Paralleled battery chargers and instability?

[Flyback]

Hello,
If four identical 1.75kw Full bridge SMPS's are each supplying 5 Amps into a single 350V battery, (each Full bridge SMPS is output current regulated), then is there any chance that there feedback loops will interact with each other, and that they could go unstable?

 

[Diver300]

Yes, they could become unstable. They wouldn't load share, so you would tend to get the power supplies taking different loads.

Power supplies of that size can usually be linked, with one as a master and the others as slaves. The master would decide the voltage/current, and the slaves would each generate the same current as the master.

 

[Tony Stewart]

I have seen many current sharing SMPS be unstable with less than 10% dummy preload even after matching Vout. The loop gain margin has a lot to do with this.
Separate passive current sensing can be more more stable at the expense of load regulation and inefficiency at max load. Low RdsOn linear current sensing can achieve better efficiency
and stability with automatic master/slave architecture.

 

[Flyback]

Each SMPS will be programmed to deliver 5 amps, so they will surely do that?....each smps wont know or care whatever the other current source smps's are doing, surely?...its just like stacked voltage output smps's in series, they never go unstable.

That is, stacked voltage output regulated smps'S never go unstable, the same goes for parallel current output regulated smps's.

Ive worked in loads of places where they wanted say a 96V power supply , so they just stacked two 48v isolated psu's...they never went unstable....so why should parallel current source smps's go unstable?...its an analogous situation

 

[AnalogKid]

No, it isn't.

At it's core a switching power supply is a constant voltage, variable current device. Stacking constant voltages can work (batteries are stacked all the time) but also certainly can lead to system instabilities and massive reverse-polarity fault currents; you guys were lucky. But combining variable-current supplies, which is what you have in your parallel application, is nothing like combining constant voltage supplies.

ak

 

[Flyback]

At it's core a switching power supply is a constant voltage, variable current device

thanks, but not ours...ours regulate their output current...they do not regulate their output voltage

 

[Diver300]

While the power supplies are in constant-current mode, you should be OK. I would guess that your battery can absorb 7 kW, or you wouldn't be going to the expense of several power supplies. There will be slight differences between the currents from the different power supplies, but that doesn't matter. The situation should be stable even if the currents were very different. Each power supply is a high impedance current source feeding into a low impedance load.

There will be a potential problem when the battery is fully charged, which I assume it will happen at some point. The cut-off voltages of the power supplies will differ slightly, so they will all start to cut back the current differently. As each power supply starts to shut off, its output impedance falls. The situation is quite complicated, and modelling it is beyond my abilities, with four power supply impedances all falling from near infinite to near zero, all possibly at different times. The battery impedance will also have to be considered, and that will get larger as the battery ages.

You should also consider what happens if the battery becomes disconnected.

I'm not saying the set-up won't work, it's just difficult to be sure. Even if the power supply loads are unstable, which is far more likely when the battery gets charged, it may not matter. Certainly there are plenty of smaller battery chargers, and this includes car alternators, where the voltage control as the battery gets fully charged consists of turning the charging on and off quite quickly. (On car alternators it's damped by the field winding inductance so the alternator current changes little). I don't know what the effect of one or more 1.75 kW chargers turning on and off rapidly would be. Instability could be just a small variation in current, and it may not cause any adverse effects at all.

 

[tomizett]

My own feeling (based on virtually no facts at all) is also that the problem is likely to be harder as the battery approaches full charge and you start to want to switch to constant-voltage charge.
In this situation, it may be preferable to switch individual chargers completely off in sequence, so that - eventually - only one is left to do a constant-voltage float.

While I'm sure that ideal current sources in parallel should be perfectly fine, I'll yeald to others' opinions that SMPS control loops may not behave well in practice.
My instinct would be that parallel capacatance accross the load would help encorage stability, just as you'd use a series inductance to isolate a voltage source from its load.