endless sub-c nicd soldering

[dirty_d]

i got most of the nicd cells i ordered off ebay, the same ones safe is using in his tubes, i opted to just solder them all together with the tabs. i got 110 of the 160 i ordered, not sure why i didnt get all 3 packages the same day.

the first string of 30 cells was a nightmare, but after the second one, i turned into a regular cell soldering machine. you can bang them out pretty quick once you get a routine down.

here it is so far, there will be five 15x2 chunks in the pack, ill probably change the shape of the 5th one so it stays in a square. you can see three chunks in the picture.

the wires are only 20 gauge, this is all i have left, i figured it will be fine, with the 30A controller the most each wire will have through it is 6A.

 

[D-Man]

Cool. Looks like your almost done. Looks like you might beat Safe in finishing it and doing the first tests.

 

[dirty_d]

i finished it this morning, still not exactly sure how im going to put them on the bike, i think im going to make 2 pairs of two 36V string as one pack, then make a thing out of that nylon fabrc that some backpacks are made of, and have 2 pockets in it put those 2 packs in the pockets and hang it over the top tube, and have the 5th single 36V string hang in the middle somehow.

i still need to finish the charger, i just need 2 lm324s the radioshack around here didn't have any left, im gonna go to another one today and see if i can get some. im curious as to how much range ill get with this pack on a bike with the crystalyte 409 brushed motor, anyone got any ideas? the pack is 36V 12Ah

 

[dirty_d]

well i finished the charger, it worked for about 5 seconds as i had planned with my multimeter in ammeter mode and after about 5 seconds one of my home made 1 ohm resistors shorted against one of the fet cases which is also the drain, sending 67V to the inputs of the lm324 subsequently sending 67V to ALL of the fet gates which fried them all. DAMNIT!

 

[pwbset]

Awww man that sucks. Sorry for your loss.

 

[dirty_d]

i think im just gonna get set up to make pcbs before i make another, this perfboard stuff is a major pain in the ass. right now im trying to do it bad boy style, im using 2 120V bulbs in parallel but im still only getting .5A though the entire pack, i need 1.2A for C/10, i need more light bulbs!

 

[safe]

Solderless Tubes?

You should have done the solderless tubes approach. It's so much easier than soldering and if there is a problem with a cell you just replace it. I haven't done the "on the bike" testing yet (though I might throw some on #001 just for some longer range rides) but they've been flawless as far as providing connectivity.

I'm kind of surprised... after doing a thread that showed how easy it was I would have figured you would have done it too. :?

 

[safe]

Looks like you might beat Safe in finishing it and doing the first tests.

I have so many concurrent projects going that you can't measure time in a linear sense with stuff I'm working on. The solderless tubes are "done" in the sense that they are assembled and work. I've had a 100% success rate with conductivity with the cells in the tubes and they charge up just fine.

This was not something that took a lot of "linear" time... if all you were working on was just making up solderless tubes it could be done in a day or two.

 

[dirty_d]

id rather just spend more time soldering them together than spend more money on tubes. they take up more space too and are heavier.

i think youre underestimating the cheapiness of my cheapness.

 

[safe]

...and are heavier.

i think youre underestimating the cheapiness of my cheapness.

They weigh ounces, not even a pound in total. The weight is next to nothing and the added safety of having the cells safely packed inside an indestructable shell means that no matter how much bouncing takes place you never have to worry about a soldered connection coming loose. There's nothing to break... everything is spring loaded so shocks have no effect on them.

Well... just wait... I'll give feedback eventually on them... what I need is a team of fabricators and engineers to complete all the stuff I'm involved with. :lol:

(it might have been easier if I was more focused on just one project rather than three at once. )

The total extra cost was about $50 or so... it does add a little...

In the long run I'm hoping the improved durability and ability to identify and replace bad cells will make the long term cost a good investment. Only time will tell.

 

[D-Man]

So I guess the problem with going this cheap route here is the chargers? If "off the shelf" chargers were to be used, it would cost way more then the battery pack because too many would be needed due to the low ah of the battery?

 

[dirty_d]

right, im going to make my own, it will probably cost like $15 to $20, for a C/10 15 hour charger. i already made it with expensive radioshack crap for about $35, i destroyed that one though so im going to the parts from digikey next time, much much cheaper there.

 

[D-Man]

How do you make this charger and how does it charge all strings at the same time? Seems like nobody has done this successfully yet.

 

[dirty_d]

basically its a 24V transformer with a voltage doubler to get about 67V DC, then there are 5 constant current sources to charge each of the 5 series strings. nothing great, its only about 50% efficient since the extra power is dissipated as heat to keep the constant current, the efficiency would be higher if the voltage was a bit lower, i only really need about 45V i think to charge a 36V nicd pack. im gonna have to do some serious heatsinking, the dissipation on the current limiting fets is gonna be about 10W each, so 50W total needs to be handled.

 

[John in CR]

I really appreciate you guys leading the way on this stuff. I've got 130 D cell nicads that I need to use, though I'm leaning toward tubes and want to somehow make them part of my next bike's physical structure, but retain the easy access that Safe is using. It will be interesting to see how your charger scheme plays out.

John

 

[TylerDurden]

I really appreciate you guys leading the way on this stuff. I've got 130 D cell nicads that I need to use, though I'm leaning toward tubes and want to somehow make them part of my next bike's physical structure, but retain the easy access that Safe is using. It will be interesting to see how your charger scheme plays out.

Somewhere, we discussed using mesh tubes (or at least perforated), so cells can be probed without disassembly.

One feature of shrinkwrap, is the ability to get individual cell readings with sharp probes. Solid tubes lack that convenience.

 

[dirty_d]

well i made V2 of the charger after i destroyed the other one with my stupid homemade resistors from coiled iron wire, the inulation broke or something and shorted against a fet case just destroying everyhing.

this one is a little better with better rated components. i am having a really wierd problem though what i have is a fet an op-amp adn a current sesne resistor, the -input comes from the voltage across the 1ohm current sense resistor attached to the fet source, the +input is a fixed 0.250V to get a charge current of 250mA, the output goes to the fet gate. when i connect one of the single series strings to the charger it works as expevted, the op-amp raises the voltage of the fet until the -input equals the +input and you have a constant current of 250mA. but when all 5 series strings are hooked up the current is off, the +input is at 0.250V like it should but the -input is at 0.200V, and the output is only about 3V, it doesnt make sense to me. how can there be a 50mV difference and the output is only 3V?it should rise up to Vcc - 1.5V. it almost seems like the op-amps are duds or something. but then why would it work with only one series string? i cant even think of any reason why its doing this.

i added 1k resistors to the gates to stop oscillations, someone suggested it.
i put an ammeter to measure the total battery charge current, if i hook up any one single series string i get 250mA, perfect. but if i connect any two series strings i get 540mA, with three i get 780mA with four 1.2A, and five 1.55A, it should go 0.250, 0.500, 0.750, 1.000, 1.250.


heres the picture before i realized that heatsink wont work, they all need to be separate so there drains are not connected. i already separated them all.

 

[dirty_d]

here it is as of now, the caps sticking out of wierd places is just stuff i was screwing around with to see if could fix it.