# Supercapacitor bank charging process breaks down dc/dc converter

I'm an undergraduate student and I'm member of university's team, which deals with the development of an electric vehicle. One of our projects, is the implementation of a complete regenerative braking system. We have almost achieved that but there is a problem with our charging system. To be more specific let me tell you some basics about our system:

1) during decelaration of the vehicle, the 3-phase electric motor functions as generator and transforms kinetic into electric energy,

2) the ac of the motor (generator) passes through an ac/dc inverter that produces a dc output voltage with range 0 to 34 Volts,

3) that dc passes through a dc/dc converter (buck converter) which produces a fixed output of 15 Volts,

4) that dc/dc converter (buck converter) is used to charge a supercapacitor bank (156 Farads) with rated voltage 15 Volts.

You can see a graphical depiction of the topology in the following image. The green arrow shows the direction of the dc current that flows from the motor to supercapacitors via the ac/dc inverter and the dc/dc converter during charging process.

During laboratory tests, we increase gradualy the rpms of the motor (by using another motor) and then the charging system smoothly starts to charge the supercapacitor bank.

Unfortunately when we took our vehicle for a ride to test the regenerative braking system on a real deccelaration, the dc/dc converter (see the orange box in attached pdf) burned.

The difference, in relation to laboratory tests, is that during the real deccelaration, the vehicle is moving with a high speed and suddenly the charging switch (see attached pdf) turns on for the charging process. We think that at those moments (while the charging switch is turning on), instantly some dc current spikes are produced with result the disaster of the dc/dc converter.

We think that a solution to that could be the placement of a coil either at the input of dc/dc converter or at the output, so as to eliminate the spikes. Is our thought right or something else happens?

-
you should use a supercapacitor charging IC to charge supercapacitors. Driving your DC/DC converter directly into a cap bank is most likely causing overcurrent issues. – scld Jan 18 at 16:22
Perhaps you should consider some kind of 'soft-start' which would make the DC-DC converter gradually ramp its output current up from zero when your charging switch is closed. In your real-world case you have a lot of energy available from your motor/generator and a big empty hole to put it into (your supercap bank). This sudden rush is probably what's causing trouble for your DC-DC and if you cause it to increase in a controlled way it'll probably thank you... – brhans Jan 18 at 16:27
In practice you'll know its safe when your DC-DC converter can survive a load-dump from the input with a dead-short on its output - it must be able to current-limit to protect itself. – brhans Jan 18 at 16:30
@scld thanks for your response. Have you something in mind when you say IC? – giannis.t Jan 18 at 19:53
@immibis I suspect its more complex than just a rectifier given the fixed output V -- presumably a rectifier followed by a switch-mode convertor, and assembled into a module. – Chris H Jan 19 at 8:56

The problem is the fixed 15v output voltage of your DC-DC converter. Capacitors should be charged with a current output, not a voltage output.

If the voltage on a capacitor is less than the output voltage of a power supply feeding it, it will take as much current as the supply cares to supply, in other words it behaves like a dead short.

The DC-DC converter can still be buck, can still use the same power stage, but the controller should be changed to control the maximum current output to safe levels, as well as the maximum voltage to be safe for the supercaps.

With regenerative braking, ideally you would program the converter to put a braking torque on the wheels proportional to the position of the brake lever/pedal. This is naturally what you would get if you try to mimic the response of how mechanical brakes would behave, I presume you want the electric system to behave as a driver would expect it to behave.

-
Thank you @user44635 for your answer. You are totally right about the current source. We used a voltage source in order to build something simple. For that reason we use supercaps pre-charged at 12V. After our lab tests we figured out that our dc/dc converter (buck converter) is able to handle the maximum continuous current that the motor produces in 2000rpms. That's why we thought about current spikes and not continuous current. So what you think about this? – giannis.t Jan 18 at 19:44
Shouldn’t the current be highest when you short-circuit the generator? – Michael Jan 18 at 20:08
I am not sure your test shows that the "dc/dc converter is able to handle the maximum continuous current that the motor produces at 2000rpms" because you accelerate the motor-generator slowly. This may allow the voltage across the capacitor to reach 15v (the dc/dc converter's maximum output voltage) by the time the motor-generator reaches 2000 rpm (especially as it is pre-charged to 12v). If so, there will be essentially no current at that time. This test tells you nothing about what will happen in actual braking. – sdenham Jan 18 at 23:20
Do you only pre-charge the capacitor to 12v for testing, or is this the minimum voltage there will ever be across it in actual use? If the former, your test is even further removed from representing actual use, and if the latter, you will only be using a fraction of the capacitor bank's regenerative braking potential. – sdenham Jan 18 at 23:41
Supercapacitors are pre-charged to 12V in every case. With maximum 10A at 12V we have seen that the system needs about a minute to charge the supercaps to 14-14.5 Volts. Thats why I said that we think our buck converter can handle the continuous current. I don't know if this thought is right. – giannis.t Jan 19 at 13:07

I blew a buck converter up (42 volt rated) when I had an input supply inductor that was intended to smooth out voltage pertubations. Ironically it worked just the opposite; an output short circuit (intentionally applied during a test) when removed caused such a sudden change in the input current that the input inductor produced a voltage in excess of 42 volts and bingo, the Linear Tech device failed and went short circuit from the input to the output i.e. the MOSFET that did the switching broke down!

So, what might be happening here is that when your switch kicks in, you rapidly charge the chip's input supply decoupler cap and a high current pulse exists as this cap charges up - if there is any appreciable inductance in the line feeding the switch from the AC/DC converter, as the cap input current starts to fall off (because full charge is approaching), the inductor will attempt to keep this current flowing by raising the potential on the buck-converter side. I can envisage that same scenario happening with my old design but, I got rid of the inductor and never tested that scenario specifically.

Most decent buck regulators have got short-circuit protection built in so, if you don't think the problem lies in the output try and think about the line inductance from the AC/DC converter, thru the switch to the buck regulator.

-
Thank you @Andy aka for sharing your experience. Our buck converter has short-circuit protection but it didn't help in our situation. As I can understand from your answer, you suggest to place a coil after AC/DC converter and before DC/DC converter. Am I right? – giannis.t Jan 18 at 19:51
Quite the opposite - Andy is suggesting that you may have too much inductance there. – brhans Jan 18 at 20:05
Long cables and coils in the generator -> lots of inductance. Which makes the scenario quite likely. Charging the capacitor to its rated voltage also sounds dangerous, a small ripple could damage it (probably the next component which will fail). – Michael Jan 18 at 20:18
@giannis.t Inductance before my switcher killed it. Please provide details of the switcher and the length of wires feeding it plus any output inductance in the AC/DC converter. – Andy aka Jan 18 at 21:51
Hi @Michael, we never reach exactly the rated voltage. The maximum voltage that the charging process has reached is 14-14.5V. – giannis.t Jan 19 at 13:15