Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: diyengineer on March 23, 2011, 04:06:34 AM
-
I have 2 practically new computers and 1 older computer.
Computer 1: Micro Atx intel motherboard, Intel E1400 2Ghz, Nvidia Quadro 3400, PCI parallel port card. (windows XP)
Computer 2: Mini ITX, Intel motherboard, Intel Atom 1.6Ghz, On board graphics, On board parallel port. (windows 7)
Computer 3: ATX, Intel Motherboard, Pentium 4 1.8Ghz, Nvidia Geforce 6200, On board parallel port. (windows XP)
I would like to run as fast as possible. My Viper drivers are capable of 125Khz max input so i wont max them out. Currently have a smooth stepper running them but wanted to see how the parallel port works for me, and backlash compensation.
Attached are the pictures of each driver test screen shot. The numbers really didn't move much, nor did the graphs on any of them so each picture should accurately give everyone a decent assessment.
I built computer 1 to be the controller of my machine. I would just like someone to let me know how the numbers stack up, and if everything looks good! Thanks everyone!!
Computer 1: "microatx windows xp.JPG"
Computer 2: "mini itx atom windows 7.JPG"
Computer 3: "old pentium windows xp.jpg"
-
Also for a quick note! I have been using computer 3 (or the old pentium 4) with my current setup with no issues (with a smooth stepper). I wanted a little bit more horsepower and reliability from a newer computer so i built computer 1 as the replacement. Computer 2 is almost a exact duplicate of what tormach uses for its controllers.
Anyways! thank you!!
-
All look decent with the first being the best in my opinion.
Did you know there is a plugin available for the SS which has backlash?
The few that have tested it seem to think it works well, shame only a few tested when so many moaned for so long about it not having backlash, then again thats just normal.
Hood
-
shame only a few tested when so many moaned for so long about it not having backlash
Off topic but just had to comment.........
Glad Greg finaly implemented BC......thanks Greg. I am one of those moaners and haven't had a chance to try the new plug in.
After waiting a few years I fixed my backlash problem by removing it, but, that will not be the case for a lot of users.
RICH
-
Rich there were many moaners and some a heck of a lot worse than simply moaners and they were doing so in the last weeks leading up to the release, so that is the ones I am aiming my comment at.
I have never asked for backlash, have never wanted it but I still tested it for Greg but I see none of the most vociferous complainers has even had the courtesy to thank Greg, you did so you are not in that group.
Hood
-
Yes, the 1st computer i plan on using. It is the newest and has the lowest maximum variation it seems like. The highest max variation i have seen it reach is 2 US. I will have to strip the computer down and see how clean i can get it running. I assume lower uS is always better. (i had about 23 processes in the background running in windows XP pro).
I have been a new active user over on the smooth stepper forum. I did see the new plugin that Greg has come out with! Pretty cool stuff! He has worked really hard on getting it up and running.
I have never ran my cnc machine without the smooth stepper so it will be interesting to see the outcome. I would like to bypass the SS temporarily, eliminating any extra realms a problem could be created. The smooth stepper has not given me a single problem at all and has been a joy to use. My viper drives can only accept a 125Khz signa (max)l. Mach 3 can only output 100Khz through the parallel port. So i am getting an extra 25Khz by implementing it (SS) over the parallel port. My machine uses servos, and the rapids are already scary fast (for me). I figure if i use the parallel port @ 100Khz i will be down 20% in speed from the smooth stepper. My max jogging speed is 500ipm, and 800ipm (X,Y). If those get cut down 20% i can't complain since i will be rarely flying around at those speeds (worse case scenario, i haven't tested what my MAX speed is even @ 125khz and im still getting scary fast rapids so im doing well.)
For the time being i was just want to cut out any unnecessary things between the controller and machine. This way i will hopefully be able to find and address problems faster. I currently dont need USB, or 125Khz. I can live with 100Khz and the parallel port until i get a feel for the machine. (its a new machine, and im a noob).
By the time i get a feel for how it runs and what its doing, i can convert back over to the smooth stepper. By then im sure more people will have tested out the new driver and backlash plugin, furthering its success. :D
-
I am re-doing my Bridgeport at the moment and normally I use a Via Pico mobos (tiny things :) ) but as I am needing dual Lan this build I decided to go with an Asus micro ATX with a 1600 Atom. Although I wont be using the parallel port I ran the driver test and was amazed at how good it was even at the 100KHz. I scoped the output and it was also quite nice but not as clean as the SS is at that or faster. I think however it would run fine at 100K from the PP if I was to go that route but I need almost 500KHz on this setup so it wont happen ;)
Hood
-
I am re-doing my Bridgeport at the moment and normally I use a Via Pico mobos (tiny things :) ) but as I am needing dual Lan this build I decided to go with an Asus micro ATX with a 1600 Atom. Although I wont be using the parallel port I ran the driver test and was amazed at how good it was even at the 100KHz. I scoped the output and it was also quite nice but not as clean as the SS is at that or faster. I think however it would run fine at 100K from the PP if I was to go that route but I need almost 500KHz on this setup so it wont happen ;)
Hood
I tried to follow along to see how many Khz i would need to find a theoretical max speed in the mach manual but the wording is very confusing and frankly the example they use SUCKS. Anyone wanna give me a run down? I have 3200rpm servos with a 10:1 gear reduction on them followed by roughly a 1" output spur gear on a rack. I have 1000CPR US digital E6 encoders mounted on my servos. anyone have the equation?
Technically 3200rpm/10= 320rpm output MAX.
circufrence of 1" output gear is equal to pie (3.14). Take the 320rpm x 3.14 = 1004.8 inches per minute. This is my physical max limit.
Now with 1000CPR encoders with differential output, @ 1000 IPM what would i need for Khz?
Thanks!
-
For your KHz you are
3200rpm times 4,000 encoder counts = 12,800,000 per min so divide by 60 and you will get per second or Hz, so it equals 213.333333KHz that you would require, so you would need something like the smooth stepper to get full motor RPM.
If using the PP 100KHz is the max which would mean your motors RPM would be restricted to 100,000/4,000 = 25 revs per second or 1,500 rpm
Hood
-
That is really odd. I have been running my X axis @ 800IPM, and Y axis @ 500IPM @ only 125Khz on the smooth stepper. My viper drivers only accept a maximum of of 125Khz input. I was told by viper to make sure that i don't exceed that number so i set it @ 125khz and left it. I believe the smooth stepper was set to 1Khz or 512mhz speed, and each axis was set to specifically 125Khz. How would you explain that? Thank you by the way for your time, it means a lot Hood!
-
Ok your drives will be the limiting factor if their max is 125KHz.
The numbers I had above were motor RPM so at 125KHz your max motor RPM would be (125,000/4,000) x 60 = 1875rpm
Your axis travels 3.14 inch per rev so 1875 x 3.14 = 5887.5 IPM
Hood
-
I have a 10:1 gear reduction on each servo. So would it be 1875RPM/10=187.5 RPM x3.14 = 588ipm max?
-
Sorry doing to many things at once here ;D it would be my number above divided by your gearing (10) so 588.75 IPM I think ;)
Hood
-
Let me say it another way.
Ok we know the max RPM of the motor at 125KHz as it is (125,000 / 4,000 ) x 60 = 1875 rpm
It takes 10 turns of the motor to turn the outoput shaft 1 turn so 1875/10 = 187.5 rpm on the output of gearbox. Each rev of the output moves 3.14 inch so 187.5 x 3.14 = 588.75 IPM
Hood
-
Is there a way to calculate the gear ratio if i know my exact steps per inch?? I went out to measure the gear but cant really get to it, and even so i can't measure the Diametral Pitch, which is obviously less than the overall diameter which i estimated. With the following below is there anyone i can reverse calculate the gear pitch?
X axis: 3191.931756 steps per inch @ 700IPM (10:1 gear reduction) (1000CPR encoder)
Y axis: 2491.704346 steps per inch @ 499.98IPM (10:1 gear reduction) (1000CPR encoder)
Z axis: 10,040.73425 steps per inch @ 49.998IPM (No gear reduction directly linked to lead screw) (1000CPR Encoder)
-
Not sure exactly what you are wanting but heres what I think ::)
X Axis
CPR = 1000 X 4 = 4000 per motor rev
Gearbox is 10:1 so motor rotates 10 times for one output rev so that is 4,000 steps per output rev
3191.931756 steps are needed to move 1 inch so 4,000/319.1931756 = 1.25316inch per rev that the axis moves per rev so divide by Pi and you have the dia which is 0.399 inch.
Y Axis
Same encoder so
Gearbox is 10:1 so motor rotates 10 times for one output rev so that is 4,000 steps per output rev
2491.704346 steps are needed to move 1 inch so 4,000/2491.704346 = 1.605 inch per rev, divide by Pi and you get 0.511 inch
Z Axis
Same encoder so still 4000 per rev of motor, no gearing so again 4000 per rev of screw. 10,040.73425 per inch so 4000/10,040.73425 = 0.398 pitch screw
Think thats right ;)
Hood
-
Well i know my output gear for the x and y are somewhere around 1" i will have to measure them somehow. Really hard to get a caliper in where they are.
The z axis screw is a .1 pitch kirk leadscrew, its brand new.
-
Ok so just going by the Z, 0.1 pitch with no gearing and 4000 pulse per rev encoder that would mean 4000 pulses move 0.1 inch so 40,000 would be needed for 1 inch. So something mixed up somewhere.
Are you sure you dont have some electronic gearing set in the drives?
Hood
-
Seems like if you divide 1000CPR/10,040.73425 you get .0995943, or roughly .100 leadscrew pitch which it was advertised as. My mind is fried haha.
-
I will have to go check to see if i have the pins set for electronic gearing. I'm fairly certain i do not. be back in a moment.
-
If that is the case then your encoders are 250CPR and thus 1000 pulses per rev.
Hood
-
I checked my Viper 100 Drivers:
The only thing to "electronic" gearing i found on the drive is the step multiplier.
The options are:
1. Programmable multiplier
2. 4x
3. 6x
4. 10x
I selected the first one, by not using any jumper on the pins. I thought this was suppose to be a 1x, but doesn't actually state if it is or isn't..
I bought all of them from U.S. Digital directly. They all came sealed in the bags and cases that read 1000CPR.
Even the product number i double checked on there website just now: E6-1000-250-I-D-E-D-B (1000cpr, .250 shaft, options I-D-E-D-B-)
-
Sorry had calcs arse for tit, talking on phone and typing at the same time is not my strong point ;) , heres the new calcs with 1000 PPR
X Axis
1000 per motor rev
Gearbox is 10:1 so motor rotates 10 times for one output rev so that is 10,000 steps per output rev
3191.931756 steps are needed to move 1 inch so 10,000/3191.931756 = 3.13289inch per rev that the axis moves per rev so divide by Pi and you have the dia which is 0.9977 inch.
Y Axis
Same encoder same gearbox so
1,000 steps per output rev
2491.704346 steps are needed to move 1 inch so 10,000/2491.704346 = 4.0133 inch per rev, divide by Pi and you get 1.278 inch
Z Axis
Same encoder so still 1000 per rev of motor, no gearing so again 1000 per rev of screw. 10,040.73425 per inch so 1000/10,040.73425 =0.0996 pitch screw
So looks like the Z at least is out as I presume you used the auto calc and your measuring was a bit out.
Hood
-
Not sure how US Digital marks their encoders, maybe the 1000 is pulses per rev rather than the line count?
Hood
-
Actually hang fire on my calcs, I will think about this before I type, its always a good idea to do that, just I often dont ;)
Hood
-
Ok so start with Z, to calc properly we know 1000 pulses per rev of encoder so 1 rev rotates your screw 0.1inch as we know its 0.1 pitch so steps needed to move 1 inch by calculating would be 10 times 1000 = 10,000
You have 10,040.73425 so divide by 1000 and that would mean 10.04073425 turns per inch so divide 1 inch by that and you get the pitch working out at 0.09959 which is close. Best using the calculated value though unless your screw is not precision.
Hood
-
Your new calcs look dead on. The Z pitch is .1" for sure. I'm 99% sure the X axis is 1" (i bought it new from mcmaster, just lost the paper work but swear its 1") and the Y axis seems to be 1.25" last time i tried to get a caliper in where its housed.
The US digital site has now confused me further. In one sentence they say " we offer 256 to 10,000 pulses per revoultion (PPR) ". But on the drop down menu when you actually select the count it clearly states 1000 CPR. The ones i have are 1000CPR, what the heck is the difference between CPR and PPR.
http://usdigital.com/products/encoders/incremental/rotary/kit/e6
http://usdigital.com/assets/general/90_e6_datasheet_1.pdf
-
So X we start on now ::)
1000 pulses per motor rev
Gearbox is 10:1 so motor rotates 10 times for one output rev so that is 10,000 steps per output rev
3191.931756 steps are needed to move 1 inch so 10,000/3191.931756 = 3.1329inch per rev that the axis moves per rev so divide by Pi and you have the dia which is 0.9977 inchr.
Hood
-
It is possible my leadscrew is not perfect, or it could be that my dial indicator is not perfect either. I will use the auto calculator either way. I need a better dial indicator!! Mine reads .0005" but its a cheap digital 0-1" one from harbor freight ($35USD). It seemed to work well but i would feel better with an expensive indicator.
-
Now Y
1000 pulses per motor rev
Gearbox is 10:1 so motor rotates 10 times for one output rev so that is 10,000 steps per output rev
2492.704346 steps are needed to move 1 inch so 10,000/2492.704346 = 4.0133inch per rev that the axis moves per rev so divide by Pi and you have the dia which is 1.278 inch
-
So are my differential encoders truly 1000 count encoders, giving me 4000 pulses per revolution or 1000 pulses per rev encoders (really 250 count encoders x 4 quadrature). The wording on there website is slightly tricky...
-
Could be both. I never really like using the steps per calculator in Mach as it will only ever be as accurate as you can measure. Using longer distances to measure over will reduce the error but you still need an accurate way to measure that distance. Gauge blocks are good and usually accurate. DTI's are usually accurate but are only measuring over small distances.
Hood
-
I would say they are 250 line encoders giving 1000 PPR
This seems to be borne out by the Z Axis as that is the one we know all the facts about, ie no gearing and 0.1 pitch screw.
Encoders have always been a PITA as some quote the PPR some the line count etc.
Hood
-
According to the manual i have my multiplier set to 1x. If i turned it to 4x would that give me the right numbers?
I found this in my viper driver manual:
"- 4 X o r 1 X q u ad ra tu re e nc o d e r c o u ntin g"
"Encoder Line count is the number of lines per turn. With-in the viper drive the line
count is decoded 4 times finer(quadrature), so with a 1000 line encoder, the viper drive
sees 4000 counts/turn. This is not to increase machine resolution, but to give the drive
more counts to reduce error.
The Step Multiplier setting in the viper sets the ratio of incoming Step/direction pulses
from the control software to the 4X encoder counts.
So with a 1000 line encoder and a step-multiplier of 4 you would have 1000 steps/turn.
With a step multiplier of 8 you would have 500 steps/turn.
With a step multiplier of 1 you would have 4000 steps/turn, but don’t be fooled thinking
your machine will be more accurate. A servo system always runs with a error of at least
a few encoder counts. Also it would require a very high input step rate to make the
motor spin fast.
We recommend a 1000 line differential encoder for most applications. Encoders with
less than 400 counts/turn should be avoided. The viper can handle up to 625Khz
encoder frequency (assuming good cableing )"
-
Also he recommended 1000 line encoders. "anything less than 400 counts per turn should be avoided!."
-
1xI would imagine means no electronic gearing so that would mean if the steps per you have set are moving the correct distance then your encoders are 250 line count and thus 1000 pulses per rev needed from Mach for 1 rev of motor.
If you increased the drives gearing to 4x then that would mean only 250 pulses from Mach would be needed, but you dont want to do that as it reduces the resolution as each step sent by Mach would move the motor 4 encoder pulses.
The higher encoder count would be better for accuracy and resolution but really I wouldnt worry too much unless you are not getting the accuracy that you are requiring with the encoders you have.
Hood
-
I guess we need to figure out what they mean by "CPR" or counts per revolution. They only offer 64-2500 CPR wheels. I picked 1000 CPR because i read the viper manual and took his suggestion of 1000 lines per rev.
-
I figured 1000 lines per revolution = 1000CPR. I wonder if they automatically calculate quadrature into the "CPR" really giving me, as you suggested a 250 line wheel.
But then again they offer a 2500CPR wheel. If they did in fact count quadature in there "CPR" then it would really only be a 625 line encoder, which doesn't sound right since thats there biggest model.
Also there 64 count encoder would only really be a 16 line encoder on a 1" disk.
I will call them tomorrow and let you know what they say. Should be interesting now that im fully confused! haha
-
CPR and PPR can mean the same, I have seen all sorts from different places and it can get quite confusing.
As I said your Z we know has no gearing and you know the pitch is 0.1inch
So to move 1 Inch you need to rotate the motor(and screw) 10 times.
As you have the Steps per in Mach at 10,000, and I presume when you command a move of 1 inch it moves 1 inch, then that means your encoders are 1000 pulses per rev or 250 line. The only other factor would be the drive, if you had it set to 4x that would divide the pulses required from Mach by 4 so that would mean your encoders are actually 1000 line 4000pulses per rev.
So double check the drive setting and make sure you dont have it set to 4x.
Hood
-
Awesome i will let you know asap tomorrow. Thanks for helping me out, its 5am time for bed haha.
-
:D
Best to concentrate on the Z though as it has known values so we can work out where the issue is depending on what US Digital say.
Hood
-
just for some more info: If i wanted to calculate my minimum step length.
1000 Line encoders (4000 count quad)
X: 1"/ 4000=.00025"
Y: 1"/ 4000=.00025"
Z: .1"/4000 = .000025"
250 Line encoders (1000 count quad)
X: 1"/1000= .001"
Y: 1"/1000= .001"
Z: .1" / 1000= .0001"
(if i changed to a 2000 line encoder (8000 count quad))
X: 1"/8000= .000125"
Y: 1"/8000= .000125"
Z: .1"/8000= .0000125"
-
Yes thats correct but keep in mind the limitations of the Viper, that is if it is 125KHz then it will limit your rapid speed the higher the encoder count you go.
Hood
-
I remember sometimes seeing a .00025" deviation from zero or a position i commanded it to go sometimes on the mach 3 dro! Is that because it can only take that small of a step?
I would command say, X10" and sometimes it would land exactly @ X9.99975" Hopefully tomorrow will shed some light on this.
I wouldn't mind running a really high resolution encoder if i had a really small mill/table. My table is about 60" wide usable and 136" usable long. It would take forever to get anywhere haha. We will see! :D
-
I will probably stick with the smooth stepper so i can take advantage of the 125khz. Might as well get as much speed as i can for free since i already bought it. I have yet to try out the new backlash plugin! Excited!!
-
Actually what you said above is not true as you have not taken gearing into consideration so X and Y would be 10 times smaller. for a min step.
It all depends what you need, on a wood cutting machine do you really need very high resolution?
Hood
-
I will be machining aluminum sheet/plate.
OOO so take all those numbers then divide them by 10 i see. opps! :D
1000 Line encoders (4000 count quad)
X: 1"/ 4000/10=.000025"
Y: 1"/ 4000/10=.000025"
Z: .1"/4000/10=.0000025"
Since the Y axis has a 1.25" gear instead of the 1" gear how does that affect things?
-
Actually you would also have to take into account the value travelled by one rev of the output so your step value on X and Y would be different than what I said in the last mail.
What you need is the step per inch, that will give you your min resolution. So really you need encoder counts, any gearing, any electronic gearing and finally pitch to work out what the min step will be.
Hood
-
Support from US digital got back to me today. I have 1000Line count encoders, with 4000 quadrature counts.
"The "CPR" or counts per revolution is the number of lines on the disc. The "PPR" is the number of pulses in quadrature (x4) it gives per one revolution".
Regardless if i use a smooth stepper or parallel port my Viper drive max is 125Khz.
That means my servo speed max would be 125,000khz / 4,000 pulses per revolution x 60 =1875 RPM
My X and Y axis both have a 10:1 gear reduction so the output gear would turn @ 187.5 RPM.
The Z axis has no gear reduction and is linked to a .1" pitch lead screw: 1875 rpm x .1" = 187.5 ipm max.
My Y axis has a 1.25" output gear. 187.5RPM x 1.25" x 3.14 = 735.9375 IPM MAX.
My X axis has a 1" output gear. 187.5 X 1" x 3.14= 588.75 IPM MAX
-------------------------------------------------------------------------------------------------------
Max servo rotation speed: 1875RPM with 1000 line encoders @ 4000 counts per revolution in quadrature. 125Khz, limited by driver.
X axis: 588.75 ipm
Y axis:735.9375 ipm
Z axis: 187.5 ipm
These are my current mach setttings: They don't seem to match what i have calculated.
X axis: 3191.931756 steps per inch @ 700IPM (10:1 gear reduction) (1"output gear *i think*) (1000CPR encoder)
Y axis: 2491.704346 steps per inch @ 499.98IPM (10:1 gear reduction) (1.25" output gear *i think*) (1000CPR encoder)
Z axis: 10,040.73425 steps per inch @ 49.998IPM (No gear reduction directly linked to lead screw) (1000CPR Encoder)
-
In the Viper manual it does state this:
125Khz max step rate.
"The viper can handle up to 625Khz
encoder frequency (assuming good cableing )"
Even more confused!!
-
Lets just stick with Z as all details are known about it, ie pitch of screw, direct coupled and the only thing is encoder counts or drives electronic gearing.
Ok so if it is 4000ppr then that would mean each 4000 pulses would move the Z 0.1inch so 40,000 pulses would be your steps per inch. You have 10,000 set however so if you zero the Z axis and command a G0Z-1 move from the MDI line does it move 1 inch?
If it does then either the encoder is 250 line (1000ppr) or you have 4x electronic gearing in the drive or you have a scaling value of 0.25 set for the Z axis.
Hood
-
In the Viper manual it does state this:
125Khz max step rate.
"The viper can handle up to 625Khz
encoder frequency (assuming good cableing )"
Even more confused!!
Two different things really as the pulses are from Mach via the parallel port and the encoder is direct to the Viper. Could be the opto used for the Step/Dir input is only capable of 125KHz and there is no opto on the encoder inputs. Could also be just a recommended max of Step/Dir and you may find you can get more, you would really have to ask the Viper guy to find out for sure.
Hood
-
AH HA! Look what i found. Remember how i said i had jumper in the step multiplier?
Well apparently programmable step multiplier must not mean 1X. Below it states "any size"!
It must be a default setting in the drive that i have to configure i bet.. I cant find the command to do it though.
You have to use hyper terminal and RS232 to program the drive.
- 1 ,2 ,4 ,8 s te p mu ltip lie r (o r a n y s ize
-
*i had no jumper (sorry typo)
-
Page 30 for commands and 32 for step multiplier.
http://www.larkencnc.com/dloads/viper-manual.pdf
Would the "enter step size" command be what i need to look @?
-
It does look like it but I would have to read the manual properly to know for sure, even then it may be ambiguous as manuals often are. Give Larry a quick email and see if he can advise..
Hood
-
Just had a quick skim through the whole manual and really there is not much info on it but it does look like it is that setting. If you connect to hyperterminal and look you will see what the J (ju) is set to, my guess is it 0004.
Another thing you could do is put the jumper on the 4x pins and see if it is exactly the same, if it is then you know its set to 4x in the drive.
Hood
-
Larry saved the day again! He confirmed what we thought. It was set to 4x.
With no jumper on , you set the j command in the drive.
With jumpers on the j setting is set by the jumper.
EG : with no jumpers on, j5 sets 5 encoder counts/step.
I don't recommend a 1X setting as the drive error is usually a few counts. 4X gives the same resolution as the Line count.
since 1000 line, and with quadrature they are 4000 counts per revolution.
-
I think i will keep it @ 4x. It seems to give me a good amount of speed and the resolution as you stated is very very small.
I also remembered the "top speed" in the servo spec sheet is listed @ 3200rpm @ at a voltage constant of 90VDC. I have a 105VDC power supply.
That is an additional 16% voltage, which i have been told by a few industry professionals is a common practice to do.
If voltage and speed are a perfect linear relationship then that would mean i would have an extra 512RPM, giving me 3712rpm max.
What would my new servo max speed be if limited by my driver? I tried to do the math but it gives me a funky number in return.
Larry says-
"I don't recommend a 1X setting as the drive error is usually a few counts. 4X gives the same resolution as the Line count.
since 1000 line, and with quadrature they are 4000 counts per revolution,"
So even with 4x setting on, i am still getting 4000 counts per revolution or only 1000?
-
I also remembered the "top speed" in the servo spec sheet is listed @ 3200rpm @ at a voltage constant of 90VDC. I have a 105VDC power supply.
That is an additional 16% voltage, which i have been told by a few industry professionals is a common practice to do.
If voltage and speed are a perfect linear relationship then that would mean i would have an extra 512RPM, giving me 3712rpm max.
What would my new servo max speed be if limited by my driver? I tried to do the math but it gives me a funky number in return.
I dont know what servos you actually have but the 80's vintage DC servos that were common on milling machines usually had voltage ratings of 160v but more often than not the power supply was around the 100v and then with the losses in the drives the actual voltage the motors got was 90v or so. Your motors should have a max voltage and current rating and I would not go over that.
One thing however is the problem you have at the moment is the drives limit on its step/dir input, so going faster would require you to reduce the resolution, this may or may not be a problem for you but something to think about.
Larry says-
"I don't recommend a 1X setting as the drive error is usually a few counts. 4X gives the same resolution as the Line count.
since 1000 line, and with quadrature they are 4000 counts per revolution,"
So even with 4x setting on, i am still getting 4000 counts per revolution or only 1000?
You are only getting resolution based on the 1000, Mach will put out 1 step and the drive will treat that as 4.
I personally am not a fan of electronic gearing and was why I was overjoyed when I got the Beta SS as it allowed me to get full potential from my drives. I have 2000 line encoders so thats 8000 per rev but as my drives can handle 2.5MHz input that is not a prblem, I am not even using a fifth of that.
Hood
-
What drives do you have that will handle 2.5MHZ??? That's crazy! It seems i don't really have a choice. Like you said if i did a multiplier of 1 (instead of 4) that would absolutely kill my speed and require a lot of bandwidth (4 times more). The machine works really well how it is, so i can't complain at all!! The resolution is amazing, and i cant complain about the speed cause for how big it is, its speedy!!
I have these servos:
http://www.kelinginc.net/KL34-180-90.pdf
According to this cnc machinery building manual by Siemens, many power supply builders will go 10-15% higher voltage to accommodate like you said for lack of efficiency, voltage drop over long wire runs, etc. I do however have the amperage below the maximum.
Let me know what drives you have. They sound expensive haha.
-
There may be an advantage to going quite high on the encoder counts and using electronic gearing when comparing a max input step/dir like your drives have. Waht I am meaning is really you have a min resolution you can really get due to the limitation so going with a higher encoder count and using a higher electronic gearing would get you the same resolution. The advantage may be is with the drive itself as it can then read the ungeared encoder input and thus compensate better within its tuning thus reducing the error.
The drives I use would be expensive if bought new, the 22KW one I have on the lathes spindle comes in at around £4700 ($7500 ? ) but I paid £180 for it and also picked a spare up for the same price and both were new, well one had been powered up for 15 mins, I think, it said in the drives software.
They are Allen Bradley DSD AC drives and I use various Allen Bradley AC servos. As said the most I have paid is £180 for a drive and the most I have paid is £250 for new 12.5KW motor that I use on the lathes spindle, it too has a retail of about the £4500 mark, isn't ebay wonderful ;)
Hood
-
WOW! Awesome Ebay Find Hood! :D
So you use the smooth stepper with that sucker? Wow i bet it really flys!!!
-
I use the SS on all my machines although presently redoing the Bridgeport and I am going to be using it with some new control although cant say too much yet and to be honest I have a few options.
The lathe is big so I have limited it to 10m/min rapid although I could crank it to 20m/min. Likewise my Beaver mill I have at 8m/min but could theoretically get 40m/min but 8 is plenty fast for that type of machine.
Heres the progress so far on the Bridgeports panel.
Hood
-
Hood, that work is just sooo neat :) (why does all my stuff look like a rat's nest ?)
Tweakie.
-
I dont really know how I manage as I am a very untidy person, just need to see my workshop ;D
Hood
-
holy cow that is a very very clean install!! Got a picture of this lathe? Must be equally impressive!
-
Sadly my workshop is narrow so getting a good pic of things is difficult but heres the lathe thread http://www.machsupport.com/forum/index.php/topic,1234.0.html
Hood