Problem using stepper for high-momentum load... (blowing drivers!)

[Riceburner98]

Figured I'd ask here and see if anyone has input.. We're using Stepper motors (Probably should be servos with encoder feedback, but not my call..) to drive turntables for product testing. Basically spinning a group of 16 targets at a set rate to test cameras. We need to have exact triggers and exact speeds, hence the steppers. Anyway, with the new high speed cameras we need to go faster. The current turn tables top out at 675 RPM or so, (180 frames / second) using IM483 drivers on Pacific Scientific motors. I use the same drivers on my Taig CNC machine, so I knew they'd go faster.

Well, I hooked up one of my spare (black Xylotex 269oz) steppers to my spare IM483 driver, and popped the platter on top. (11" diameter plexiglass circle, about .25" thick) 1200 RPM seemed to be no problem for it, which is all we *really* need to do, but being how I am I wanted to find the max speed. I got the thing up to 3,000RPM several times, then the driver chips on the IM483 exploded. Turns out accelerating (slowly, using mach2) is no problem and it'll sit at 3,000RPM no problems. But on deceleration, the motor turns into a generator and I was seeing 90+V at the power supply! (48v supply) So basically I was dumping 90+V through the IM483 driver chips, and it didn't like it. (rated at 48v, surprised it lasted as long as it did!)

What I think I'd need is either A) a brake on the stepper that stops it quickly (which would likely loosen the nut on the platter eventually) or B) a way to disconnect the motor from the driver immediately instead of spinning it down. I was thinking of 4 heavy relays in line with the phase signals going to the motor, that would open up when you want the motor to stop. But what would that do to the driver? It would be like unplugging the motor at high speed, which is supposed to be a definite No-No..

After I blew the IM483 I tried some Centent CN0165 drivers (basically old Gecko's from what I hear) and they dumped the full 90+v back into the power supply instead of soaking up most of it like the IM483's, so the power supply went into over-voltage and shut down. (which is when I realized why the IM483 exploded..) So..... Anyone ever do anything similar, where you need to slow down a 3,000RPM (well 1,500 RPM really..) stepper spinning something heavy?? Or am I just nuts...?

They are designing newer "high-speed" tables with new drivers and motors (over $500 for the driver boxes alone!), but I have a feeling they'll run into the same problem, especially since the platters are going to get thicker too.. Maybe using the relay trick, but instead of disconnecting the motor, using them to drop a low-ohm resistor or Zener across the coils instead to soak up the excess? Maybe try the 50ohm heat-sunk resistors I have lying around... What's the worst that could happen, I blow up a couple more IM483's? LOL

[Mariss Freimanis]

Riceburner98,

You're not nuts._) Without getting into the physics of step motors, look at the problem this way: You store a substantial number of joules in the disk and that stored energy must be removed to decelerate the disc to a stop. There is no mechanism in either the motor or the drive to dissipate this energy, so it is returned (motor acts as generator) back to the power supply. A circuit must be built to dissipate this returned energy, usually as an active clamp across the power supply DC voltage.

Unless I slipped a decimal point, my calculations show 230 Joules (Watt-Sec) of stored energy at 3,000 RPM. That is a substantial amount.

Please see the attached .pdf which goes into a little more detail on this subject.

Mariss

[Riceburner98]

Yeehaa! That's a lot of energy.. The clamp circuit would work great for protecting the power supply, but unfortunately the IM483 driver board I think absorbs most of that before it gets that far, and it's only rated at 48-ish volts.. The Gecko / Centent would seem to be a better choice as it appears to dump the excess voltage back to the power supply, then the clamp would absorb it. We have a lot of the IM483-driven motors though.. Wonder if a sufficiently large suppressor installed on each of the 4 coil wires would work similarly? Assuming the coil drive voltage doesn't exceed 48v (since that's all the power supply can give) if I used say 52v suppressors, any returned voltage over 52v from the coils would be dumped to ground.

I think I'm going to have to attach some of these steppers to my bicycle and power my TV off it.

[Mariss Freimanis]

Well, 230 Joules will run a 100W TV for 2.3 seconds.:-)

By an 'active clamp' I mean a comparator that goes true at a voltage higher than your supply voltage, turns on a MOSFET to gnd that has a 5W to 10W wirewound resistor scaled to support 5A at your supply voltage in its drain path. MOSFET crowbars a 5A load across your supply while supply bus > supply voltage. No harm is then done to any motor drives.

Mariss

[Santa Fe Al]

Hi Mariss & Riceburner,

My apologies for entering this conversation, but I am confused. I'm using a regulated Power supply with an output of 28 VDC @ 12 Amps maximum. My drivers are G251s and the three motors are Nema 23 @ 425 oz from Keling Inc.

I looked at the Gecko sight about Stepper basics, looked at CNC4PC's website and a few other places. After reading a few threads, I am totally lost.

It seems many people don't seem to worry about this situation. I, on the other hand tend to worry about everything I know little about when it might affect me in an unfavorable way.

My problem is that I am confused as to how to hook up the components, and where, to prevent blowing up my PS or drivers when the steppers wind down.

I guess I should say that I need a diagram showing me where and how, referencing polarity, to hook up the components to prevent a disaster such as you guys have mentioned. I know to hook up positive to positive, but diode orientation eludes me in this case.

It's been a long time since I played with electronics and am doing much studying trying to catch up but, alas, my brain is old and much time is needed to absorb all this data. Pictures are the easiest for me. Words tend to confuse me for quite some time until I finally get a handle on it.

Therefore, a request for a visual hookup diagram.

Thanks for any help - info - you can provide.

Al

[Riceburner98]

Active clamp, got ya..

Al - I doubt most (if any) milling machines would ever run into this problem, there just isn't that much rotational mass (only the lead screw / motor knobs if any) to worry about. As Mariss stated, the energy that's killing our (my) drivers is being stored in the spinning 1/4" disk at 1,500-3,000 RPM, then being dumped back through the system when it spins down. On the milling machine, the motors accel and decel very fast since you don't have that mass to "spin up" or "spin down". Maybe if you had some big dampers on the ends of the motors to absorb resonance, but still doubt that's as much energy storage as our big spinning disks. (and they're building some new tables with 2' diameter disks! Hope they're planning on servos for those..)

I had my new CNC slide tables running at 600 inches / minute with no worries of this type of thing, so you'll likely be fine without worrying about it. Anyone else feel free to chime in of course!

- Bob

[Santa Fe Al]

Originally Posted by Riceburner98 Active clamp, got ya.. Al - I doubt most (if any) milling machines would ever run into this problem, there just isn't that much rotational mass (only the lead screw / motor knobs if any) to worry about. As Mariss stated, the energy that's killing our (my) drivers is being stored in the spinning 1/4" disk at 1,500-3,000 RPM, then being dumped back through the system when it spins down. On the milling machine, the motors accel and decel very fast since you don't have that mass to "spin up" or "spin down". Maybe if you had some big dampers on the ends of the motors to absorb resonance, but still doubt that's as much energy storage as our big spinning disks. (and they're building some new tables with 2' diameter disks! Hope they're planning on servos for those..) I had my new CNC slide tables running at 600 inches / minute with no worries of this type of thing, so you'll likely be fine without worrying about it. Anyone else feel free to chime in of course! - Bob

Bob,

Thanks for relieving some of my worries. I really appreciate that.

Al

[neilw20]

I believe I downloaded this file direct from Gecko site.
You may need to upgrade it to absorb more power depending on energy.

[Riceburner98]

Well there ya go now! That explains why the Centent / Gecko just shut down the power supply and the IM483 blew up. According to the documentation, the Gecko dumps the excess current back to the supply (as theorized before), and the IM483 absorbs it. (I checked one of the working rotary tables with a meter, voltage @ the supply doesn't go up more than .01v on decel) The first new "high speed" table is up and running today, at almost 1500 RPM. It's using newer IMS drivers, hopefully they're like the Gecko's and dump the excess back to the supply side, and maybe have the clamp circuit built in. That'd be nice. I'll have to monitor the supply while the motor ramps up as well, I have a feeling there's going to be quite a bit of back-fed voltage even on ramp-up as the unused coils are de-energizing. Would explain why some of the drivers can't get up to speed as well..... Interesting stuff, thanks for all the info guys!

[neilw20]

Dumping the energy back to the supply is what must happen, (because where else could it go unless it has an inbuilt clamp).
Because most supplies (other than batteries) are fed from a rectifier, the charge goes back on to the filter capacitor (or battery is OK) and pumps up the supply. The circuit in my previous post is very easy to breadboard, in 3D birdsnest mode. Put a LED with a 1K in series across the dump resistor and you will see the problem.

[Riceburner98]

The IMS drivers are using 2 of these H-bridge driver chips to do the dirty work: http://www.st.com/stonline/products/...1373/l6201.pdf There appears to be a bank of diodes (or TVS's? -- marked EDG6 or similar) connected to the coil outputs as well, 2 per pin. (one -> GND, one -> 48v) I have some TIP125's on hand that I can try the circuit with, which will likely work for the Centent / Gecko's, but since I didn't see a voltage rise at the PS during spin-down with the IMS drivers I'm guessing it's not going to work for those. Was going to try and get some spare chips, but they're $15 each and the drives only go for $20 used. (actually, Mouser / Jameco has the chips for $10, almost worth it...)