Power Ultracapacitor Series/Parallel board

[Doug Goncz]

Hello. I am new here, but not new to Usenet.

I just purchased one 12x12 inch 1/16 inch thick 2 ounce / sq ft single side
copper clad circuit board blank from Digi Key. It will make two roughly 6 x 12
inch blanks.

I have sixteen Maxwell Technologies PC 2500 2700 F 2.5 V ultracapacitors and
thirty two 10/5 A 12 V SPDT relays. My ammeter only goes to 10 A, 3 1/2 digits.
My bicycle is equipped with a six pound servo motor serving as motor/generator,
putting energy in the caps and taking it out. So far, ride tests have been
encouraging, but to fully use the caps I must keep them balanced.

I finally figured out that by switching them all in parallel they'd balance
nearly instantly. The ESR is 1 milliohm or so. It varies, as does the EPR and
capacitance. These combine to produce imbalance.

In the future I will operate high current switches with a pull on the brake
levers but for now I want to take the 12 V supply and use it to switch the
relays to put the bank in either parallel or series mode. Parallel for
balancing and braking and series for accumulating up to around 40 V.

Those little PC mount pins on the relays look awfully slim to take 10 amps.

My question is, how wide do my 2 ounce thick traces have to be to pass 10 amps
with, say 1600 milliohms over like I am guessing around 32 inches of trace or
maybe much less in series?

Potter and Brumfeld PB-380-ND at DigiKey. The resistance of the traces depends
on the resistance of the relay contacts.

http://www.pandbrelays.com/datasheets/PC_Board_Relays/T7N_DS.pdf

Contact resistance 100 milliohms, so what difference does it make? I don't know
how to find square mils from ounces per square foot.

I figured 16 contacts per pack of eight caps, 1600 milliohms contact
resistance, the same in trace resistance.

Do contact resistances come down with time and use?

I know I could have cut that to 700 milliohms in series but it would complicate
the design and this costed out well. The PCB is more than the relays!

Are the relay PC pins standard so I can upgrade to a higher current version
later? Right now I am into measuring, not performance.



My senior project at ODU:
Google Groups, then "dgoncz" and some of:
ultracapacitor bicycle fluorescent flywheel inverter
Equipped with BoBike Mini removable child seat, too!

 

[Spehro Pefhany]

Hello. I am new here, but not new to Usenet.

Hi, Doug:-

I just purchased one 12x12 inch 1/16 inch thick 2 ounce / sq ft single side
copper clad circuit board blank from Digi Key. It will make two roughly 6 x 12
inch blanks.

So far so good. ;-)

I have sixteen Maxwell Technologies PC 2500 2700 F 2.5 V ultracapacitors and
thirty two 10/5 A 12 V SPDT relays. My ammeter only goes to 10 A, 3 1/2 digits.
My bicycle is equipped with a six pound servo motor serving as motor/generator,
putting energy in the caps and taking it out. So far, ride tests have been
encouraging, but to fully use the caps I must keep them balanced.

I finally figured out that by switching them all in parallel they'd balance
nearly instantly. The ESR is 1 milliohm or so. It varies, as does the EPR and
capacitance. These combine to produce imbalance.

In the future I will operate high current switches with a pull on the brake
levers but for now I want to take the 12 V supply and use it to switch the
relays to put the bank in either parallel or series mode. Parallel for
balancing and braking and series for accumulating up to around 40 V.

Those little PC mount pins on the relays look awfully slim to take 10 amps.

They're fine.

My question is, how wide do my 2 ounce thick traces have to be to pass 10 amps
with, say 1600 milliohms over like I am guessing around 32 inches of trace or
maybe much less in series?

The weight is per square foot, so you could work it out from the CRC
handbook, but John Larkin's rule of thumb is "500 micro-ohms per
square for 1 ounce", so halve that. You know that 32" x 0.1" is 320
squares or 80m ohms for 2oz Cu.

Potter and Brumfeld PB-380-ND at DigiKey. The resistance of the traces depends
on the resistance of the relay contacts.

http://www.pandbrelays.com/datasheets/PC_Board_Relays/T7N_DS.pdf

Contact resistance 100 milliohms, so what difference does it make? I don't know
how to find square mils from ounces per square foot.

I figured 16 contacts per pack of eight caps, 1600 milliohms contact
resistance, the same in trace resistance.

In series?

Do contact resistances come down with time and use?

IIRC, the actual resistance increases slightly during life, but the
100m ohms is a maximum, not a typical. I suspect it's typically a LOT
less since 100m ohms would be dissipating 10W at 10A, which would burn
up the contacts nicely (if you like, break a dispensable one open and
look inside- there isn't any way to dissipate that much heat).

I know I could have cut that to 700 milliohms in series but it would complicate
the design and this costed out well. The PCB is more than the relays!

Hmm how much did they gouge you for the PCB? Ouch. $22 US for bare and
$41 for photosensitive. 8-(

Are the relay PC pins standard so I can upgrade to a higher current version
later? Right now I am into measuring, not performance.

10-12A (*maybe* 15 with reduced life) is about the highest rating you
can get in that style, though there are a thousand choices from 1A to
~12A. For higher currents, consider the dirt-cheap T90 series, but
they burn more like 900mW rather than 360mW.

My senior project at ODU:
Google Groups, then "dgoncz" and some of:
ultracapacitor bicycle fluorescent flywheel inverter
Equipped with BoBike Mini removable child seat, too!

Best regards,
Spehro Pefhany