Voltage limiting in Hybrid traction system

[JinPA]

In an AC traction drive system braking the vehicle can be done and much of the energy recovered using the 'regen' feature. This lets the traction amplifier feed dc power back to the battery.
In an application I am looking at, we are using batteries that are smaller than usual in combination with a fuel cell to power the vehicle. The batteries accept the current, but the voltage rises above the 45v limit that caused 'faults' in the traction amplifier.

The system's nominal voltage is 40v and the energy involved in the regen can be 20,000Joules, over about 2 seconds. Peak currents can be 450amps.

A voltage 'clamp' has been proposed to prevent the faults and I need to understand some of the ways this is done.

Any insight you can provide this mechanical engineer would be welcomed!
Thanks.

 

[Styx]

one way if you cannot regen that energy fast enough is to use a brake circuit (basically a resistor and FET/IGBT in series across the DC-link), this basically dissipates the extra energy in the brake resistor

wasteful, but sometimes the only way

 

[Sceadwian]

A Zener diode is often used to clamp. If shorted to ground in reverse it will stay non-conductive until the voltage reaches the zener voltage and then it begins to conduct, in this case to ground, so it'd effectivly bleed off the over voltage. Not sure if they make zeners ment to shunt currents that high though.

 

[Styx]

show me a Zener that can take an avanlanche of 450A then...
the dirty little "hacks" that can be used in 12V hobby systems cannot be used in real power systems

 

[Nigel Goodwin]

show me a Zener that can take an avanlanche of 450A then...
the dirty little "hacks" that can be used in 12V hobby systems cannot be used in real power systems

Most things are scalable though, in the case of a zener like this simply make it an amplified zener - using the zener to turn on transistors or FET's which switch a low resistance, high wattage, load across the motor output.

 

[Styx]

but then the Zener is being used as a voltage-sense to then switch the load (and not as a clamp as mentioned).
Problem with that is Zener's are slow and a break cct need to be engaged pretty fast (esp with these kind's of re-gen currents mentioned here - will jack the DC-link cap up fast!!!)

a fast (lm319) comparator coupled via suitable drive to a FET to switch in the resistor (with a free-wheel diode across the resistor) will chop the link and control the DC-level to below a safe thresh-hold

On the drives I work on with a nominal link of 270Vdc we usually turn-ON at 450V and turn-OFF at 370V

 

[Roff]

but then the Zener is being used as a voltage-sense to then switch the load (and not as a clamp as mentioned).
Problem with that is Zener's are slow and a break cct need to be engaged pretty fast (esp with these kind's of re-gen currents mentioned here - will jack the DC-link cap up fast!!!)

a fast (lm319) comparator coupled via suitable drive to a FET to switch in the resistor (with a free-wheel diode across the resistor) will chop the link and control the DC-level to below a safe thresh-hold

On the drives I work on with a nominal link of 270Vdc we usually turn-ON at 450V and turn-OFF at 370V

Styx, I have trouble believing that a series zener (as is used in an amplified zener) is slower than an LM319. Can you point to some documentation?

 

[JinPA]

Voltage limiting follow-up

Thanks for the info. Styx's suggestion is pretty much what we will try - An IGBT based circuit will be used to dump power through an immersion heater with the amount controlled by PWM. We wanted to use an ultracap, and plan to do so in future designs, but there was just not enough room.
As to the waste, it looks like we will dump about half of the power we might otherwise have recovered.
Thanks again.

 

[Styx]

Styx, I have trouble believing that a series zener (as is used in an amplified zener) is slower than an LM319. Can you point to some documentation?

just my experience... I did try using a Zener as a threshold sensor for 28V motor-drive system but it took longer (ie the DC-link crept up to higher potential then we wanted it to chop at) and it is un-controlled

comparator, Gatedrive, IGBT,Dump resistor and freewheel works, has worked and continues to work as expected.

Not saying a Zener can't be used but when do deal with real amps (ie a couple of hunded for starters not a couple on one's) its not as simple

 

[Roff]

just my experience... I did try using a Zener as a threshold sensor for 28V motor-drive system but it took longer (ie the DC-link crept up to higher potential then we wanted it to chop at) and it is un-controlled

comparator, Gatedrive, IGBT,Dump resistor and freewheel works, has worked and continues to work as expected.

Not saying a Zener can't be used but when do deal with real amps (ie a couple of hunded for starters not a couple on one's) its not as simple

Yeah, I don't have any experience with the real issue here. I was just addressing the question of the speed of zeners.

 

[Cabwood]

How much over the 45V limit did you experience? 46V? 60V?

This information would answer the question: how much energy the cells are absorbing to recharge with, and how much is left over that needs to be dumped. Then we can figure out how sturdy to build the clamp.

In addition, I am assuming that at 450A the voltage at the battery terminals rises above 45V. If this is the case, and you can also tell us what the battery current is at, say, 40V, this will help to determine what currents to expect through the clamp.