DC electronic load polarity

[lukier]

Hi,

I have a question on these cheap DC electronic loads like BK Precision 8500, Maynuo M9712, RK8511 and so on.

Can they operate with reverse polarity (2 quadrants)? In BK Precision datasheet they only show V+/I+ quadrant power curve.

I build various motor control drivers, but so far I've used a lightbulb/resistor for test load or the actual motor forcibly mechanically stalled (not very nice for the motor). I wonder if I could use any of these DC loads to test the performance and limits of H-bridge power stages more accurately. I suppose the DC load behaves like pure resistance, so no inductive spikes or back EMF effects of the actual motor, but should be better than just a lightbulb.

[fcb]

My understanding that the loads you mentioned are unipolar (single quadrant) sinks.

You could but a bridge-rectifier on the front end if you can cope with the Vfwd voltage drop (~1.2V).

[Simon]

Then the probably can't, as far as i know your load has a circuit that dumps heat at various wattage's in asemiconductors. Now if you use mosfets in theory you can put current through either way but the control circuit won't have any negative sensing so either nothing will happen or in order to try and sink the required load it will go to maximum.

Sounds like you need a load specifically for this kind of work or you use two, one for each polarity maybe with a diode but that could get messy.

[Kjelt]

I have a question on these cheap DC electronic loads like BK Precision 8500, Maynuo M9712, RK8511 and so on.
Can they operate with reverse polarity (2 quadrants)? In BK Precision datasheet they only show V+/I+ quadrant power curve.

Hi I recently bought a M9812 and it can not, it gives a warning if the polarity is reversed.

[lukier]

Thanks for your comments. I don't know if it is feasible to build a 2 quadrant load myself  As Simon said, it might be tricky to control the MOSFETs in this scenario. Maybe some isolated/floating/level shifted driver setup (like those used sometimes for the high side N-MOSFET/IGBT) would do the job, need to think about it.

It's a pity, because these 150/300W DC electronic loads are good value for money from what I read here and seen in various review videos.

[Simon]

basically you need 2 loads one for negative and one for positive, infact if you rectify and smooth the feedback voltage you should be able to get the two to work in harmony, I think diodes for each power branch would be a minimum to ensure no wrong conductions. I thiunk what we are also looking at here is that you need a constant power resistor load that can vary. What most of us assume when someone says dummy load is a constant current sink.

[lukier]

basically you need 2 loads one for negative and one for positive, infact if you rectify and smooth the feedback voltage you should be able to get the two to work in harmony, I think diodes for each power branch would be a minimum to ensure no wrong conductions. I thiunk what we are also looking at here is that you need a constant power resistor load that can vary. What most of us assume when someone says dummy load is a constant current sink.

Interesting idea, although doubles the price. Constant power is useful for things like overall performance of the H-bridge and verifying thermal dissipation, but maybe I will have to build something myself.

I'm starting to think that I would be better with something custom, more like a motor simulator, maybe with more motor-like effects (charging slope of the inductance first - BK Precision simulates that, then spike when switched, back EMF when simulating no-load spinning etc). I suspect this might be very difficult and it is a very niche market. People usually go the other way around with some motor with a driver connected to MATLAB/Simulink where they tune control loops or characterize the motor.

I often use ACS712/ACS758 current sensors, mostly for just over-current protection and to the ADC of a MCU for monitoring. Now I'm thinking about doing a constant current (therefore torque) H-bridge drive purely in the analog domain (driven from a DAC, with outer control loops on speed/position done in the MCU). I don't know if this setup with 2 DC loads will work well for tuning such circuit.

[fcb]

If you use two seperate loads (as someone suggested) and then use diodes to link them - you might as well just use a bridge rectifier on the front end of a single load..

[zapta]

If you use two seperate loads (as someone suggested) and then use diodes to link them - you might as well just use a bridge rectifier on the front end of a single load..

And make sure not to have ground loops.

Alternatively, since you want to test H bridges, can you test one side a time?

[NiHaoMike]

http://www.linear.com/product/LT4320
One way to rectify the input without the diode losses. If only there was an externally biased version that would work down to near 0...

[tom66]

The vast majority of electronic loads operate using MOSFETs, which have integral body diodes built into them. They will conduct in reverse regardless of load set point. I have accidentally reversed the polarity on a load unit and it essentially acts like a diode. And it beeps at me, flashing "REVERSE POLARITY" on the LCD... I am guessing that under significant load current, damage is definitely possible.

[lukier]

LT4320 looks like an interesting chip, I didn't know about it.

Testing one side at a time would be effectively just testing one MOSFET (first low side, then high side). In a full bridge motor control application the most important effects happen when the direction is changed. There might be a spike, then rapid reloading of the inductance or using synchronous rectification. In all these cases there might happen to be a short shoot through that should be avoided and so on (hence all this testing I'm thinking about).

Correct me if I'm wrong, I suppose it would be possible to build such a motor driver testing load, it would need to employ H-bridge as well and operate in the opposite manner to the driver under test, at an order of magnitude higher frequencies (response) I think?

Aren't SMUs using full bridge setup for their 4 quadrant sink-source operations?