Question Preventing motor drivers from burning up from reverse curreng (regenerative breaking?)

I'm scratching my head over this problem. I'm building a 6DOF with 24V 250W wheelchair motors and Cytron MD13S motor drivers. I assume those motors can generate reverse current when they break/reverse direction. I read about people protecting their power supplies from this either via a battery or via KBPC5010 Bridge rectifier diodes. I got those rectifiers, but obviously, I cannot connect them between the driver and the motor, only between the PSU and the drivers.

Each PSU will supply power to 3 drivers / 3 motors. The MD13S has "regenerative breaking" in its features, so I guess it can handle the reverse current, but the question is what happens next? The rectifier upstream won't let it through to the PSU, so will the driver burn up?

The power is supplied from the PSU to 3 drivers, so if one of them lest through the reverse current, but the other two are connected to the same wires, what will happen? Should I put 3 rectifiers so the reverse current can't flow to the other 2 drivers?

Here is the schematics of my 6DOF build. What should I do in order to protect both the PSUs and drivers?

I already had a debate about this topic here. Nobody seems to care so so I think the returned energy is not a big problem in most cases. 2DOF seats don't move vertically and worm gear motors have a low efficiency and are self locking. So there is also no back feed power to be expected. In your case however, I think this could be different. I might be wrong. In this case no protection is necessary and all discussion about it is futile. But if not it' going to get expensive so it's worth to take some precautions.

It would be good to know how the MD13S reacts when doing regenerative braking. The possibilities are

1. It does not monitor the supply voltage at all
2. It has overvoltage protection and shuts down the motor current if the threshold is exceeded
3. It has active protection like a bleed resistor or brake chopper

Case 3 can be ruled out since it would require extra components and/or terminals. I don't see any on the board and it also costs money which the manufacturer most likely avoided. Case 2 at least avoids damage but would mean the torque is also cut off. In the best case this leads to a momentary false motion cue or in the worst case the driver shuts down completely. In case 1 the regenerative energy charges the small caps on the board until they blow up (50V). The driver H-bridge acts as step-up converter so the DC bus voltage can go higher than the motor voltage.

But speculating about the behaviour doesn't help much. You could ask the manufacturer. But in any case, if we need BEMF protection it's going to get more complicated, anyway. One or two diodes don't help much. It would isolate the problem and avoid one driver blowing up all others and the PSU.

Before overcomplicating the circuit let's first find out if we actually need external protection.
The easiest way to do this would be running the drivers from two car battieries in series (24V). They can handle large currents in both directions so you don't need to worry. To be super-safe you can put 25A fuses between the battery and each driver.

Then run the motors full speed up and down with the seat loaded. Measure the current at the battery. You need a clamp amp-meter for this. If the current becomes negative at any time then you now you have to add some protection.

Sounds like a plan. But don't have a clamp amp-meter, and my multimeter can only handle 10 Amps.
I do have 3 of these (and 3 more are coming):


My idea is to put it inline for each motor. The specs say "DC 0-100V 10A 1000W" so I assume 10A @100V means it should handle 41A @24V. The way it's wired is this:


I can see the images in the editor, but I don't think they are loading in the topic, so here's the page where you can see the images, specs and wiring diagram for Deek Robot BL-02

I got them at $3.50 a piece so I thought why not. It will give me usable info about the load on the motors and voltage changes etc. and will look nice. I don't think it can display negative values but I'm not sure. But I guess that if the current reverses, the voltage will fluctuate in some way, and if there is an overvoltage that we are discussing, it would jump up, right? If the current is reverse, would polarity swap? If so, the voltage would drop to 0 or at least drop sharply, and then I will know that is reversed. Or would the reverse current have the same polarity but just in reverse direction?

Cytron support has responded quickly and stated that MD13S does NOT have overvoltage protection. I does have overcurrent protection though. They suggested considering this shunt regulator if I want protection of the driver and PSU. I don't see people using it though, I saw one topic when it was discussed but it doesn't appear to be widely used.

The people describing these rectifiers say they heat up a lot, they basically converting reverse current to heat, so that could protect the drivers, hopefully. I do have 8 rectifiers (they are cheap - $3.66 for 5) so I think I will put one in front of each driver instead of just each PSU. That way they also will have less load so can convert it to heat more easily if needed. I guess they won't hurt, they just cause negligible efficiency loss if any. They was people described the problem at that topic is that PSU would shut down due to reverse current protection, and using these rectifiers eliminated the problem. Nothing really burned up, drivers were fine. But every rig is different of course, I wouldn't want to burn my drivers if I can avoid it...

For more safety and better heat dissipation capacity I think I'll use 6 of the KBPC5010 Rectifiers, having one before each driver, so the back current from the motor on one driver won't be able to get to other drivers. Hoefully if any occurs it will be converted to heat by the rectifier. Here's an updated schematics:

(I'm also cross-posting it to my main thread to keep it current)

Should I test with 2 batteries before the rectifiers are connected? I guess there wouldn't be any reverse current on the battery if the rectifiers are present.

The rectifiers are like check valves. The let the current through in one direction and completely block it in the other. This prevents damage to the suppy and to other drivers. But it doesn't dissipate the energy. If the diodes get hot then it's due to switching losses from the high frequency of the PWM.

For the test with the car batteries you don't need diodes. They would in fact stop the test from working at all. The battery can take reverse current.

The shunt regulators are the ultimate cure. But you'd need 6 of them as a single one can only handle ~10A peak and 15W continous. But I think car batteries are less expensive than 6 shunt regulators. Used ones are OK. They don't need to supply current for an engine cold start. But they must not be damaged. Connect one of your motors. If it runs and the voltage stays above 12.0V then it's ok.

It would be nice if the shunt regulators had an LED that flashes when the regulator actually draws power. We could then try if they are actually required.

BTW, I don't see most of your pictures. The schematic is visible but the other pictures are only displayed as "X". How did you upload them? I drag pictures over the message editor and then select "insert thumbnail".

So let me get it straight: I connect the battery and monitor motor voltage with Deek Robot BL-02 (or a multimeter) and it shouldn't dip below 24V (you said 12v but I think you meant 24V) . If it does, it means there is some reverse current going on, right?
I need to winterize my sailboat so I'll be taking out 2 good marine batteries, which I can use to power a motor for the test. I also have a couple of old car batteries, but I think they don't even hold 12V anymore, so they aren't going to be usable, unless just used in parallel with the PSU.
BTW not sure why those images didn't load, I did the same as I always do. But the link has the details.

In DOFReality H6 review the internals of the control box is shown. There are 2 identical boxes. each has 3 24V PSUs, Arduino with a dev shield and 3 of the same MD13S drivers I'm using (that's where I got the idea from), with added heatsinks. Note that beside the relay for the emergency stop button there is nothing else there. The top 3 things are the backs of the AC plugs. So they use no capacitors, no rectifiers, no battery and it's a commercially sold and popular motion system. They do have different motors than I'll be using, so there's that. But they are not less powerful, most likely more.

No I meant 12V. This is to test if the battery is still usable. Old batteries can have contaminated electrodes or a too low electrolyte (acid) level. This causes the internal resistance to rise which means if you charge them the voltage rises quickly but the current doesn't. This would render them useless as voltage limiter during reverse current conditions.

To test the resistance put a load to the battery (one of the 24V motors is perfectly fine for a single 12V battery). If the voltage drops below 12V is no good anymore.

OMG! What a complicated piece of gear! Prop shafts with U-joints and ball screws to convert rotational motion to linear and then back to rotational and with the lever arm back to linear, again! I wonder how they can make this for an affordable price.Why not a toothwheel gear and drive the arms directly?

Maybe their overcomplicated mechanics has enough friction so regenerative energy is not a problem. Maybe I'm just overly cautious and all this is not necessary.

Ok, I have an idea how to test if protection and/or batteries are really necessary. No need to buy expensive stuff just to find out that we eventually don't need it...

All you need is the rectifiers you already have, two car or marine batteries and an LED and a resistor.

The purpose of this circuit is to generate a small voltage difference (4 diodes in series = ~2.4V) when current flows back to the supply (batteries). The LED indicates if/when this happens. The third bridge rectifier BR3 allows current to flow in the normal direction (battery to driver). Use a bright red or yellow LED. green or white LEDs need to much voltage.

So the test goes like this: Connect 5 Motor drivers directly to the batteries (with a fuse if you want). Connect the 6th to the batteries with the circuit shown above. Run all motors simultanously with max speed up and down. If you see the LED flashing then there is reverse current. If not we don't need to worry.

What a great idea! I'll definitely do that. I have 2 relatively new marine batteries - they are good for sure. Just started the outboard yesterday to drain all gas from the carb before storing it. Is there a reason to connect all motors? The rig is not ready yet, so I can connect a single motor with no load on it. Or I can add a short arm to it and attach some weight to it to create inertia when changing direction. I can just connect a single motor and driver with your circuit and feed it the square wave in SMC3Tools. There won't be a heavy load on the motor, but I guess it still would generate reverse current when reversing direction, if it's prone to that.

Not entirely sure about KBPC connections: KBPC has 4 pins marked + - ~ ~ so the pair is for AC current and a -+ pair. Recommended connection for protecting the PSU is to both ~ from PCU and then + to the driver. From the image below I assume that top 2 rectifiers are connected PSU + to BR1 + , BR1 - to BR2 + , BR2- to driver +. And PSU + to BR3 both ~ inputs, then BR3+ to LED and BR2-. Am I correct?

Or, in simpler words, in your diagram on rectifiers 1 = +, 2 = ~, 3 = ~, 4 = -, correct?

The original model used a geared motor, but suffered from backlash. This was mitigated somewhat by the use of gas struts and dampeners which were added later as part of a redesign. The current design is meant to be more stable still though I haven't any experience with it personally. I thought it was quite a clever little design to get the benefits of a ball joint for precision but without needing a long actuator, but I don't know how effective it is in reality.

I do enjoy the innovation of new designs and ideas. I love this one in particular, though it looks very complicated to control:

https://www.xsimulator.net/community/posts/249057/

Simply put a 12v battery or a pair of 12v batteries across the power supply

they can provide hundreds of amps and will absorb any regenerative voltages like a capacitor

your power supply only has to keep them charged

My system is 24V and I'm using 2 pairs of PSUs, so I think I'd need 4 12V batteries for that setup, which is a bit extreme, and I don't have that many anyway. I have 1 double UPS battery - it's old and can't handle UPS loads but I think it's still alive. But it's only one, and I'd need to of those. Besides, I don't know if it's large enough to handle the spikes...

Yes, you can also test it with a single motor but the load is important. Most of the returned energy will come from the rig weight moving up and down, not the rotational energy of the rotor spinning. So you should at least mount an arm on the gear output shaft and hang a bucket of sand on it as dummy load.

Yes, exactly.

Yes, this is the easiest way to protect the power supplies. But we like to find out if it's necessary at all. The majority of motion rigs seem to get away without protection. With commercial AC servos it's easy. You can try out without risking damage to the drives. They fault out with an overvoltage error but don't blow up. So you can add a brake resistor if (and only if) necessary. The cheap DC drives don't have built in protection. So the "try out" experuiment might end destructively.

Ah, interesting. Without gas springs backlash should not be a problem as the weight usually bias loads the gear so it's always on the positive force side of the backlash. Only in extreme situations (crashes, loopings) you have negative g forces and then you don't care much about a little bit of backlash.

But with the gas springs the forces of the actuators chang direction quite regularly so backlash becomes an issue. Maybe the gas springs should be adjusted to not support all of the weight but only half to three quarters. This way the gear is kept biased a bit in most cases.

Yes, looks interesting. What are the benefits? Bigger range with shorter struts and unlimited yaw? But the drawback is the motors mounted below the curved racks take a lot of space, vertically. The curved rails and teeth racks also look expensive. And you also need low-backlash gears.

There are videos on YouTube about DOFReality H6 having not only backlash but large shuddering during slow movements. Struts were helping with that as well. I guess that's why they changed the design to that complicated arrangement. I'll see if my rig would have similar problems or not... I got 2 struts 120lb rating each. That means they should hold 109kg which is definitely less than me (80kg) + top platform weight with seat and stuff (?) but not by that much (20-30kg I'd guess). I'm a bit worried that they are too stiff. I can even compress them by hand at all. But we'll see...

the shuddering comes probbebly from the wormgearboxes becauce they are sliding they can make little jumps turning slow. Youre gearbox is a rolling one and wont have this problem
Regards Ad