So it begins...... Battery spot welder

[sternwake]

There was a 2000mah 'roofer' wrapped 18650 battery which came with a Cloned SOFIRN HEADLIGHT,  which i soldered into an even cheaper headlamp whose 18650 cell was a  pathetic 700mah capacity.

I reopened that light, removed the roofer battery, and found its  shrink wrap was all rough, like rust bubbling below paint.
Sure enough.
    .

I do not recall ever letting this light get wet.
oh well.

I set up old red with the two batteries, one for welding one for driving mosfets, and soldered  wires from LED driver to some 0.10 nickel strip provided with the Black BSW, then welded nickel strip to the 4th LG cell pulled from the failed Ryobi pack.  A longer protected cell or one with a  button top would not fit in cheapo headlamp enclosure.

I used a super low welding setting and it was shooting sparks, and woukd bump it up a  'gear' and get a better weld.

I have to assume the black BSW provided nickel strip is nickel plated steel.  did many little weak welds, lifted hard enough to insure adequate adhesion, and reassembled  cheaPo flashlight, and placed some kapton tape over the bluish leds to make them a weird green tinged yellow glow, which is far less annoying than bluish tint to my eyes.

So those are the first spot welds  have put into service.

Then i tested old red with new XT90d lipo battery at 12.12v  with mosfets driven by 13v ub12220 via another XT90 antispark.  I was able to weld 0.15mm nickel strip to a 0.24mm  razorblade at 13 ms, where it ripped the nickel strip from razorblade, but tiny pinprick welds with the relatively sharp welding pen tips.  Same with the roofer baTtery -.

Id cut out a slit in the nickel strip and it really does imlrove the penetration, and allows a much shorter pulse.  13ms vs 45 to 50ms with generic xt 60 and no second battery.

Two.layers of 0.15 nickel strip were welding well at 65ms, were 95 was required for not quite adequate welds before dual battery mods and the slit nickel.  the warmest oart of the whole  circuit was the 12awg from battery to xt90, but was far cooler than earlier testing with factory fake xt60.

So whohoo.  Improvement in Welding power at shorter pulses, with mods, in combo with better technique.

Welding 2 layers of 0.15 nickel to 0.26mm razorbkade, well, there was barely any discoloration on backside of razor blade.  razor bkades are a different alloy and are heat treated, so its not a great comparison to a battery can of a softer, not hardened steel - said to be 0.3mm thick.

Ill should openn the cell and see if there are burn marks opposite the welds.

Last night i was looking closely at the 'INFINEON ' MOSFETS oN new purple, and saw each ones printing was slightly different, and looked nothing like the Infineon Data sheet, and i figured they are  counterfeit and have little chNce of being matched  counterfeits and lost faith in new purple's ability and potential longevity.

I transfered all 10awg keads to purple from red, 2 batteriez, lipo welder, agm for triggering mosfets.
Long story shorter, it is more powerful.  2 layers of 0.15mm nickel strip welded well  to razor blade and battery -, at 55 ms pulses. Not as well on the battery + though, to my surprise.

Looks like i got plenty of overhead with old red and new purple BSW, with split strip.
The true test of ability will be if i can split 0.15mm copper strips, and caPped with 0.10 nickel.plated steel.
the longevity test will be a time will tell thing, but at least i have 2 functioning BSWs.

whn  i was slowly increasing the ms duration, i was using ir temp gun, one mosfet was 5'f warmer than the other 4, but the 12awg at battery was 25f hotter.

The longer pulses  really seem to take some of the kick out of the lipo.  I had returned it to 12.2v power supply  when taking balk throwing breaks between earlier tests, and it did  not gobble too many amps on initial hookup, but did after testing those longer pulses.

The purples beeper is asininely loud, and basically pointless. it only beeps on power on, or auto power off, the latter being alarming, and it beep when changing power levels..

Old red beeps at half the volume and beeps twice before welding and once during.
Im.gonna disable the purple's buzzer.

I can achieve welds using shorter duration pulses with the purple, vs red, but red more pleasant to operate. 

The  smoked black bsw's enclsure is.on the purple, but not quite dialed in yet

There are high quality BSWs in the 259$ to 350$ raNge..  Kweld and malectrics are built much burlier.

Full charging the lipo to 12.6v will likely increSe welding performance, or at least allow shorter pulses to achieve the same quality of weld.

Buying 10 matched infineon mosfets, and putting all new on the purple shoukd  yield burlier ability and much extended  longevity, but SMD soldering ability is not something i have  loads of.


the welding pen tip shape has huge effects on the weld, as does the pressure one uses to hold them in down,and  human error factor is quite apparent.

cutting slits in the nickek strip is very effective, but time consuming.
me thinks sta king several and dremelling a slit be the way to go.
 It is very obvious that the closer one gets to the crotch of the split strip, the worse the weld strength to battery is.

10mm wide nickle strips are too wide for battery + terminals, and the fishpaoer insukator rings become more appreciated.

[sternwake]

The beeper had to go, it was far too useless, and obnoxious. 

Luca showed which trace to break to deactivate it.. 
Done and done.

I modified the black BSW acrylic enclosure to fit the Purple, and flowed a LOT more solder onto under and over  the brass bussbars on both sides of the Mosfets.

Here is a 50ms pukse, hakf power,  solidly welding 0.3mm of nickel to a razor blade with a 12.55v LIPO and freshly shined and rounded welding pen tips on a freshy cleaned razor blade. 

   

0.15 Copper aheet supposed to arrived monday.
If it can reliably weld that with a 0.1mm nickeled steel cap, at less than full power, ill be quite happy.

If it can't, then perhaps 10 matched infineon mosfets might be ordered, as well as a second lipo, and fatter welding leads..fabricated.

or.i can call 2 layers of nickel good enough, even though 4 would be required to match the resistance of the 0.15 copper.

[sternwake]

I rewatched the video of the guy who beefs up his purple BSW with 5 more mosfets, and sucessfully tests welding on 0.22mm copper with 0.20nickel cap, at 99 milliseconds, with no slit.

He did not do luca's modifications to driver  circuit or use second bTtery driver technique, but did have some interesting leads to and from the BSW.

Noticed his welding pens  were way burlier than mine.

So I burlified mine, cutting off about 14mm of 14 awg  copper, and dremeling a roinded cone with very slight flat top.

   

Then the  0.15mm copper sheet arrived.   And both new ESC's.

8 capped  a 10mm wide strio with 0.10mm nickel plated steel with welding slit cut into it

Achieved very good welds at 35 ms
Achieved excellent welds at 40 ms.

The copper nickel plated steel sandwich is much harder to peel from the bTtery than is dual 0.15 pure nickel

   


Can you hear manaical laughing in the distance?

Or is it just assumed ?

[rvpopeye]

As long as it's not Murdoch ...

[sternwake]

I tabke tested the newer Puaida ESC which is a 10s 36v nominak controlker.
But I dont have any 10s batteries.

But i do have 2 5s batteries, makita or Ridgid, so i hooked the ridgids in series, hooked everything up, turned it on, the remote paired on its own, pushed throttke forward, and heard my hub motors spin up to ridiculous speeds. the remote was displaying something like 47 kmh, bjt thinks they are 100mm diameter wheels.

but satisfyingly they were smoothly gaining speed, and maintaining speed.

The ridgids BMS wilk not allow charging through output pins, so i did nit bit the brakes, fearing magic smoke release, as they are regenerative brakes, upto 10 amps charging

the other ESC is a LingYi. same as on the mini but dual motor. same remote.

I had soldered the 7s bridge but the ESC was just too big to fit inside enclosure despite a little plastic hacking.
The puaida ESC is same size as the shitty one the 7s board came with, but it needs a 10s battery.

So my bigger skate is nearly completely disassembled. I could return shitty esc. to it.
but .....
Need 10 s battery, 10s battery enclosure which can fit everything inside and allow for flex.

I could likely fit a 10s 3p battery, maybee even 21700s. which means i need 10 more dmegc 18659s or 20 more molicel p42a.

So.im gonna make enclosure as big as possible to potentially fit 30 2!700s.
but will likely just make a 10s 2p battery for it.

I wanted to test the New lingyi esc, on 7s, but it has different motor connectors, and no easy solution presented itself with materials at hand and soi gave up for the night.

Kind of wish I didnt disassemble skate as I might need to paddle tomorrow for sanity's sake.
could reassemble relTively, it but want to design enclosure while it's off.

my

the thick thermal grease Id applied, too thickly, did not really spread and squeeze out as intended, meaning heatsinking was improved, but not nearly as much as it should have been.

Wonder if 10 molicels 21700s can fit existing enclosure that was designed to house 14 18650s.

[sternwake]

Im stuck in the evil loop of analysis paralysis.

ramblin

dot


dot
.. surprised?

All l i need, plus generous buffer ,

vs

why not maximize future expansion potential, even though I currently have no need for extra.

Im staring at 30 cells placed on my 538sk8, hull, with the large Yingli easc and a 7s bms between truks

10s of 3p, of high end Molicel 2!700 cells , vs building a 10s 2p of 18650s that woukd greatly outperform my 7s 2p pa k.

I will be maKing a fiberglass enclosure that should shame anybody without tons of experience doing similar.

But Im about to form a 10S3P enclosure upon a mass produced deck.. whicb h bothers me,

But the actual chinese plywood deck, with thick bamboo veneer on both sides, is surprisingly , very well done, i just gotta make it different enough to be mine.

[sternwake]

after staring at the hulk with loose 18650s and 1!700 for far too long. i de idex to just buikd as big of an  enclosure as i  coukd, for a padded singke stack of  cells.

Will a 10s3p battery of 21700s fit?  i wont know for sure until i have 30 21700s.
but 10s3p of 18650 certainly will.

But I dont have 30 18650's, i have 20.

36v nominal 5200mah.
vs 25.9v nominal and 4000mah,..... how the skateboard arrived.

That 0.4v difference between input charge voltage and output battery voltage scared me enough to cut away the heatshrink where the BMS should be, and am  c.onsidering never returning this 7s2p pack to service.

There is a tiny bms with aplooge all over it, and the balance wires, which are all live and can pote tially short and cUse a fire, were just willy nilly  compacted into the bms void, with a piece of nickel strio and its sharp edges wbich is alzo the main - output, a mm or 2 away.

the fact that what appears to be 8mm wide and 0.12mm thick nickel is designed to carry the  entire load of the battery, means it has to be getting scorching hot when towing the FYak, which must be close to a  continuous 20amp load. 
 0.15mm 10mm wide pure nickel is rated to safely pass about 7 amps, and the strip they used is likely nickel plated steel.

The BMS is bypassed for discharge, the outout leads going right to  fake brass xt60. the esk8 reliez on the ESC for low  voltage  cut off, which is kind of standard on Esk8, as the bms tripping  will happen under full load which sends rider flying off.

Ive zero confidence in this provided  7s2p battery now.  
Ignorance was blissful, as usual.

It might be torn down to make a 3s4p 11.1v nominal, 12.6v max 8.0ah battery.

but not today.

As i was laying out my 20 Dmeg cells, and frustrating myself with the options, i put them all away and cleaned up the 6 good 2000mah 10 amp cdr cells from 38e$k8, used three of them to make a 3s0p battery as my first spotwelded battery.

they all measured the same vtage as i wrote on them 6 weeks ago, and were within 0.002v of each other

its significantly smaller than my 3s battery inside the 3s holder with leaf spring contacts. same bms attached at end.

Its got fishpaper between anytbing which coukd chafe or short under impact, amd is wrPped in several layers of kapton tape.

But i have no good way to check individual cell voltages.  I can easily probe the 3s holder.

the 3s Lipo and the one display i have shows individual cell voltages via a JST  connector.

so i ordered a bunch of different jst connectors with pigtails.

I will use them to  heck cell balance and charge lower cells individually on future battery oacks, as i si.y dont trust that a BMS is actually going to do a proper job, and since i was the human BMS with lead acid, guessing at cell voltages as an average of total voltage, it just makes sense to not leave human dutiez solely to some magical device which claims it can replace me, designed and built by people who are trained to despise me.

   

[sternwake]

Building the 10s3p esk8 enclosure hit some snags. wjth the mold, forcing a two stage figerglass layup, but then i used the poor mans vacuum bag technique, with the wrong plastic, which then bonded to the epoxy instead of peeling off easiky and aLlowing that second layup during the chemical bond window.

Now i'll have to sand it before that second layup, and hope i can get all the wrong clear plastic bits off.
i spent about 2 minutes trying to remove the wrong plastic before having to walk away.

The JST connectors already arrived, ordered on ebay, delivered by amazon prime. The 10 amp dc to dc booster which can exceed 42v, also arrived in an amazon prime envelope, but via usps, and i only got an update this AM saying it shipped.


The Lipo battery and the little voltage monitoring device i got for it, is quite interesting comparing individual cell voltages whike charging, under load, and at rest. ive on ever charged all three together.

the JST plugs allow a 'hobby' charger to chaRge one individuL cell.
with 12v lead acid there is no way to easily discern individual cell voltage, and when a battery staRts behaving badly the answer is to abusively overcharge 5 cells, hoping that 6th one catches up and 'equalizes'


With Esk8, the BMS is basically wired as charge only, unless they have buck converters to use 12v led lights, and then only the buck converters and charger leads are wired to the BMS, the esc gets power right from battery and has its own LVD protection.

Reading the esk8 forums, i think some of the pros believe the BMS, for sure the smart bluetooth bms but also the dumb bmss too, is always actively balance charging each individual cell group.

Watching Andy, from Offgrid garage, a german guy living in australia, who does extensive bms testing, ahows even the smart ones dont realky balance all that well, and when they do, it is with tiny currents, bleeding off the top cells or charging the lowest with as little as 0.02amps.

The BMS might not even attempot to apply any balancing current until the cells get to very high states of charge, and those who onky ever charge to 4.1v, can easily have parallel packs way out of balance.


I think I m going to wire uo my bms for charge only and this means a balance wire for every series connection, for the bms, bjt im also going to wire up two 6 wire jst plugs, so i can monitor each individual parallel pa kmof cells and drain or charge them individually.

a smart BMS would monitor voltage of each parallel pack of cells, display nicely on the phone, but what about proper top balancing.
when does it occur?
what starting voltage?
with how much current?
and if uktra tiny currents, will one need to hold battery at 4.2v per cell for days to get low parLlel packs up inline?

If i charge to only 4.1vpc will my dumb daly bms even attempt to balance?

Ignorance is bliss, but i think the bms's are kften thought if as taking care of everytbing, but i reLity onky stop over temperature, over or under voktage charging or discharging.

I do intend to make a 10s3p battery with $$ 21700s and a smart BMS in the future, but i also believe i will have a separate JST connector so i can charge or discharge individual cell groups and not hooe and pray the bms is doing so on its own.

[sternwake]

Used the last 3 18650s from 38e$k8  to make another 3s 11.1 nominal 12.6v max 2.0 ah battery.

Decided i needed more practice with copper sandwich spot welding even tbough these  cells are 10 amp max, and unlikely to ever be asked to provide that and 0.15 nickel more than adequate, copper nickel sandwich it got.

Instead of cutting a slot in  10mm wide nickel and copper strips, i cut 2 5mm wide strips in the 'infinite slot' method.  It was very hard to position all 4 acrosz two cells so i turned power way down and joined nickel to copper first, then placed sandwich  on batteries, cranked it uoto 45ms, and welded the sandwich.

The first BMS was inop, and after  ursing uo a storm, then  replacing it, and testing, i added the JST balance plug as fiund on Lipo batteries.

Here is the pack powering some red Leds and the 'capacity checker' showing a 0.062 volt difference between highest and lowest cell

The 3s holder with loose LG ryobi pulled  cells would have as much as 0.4v difference after a few cycles, always the middle cell being lowest and Id emove them and charge individually.



   

[rvpopeye]

That's an interesting little display unit.
 
These descriptions are almost like looking over your shoulder.
Some day I'd like a shop/shed area to play in.
In here it's on the floor , bed or more usually on my lap ....
and I can't wait for John' s comments about that. Skuh kuh kuh kuh

Both of my house batteries are getting noticeably weaker . Have started looking at  LI ...Wish the new sodium ion ones weren't still vaporware , doesn't look like I'm gonna make it that long and it's getting colder.

Thinking move them inside for a little more life ?  Maybe .  But looking now.  Now all I need to do is figure out WHERE ?
Probably would put a LI inside anyway so staging for the feature in advance .  It will be really close to the controller and distro box ..  Even if I just get another L/A  think I'll put it inside. Finally can get wiring finished,,, shorter big 2 ga with less needed too. Only one batt and no selector switch ! It's nice to not need more of that stuff instead. $$$$$$$$
 
Give that cutie poochicle a scratch for me and how 'bout another mangrove pic? , it's gettin' cold like I already said !