can Soliton handle wide battery bank voltage changes

michaeljayclark said:

when the caps drain the pack goes from 288 to 144.

doies the controller see the different voltages and freak out?

Click to expand...

Depends. An ultracapacitor bank will have a much higher ESR than a typical EV battery pack so the voltage will drop more for a given amount of current draw. It will rebound just like a battery pack when the load is removed so the "minimum battery voltage" function in the Soliton1 will still work, but you will likely have to keep the motor current slew rate below 1000A/s (or perhaps even slower).

Of course you could also set the minimum battery voltage to 10V which will more or less disable this function... Note that there are two things you absolutely don't want to do to any motor controller (not just the Soliton1): 1) interrupt the motor circuit when it is under load; 2) apply a step change in voltage to the controller input without going through precharge (in other words, a continuous change in voltage is ok, but an abrupt one is BAD).

It's not real clear what you are trying to do here, however. Are the battery pack and capacitor bank in parallel? How is the capacitor bank recharged if they aren't in parallel? You have a lot of unknowns coming together at once - it's going to be difficult to sort out what is causing any problems should you have any.

michaeljayclark said:

another question is a dcv to dc converter. would the voltage change freak it out as well?

Click to expand...

Most of the dc/dc converters I've seen used in EVs won't handle this too well, especially if the change in voltage occurs rapidly (over a period of 1 second or less). The Vicor "bricks" seem to be the lone exception to this experience.

This is doubly problematic if you don't use a 12V battery to help stabilize against changing loads on the output side of the dc/dc converter. I highly recommend retaining at least a small 12V battery in all EVs.

I know of many who are using them in EVs and seeing a difference but yes in parallel. I just saw the MIT paper and wondered why not increase the amount of caps.

The amount of energy in these caps is far from minuscule.

54 caps starting at a working voltage of 145 volts can be drawn by 600 amps continuous over 11 seconds and still hold 75 volts. they can supply 584,000 joules over those 11 seconds.

the caps would be charged in the circuit and then as the energy from the caps is drawn out wouldn't the batteries replace that energy?

Im guessing not fast enough to keep the controller from thinking there is a disconnect and act according.

how about if no motor controller? (Not like I would actually try it but be nice to know )

michaeljayclark said:

wonder how it worked for MIT

http://web.mit.edu/first/kart/everpres.pdf

I know their circuit was much smaller yet the circuit worked fine with an ultracapcitor module in series with the batteries and motor.

why wouldnt a much larger string work when it worked in their experiments?

Click to expand...

That is a very well done project, but you missed its entire point.

You omitted all the switching and power conversion components, by taking functional diagram and converting it into electrical diagram, while dropping key components from it. The key in this approach is all the switching going on in real time, using sophisticated software and power switching IGBTs or FETs. By the time you figure all the costs involved in switching hardware/software, the whole idea becomes utterly useless in practice, although a nice school project, I have to admit.

Also, the cap in series model is getting its energy from regen, not from batteries. It would make no sense to feed the cap from the battery since there is no gain of energy, just loss. As most of us know, regen is basically a unicorn, so if you are really chasing regen, then just get an AC drive system and drop caps entirely.

The best approach to actually showing benefits is experiments with proper documentation.

I agree, the proof is the pudding, although pudding would mess things up a bit in electronics!

since I have the caps and will be using them, and as it seems parallel seems to be my first way to go which will be the easiest first step.

What instrumentation and documentation do you believe would show you the caps usefulness? Id be willing to do it, and if the caps show no benefit then we can finally put that in the archives.

But I just wonder why ultracapacitors are appearing in more and more applications and are now planned to be integrated in with batteries by DEKA for example?

Also, for many years now, capacitors placed in parallel with car stereos in cars have eliminated voltage sag in these loud cars. There are existing battery companies integrating capacitors into their batteries for car stereo use already. I wonder why not for EVs?

The devil is in details. Energy levels in audio application and motor application are not comparable. To get meaningful results you need to operate in similar energy units. When you find out how big/expensive the cap must be to push the car, then suddenly the idea becomes impractical.

Caps can be used to smooth the spikes and ripples, which arguably helps the battery lifecycle, but amount of such help vs. the cost of the cap bank is questionable and very difficult to measure.

People use caps as buffer with Lead Acid batteries because those batteries suck at instant power. Since we are now fortunate to have LiFePO4 batteries available, why bother with caps? Spend the money on more battery.

I think you are looking at the problem from wrong angle. You start with having caps and trying to get something out of them, but you don't really know what you are trying to solve, so you are searching for a problem to fit the solution. Instead, define a problem first, then pick most effective and practical solution for it.

michaeljayclark said:

wonder how it worked for MIT

http://web.mit.edu/first/kart/everpres.pdf

I know their circuit was much smaller yet the circuit worked fine with an ultracapcitor module in series with the batteries and motor.

why wouldnt a much larger string work when it worked in their experiments?

Click to expand...

You're drawing has nothing in common with the block diagram on p.5 of the above referenced document. In that document you have a battery pack and ultracapacitor bank in parallel via a diode and a bidirectional dc/dc converter. The two are not in series. In fact, it makes no sense to put capacitors on series with batteries because the total power delivered by a "battery" pack is limited by that of the lowest-capacity "cell". Since ultracapacitors have very low equivalent Ah capacity you've basically chopped your battery pack capacity off at the knees (ankles, even!).

In the MIT go-kart the lead-acid battery pack charges up an ultracapacitor bank via a bidirectional dc/dc converter, but can also provide current directly to a bidirectional motor controller via a bypass diode. The motor controller can either supply current to the motor, or charge the ultracapacitor bank (regeneration). The bidirectional dc/dc converter can then drain down the ultracapacitor bank to recharge the battery.

All in all, this is a very complicated setup that is unlikely to deliver anywhere near enough gains in efficiency or range to justify its cost/complexity.

Sorry to be a wet blanket, but this is the sort of stuff I think about for a living, you know...

a wet blanket is needed occasionally to put things in order, but a wet blanket over a battery pack with caps would be an interesting and colorful explosion !

afterall this is going to be a soliton jr testbed so it better be down properly!

I'm going to just go with the caps in parallel with no switching circuits or anything else complex. Well just put them in parallel as the first test because it will be the easiest way and would be the simplest way.



what instrumentation would be needed to track the data in a caps and batts parallel circuit.

I think we have come to the conclusion that without any real data to say caps work and deliver their promised benefits we do not really know unless you look at formulas and try to decide that way. On paper says one thing but real life has soo many different variables I dont think we will know the true answer unless an experiment with proper documentation is set up.

michaeljayclark said:

what information do you require to prove/disprove the effectiveness of ultracapacitors in electric vehicles?

Click to expand...

Define "effectiveness". In EV, battery is energy source and motor is the load ( regen is an optional source with several important conditions attached ). So, what exactly are you expecting cap bank to solve/improve/effect ?

Theoretically, smoothing ripple currents should improve battery life, BUT there are so many variables in this theory that its almost impractical to prove one way or another. This applies more to bad batteries like Lead Acid, so even if you add cost of a cap bank to prolong life of a bad battery, then you are essentially spending money to drive bad EV longer. There is your effectiveness, if that's what you were shooting for.

Just to be clear, when I say "life" and "longer" I don't mean range on a given charge, since caps don't add range, I mean you might get more charge cycles to your bad battery, sort of like keeping a terminal patient on life support for few more days to let him suffer longer.

To measure this "effectiveness", you'd need 2 identical EVs in same driving conditions with same batteries, one with cap bank and one without. Use those EVs in same driving pattern until first battery croaks and see which one croaked first and how long after the other one croaked. Meanwhile, everyone else will be enjoying their Lithium packs which don't need any caps to begin with. LEAD IS DEAD. The sooner you get this, the better.