hello - new forum

[jmelson]

Hello, everyone! And, welcome to this new forum.

If you have any questions about the products made by Pico Systems, please ask here, or send me an e-mail at : elson ( at ) pico (dash) systems (dot) com

If you have a brushless motor and are looking for a suitable, low-cost drive for it, let me know. I may need to borrow your motor to check it out and find out the correct wiring and settings. I have been testing motors with this drive, and just got excellent results with a Panasonic Minas S series motor (MUMS042A1A, but it should work for the whole series).

Thanks,

Jon Elson

[Dan Falck]

Hi Jon,

I'm glad to see you have a forum here. Maybe you could post a few pics of machines running your systems.
I'm hoping to buy some of your PWM brush servo drives after the holidays. Christmas is eating away at the budget currently, but there is always hope : )
I have that Bridgeport with the big SEM motors on it.

Thanks,
Dan

[jmelson]

Originally Posted by Dan Falck Hi Jon, I'm glad to see you have a forum here. Maybe you could post a few pics of machines running your systems.

Hmmm, I'll see what I can do. I do have just a FEW pictures on my web site at
http://pico-systems.com/minimill.html
and
http://pico-systems.com/bridgeport.html

I'm hoping to buy some of your PWM brush servo drives after the holidays. Christmas is eating away at the budget currently,

yeah, I know about that one!

but there is always hope : ) I have that Bridgeport with the big SEM motors on it. Thanks, Dan

Well, the only problem with the PWM system is it needs fairly high encoder resolution to make up for not using tachometer feedback for velocity control. The higher the encoder resolution, the more precise the motion and velocity control will be. (Higher resolution decreases the sensitivity to the PID's D term magnifying the quantized nature of encoder counts.)

My own Bridgeport has my own velocity servo amps, which I made available as a kit - I supplied the boards and a few parts, the buyer had to supply the rest. I sold about 60 boards, I know of NOBODY who is actually using them.

Jon

[Dan Falck]

Hi Jon,

You might want to get your main URL placed in this forum.

I'll put it here:

http://pico-systems.com/motion.html

But maybe you should place a 'sticky' message somewhere in here.

Thanks,
Dan

[cncdreamer]

Hi, Jon.

Nice to see you got a forum going here.

Previously, you said---

Originally Posted by jmelson Well, the only problem with the PWM system is it needs fairly high encoder resolution to make up for not using tachometer feedback for velocity control. The higher the encoder resolution, the more precise the motion and velocity control will be. (Higher resolution decreases the sensitivity to the PID's D term magnifying the quantized nature of encoder counts.) My own Bridgeport has my own velocity servo amps, which I made available as a kit - I supplied the boards and a few parts, the buyer had to supply the rest. I sold about 60 boards, I know of NOBODY who is actually using them. Jon

I don't see any data either here or on your website regarding "fairly high encoder resolution" and "the higher...the more precise...." It would be useful to see numbers here or there or both.

Regards,
Kent

PS - my ears are burning. I bought a USC board from you several years ago, just before I suffered back-to-back life-and-death events. It, along with the rest of the parts I needed, sits unused next to a partially completed CNC mill I hope to get back to real soon now. My development plan is still to swap to servos after I get my feet wet with junk-box steppers.

[jmelson]

Originally Posted by cncdreamer Hi, Jon. I don't see any data either here or on your website regarding "fairly high encoder resolution" and "the higher...the more precise...." It would be useful to see numbers here or there or both.

OK, the Bridgeport has 1000 cycle/rev encoders and a 5 TPI leadscrew. So, the encoder resolution is 1000 x 4 (quadrature) x 5 (TPI) = 20,000 counts/inch. (I have a 10 TPI ballscrew on Z, so that gets 40,000 counts/inch.)

The minimill has 16 TPI Acme screws, a 4:1 belt ratio and 500 cycle/rev encoders. So, that is 500 * 4 * 4 * 16 = 128,000 counts/inch.

PS - my ears are burning. I bought a USC board from you several years ago, just before I suffered back-to-back life-and-death events. It, along with the rest of the parts I needed, sits unused next to a partially completed CNC mill I hope to get back to real soon now. My development plan is still to swap to servos after I get my feet wet with junk-box steppers.

Well, I know about that. I had some CNC conversion parts sitting around for many years before I did the Bridgeport. I was then kicking myself for not starting sooner!

The last few projects done on the manual machine really stand out with hours of effort trying not to screw up on some coordinate and ruin the whole piece.

[Dan Falck]

Hi Jon,

Maybe you could post some pics or info on customer's setups that you have done. I'm always interested in machines and different setups.
I seem to remember Kirk Wallace using your interface boards.

Thanks,
Dan

[jmelson]

Originally Posted by Dan Falck Hi Jon, Maybe you could post some pics or info on customer's setups that you have done. I'm always interested in machines and different setups. I seem to remember Kirk Wallace using your interface boards. Thanks, Dan

I don't generally get on site to install my stuff on customer's machines. There is one setup in town, but I'd be a bit embarrassed to show it. It was set up by Mechanical Engineering students at a university, and then my friend bought it. It was running Mach (I think, might have even been one of the DOS packages from the DARK ages). He switched to an old BDI EMC version, but had problems with losing position on occasion. So, I set him up with a USC and Gecko interface.
He has had good luck with the system, but after that many change-overs, it is a mess.
Also, the machine is a bad fit for him. It is one of the geared-head machines with 1250 RPM max on the spindle, and he carves up 6063 extrusions with a 1/8" end mill. He really needs a higher speed spindle. Other than mine, it is the only machine in town.

Now, there is Stuart Stevenson's 5-axis Cincinnatti mill, running with my PPMC hardware,

YouTube - EMC2 5 axis cinci at MPM

YouTube - EMC2 5 axis cinci at MPM

and there's a silk screen printer in Brazil running with my USC board and EMC. I can't find the link to the video at the moment.

Chris Radek has a Hardinge HNC using my resolver converters, so he didn't have to retrofit TINY encoders to the machine.

YouTube - Hardinge HNC conversion running EMC2

YouTube - Hardinge HNC conversion running EMC2

There are at least two other HNCs running with the UPC/PWM servo amps. Kirk Wallace is one of those.

There are a couple machines at Hogeschool Antwerpen (Belgium) that are running with various Pico Systems boards, but I'm not too clear which one use my boards, and which are using other stuff.

There are a couple of PPMC board sets at Nanyang Technical University (Singapore) that have been used in student retrofits of various milling machines.

Jon

[cradek]

Originally Posted by jmelson Chris Radek has a Hardinge HNC using my UPC controller and PWM servo amps, and also using my resolver converters, so he didn't have to retrofit TINY encoders to the machine.

Oops there are so many HNCs running EMC2 that you confused mine with another one. Actually I'm using the original velocity-mode amps and servos, and Mesa hardware for the interface with the PC. Mesa does not make a resolver input board; you're right that I'm using your product for that part of it, and it has worked very well for me.

The resulting resolution with the resolver board is 204800 counts per inch so the motion is extremely smooth. I should post a video of my tenths indicator with the machine moving at .00005 inches per second. It is perfectly smooth. If not for the indicator you would never suspect it is moving. To go from that to full speed rapids (conservatively set at 200 inch/min on mine, limited by the amps and machine, not the Pico or Mesa devices) shows just how good these servo setups can be.

[jmelson]

Originally Posted by cradek Oops there are so many HNCs running EMC2 that you confused mine with another one. Actually I'm using the original velocity-mode amps and servos, and Mesa hardware for the interface with the PC. Mesa does not make a resolver input board; you're right that I'm using your product for that part of it, and it has worked very well for me.

OK, I should have known you didn't have my PWM servo amps. Yes, I have gotten confused who has what.

The resulting resolution with the resolver board is 204800 counts per inch so the motion is extremely smooth. I should post a video of my tenths indicator with the machine moving at .00005 inches per second. It is perfectly smooth. If not for the indicator you would never suspect it is moving. To go from that to full speed rapids (conservatively set at 200 inch/min on mine, limited by the amps and machine, not the Pico or Mesa devices) shows just how good these servo setups can be.

Wow, 66666 : 1 speed ratio! That is pretty impressive.

Jon

[CNCfun&games]

JON,

This weekend I was blogging around the Zone and found a reference to your
UNIVERSAL STEPPER CONTROLLER, which I downloaded.

To minimize the Confusion around this generation of products could you define
the similarities & differences between your product ($250) and the Smooth Stepper ($150), the Pod ($???) & the Brain($499) for neophytes like me who are a little performance crazy?
I want my Gecko Drives & SQ. Steppers to run silky smooooooth!

Thanks…….W. Smith

[jmelson]

Originally Posted by CNCfun&games JON, This weekend I was blogging around the Zone and found a reference to your UNIVERSAL STEPPER CONTROLLER, which I downloaded. To minimize the Confusion around this generation of products could you define the similarities & differences between your product ($250) and the Smooth Stepper ($150), the Pod ($???) & the Brain($499) for neophytes like me who are a little performance crazy? I want my Gecko Drives & SQ. Steppers to run silky smooooooth! Thanks…….W. Smith

Well, I don't know the competitor's products as well as my own, so I may not be comparing things exactly. First, my boards only work with EMC, not Mach (unless somebody wants to write a driver for it). The whole concept of it was to treat steppers like servos, which was the only paradigm EMC knew back when I designed this. So, EMC sends a velocity command to the controller, the controller sends out step pulses at the requested rate, and counts the pulses as they go out (or can use encoders if available). EMC then reads the position count and figures out a new velocity. It repeats this 1000 times a second. Having to only do something 1000 times a second, it greatly relieves the burden on the CPU.

I tried to make these boards all-inclusive, so they have E-stop logic, 15 additional digital inputs and 8 digital ouputs. You can plug up to 8 solid state relays directly into the board, so no external SSR mounting boards are needed. The boards have individual wire terminals so you don;t need additional break-out boards. There is an option for a spindle speed DAC if you have a VFD or other speed controller on your spindle motor. Multiple boards can be daisy-chained on a single parallel port, and they interchange position sampling, watchdog and E-stop signals on the parallel port cable.

As for performance, I tried the first prototype of the USC at the 2002 NAMES show on a Sherline with a Microkinetics driver, which is quite pitiful compared to a Gecko. I was still able to get about a 50% increase in performance. I think it was 2003 that I put some NEMA 34 motors on a 150-Lb minimill with Gecko 201A drivers, and the production USC board. I demoed it the NAMES show running 90 IPM! The motors were spinning at 1440 RPM to do that 90 IPM. I think a major reason such performance could be obtained was the very smooth step pulse timing.

I hope this answers your question, if not, ask some more.

Jon