guy "http://www.ebay.com/itm/Sanyo-Denki-BL-Super-P5-Servo-Motor-P50C08075HXS21-750W-/400251738520?pt=LH_DefaultDomain_0&hash=item5d30dcd998" and after this I have used Teknic servo drive witch made this system works. Smooth stepper is necessary to be in this configuration, and you can have fast machine (3.5m G0) and precision of 0.00025mm theoretically. In practice I have checked out to 1 micron repeatability in all axes. Now, I'm working on 4th axes and results will come in the week. If you need some advice just send post. Best regards Sasha