Fusible Link Wire for Voltmeter

Hello, sci.electronics.design readers.

I have been working on an "electric flywheel bicycle" for five years now. In various versions, a road-wheel or crank-coupled brushed DC electric motor/generator is always coupled to an ultracapacitor pack. Energy stored is proportional to voltage squared, and voltage a linear function of speed, so it acts like a massive flywheel with fluid clutch.

I have 6 Maxwell Technologies manufacture PC 2500 model 2700 F, 2.5 WVDC rated ultracapacitors on the current bike, which can source 4500 amps into a short. I'd like to fuse the voltmeter with fusible link wire. The last time I had a voltmeter short I was using 24 gage solid wire and it didn't open. It smoked the nylon fabric capacitor pack. I'd like to avoid damage to the surrounding equipment.

Where do I get flat cable, 24 conductors, that will fuse and part on a short? Would ribbon cable for computer equipment do the job?

I don't want to just fuse the return wire; I will be metering 12 caps soon to establish the standard deviation of voltage and so the capacitance distribution, with an eye to avoiding cell overcharge by an appropriate pack charge limit. A DMM with serial comm gets the measured values into a spreadsheet to do the standard deviation.

I have some 12-contact rotary switches to monitor cell or pack voltages. One will be wired to 12 negative terminals, one to 12 positive terminals, and a mis-switch would short a cell. Right now that's the way to do it. In the future, an NI.com USB DAQ with auto range will monitor 12 voltages above ground with no short potential. The onboard voltmeter is a 15 V analog panel meter with 50 mv internal full scale capacity. Two bannana/probe jacks interface the DMM when balancing is undertaken.

The 12 caps will be in two banks of 6, series within packs, series/parallel pack by pack, to give regeneration (parallel) and boost (series) modes of propulsion. Power is also taken out by an inverter, but when the packs are in parallel, this imbalance is rebalanced..

It's not a problem to replace all the wires right now if one opens under a short.

There is a 100 ohm, 1/4 watt resistor across each cell to establish and maintain balance by voltage division. I will go to 1/2 watt soon for more durable lead wires.

When the characteristics of the cells are known, the whole tap system will be pulled and replced with clip leads for use only when parked.

Doug Goncz Replikon Research Seven Corners, VA 22044-0394

You could just put a resistor in series at the capacitor end to limit the short-circuit current to a milliamp or so. Most DVM's have many megohms of input impedance, so yuo'll only lose a few microvolts across the resistor.

You might want to consider som kind of automatic charge-balancing scheme. A simple logic-level MOSFET, one that turns on around 2.5 volts, would make a very simple voltage limiter. Source to minus, and gate and drain to plus. That's it!

Or if you're more into bipolars, a small power transistor and three or four diodes.

The Dougster wrote: [snip]

A voltmeter will (should!) pull virtually zero current (it has a very high input impedance), so you can put a resistor in series with the voltmeter sense wires. That will limit fault curent, but shouldn't affect the readings. Say you want to limit fault current iin the event of a voltmeter short to 1mA. 6*2.5V = 15V. 15/0.001 = 15,000 ohms. A 15k resistor should be insignificant compared to the input impedance of the voltmeter, which will be a megohm or more. With the fault current limited to 1mA you won't need a fuse (and you'd have a job finding one which would blow at 1mA!).

I don't see why any of that stops you using purpose-designed fuses. You can use one fuse per capacitor if you want to be able to safely use any series/parallel combination at will. With caps capable of delivering such huge currents I'd probably do that anyway, at least during development when they won't be safely sealed in an enclosure. Make sure you can't accidentally touch the bit between the fuse and the capacitor (use heatshrink) and you're protected against wiring faults, component shorts, dropped spanners and virtually any other eventuality. Fuses are much cheaper and simpler to replace and you get a far better defined fusing current than using undersized wiring, which will likely have to be so undersized it'll have a noticeable cost in resistive heating.

Tim

That's an overcharge limiter? That could be useful. I will consider it.

Charging caps that vary by 20% all to the same voltage induced a run-down discharge problem of reverse charging some of the caps, but these are reverse polarity protected. I already operated them in reverse for two rides with a wiring error.

So that tip might just be real handy, and of course a resistor or two > You could just put a resistor in series at the capacitor end to limit

Yes, and I think this is the best policy.

I just had a little trouble finding the 2-gang 12-position non-shorting switch. I will find one soon. They are in a 6-cell series bank now. Two banks of 6 in series will be switched series/parallel soon. See below for an alternative.

I do this because Maxwell recommends it at the minimum. What I really want is a generalized solution to putting N ultracpacitors in series/parallel all they way from N series for maximum voltage to N parallel for maximum capacitance, including all JxK configurations where J and K are complementary divisors of N.

What I have towards that goal are some 8A rated relays and some boards I had made but the relayss don't latch. That was the best I could find at Digi-Key. I am looking for 15 A relays that latch, in a DIP configuration, with 2.5 V coils. Those are like hen's teeth.

I had 6 boards manufactured by pcbexpress.com. I machined the vent cap holes in the boards to 17mm and the supplied plated holes to 6mm for the capacitor terminals with a drill press and reamer. I have not sawed the 2-up boards apart. I still have the design saved. Each board mounts to a capacitor to allow the 2x3 configuration for the rack trunk and the 1x6 configuration for the seat bag.

I had designed a 4x2 relay PCB in Windows Paint but never made it up. I sent it to a local shop for vinyl masking but they flubbed it. Also, as a small producer of waste, I might pollute more than necessary with a small job, where a large company would probably deal with waste in volume, under regulation, and efficiently. So I gave up on making a board myself.

A local design shop offerd to build a board with MOSFET balancing and series-parallel switching for thousands of $$$$. No way was I ready to do that; I am not doing that kind of volume. That is for much later.

I believe the way I designed the boards they can do JxK switching, although I only intended 1xN and Nx1 configurations at the time. It has a lot to do with reducing the many abitrary and equivalent JxK arrangments to just one. There is only one 1xN and one Nx1 confiuration, barring permutations.

Doug

any wire will fuse and short eventualy with high enough current density, however what you need to avoid is the failure of the pvc type insulation and further shorts and inevitable fire/smoke cuased by the very high temperature at wich the wire fuses >1000'c This is why fuses are encapsulated in glass ceramic etc. fuses also use a lower temperature fusing wire wich probably isnt any good at all for ordinary wire.

many hi power battery packs come with reseting solid state fuses, wich would probably be ideal here.

as for load balancing you could repeatedly switch a single capacitor accross two batteries in turn. this would lose the least amount of charge. obviously the rate and size of cap depends on the rate at wich the batteries are being unbalanced. maybe you could use a charge pump IC to do this, some are quite flexible.

Colin =^.^=

----------- SNIPped --------

...

...

Several people mentioned that voltmeters have high-impedance inputs, so some series resistance is ok. Ie, you would have voltage sensing leads separate from the high-current leads. Each voltage sensing lead could have a 10K (eg) resistor at its capacitor, after which it would not need a fuse. The high-current leads might need fuses, but that apparently won't involve your rotary switches.

...

One can make small boards with negligible waste of chemicals ~ just time ~ eg see

-jiw

Yes, I understand that now.

That's how I did my board in Paint; I flooded the traces to avoid waste of etchant and pollution of copper products. I would go back to it, but now I realize the bike needs two separate packs, each 1x6 electrically, but one 3x2 mechanically and the other 1x6. Maybe I will go back to it.

Doug

Don't worry about "waste", if you're talking about etching a copper PC board.

Use ferric chloride (FeCl3) for the etchant, and when it's used up (i.e., so much cupric chloride that it won't etch any more), then pour washing soda (sodium carbonate - NOT bicarb) into it, and it neutralizes the chlorides by bonding the Cl with Na, making NaCl, and the iron and carbon make carbonates, which are essentially dirt. You could throw it in your garden, or flush it totally harmlessly. But flush well, it _is_ mud, and will leave sediment.

Cheers! Rich