User content is not official advice; where it conflicts with official documents, the official documents prevail.Safety & Use Guidelines ›

© 2026 AWithZ
Topic

#Lithium battery assembly

16 posts · 14 participants
Follow topic
Me
#Lithium battery assembly Share your welding progress, a question or a finished build…
Public Post
LV5
Battery Master Zhang@lidianzhangs82 ·
Master Zhang opens a follower’s ‘48V15Ah Panasonic’ e-bike Li-ion battery: feather-light, foam stuffed down both sides of the box, actually unbranded 18650 knock-off cells in 13S3P; rebuilt as 15Ah with 78 × EVE 2.5Ah 18650 (13S6P), 0.2 nickel-plated steel strip, capacitor spot welder with 4 welds per cell, checked for wrong / missed welds, BMS fitted and charge/discharge tested; warning: any pack labelled CATL / Panasonic cells is fake
A follower from out of town sent in a 48V Li-ion e-bike battery, label says 48V 15Ah Panasonic cells, made 2023-12; it feels feather-light in the hand, clearly not 15Ah, range is very short and nowhere near the kilometres the seller promised. Off with the top cover: the cells only take up a little space in the middle, the rest is foam padding, the block is very small; inside it’s 18650s, 13S3P, the wrappers carry no brand or capacity info — typical knock-off cells, you can imagine the range; this battery is worth nothing. Rebuild as 15Ah: brand-new EVE 18650s at 2.5Ah each, 6P13S; place them in the holder alternating + up / − up, lay one row first as the reference, mind the orientation and don’t get it wrong; snap the top holder on to make one block; the connecting material is 0.2-thick nickel-plated steel strip cut to the needed lengths; the welder is just an ordinary capacitor spot welder — 18650s don’t need much current; 4 welds per cell to guarantee current-carrying strength; weld the parallel and series links in order, then check the whole block for wrong or missed welds; fit the BMS, charge and discharge test, no problems and it’s done. Warning: when buying a battery, any pack whose cells are labelled CATL, Panasonic and the like is fake. Comments (343): how much for a 48V30A build (2); 12C 25P cell config is overkill, 5C is enough (2, doesn’t match this video); selling fakes is a crime, jail (1); which courier did you use (1); how much for 48V25A; how much to rebuild a Sunra; you can tell it’s only a few Ah just by looking.
|
Why does a 48V e-bike Li-ion battery have such short range?ProjectsBattery Master Zhang
LV5
Battery Master Zhang@lidianzhangs82 ·
Sized to the customer’s old case: 78 × EVE 2.5Ah 18650 cells in 13S6P for a 48V 15Ah e-bike pack
A customer sent in a 48V e-bike battery case whose cells were dead. The creator measured the inside (130 × 370), worked out that 13S6P would fit, and chose EVE 2.5Ah 18650s — 78 cells for 48V 15Ah. Cells go into the holder alternating positive-up / negative-up (parallel across, series down), get spot-welded into one block with 0.2mm nickel-plated steel strip on both faces, with the main positive and negative brought out separately. The pack is wrapped in strapping tape, the BMS is fitted with its balance leads soldered to each series group in order, the input/output leads are connected, and a charge/discharge test (49.14V / 6.50A on screen) is run before delivery.
Building a 48V e-bike battery from 78 × 18650 cellsProjectsBattery Master Zhang
Who to followRefresh
BlindFix@71746738761
5Projects1896.0kFollowers2012.0kLikes
Follow
Laodanan@Lda_N
1Projects1618.0kFollowers27.4kLikes
Follow
YLV1
Yuzu@37548705751 ·
Youzi Teardown (Chinese/English subtitles) builds a 5S Samsung 40T 21700 pack in a Dayi A3 pin-terminal kit with the AWithZ UK1 (nameplate 5V/2.4A in, 8.4V/1380A MAX, adjustable mode / gear / trigger / continuous, the pen only welds when its switch is pressed): cells go in by the holder to locate main − and main +, a small nickel piece joins the main − first, flip over for the main +, gauge red to L+ and black to L−, fit the spring and lid with four screws, test on an A3 tool
Building a 21700 five-series pack with Samsung 40T cells (INR21700-40T); the welder is an AWithZ UK1, 5V (2.4A) input, 8.4V output voltage, 1380A current, Shenzhen Juequi Technology; on power-up it has adjustable mode, gear, trigger and continuous; its pen doesn’t weld the moment it touches, you also have to press a switch — he’ll show that when welding. The kit is a Dayi A3 pin-terminal one, open it, there’s nickel inside, tip it out. How the cells go: the bent part faces this way, this is the main −, so the negative goes on this side; once in, locate the main − and main +, this is positive; join the main − first, the small nickel piece goes here; unlike the previous pen that welds as soon as it touches, this one only starts when the switch is pressed, very convenient; the large nickel goes on; flip it over, weld the main + first, then put it in; this pack is now connected. Gauge: by the way the case sits, the gauge goes at the back, red to L+, black to L−; set it in; pack welded, lid on, there’s a spring that must go in or the button won’t come back up, four screws in; find an A3 tool and try it; have you learned to weld a pack now? 159 comments: where to buy brand-new genuine cells, JD / Taobao / Pinduoduo are full of fakes (3); are 5 × 4.0 reliable, EVE 40P for about ¥100 (2); how much is the welder; tabless vs Sony C6; which welder for copper shims.
|
Got it now? Building a 21700 Li-ion pack with an AWithZ welderProjectsYuzu
LV4
Doomsday Lab@morilab ·
Repairing a ¥30 scrapyard Boston Swing 5300 8P6S 24V pack: two dead cells swapped for new, each parallel group 2.9–3.5mΩ, nickel strip welded with a portable welder (the pen tips need regular cleaning of oxide), total 24V at 19.8mΩ; the salvaged BMS was water-damaged and broken so a new one is on order; two H4 bulbs pull 10.8A / 261W fine, capacity test waits for the BMS
Bought a complete Boston Swing 5300 Li-ion pack at a scrapyard for ¥30 (a single cell normally goes for ¥8–9); 5 of the series groups are fine, each parallel group about 3mΩ, one group has two dead cells; bought two replacements on Douyin and there is a complete 24V pack; it is no good without a BMS, and he happened to pick up another pack with a BMS to fit to this one. Repair: weld the nickel strip on with the welder (a bit crooked; the welder is very easy — press it down and one click, bang); push the cells into the holder in order, insert from the front, flip it over and weld the underside too, leave some room to move; very solid; the portable welder is just like a power bank; not a neat job but it works; the welding pen tips need regular cleaning, they oxidise and cause poor welds (the pen got hot); ugly after welding but very solid. This parallel group measures 3.5, the others are all 2.9–3.5; three are welded, keep welding the rest into one strip, rough but solid. Total voltage 24V, internal resistance 19.8mΩ, nearly 20. Fit the BMS: the salvaged one, opened up, was cracked and water-damaged, no good, buy another. Discharge test: multi-function tester clipped on at 25V, two H4 bulbs (yellow + green together is 24V, black-yellow / black-green is 12V, don’t mix them up or they burn out), sparks flying, discharge current 10.8A (about 261W), no problem; the voltage drops fast because one group isn’t full — charge it up and test total capacity when the BMS arrives. Comments: with hands like that, thumbs-down for not building your own welder; AC or DC resistance meter (seems to be AC); it’s got some years on it (five or six); what resistance and what C rate (single cell 22mΩ, new 13mΩ, 3–6C discharge, mine I reckon 1C is fine); what is it for (storage / backup power); this kind of welder won’t hold on these cells, goes intermittent after a while; chicken-poop welds.
|
Boston 5300 Li-ion 24V 8P6S pack picked up at a scrapyard, repaired successfully — but no BMS; ordered one online and will test capacity when it arrivesProjectsDoomsday Lab
Scroll for more · 16 rows