I’ll invite you to the private group chat, where I have shared more detailed info and testing data on the structure.
Decently stiff yeah, Canadian Rock Maple.
I’m guessing no more than 40A, maybe up to 60A burst.
I’ll be using Lishen LR2170LA
1 Like
I’m also very interested in this.
There is a JST-PH ready place on the PCBs to just plug in a balance wire harness.
Although that being said, due to your pack being in a single file there isn’t any place/gap/corner where I could sneak the connector into and also running the traces to connect to the correct S-groups (at least for these initial 2-layer test PCBs).
In your pack’s case I would solder the balance wires directly to the PCBs themselves.
40A is no problem, and the 60A burst should be fine as long as it’s <10s
1 Like
I see this, and I’m wondering if the JST could be right angle to make the plugs not stick out at much? Or have the JST face into the cells so it doesn’t stick out of the module? Or maybe I’m seeing it wrong.
One 1 balance wire would be needed per 2 cells. Would a discreet solder tab be able to be made so I don’t need to solder where the cells need to sit?
And for arrangement, so you think I could fit this using 2x PCB mount sets for 2x 10s2p modules, then series the modules for 20s2p? Obviously balance wires set appropriately.
I believe that’s the case. Would be pretty silly otherwise
2 Likes
Yes it does
1 Like
The JST-PH connector are pointing “inwards” towards each other when the pack is assembled. Of course the end user has the ultimate decision on how they want to mount them.
There are discreet solder terminals available at every cell’s terminals.
Based on this picture, the provided measurement and the 7 cells at the bottom.
The 7 cells back-to-back as shown is at minimum 7*70mm = 490mm in length. The single 10S2P straight pack is ~454mm long, so it’s less than what the 7 cells in the image use, so that direction should be okay.
The marked width of the enclosure is 185mm, the packs are 76mm in length (in cell’s length direction), so with a 10mm free space between the 2 long segments should need 2*76mm + 10mm = 162mm, which should also fit no problem and leave good space for wiring and vibration damping material.
Everything should fit ok. You want to move forward with this design?
1 Like
Heck yes.
Neat. How’s the compression on the cells done?
That’s a secret for now 
Not anything too complicated though.
1 Like
If they can support the high discharge of the latest gen molicel I’d definitely be interested. Pack welding is satisfying when finished but the act of pack welding is tedious. I gave up on easy cylindrical cell pack design and went with lipo. I didn’t want to find out that there was no way to milk 45 amps out of a clamped connection after purchasing the cells. My idea was a pcb making the connections with tons of vias to the other side where you would place a somewhat thick aluminum plate to sink heat off those vias. That plate on each end of the cells gives it the structure and clamping pressure needed after bolting them together (On paper).
What kind of pull time are you thinking/wanting for that 45A current? 10 secs, 1 minute?
TLDR… You’ve made an improvement on Trampa’s PCB-based pack, which they no longer sell? If so, PLEASE take my money. I’ve used a batteryblocs 10S6P/12S6P set since late Summer 2019 filled with 60 or 72 Model 3 2170s, have never been able to pull more than 10A/cell. He no longer makes these, I’m too lazy to print more (he was nice enough to give me a somewhat workable .stl file for the 12-cell 21700 module I bought.) The second set of cells in these holders are now wildly out of balance, time for a rebulid or re-think. I’m looking for a 12S8P brick to mount topside. Feel free to message me directly to arrange.
Update.
Based on feedback from @rusins and @Plurf field testing, the current design, doesn’t work adequately with shorter cells, which they are using (Molicel P42As, around ~70.0mm from terminal to terminal) and is getting too tight with longers cells like when I tested with Samsung 50Es, which are about ~70.6mm in length.
They have reported arcing, meaning that under higher vibration use, the cells are able to bounce around too much and break the contact with the PCB side contact, which causes an arc. And this then slowly causes the contact quality to degrade over time as more arcs occurs.
I’m working on new iteration of the design and I’m in contact with a chinese fab to get a fully custom spring contact made for this use. I’ve designed the spring contact myself and it’s meant to solve many of the issues I have found along with the feedback from the testers. I’m keeping the finer details for now, but the contact will be made with/in a progressive stamping and the starting costs to get the tooling made will be 1.5kUsd, which I’m willing to invest, as I believe the design will work reliably once I get these problems solved.
I have offered to send free replacements with the new design to the testers, if they’re interested, once and if I get it passed through my own testing first.
I’m gonna keep cooking, as the current state is still a bit raw, so I can’t with good consciousness start selling this to customers.
7 Likes
If you’re going to do some moulds, might want to consider atleast 2mm walls. Then they’d be compliant with a bunch of regulations and laws as well 
Moulds?
Do you mean encasing the whole pack in resin/epoxy/polyurethane?
this is probably what caused the trampa battery box to catch on fire, which trampa pretend nothing happened and just discontinued the product
I was under the impression that there were cell holders, I might have missunderstood
Pretty sure it was one of thier SMD nickel terminals that shorted plus and minus on the topside of the cell. If it was bad pick n place or if it deslodered itself I’m not sure of tho.