Signature Solar EG4-WP 48V teardown

[Johan]

This post shows how I opened the case of the EG4-WP 48V 100Ah (5.12kWh) battery without damaging any cells or other electrical components. This voids your warranty. First try to contact Signature Solar to see if a less invasive solution exists. In my experience they do their best to help you.

Read this post if you want to:

• reset the BMS from being stuck in "protect" mode back to "stand-by" mode*, or
• replace the BMS by a different BMS**, or
• see what's inside

Procedure:

1. First remove both plastic handle ropes (can be done fully non-destructively);
2. cut the case all around using an oscillating saw (the famous Will-Prowse tear down method),
   • Carefully saw right above (!) the seam (in my case this worked without touching cells - in your case it may not - it's your risk!
   • The case wall thickness is 5/8in (8mm) all around (see photo).
   • Apply multiple saw passes, gradually increasing the kerf depth.
3. carefully pry the lid off with chisel (re-solidified debris may "weld" the lid a bit to the case's body locally).

Clearances (as measured):

• from cutting plane to cells: 1/4in (6mm); ? OMG
• between case long inner wall and cells: zero
• between BMS circuit board green top surface and lid bottom surface: 3/4in (20mm)

There is a brown small microswitch on the BMS that resets the BMS. After briefly pressing, the on-board LEDS of each cell will briefly turn on and off and a few seconds later, the BMS has reset and should now be in stand-by mode. Pulling out wires is not needed.

If you only want to reset the BMS:
Instead of sawing the entire lid off, You could also make a small hole in the lid of the case. I recommend not using a drill (unless in a very controlled manner) because the clearance between BMS and lid is small. Instead, use a soldering stick or a router if you have one. Where to make the hole... this may very well vary 0.25in (6mm) per battery!:

• 7 5/8in from the handle outer edge (see photo),
• 4 3/4in from the case long outer edge on the negative terminal side (see photo)

Does anybody recognize the manufacturer of this BMS and if so, if there a way to access it and read/change settings? The BMS board contains:

• Panasonic AN49503A Battery Monitoring IC (link)
• Microchip 18F45K22 (link)

@Will Prowse @electric @Trkarl @RichardfromEG4

Notes
*This happens, for example, after connecting to an inverter without pre-charging. In my case, the BMS retried connecting roughly five times (each followed by a 4-min protect-mode-time-out) with a Cotek SD-3500-148. I tried reviving the BMS by charging is 24h using a an EG4 charger: unsuccessful. SS sent me a replacement.
** the Overkill BMS may have the same cell tap connectors, but the JK BMS may be a good alternative too.
This is what it looks like inside.


What BMS is this? How to access?


Distance from reset switch to handle outer edge (the ruler's zero is right above the switch center!).


Distance from reset switch to case long outer edge.

 

[Trkarl]

Have you decided to replace the bms? Are you going to try to top balance the cells like they should have been before the pack was assembled in the first place? I may get adventurous too because two of the three I ordered still have about a 500mv differential if I charge them full. To me that is unacceptable.

 

[Johan]

Have you decided to replace the bms? Are you going to try to top balance the cells like they should have been before the pack was assembled in the first place? I may get adventurous too because two of the three I ordered still have about a 500mv differential if I charge them full. To me that is unacceptable.

I am trying to balance first. See below the LEDs that may indicate balancing cells, this was yesterday, 7 of the 16 are on. This morning only 3 were on. The PCB is 10-20F warmer around those LEDs where the balancing resistors likely are. I manually keep a cell-level voltage log in Excel to monitor changes through time. I CC-CV charge using a bench top power supply to 3.45V per cell until the charging current is near zero. But right now it is too early for me to say if the balancing function really works properly. I have to apply a few more charging cycles I guess.

 

[rhino]

What is that goop on the battery connectors?

 

[Johan]

What is that goop on the battery connectors?

It feels like a soft rubbery caulk (so not hard like cured general purpose epoxy), like white silicone caulk. Maybe it is intended to prevent or delay the onset of loosening from vibration as discussed here:

• https://electronics.stackexchange.com/questions/236457/white-glue-on-circuit-board
• https://electronics.stackexchange.com/questions/29787/whats-this-white-stuff-on-my-board
• https://www.reddit.com/r/electronics/comments/icfza
• https://electronics.stackexchange.c...or-pcb-mounted-components-to-avoid-vibrations

 

[Trkarl]

Are you able to determine at what cell voltage the balancer turns on?

 

[Johan]

Are you able to determine at what cell voltage the balancer turns on

So far I can only tell you this (quick and dirty):

The balancing LEDs sometimes turn on when I do CC-CV-charging to 54.26V (roughly 3.4V per cell on average). This battery voltage was read from the app, which matches my power supply @ 0.1V difference. However, sometimes they did not turn on at that battery voltage or they just blink a bit. So far I always observed the balance LEDs to be on consistently when CC-CV charging to 55.2V (3.45Vpc), so for now I balance at that voltage. But even then it still goes into protect mode sometimes because one cell hits 3.9V .

 

[Trkarl]

If you can get at the terminals you can manually discharge that cell a bit with a pair of alligator clips wired to a 12 v light bulb like a turn signal bulb or something. Just watch the voltage of that cell in the app so you don’t over do it. You could also use a resistor but I’m not sure what size. Hopefully someone else could suggest a resistor to use.

 

[Johan]

Update: The BMS increased* the imbalance (wtf) instead of decreasing it, so I decided to take the cells out of the case and manually balance them individually using two parallel 3 Ohm (actually measured less 1.8 Ohm) resistors. Took me a whole bunch of manual actions spread out over a couple of hours. Better had a BMS do this for me . But I wanted to experience this myself once. Result: 10mV unbalance at ~3.65Vpc. I kept the BMS attached all the time. It was easy to get the cells out, the case is undamaged.

*Cycling from 100% SoC (3.45Vpc) to 95% and back - imbalance was compared after charging at roughly same battery settling voltage.


Photo: Cell number 1 [edit: actually number 2] discharging.

 

[Trkarl]

I hope the BMS doesn't just undo the balancing you did. If it was making it worse before it might just do it again.

 

[Johan]

I hope the BMS doesn't just undo the balancing you did. If it was making it worse before it might just do it again.

Unfortunately I do not have time right now to swap BMSs. I sold the EG4 charger and bought a Wanptek KPS6020D (max 60V, max 20A) on Alibaba for charging to 3.40Vpc. *So far* no cell exceeds 3.65Vpc, the SoC "only" reaches 99%, which is fine for me now.

 

[OffGridPhilosophy]

Thank you for doing this so we didn't have to! ?
I was wondering what it looked like in there!

 

[Microbiologycory@gmail]

Hey guys,
Having some difficulty with EG4’s 48V-100Ah WP battery and SS customer service. I received the battery 2 mo. Ago, app reported 54% SOC, I installed into an ATV that pulls no more than 100Amps and made a 200yd circle, battery died. BMS tripped hard and EG4 smart charger wouldn’t wake it up, even days later. Got a buddy’s 48V “non-smart” charger and it did the trick. Battery will only charge up to 3.333V (53.2V) before smart charger turns off. Cells balance perfectly within 24-hrs to within .001V but that doesn’t matter when I’m only able to get 1/2 capacity out of this battery! Customer service was notified the flat I ghad t the battery and it died, they sat on the tech support ticket for weeks before scheduling a call, which they didn’t show up to and didn’t bother trying to reschedule.(by that time I’d woken up the battery but not realized it was severely handicapped). So now they’re having me troubleshoot their apps reported voltage with my multimeters, and discharge/recharge with another charger to erase me meet effect…no dice. The tech said “it’s tricky to know what’s up without knowing how many cycles on the battery” molting maybe I’d abused the battery during the 2 mo I’ve had it…which couldn’t been avoided if they’d FUP with me in the first place when the battery died. I have a hunch it’s a faulty BMS. I’m comfortable cutting into it and removing the power flow (charge/discharge) to BMS if they won’t repair/replace it. Maybe just use the cell balancing occasionally when it’s sitting topped up.

Any ideas?

 

[Dave_11]

Anyone know if the newer version with 200A BMS still has the balancing issue?

 

[Johan]

Update:

• Only done a couple dozen cycles so far.
• Always kept SoC between 50% and 90% SoC
• while stored in climate-controled space.
• I still have the version with the 1st gen (?) BMS.

Result: Extremely out of balance, only 3.6 out of 5.1 kWh available.

• Replaced with JK BMS 200A 2A active balancer, model B2A20S20P (not sure if same model that EG4 used as replacement).
• Took somewhere around 7-10 fulltime days (I lost count) of 2A active balancing to get the cells withing 10mV.

Available capacity after this balancing and charging to 3.45Vpc: 4.8kWh.

 

[LydMekk]

JK for the win ^^

 

[MURPH-COMPLEX SYSTEMS]

So far I can only tell you this (quick and dirty):

The balancing LEDs sometimes turn on when I do CC-CV-charging to 54.26V (roughly 3.4V per cell on average). This battery voltage was read from the app, which matches my power supply @ 0.1V difference. However, sometimes they did not turn on at that battery voltage or they just blink a bit. So far I always observed the balance LEDs to be on consistently when CC-CV charging to 55.2V (3.45Vpc), so for now I balance at that voltage. But even then it still goes into protect mode sometimes because one cell hits 3.9V .

Where can I still buy the refurbished ones bc I have four that I just bought in April as part of an 24v to 48v upgrade and I was on a budget so that is the battery that signature solar company in Texas recommended telling me they had plenty In stock and that since it was a legacy product that EG4 would manufacture thay particular product forever. As I've come to find out this is the farthest from the truth..

 

[Johan]

Update: EG4 BMS low-voltage protection does not work. BMS allowed battery to discharge to 30.5V / 16 = 1.9Vpc (avg) while Specs claim LVCO @ 43.2V / 16 = 2.7Vpc and cell protection = 2.3Vpc.

[Edit: The discharge process happened during the last half year, starting @ SoC = 100% @ usable capacity of ~2.5kWh or ~50Ah. I accidentally left my Victron Smartshunt connected (draws <1mA according to specs). The Smartshunt was still operating! So, the BMS did not perform LVCO. This load alone should not have discharged the battery during half a year (runtime should have been 50Ah / ~0.001A = ~50,000h = ~5 years, unless Victron's specs are inaccurate). Maybe the small current drawn by the Smartshunt kept the BMS from going into a low power mode so that actually, the BMS was the biggest load. Note that my other identical EG4-WP (but with BMS swapped for JK BMS, without smartshunt or ather external loads) was not discharged at all.]

[Edit2: Maybe 1mA is below the discharge detection treshold of the BMS just like TheOffGridGarage noticed at some other BMS]

 

[42OhmsPA]

Update: EG4 BMS low-voltage protection does not work. BMS allowed battery to discharge to 30.5V / 16 = 1.9Vpc (avg) while Specs claim LVCO @ 43.2V / 16 = 2.7Vpc and cell protection = 2.3Vpc.

Did it sit for a long time in storage?

 

[fromport]

Photo: Cell number 1 discharging.

Isn't that #2 discharging ?

 

[Johan]

Did it sit for a long time in storage?

Roughly half a year, but that should not have caused this, see also the [edit] section in my previous post that I just added.

 

[jorby]

I have opened 16 of this type of battery case by wedging/ prying at the corners (one corner at a time) with screwdrivers. Start with a thin screwdriver then wedge in thicker ones. Once the corner starts to let go, work halfway down the end then along the long side. All of the cases came open and the cover goes back on without looking hacked up. Some were pretty easy and some were a lot of effort, depending on how much glue was used. One of them had no glue on one whole side. So much for being waterproof...

 

[42OhmsPA]

This happened in the last half year. I accidentally left my Victron Smartshunt connected (drawing <1mA according to specs). This load alone should not have discharged the battery during the past half year (100Ah / 0.001A = 10,000h = 11 years), but maybe the small current kept the BMS from going into a low power mode so that actually, the BMS was the biggest load. Note that my other identical EG4-WP (but with BMS swapped for JK BMS, without smartshunt or ather external loads) was not discharged at all.

Sounds like a crap BMS to me.
Any warranty?

 

[Johan]

Sounds like a crap BMS to me.
Any warranty?

The 2nd EG-4 WP was already given to me under warranty but is the same.

 

[Hanzo86]

I have a similar issue with my EG4 WP. It will go into protection mode when connecting to inverter/charger every so often. Hook it up to a bench power supply and start with from 48-52v @ 1amp and slowly increase until it wakes and increase amperage to desired charge rate.