Appliance repair ep. 258: Xiaomi car washer battery replacement

Repairman Xiao Li ep. 258, Xiaomi car-washer battery swap (10 minutes): a Xianyu customer supplies 3 × 18650 for swapping; the Xiaomi shell is hard to open, clips snap when prised; old cells read 1.2/0.7/1.8V but resistance 14–16, not actually dead, just need waking up, but the customer bought cells so they get swapped; the new cells' resistance is higher than the old ones', 'useless'; AWithZ welder on 0.15 nickel strip, gear 40→50→70 before he trusted it; the original strips had little tabs that can't be copied, so hammered flat and welded on; wired the polarity wrong (even with photos), a friend pointed it out, rewired to 12.6V and fully charged, 'lucky it was wired to the BMS and not the main unit'

A Hangzhou customer found him on Xianyu, bought three 18650s himself and asked for a swap; thought he'd have taken it apart but he hadn't, 'lost out on that'. Screws out and the Xiaomi shell is still clipped on really tight, the clips won't prise off and snap as soon as you try; more or less got the battery out (XCJ002 MB_26_240311 three-cell pack). Old cells read 1.2/0.7/1.8V; RC3563 tester puts resistance at 14/15/16, 'not actually dead, they just need waking up'; but the customer bought cells, and if you don't swap them he might get ideas, and if the swap doesn't work it's not on me; do the new cells need balancing? new ones are generally fine; the new cells' resistance is higher than the originals', 'useless', but no big problem, fit them anyway. Get the AWithZ soldering iron — no, spot welder — out, gear up to 40; 0.15 nickel strip feels weak so 50; can't tell which gear is right so 70, worried it won't hold so higher still; welds beautiful; each original strip had a little tab soldered to the board, no such tabs here, should have prised the small ones off, without tabs just hammer flat and weld on, 'tell me if I'm doing this right'; feels like the welds are big; weld one corner solid then flip; clean up the pads and board, fish paper along the edge for protection. It was actually fixed, but the cells got wired wrong when connecting (even with photos): drew it out and found the + − + alternation had two negatives together, a friend pointed out that if this side is negative the other can't be negative too; strip and rewire, one side positive one negative, 12.6V now right, fully charged, fixed; lucky it was wired to the BMS, the main unit would have burnt out; from now on take clear photos. Comments: 20mΩ is quite normal for capacity cells; no need to strip it, just wake it with an adjustable supply; why is each cell's voltage different (quite normal); low resistance doesn't mean a good cell, no voltage means it's dead; you're hilarious, greener than me; brute force works miracles.

Outcome

Steps8 steps · click a frame to jump

1. 
   1. Xianyu customer supplies 3 × 18650; not stripped, 'lost out'

   Hangzhou customer found him on Xianyu, bought three 18650s himself and asked for a swap; thought it'd be stripped but it wasn't, feels like a loss; pick out the screws.

   00:00 – 00:26
2. 
   2. Xiaomi shell hard to open: all screws out, still clipped tight, clips snap, battery barely out

   All screws out; one more; prise with a utility knife; everyone says Xiaomi's are hard to open, and they really are, clipped on really tight; clips won't prise off and snap as soon as you try; won't come off; more or less got the battery out.

   00:28 – 01:49
3. 
   3. Old cells 1.2/0.7/1.8V, resistance 14–16, not dead, just need waking; customer bought cells so swap anyway

   Three-cell pack XCJ002; voltages 1.2, 0.7, 1.8; RC3563 resistance 15/14/16, all around 15, not actually dead, just wake them up; but the customer bought cells and would get ideas if not swapped, and if it fails after swapping it's not on me; new cells don't need balancing.

   01:49 – 03:24
4. 
   4. New cells' resistance higher than the old ones', 'useless'; fitted anyway; out comes the AWithZ welder

   Test the three new ones: higher than the original cells, damn, useless; no big problem with three cells anyway, fit them; get the AWithZ soldering iron — no, spot welder — out.

   03:24 – 04:09
5. 
   5. Gear 40→50→70: 0.15 strip, can't tell which gear is right, turned up for fear it won't hold; welds beautiful

   Gear up to 40 and just weld; 0.15 nickel strip feels a bit weak, up to 50; can't tell which gear is good, so 70; looks welded but worried it won't hold, up again; it holds; very beautiful.

   04:09 – 05:43
6. 
   6. Original strips had little tabs that can't be copied: hammer flat and weld on, 'am I doing this right'

   Each original strip had a little tab soldered to the board, none of those here, and each is different; should have prised the small ones off; without tabs just hammer flat and weld on top, or I'd have to cut my own; welded solid; feels big; one corner done, flip it over.

   05:40 – 07:44
7. 
   7. Clean the pads / board; fish paper

   Clean up the pads and the board; very beautiful; fish paper along the edge for protection.

   07:57 – 08:22
8. 
   8. Polarity wired wrong (even with photos), drawn out + friend's tip, rewired to 12.6V, full charge; lucky it was the BMS

   Actually fixed but wired the cells wrong; photographed and still got it wrong; test fails, drawing it out shows + − became two negatives; friend says if this side is negative the other can't be negative; strip and rewire one positive one negative, 12.6V now right; fully charged, fixed; lucky it was wired to the BMS, the main unit would have burnt; take clear photos from now on.

   08:24 – 10:36

Bill of materials

Measured on camera

Creator’s tips

• Cells with very low voltage but normal resistance (14–16) are mostly just starved; wake them with an adjustable supply instead of replacing (comments say the same) 02:43
• Before wiring take clear photos and draw the wiring; + and − must alternate, never two of the same polarity side by side 08:51
• Connect to the BMS first for testing; a wrong connection won't burn the main unit 10:20

Pitfalls

• Xiaomi shell clips snapped when prised 01:22
• New cells' resistance higher than the old ones' 03:24
• Didn't know the right gear, kept going up to 70 04:57
• Original strips' tabs couldn't be copied 05:53
• Polarity wired wrong despite photos 08:24

Li-ion safety

• Reversed polarity on the main unit would burn it out; connect to the BMS first to test 10:20
• Fish paper along the edge for insulation 08:12

Quotes