ISL94203 communication error Is my SBMS dead ?

[Harold]

My SBMS   shut off every load and now displays

"ISL94203 communication error

Befor RAM read"

I had to do some modification on my circuit so I got everything prepared, so it gave the time for the battery to be fully charged (100% soc) when I disconnected every thing respecting the install instruction step 1 and step 2 from the user manual.

everything was disconnected (solar panel, loads), so  I then  disconnected the balancing connector and lastly the Batt+ cable.

I've done what I needed to do, then reconnected the Batt+, then the balancing connector.

I waited for the SBMS to boot, everything was going well, the 49% soc was displayed (normal), I then reconnected the solar panel, not much sun at that moment but I was expecting the SBMS to shortly show 100% soc as I knew the battery were fully charged before disconnecting.

So I put all the loads back on, I'm taking of 4 breakers with only small load at that moment, max 2A. The internet box and the Raspberry pi booted well, giving time to my girlfriend  to start a first search on google, but at that point, everything stopped.

The Raspberry pi stopped.

Checking what was wrong with it, I could also see that the internet box was also off.

At that point, navigating/checking the different menu/settings the SBMS was displaying 100%soc and CFET, EOC and DFET were the only flags highlighted. I'm not 100% surer but I think there was some kind of error displayed with an error code "26680" (not sure of that number) displayed.

I thought of a bug, so I decided to reboot the SBMS doing again the same procedure for disconnecting everything, waiting few second, then reconnecting, but a very bad surprised was displayed.

The "ISL94203 communication error Before RAM read" showed up, and the SBMS would never boot as normal again. I tried to repeat once more the same procedure with no more luck.

The only 2 things that happened when the "blackout" occurred are :

I was connecting the ADC1P to the External shunt (only Batt+ was connected to the inverter, minus wasn't yet connected). At that moment, my girl friend asked why the computer shut off, so I looked at  the SBMS and it was displaying 100%soc.

I did navigate the different menu and you already know the rest of the story.

Is my SBMS fried ? did I do something wrong ?

I pretty sure I've respected the connection order, and in fact the SBMS started well at least for some time, so the internet box would have time to synchronize and the RPi to boot and load google search page...so at least for 30 second I would say.

[Dacian Todea]

Harold,

There are only two things that can damage the SBMS40/SBMS120 one is the incorrect connection/disconnection order meaning 10pin connector is inserted or removed when there is something connected to Load+ or to PV input.

And second is if the XT1+ touches the battery+ at any time.   You should have had the external shunt connected before powering ON the SBMS as SBMS will require a zero offset calibration anyway for that to work correctly and there is less risk to touch anything like the XT1+ during installation.

Based on your description the XT1+ touching the battery+ maybe the current shunt is the most possible scenario. That XT1+ is for battery temperature sensor and only has 2.5V DC connected directly to the ISL94203 thus if it will touch the battery+ the ISL94203 will get damaged.

The ISL94203 will need to be replaced. I can do that for you free of charge but shipping from you and back will take some time.

[Harold]

I'm so disappointed.

I've checked my wiring and One of the shrink tube on the XT1+ line was gone and of course it is actually close to the Shunt, so ....It seems the second solution occurred.

I remember one of the SBMS owner showed on a youtube video how to replace that ISL94203, but I can't find it. It was on an RV pickup truck I think. Would you have the link maybe ?

[Dacian Todea]

I think you refer to this video showing the ISL94203 being replaced on an SBMS60 https://www.youtube.com/watch?v=7RdbXw5mJsA 

[Harold]

That's it !

You're the Champion indeed !

[Dacian Todea]

I used Sn42Bi57Ag1  / Sn42Bi57.6Ag0.4  low temperature solder paste melting point of 139C pleas do not use Lead based solder even if that is what it was used in that video is a bad idea as even small amount of Lead contamination will change the proprieties of the SnBi solder in a bad way.

  

[Harold]

is it possible to open the SBMS120 the same or in a similar way the SBMS60 opens ?

Digikey and Mouser seem to get the ISL94203 available on stock (many references, need to check the exact reference needed).

But where do you find the Sn42Bi57Ag1  / Sn42Bi57.6Ag0.4 solder paste ? I'm not able to find it anywhere (I'm in Europe/France)

Would I need a specific flux or any other specific tools ?

Would it be possible, better or worst to preheat the board from underneath ? What maximum preheat temperature if any good ?

[Dacian Todea]

Yes it will open in a similar way just remove the screws on the back they are T10

There is no need to preheat the board as the low melting temperature should make it easy for you to remove and then replace the ISL94203

Any ISL94203 will be the same the different codes after the name only refers to what sort of package they come in  https://www.digikey.ca/product-detail/en/renesas-electronics-america-inc/ISL94203IRTZ-T7A/ISL94203IRTZ-T7ACT-ND/5415112

This seems to be the lowest cost SnBi solder paste https://www.digikey.ca/product-detail/en/sra-soldering-products/SSLTNC-10G-SQ/2260-SSLTNC-10G-SQ-ND/10709833

Or you can just get a bit of solder wire https://www.digikey.ca/product-detail/en/chip-quik-inc/SMDSWLT-040-10G/SMDSWLT-04010G-ND/9836741 then add a bit of that very tiny amount with the soldering iron to the pads.

Any flux should be good if you already have some if not you can get this https://www.digikey.ca/product-detail/en/chip-quik-inc/SMD291NL/SMD291NL-ND/1160000

The flux will be helpful in removing any solder bridges or adding solder to pins if needed. You can just drag the iron tip along the side of the IC and it will most likely remove all solder bridges no need to try and remove them individually as shown in that video.

[Harold]

What about my solder tip ?

I use it without any care about lead or non lead solder, but you said :

"as even small amount of Lead contamination will change the proprieties of the SnBi solder in a bad way.

Would cleaning my everyday solder tip in a specific (need to explain how) way (need to explain how) would be enough to not contaminate the new solder ?

[Dacian Todea]

The SnBi contaminated even with small amount of lead will result in reduced melting temperature and maybe increase possibility of whiskers growing.

I will use a new iron tip if you have one if not maybe clean that by melting lead free solder a few times then clean that on a sponge.

I have not used Lead based solder in almost 20 years (probably around 17 year) so I forget sometimes that people still use that. 

[Harold]

I'm not professional electronic, and so, I usually use what I have at hand. Maybe there is no lead, but there might be. Who knows ?

I'll see what I can do to get new tip or at least good clean.

[Harold]

I opened the SBMS for a check and I wonder if the board is damaged.

What do you think ?

Do you think, the new solder will do good after a cleaning ?

[Harold]

I've chexked again closely and the XT1 (pin20) pcb trace has vaporized and poped up close to the ISL942003.

I wonder if it will be possible to make a fix in a so tiny space with a 48TQFN package that has no access to its pin from the side.

Maybe, I can still replace the chip without fixing the XT1+ trace, so hopefully I would retrieve my SBMS120 alive but without external temperature.

Another point, Vref on pin19 seems to be shorted to ground, maybe from inside the ISL94203. I hope it won't have destroyed anything else on the board.

[Dacian Todea]

Never seen that before. There where other cases where XT1 touched the battery+ but it was just at most a visible mark on the ISL94203 the PCB trace was never affected like this.

I should be able to fix it likely also be able to add a thin wire for the XT1 if PCB trace is unusable.

There will be quite a bit of internal damage and yes Vref is used for XT1+ and that will be damaged also inside the ISL94203.

Contact me by email to provide you with my address and instructions on how to ship the SBMS

[Harold]

What do you mean by "There will be quite a bit of internal damage" ?

Do you mean inside the ISL94203 or other damages on the board or other components ?

I'm already on my  way to order the parts and try a fix by myself. If I could get the parts in 48 hours, if I success, I could get back on the road again next week end.

Do you think I would need anything else from the list you gave me previously ?

[Dacian Todea]

Unlikely there will be any other components damaged except for the ISL94203.

If you have any problem repairing it you can ship it to me.

Make sure you also order some SnBi solder wire or paste do not use a soldering iron tip that was used with Lead before as that will charge the proprieties of this solder and not in a good way.

[Harold]

Hi Dacian.

Do you think I can get similar product from MOUSER. DIGIKEY don't make it easy for non company/individual to order from overseas.

I can find the ISL94203 at MOUSER, but can't  find similar references you gave me before for soldering paste, soldering wire and flux. Do you think I can find equivalent product at MOUSER.

Shipping back to you might be quite expensive and it will take time.

Is only the new SBMS0/DSSR20 available on order by now ? Don't you have any spare SBMS 60 or 120 on stock ?

[Dacian Todea]

Here is a similar product on Mouser https://www.mouser.ca/ProductDetail/Chip-Quik/TS391LT?qs=sGAEpiMZZMve4%2FbfQkoj%252BBA3qFUvi2vSKDTi7lrAzwk%3D

I do have a few SBMS40 and SBMS120 but they are not listed as I only have a few of them and they where replaced by SBMS0 + DSSR20 that have the exact same functionality.

[Harold]

I had to heat up quite high to get the chip off the board (220°C I think)

Here are some photos about what I got just after removing the chip, then after cleaning and lastly with a piece of wire prepared to be soldered in place of the missing pcb trace missing.

Do you think there is still enough solder mask in between the neighbour pin 19 and 21 so the solder won't bridge ?

[Dacian Todea]

Harold,

220°C is not a problem it depends on the hot air gun and yes temperature should be higher than melting temperature of about 140°C

Yes it should not be a problem just make sure that copper wire is not to tall preventing the other pins to make contact.

[Harold]

Hi Dacian.

I just received all the necessary stuff to make my repair. (ISL94203 replacement)

It took me many try before getting something looking ok, first try was a fail, I needed to take it apart again unsolder, and do it again.

It's now looking ok but I find pin 42, 46 and 47 shorted all together.

Looking at the schematics they shouldn't be. Is it possible I get a bridge underneath ? I can't see anything obvious from the picture I posted.

Is it possible I've already destroyed the new  ISL94203 in one of the previous try ?

[Dhowman]

This might help (?)

[Barry Timm]

Oh, my goodness......great video! That is so above my capabilities..... It's funny! 

I am in awe of folks able to work with such tiny components and get things like this fixed. 

[Dacian Todea]

Harold,

The soldering looks great. Maybe is the 43,46,47 as yes they all go to battery current shunt so they will appear as shorted but they are not.

Have you tried to powered the IC ? Do measure 2.5V at RGO and 1.8V at Vref if so all should be working.

Check the soldering on pins 25,26 and 27 if they are soldered as those are the I2C communicating with the micro controller 

[Harold]

His there a way to power it without reassembling it all so it's possible to access to the pin for voltage check ?

What voltage and how can I power it up easily (I have a small CC/CV power supply). Np special order as for connecting with all the battery , solar panel and  load ?

Here are some more pictures.

[Dacian Todea]

The solder joints look good there is just a trance looks like on pin 18 that may be cut in some way you should check the continuity of that line as it is very important.

Once you confirm that trace is OK or correct it you can power the SBMS with say 8V and current limit of 500mA using the 10pin connector pin 1 and 2 are GND and the Batt+ for positive not other connections will be needed and you should not get the ISL94203 error but there will be multiple other errors as the cells are not connected but ISL94203 is powered.

There is no need to assemble the SBMS40 or SBMS120 you can attach the LCD if you want. 

[Harold]

I powered it up and it was drawing between a reasonnable 87-107mA current on the CC/CV workbench PSU, so I plugged in 4 LTO cells and here is how it looks like (see photo).

The only point is that I don't remember the "Stat" number was 24579 before !

Is it telling some kind of other error I should be aware of ?

The cells I'm using are "junk reclaimed" cells and one of them is over 4V, ans so OV and OVLK are highlighted, it may explain why the strange "Stat" number

[Harold]

Well. Finally, everything seems to work  perfectly. I had an unbalanced cell that was creating the problem and that strange "Stat" number with OV and OVLK.

[Barry Timm]

Great news. I've been following your thread with interest. Happy you got it fixed!

[Harold]

Good to here you had some interest in the repair.

I must admit, I had to do many attempt before succeeding. I was almost to give up.

Then I tried and watched many videos to understand how these tiny chip could be soldered. It was once misaligned (1 pin staggered from 1 pcb trace on one side at on try, it was very discouraging.

But then, seeing my SBMS back to life again was very nice. Needless to say I've spent many hours on it to get it right. Finding the way to get the right amount of solder, specially on the ground  plane, was quite challenging.

So I heated the piece, then gently pressed it on the PCB to expulse the excess of solder, then release the pressure to let the tension surface of the melted solder do the rest of the job by getting the righ tamount of solder, then releave the heat.

I also preheated under the PBC to 50°C with a 10W resistor when replacing the chip. I didn't do it first when taking the damaged chip, but it took a long time to reach the proper temperature to desolder it.

[Dacian Todea]

Harold,

Glad to here it works. Yes that is the correct method to squeeze the excess solder out from the GND pad.

I'm guessing you used low temperature SnBi solder and you should not have needed any preheat resistor. 

[Harold]

I used that flux and that solder.

It made it much faster and easy (to my feeling) with preheat. The PCB seems quite massive and absord quite some of the heat !!

[Harold]

By the way Dacian.

Thanks for the help and advices provided.

[Harold]

here the pics :

[Harold]

I don't know why I must cut my message into pieces, otherwise it won't publish it !

[Harold]

Hello Dacian.

My house has been visited by burglars and they pulled out the cables.

Fortunately, Batteries, inverter and SBMS120 were still present.

But when I powered back the SBMS120, that horrible message "ISL94203 communication error" appeared.

I managed to replace the ISL94230 and now the horrible message had disappeared but the SBMS120 show a strange behavior as the parameters always show the same even after modifying them and power cycle (SBMS120 is now powered between 10pin connector 1GND, 2 and 10Batt+).

Parameters always come to this after modification and power cycle :

It also shows a lot of strange errors :

Can you says if that is normal or if there is still something wrong ?

See pics in the previous 2 messages as I couldn't the entire message in 1 shot !

[Dacian Todea (electrodacus)]

Harold,

Sorry to hear about that.

A photo of the first monitoring page will be useful.

But looks like maybe a few pins on the ISL94203 are not soldered or are shorted.

[Harold]

A few pins seem shorted with ohmeter (not visually) but it sounds normal as per the schématics of the ISL94203 :

 26 with 27, 

27 with 28

46 with 47 and with 43

Here is the 1/5 monitoring page :

 

I still need to check the 2.5V at RGO and 1.8V at Vref

This is the second replacement of the ISL94203 as in 03/2020, I blown it myself due to bad manipulation.

[Harold]

I've just checked :

U_RGO = 2.5V

U_Vref = 1.8V

What could be wrong ?

[Harold]

Hi.

I've redone and rechecked all the weldings.

Still the same behavior and errors. Parameters wouldn't stay as programmed and stored.

What could be wrong ? I'm pretty sure I've no shorts.

What could I check further ?

By the way Dacian, I got your last response in my email, but it didn't show on the forum !

[Dacian Todea (electrodacus)]

Of course parameters will not stay as you program them as there is no proper communication between the ISL94203 and micro controller.

It seems that everything is soldered except for that one pin measuring the external temperature.

Looks like you used solder paste and is possible that it did not melt properly under the IC and could short the ground pad to some of the pins.

So you can test between ground and each pin to see if you have short circuits. If you do you can just add some flux and reheat with hot air the IC and that should fix the internal shorts if you find that there are some.

I'm assuming you did not bump some other sounding passive components like maybe a pull up for the I2C

[Harold]

When I said same errors in my previous message, it was these errors as on the picture below ! Not ISL94203 communication error.

This one has now disappeared ! Communication with the new ISL94203 seems to be ok !

 There doesn't seem any shorts between GND and other pins except 24, 28, 29 and 18 which is normal I think considering the schematic.

[Harold]

Le lundi 27 mai 2024 à 22:48:20 UTC+2, Harold a écrit :

[Dacian Todea (electrodacus)]

In your last photo I can see just above the ISL94203 a resistor (top right corner of the photo) that I think is the pull-up for the I2C and it looks like it is not soldered on one side (it may be just the lighting) but please check that.

[Harold]

I think every thing is soldered.

[Dacian Todea (electrodacus)]

One thing that I did not asked is if the 10pin ribbon cable is properly connected and not broken.  That will result in the same type of errors.

At pin 48 there is a 1kOhm resistor. Can you please check if that is still OK. Many times when ISL94203 gets damaged that resistor can also be damaged (open circuit).

[Harold]

R72=1kOhm at pin 48 is good.

I powered the SBMS with 8V-100mA from CV-CC bench power supply between on 10pin connector. pin 1-2 to GND and 10 to +

Then I tried with 3 cells and same problem.

[Dacian Todea (electrodacus)]

It will not work with a power supply. It should work with 3 cells if connected correctly.

For 3 cells you will connect cell 1,2 and 8

you have 1 and 2 to GND (cell 1 negative) then wire 3 to cell 1 positive, wire 4,5,6,7,8,9 to cell 2 positive and wire 10 to cell 8 positive but also Batt+ connector needs to be connected here.

All cells need to be above 3V

[Harold]

Hi.

I give it the the power through main connector Batt+ on top of the 10pin connector with the 3 cells at 3.6V-3.9V correctly connected as you described.

But same problem !

Any idea where to check ?

Do you think the ISL94203 is already dead maybe due to a bad manipulation ?

At first try, I was very self-confident my welding were good (but, maybe they where not) and powered the SBMS through USB thinking it would power the ISL94203.

Would it be possible the ISL94203 has been destroyed when powered at that moment with USB and the welding not so good ?

I have some ISL94202 as drop in replacement from Renesas/Intersil as ISL94203 seem to be on the end of life. Can you confirm I can do the switch from 94203 to 94202.

Thank you very much for your assistance.

[Dacian Todea (electrodacus)]

ISL94203 and ISL94202 are basically the same thing.

But your solder points to all pins seems OK only if there is a short circuit somewhere or any component around not soldered.

It will not work on USB power as there only the micro controller is powered not the ISL94203 so ISL94203 could not be damaged while powered over USB.

When powered trough the 10pin ribbon cable + the Batt+ connector it should work but only is connected correctly else you will get similar sort of errors.

It makes no sense to attempt to program the ISL94203 / 94202 unless you see the cell voltages and the default parameter that will be 3s Li-ion as programed at factory.

I do not think attempting to solder a new IC will be helpful as it is unlikely that is the problem. Double check the cells connections by measuring voltages at the cell balancing resistors so you should be able to see the 3.7V or so for cell 1 and 2 and then another 3.7V for cell 8.

You can also ship it to me and I will be able to let you know what was the problem as it will be easier for me to diagnose if I have access to it.

[Harold]

Hello Dacian.

Here is a check of the voltages between each pins with 3 cells.

Have a look at the video at this link.

As I have some ISL94202 on their way, I'll try once more, then think about sending  it back to you.

Would you have an idea of the cost of the repair + shipping back to France ?

[Dacian Todea (electrodacus)]

Yes the 3s battery seems to be connected correctly.

Repair will be free but shipping to France is a bit expensive.

If you remove the LCD and just ship the main board the parcel should be under 500g so lowest shipping will be around 35 to 40CAD.

[Harold]

Hello Dacian.

I've replaced the non functionnal ISL94203 with the newly arrived ISL94202 and the SBMS120 light up almost as normal with each cell voltages on screen.

But when I swicth CFET and DFET to ON, it immediately go back to OFF.

I think the reason is the internal temperature showing 99°C.

Plus I don't fully understand the schematics on the SBMS's manual (page 44).

Is TH1_THERMISTOR in parallel with R70 ? So it is the reason R70 measure 4.6kOhm when tested with ohmeter.

[Dacian Todea (electrodacus)]

TEMPO pin 22 on ISL94203  may not be properly soldered that is why both internal and external temp sensors show +99C

[Harold]

TEMPO pin 22 and XT2 pin21 seem well welded.

But I didn't weld XT1 pin20 isn't welded right now as it's been damaged in a previous mishandling and hard to weld

Is it possible, that not welding XT1 pin20 disturb both IT and ET

[Dacian Todea (electrodacus)]

The XT1 pad missing will only affect the external temperature not the internal one and you should not set an external temperature limit in this case.

Either XT2 or TEMPO is not making contact for the internal temperature to show as +99C

Shorting XT2 to GND will show +99C also so check that XT2 and TEMPO are not shorted to GND.

TEMPO outputs pulses of 2.5V when measurement is done so XT2 will be some voltage based on the resistor divider that includes the Thermistor.  

If Thermistor is open circuit  or very high resistance value it will show -45C and if Thermistor is very low value or shorted it will show +99C

There are 0V at XT2 at all times meaning is either not soldered so open circuit, or shorted to GND more likely as you mention you can see a good solder joint.