Dear All,

I have a linear robot that moves along a single axis on a rail. This is driven by PMDC motors (100W/24V) on either side of the robot. Please see this video to under the type of product I am working on.

Problem: some times when the robot is climbing up an incline and for any reason slides back if the incline is too steep, the motors are acting as a generator and they fry the H bridge, Gate drive and also the rest of the circuitry.

Following is the H Bridge and Gate drive circuit. How can I protect against this?

 

What might be happening is the motor as generator is driving the DC buss up too high in voltage. Since the motor is the main load when being driven when the motor acts as generator there is no appreciable load to take the energy so the DC buss rises. The level to which it rises is hard to predict so you can try to measure it carefully allowing the device to slide back slowly.

You might have to apply a load to the DC buss or use some braking action with the transistors when the device slides back.

 

If the motor is reverse rotating direction on the slide-back, then the generated voltage is opposing and is the reverse polarity fed from the bridge.
The result is very high current. And at worse is limited only by the low resistance of the armature.
Some way has to be carried out to prevent this reverse generation.
Max.

 

You might have to apply a load to the DC buss or use some braking action with the transistors when the device slides back.

Mr.Al can you give me some reference on how this can be done?

 

One method often used is a very low current sense resistor in the bridge circuit that is then fed to the processor via an op amp in order to advise the proc to shut down at a given current level.
Max.

 

If the motor is reverse rotating direction on the slide-back, then the generated voltage is opposing and is the reverse polarity fed from the bridge.
The result is very high current. And at worse is limited only by the low resistance of the armature.
Some way has to be carried out to prevent this reverse generation.
Max.

The reverse transistor diodes cause any motor current, regardless of polarity, to be full wave rectified and applied to the DC buss as if an AC line input was full wave rectified and applied to the DC buss. So the only danger is a DC buss that goes too high such that the parts can not handle it.

 

Mr.Al can you give me some reference on how this can be done?

Well the main idea is you would apply some resistive load to the DC buss during times when the device is generating power, or you can look up some braking action methods done with H bridges.
For a braking action, sometimes just two lower transistors are turned on or two lower then two upper in order to distribute power between all the transistors/diodes.

 

There are three possible conditions when they slide back, are the wheels still rotating in the right direction? Are they stalled? or are they being driven in the opposite direction?
Max.

 

I think you need another pair of diodes, like D9 and D12, but going to the +V supply.
And then maybe a power zener clamp across the supply to limit the voltage.

 

There are three possible conditions when they slide back, are the wheels still rotating in the right direction? Are they stalled? or are they being driven in the opposite direction?
Max.

Oh, well i wouldnt think the system would be damaged if the wheels were still turning in the right direction but i guess anything is possible.

 

I think you need another pair of diodes, like D9 and D12, but going to the +V supply.
And then maybe a power zener clamp across the supply to limit the voltage.

Yes perhaps a shunt regulator that could handle high current high power.
Many power supplies are active high but not active low.

 

I think you need another pair of diodes, like D9 and D12, but going to the +V supply.
And then maybe a power zener clamp across the supply to limit the voltage.

Wont the body diodes of the FETs do this job? I assumed they would.

 

There are three possible conditions when they slide back, are the wheels still rotating in the right direction? Are they stalled? or are they being driven in the opposite direction?
Max.

They are being driven in the opposite direction.

 

Then the generated voltage is of an opposite polarity to the applied voltage polarity.
Max.

 

Then the generated voltage is of an opposite polarity to the applied voltage polarity.
Max.

And the diodes conduct with opposite polarity current right?
Or did he not add reverse diodes if the transistors did not have them?
If we remove all the transistors we are left with a full wave diode bridge with the motor connected to the AC input and the DC buss connected to the positive output (the cathodes) and negative output to ground (grounded anodes).

Or did you mean that the motor voltage reverses while the transistor bridge is still trying to power it with a normal regular polarity signal.

 

If the motor is reverse rotating direction on the slide-back, then the generated voltage is opposing and is the reverse polarity fed from the bridge.
The result is very high current. And at worse is limited only by the low resistance of the armature.

Agreed. So it needs a current limiter added to keep the current within the capabilities of the FETs?

 

Agreed. So it needs a current limiter added to keep the current within the capabilities of the FETs?

If in fact it does need that it also needs the power supply shunt regulator to keep the DC buss voltage within limits unless something else is done.

 

Or did you mean that the motor voltage reverses while the transistor bridge is still trying to power it with a normal regular polarity signal.

Yes, this was my implication.
Max,

 

Yes, this was my implication.
Max,

Ok then i guess the next step would be to determine if the motor ever acts as a generator or not.

Assuming there are wheels, if the normal current/voltage is still applied when the motor is going up hill and then suddenly the wheels start to slip and it actually *slides* backwards then there would be no reverse current.
However, if the wheels have enough circumferencial friction then there would be reverse current/voltage. The current would then depend partly on the speed at which the motor 'rolls' backwards.

So perhaps some more measurements are in order.

This is interesting because for many applications the motor is modeled as a load only. Maybe we could take a look at how elevators in buildings deal with this.

 

How can I protect against this?

Try this circuit:


EDIT:

Problem: some times when the robot is climbing up an incline and for any reason slides back if the incline is too steep, the motors are acting as a generator and they fry the H bridge, Gate drive and also the rest of the circuitry.

So, TS's problem is completely solved, H bridge, Gate drive and also the rest of the circuitry will not be fried.
Instead resistor BRAKE_LOAD may be connected solenoid coil, for solving some mechanical problem, in this situation.