Using ClearPath servos together with DMM DYN4 servo

[Richards]

The two SDSK-2321S-RLN motors arrived on Friday evening. They are nice. Quiet. Smooth. Powerful for their size. The goal is use those two motors and a DMM DYN4/86N driver/motor together. The problem is that the ClearPath motors (when connected as per the Centroid Schematics, i.e., Green to Input 5, Red to Ground pulled Input 5 LOW when the HLFB line showed that the Servo was ON. The DYN4 Orange Fault/Ready/Servo OK line does the opposite. It lets Input 5 float HIGH when the 86N motor is ready. So I searched through the junk box to find some optocouplers, some resistors, some LEDs, a 2N2222 transistor and an Arduino Uno. After a little tinkering, I now have signals that basically do what I need to have done.

In wizard, I set Input 5 to NC and I clicked the Direction Box for my Z-axis motor (DYN4/86N) because it rotated opposite to the ClearPath motors. The ClearPaths' HLFB (+) signals were each connected to an optocoupler and the DYN4's Orange signal was connected to an optocoupler. The transistor side of the optocouplers were connected to digital input pins on the Arduino. the 2n2222 transitor was connected to an Ardino output pin.

Here's how it works. When the ClearPath motors are ON and ready, they pull their HLFB signals LOW, which lets me know that the motors are physically on and that they are ready. The DYN4 signal is unclear. If the drive is OFF or if the 86N motor is ready, the Orange wire signal on the DB-25 connector (flying lead that Marty uses in a video), the signal floats HIGH. Only if the drive faults does the signal go LOW. That means that the signal is ambiguous. The motor's power could be off or the drive could be ready, the signal is the same. (It's like using NO connections on limit switches. No signal could mean that the limit switch has not been hit or that the switch signal is broken - not the ideal way to check a needed signal.)

Logic in the Arduino sorts everything out so that ACORN Input 5 is CLOSED when the ClearPath servos are ready and when the DYN4 drive is either OFF or at least showing no error.

I'm from the old school that was taught to only connect one signal to an INPUT, so, I'll use a microcontroller to isolate all of the input signals. While I'm at it, I'll add status LEDs to that board so that I know at a glance if each servo is enabled, it's HLFB/Fault status and also probably use low-voltage level push buttons to control 24VDC contactors for ACORN power, ClearPath power and DYN4 power, along with any other subsystem that will not be controlled directly by the ACORN board.

[Gary Campbell]

Mike...
So have you seen the amount of K.I.S.S. (read as brilliance) in those $4 opto coupled relays that I use yet? The trigger pin can be set high or low AND you can use the NO/NC relay output terminals (daisy chained into a NC feed for the input) to sort out any active/non-active logic from differing fault signals. It's lumberjack programming logic.

That, and the Canadians seem to do everything backwards.

[Richards]

Gary,
I would use the relays that you suggested, but there are several other "irritations" that I want to solve at the same time. Opening the controller door to see which device's shows an error or a caution isn't a good option for me. I normally use a door mounted disconnect switch that requires that I turn off power before opening the door. Turning off the power erases the error indication. I prefer to have a status panel that clearly shows all the information that I need with Red, Green and Yellow LEDs, where Green = good, Red = error and Yellow = caution.I also like to have manual power sequencing. The disconnect switch supplies power to the control box, but a standard Green (NO)/Red (NC) momentary pushbutton actually turns on power so that in the event of a power failure, the controller defaults to power off. Another Green/Red switch controls power to the ACORN controller. And other Green/Red switches control power to each power supply or power supply bank in the case of multiple DYN4 drivers (a single Green/Red switch with an On/Off selector switch for each driver so that a fault condition can be cleared for an individual drive. Instead of using expensive 22mm ABB brand Green/Red switchs with indicators ($75 each), all of the required pushbuttons and LEDs can be purchased for a dedicated microcontroller for less than the cost of one ABB brand switch. It's a win/win situation. Give the operator all of the information that he needs to safely run his machine without costing an arm and a leg. Using I2C communication between modules allows distributed processing at modest prices.

I did something like that for years when I computerized professional photolabs. Kodak had two options: 1. Use punched paper tape at a base price of $20,000 for the tape system and $20,000 per photo printer modification to use paper tape, or 2. Buy a DEC computer system starting at $250,000 to communicate with individual photo printers that had been modified with a $20,000 DEC mini-computer. I offered a third option. My entire system for one printer, including a computer that attached to a Video Color Negative Analyser cost less than just the electronic's drawer for a single photo printer that Kodak sold. Adding additional photo printer drawers to the network cost $6,500 each. It seems to me that Centroid has taken a similar approach with the ACORN board. For $300 you get the guts of an excellent control unit, including software. You only add the motors/drivers/accessories that you need to do your job. The problem (for me, anyway) is that mixing and matching suppliers causes opposite errors (ClearPath HLFB feedback vs DYN4 DriveOK feedback). Fixing that one problem takes about the same amount of money and about the same pile of parts that it would take to really dress up a control box. In a world that is now filled with Arduinos, Rasberry PIs and BeagleBones (as well as a number of other single-chip microcontrollers), why not use that technology?

[Gary Campbell]

Mike...
As usual I agree with all you say. The real limit here is the number of I/O ports. If you don't run out of inputs, you should be able to connect each drive to a discreet pin and have a discreet message. Each of those pins can be set NO/NC, but need to be NPN signals unless some pullup/down trickery is applied. The PLC programming available for the BBG on the Acorn may be equal to what you are doing as most of the logic is built in and is integral to/with the system.

I also use a "door lockout" of sorts on my systems, but never during testing, and neither do you!

Shortly, with the addition of the 1616 expansion board, all should be doable. My reason for the little add-on relay module was simply to free up inputs. Same as yours. One of the most difficult parts of swapping from a controller like ShopBot, where virtually everything is an add-on hack of soft or hard variety to a full featured controller like Centroid is that most everything we would want to do has already been done and implemented in the controller, we just need to learn how to access and implement it.

[Richards]

Gary,

I think that we mostly agree. The "status" display that I'm thinking about has little to do with the operation of the machine. Once power has been applied, the ACORN is in control via its operator interface. What the ACORN doesn't tell me is the enable status of each drive/motor individually. I'd like to visually have an indicator. It doesn't tell me if the Dyn4 drive is not powered on, just that the Dyn4 drive hasn't asserted an error. I'd like to know whether the Dyn4 drive is powered on. As long as I don't need some of the outputs, I could have the ACORN control power to all but the controller's power supplies, but I'll need those I/O lines for other things.

I think that what I'm trying to say is that I want the ACORN to do everything that it was designed to do, but I don't want to NOT have all of the information on a status display panel that the ACORN was never designed to display and I don't like to cycle power to the machine to clear an error when I can just as easily only cycle power to that single device. When there is only a one-time cost for programming, and minimal cost for parts to stuff a circuit board, why not add the status information? We've been in a world that embraces distributed date since the early 1980s. It's even easier now.

[Richards]

Back, pre-ACORN, when I used Gecko G203v drivers, Oriental Motor steppers, a single linear power supply for the motors and Mach 3, I had neither the capability of the ACORN nor the motor control that servos require. I didn't use the ENABLE signals on the Gecko drivers. If power was on, the steppers were enabled. The only FAULT indicator was a red LED on the G203v driver with no signal feedback to Mach 3. Life was simple, but limited.

Now, with servo motors, I can use the ENABLE line. I also can monitor the status of the servos to shut things down if a servo faults. With a minimum of 800 counts per revolution for ClearPath servos and a minimum of 2,000 counts for DYN4/86N (500 X 4), I have at least 4X the native resolution of a 1.8-degree stepper motor's 200 steps per revolution. At more than 2,500 RPM for the SDSK2321S motor and 3,000 RPM for the DMM 86N driven by 120 AC, I have the option of using belt-drive transmissions or gearboxes. (Although I'm not sure that I understand the backlash on a planetary drive gearbox where the AutomationDirect gearbox lists 20 arc-minutes of backlash. If an arc-minute is 1/60th of a degree, then 20 arc-minutes would be 20/60ths of a degree. Or stated another way, 60/20 = 3 and 3 X 360 = 1,080. Using a 25-tooth pinion gear (1.25" pitch diameter) moves an axis 3.927" per rotation and 3.927" / 1080 = 0.0036" possible backlash. That's probably acceptable on a router, but not something that I would really like, given the fact that spending $2,000 to $2,500 for servos and another $200 to $500 per gearbox makes me expect tighter tolerances. Maybe my math is off.)

So, using bottom-up design, where I start with all possible inputs and all possible outputs, I realized that there were a lot of possible I/O points in a servo system that could have power supplies incorporated in the servo drivers - like the DYN4. And given the fact that clearing a fault requires cycling the power to that motor (if I understand the manuals), I thought it would be good practice to be able to cycle power to any driver without cycling all drivers. It would also be good practice to give each driver its own circuit breakers and contactor. That adds about $50 per driver for a 120VAC system and almost $100 per driver for a 240VAC system (double breakers for 1-phase 240VAC).

That's some of the reasoning behind using an Arduino/Pi/BeagleBone or other microcontroller to get a controller box ready for the ACORN to actually control cutting done by a mill/lathe/router. Other options could be to use the OAK or All-In_one board instead of the ACORN, but when the ACORN far exceeds anything that I can imagine needing to control the cut, neither the OAK nor the All-In-One products seemed to be needed.

Pre-ACORN, I would have just used brute-force to eliminate needing to deal with faults and enables and individually driving power supplies. That would work. If I planned on using stepper motors, that's still the approach that I would take. Servos change the paradigm. Why not embrace the possibilities?

[Gary Campbell]

Mike...
Can't wait til I see the results. I've loved your bench testing for years.

I am also with you on the AD SureGear planetaries. I think your math is correct, but where we assume it is at the output shaft, I think they calc at the input shaft. In any case, I have not seen any measurable backlash on the 5:1 and 10:1 versions I have used from them. I have a guy in MN that has been cabinet cutting with them for over a year with ~1200 ozin 3ph easy servos, and just upgraded to DMM's with no issues.

As far as the ClearPath wiring goes if you look at the colors going into my 24V terminal block in this post: viewtopic.php?f=62&t=2233 You will see that I did not relay up the enables, etc. The reality is that most servos like a power cycle rather than enable cycle to clear a fault. Some do better than others, but no Aux button relays for me on Acorn til we can define the button text.

[Richards]

Gary,

Again, thank you! It makes perfect sense that the backlash would be measured from the input shaft rather than the output shaft on a planetary gearbox. I missed that when looked at specs for several planetary gearboxes. If measured on the input side of the gearbox, then the backlash would be less than 0.001 inch, which is very tight on a router. When I ran some tests on my Shopbot (5' X 10' PRT-Alpha which was advertised as having 0.002" repeatability), I found that I could only get that kind of repeatability after I installed the 3:1 belt-drives and after I added braces here and there and everywhere to keep things from flexing on the gantry. So, on that machine at least, any gearing that had backlash under 0.002" would be lost because of the flex in the machine. AND, because wood has it's own built-in flex, it makes no sense to chase extra-tight tolerances on a router.

One of the reasons that I bought the smaller ClearPath servos (NEMA 23 frame size) was to see how they worked with a 16/5 ball screw and to try a 10:1 planetary gearbox (when the budget allows).

Thanks for posting those photos. What I missed the first time that I looked at it was the type of terminal blocks that you use. I had wondered how you could build a control box with so few terminal blocks. Even on my test bench, where I have extra terminal blocks to make measuring signals easier, the terminal blocks seem to overtake the system. If I used the same type of blocks that you used in that build, I would need half as many. I'll order a box of them. Also, when I saw your use of circuit breakers, I was impressed to see that you exceed code on your boxes by using UL 489 circuit breakers instead of UL 1077 circuit breakers even for circuit that start and end in the control box. The circuit breaker that you use probably cost more than the power supply it is protecting. Seeing that attention to detail, is just another reason that anyone looking for a control box should talk to you.

[Gary Campbell]

Mike...
Thanks for the kind words. Don't know if you've seen it, but I've used the 10:1 SureGear on a rotary axis and it worked very well.

https://youtu.be/SwWIJ-d35us

https://youtu.be/SwWIJ-d35us

[Richards]

Gary,
Great video and excellent craftmanship. I always enjoy watching your videos. I've learned much from each one.