DALY BMS Teardown & Failure Review (16S 40A)

[methods]

Okay our big battery tester is too slow to reset after the BMS blows open. I'm going to set up a far more aggressive test which is very representative

* Meanwell power brick set to 20 amp charge

What we are going to do now here is check the hysteresis of the hvc protection. You can trigger hvc hundreds of times before you even notice.

HUNDREDS OF TIMES EVERY DAY

Because....
Bro.... We are not trickle charging at 3 amps over here, lol. We have megawatts coming out of the sky. And just the last few months I have pulled over a megawatt hour for my well pump alone.

When you have solar...
You charge the battery as hard as you can! My charge limits are set at 100 amps. This is for my 100 amp hour+ batteries.

So
Think about whether you trust a device like this to blow open 100 amps, 1000 times+ a day.

C O U G H

You can't have the same production line producing High reliability and low reliability.

-methods

 

[methods]

Oh have Mercy!
I just realized this turd won't read backwards!!! It won't show negative current!!!

Never buy one of these

* It can only remember with external power supply
* It's watt hour calculation rolls almost immediately
* It can only measure positive current
* And of course it is super inaccurate

JENKEM
Not a friend of the revolution



I was sitting here thinking that I couldn't get a charge current... So I hooked up the heat gun to blow off a few hundred Watt hours... And I saw nothing on the meter

* The Cycle Analyst is the dream tool for this job

-methods

 

[methods]

The unit already failed

* I burned off some charge with a 30 amp discharge
* I started some charge at like 19 amps
* The unit is already getting super hot
* I can measure the exact same failure at the output

I show it in the video

The charge mosfets have shorted. I am now charging right to the body diode of the discharge set.

There's nothing to do with these
I'm going to take all the rest of the new ones outside and beat them flat with a hammer

(Video uploading)

-methods

 

[methods]

Discharging through the failed BMS at 30 amps to check heating. In this direction it shouldn't get hot.

I'm feeling a little bit of warming but not a ton.

It's now starting the thermal soak

Oh yeah this won't do. I'm going to go back up all the rest of the hammer right now so no one else gets these

This is f****** hazardous

-methods

 

[methods]

 

[methods]

As soon as the failure video finishes compressing I will post that, then I will document the destruction of all of these units.

-methods

 

[methods]

The deed is done

The last failure I will tolerate
[youtube]sACjIbGoIXY[/youtube]

Leading up to this video


-methods

 

[methods]

And that concludes this thread.

[youtube]b3lKlv81oQ4[/youtube]

I'll see you guys on the flip side

-methods

 

[fechter]

It might start working OK after getting smacked with the hammer.

Time to move on to a better model. I have some here I used with good results, but don't remember the name. They were cheap, but at least they worked at the rated current and haven't failed.

 

[DogDipstick]

* The Cycle Analyst is the dream tool for this job

-methods

You can even datalog your results and post them here for us dumbos.


Honestly having a CAv3 shunt, datalogger, and DAQ system is priceless to me. Sitting there. Load bank and power pack. I can straight up proove my builds.. upwards of 30kW easy.... I can datalog capacity tests, discharge currents, watt hours, averages, SOC capacity vs V, ect. All that fun stuff. All neatly laid out on a little graph or presented in a spreadsheet text file to manipulate with Excell. 400v+ and 800a +.... or even 4v and 8A... etc...

IDKY anyone would buy those little wattmeters if they havea cycle analyst abailible to purchase. CAv3 setup as a battery testing tool is more useful than my 1200$ FLIR datalogger... and more channels too, more inputs. With that I can have any one of 10-15 traces, 3/sec ( 3 Hz or less,) logged, but not on spreadsheet and only look at one at once, when exported. CAv3 does it all, (albeit at 1Hz or 0.1Hz, but thats good enuf for me).

 

[john61ct]

Are you running stock firmware on that CA or a specialist DAQ version?

 

[mxlemming]

042N08 MOSFET is 80V 4.2mohm. Standard way of marking them.

The brand... Who knows.

 

[Bbbbrass]

Love the info, even if I only understand fully about 20%. I have a few earlier versions, and they pass short circuit test and don’t get hot. You said to ask, so what was it that led to this tear down that we have all benefitted from?

 

[methods]

I just realized that I had a batch of posts to respond to from much earlier. I was in a frenzy and never got back to it.

042N08 MOSFET is 80V 4.2mohm. Standard way of marking them.

The brand... Who knows.

Well... an 80V 4.2mOhm mosfet would be great.
These are not 80V 4.2mOhm mosfets. Whatever process resulted in their use sourced them from a supplier who does not perform avalanche testing. These were probably 50V or 60V mosfets - WHICH MAKES SENSE - because there is probably some metric in their testing around how hot the units get. Running lower impedance mosfet gets them past that test... and it is so much cheaper/easier to get a 4.2mOhm mosfet in low voltage than high.

Humpf - in the past
I worked a side job this weekend to offset the cost.

Love the info, even if I only understand fully about 20%. I have a few earlier versions, and they pass short circuit test and don’t get hot. You said to ask, so what was it that led to this tear down that we have all benefitted from?

I am building Solar Farms
* micro Solar Farm
* mini Solar Farm
* Solar Farm

I am using all recycled, upcycled, "bad" batteries. Every single one of them has some problem or another... so the BMS is SOOOPER important. The SPF 5000 ES is known for being very hard on BMS's so I decided to set up a qualification test (after I already had units in the field). Every test failed...

DALY has an incredible responsibility to keep people/property safe. Someone there is unqualified.
* Maybe in Management / Business
* Maybe in Procurement / Supply chain
* Maybe in Quality / Inspection
* Maybe in Engineering
* ....

Somebody made a choice to do a run of BMS's using hazardous parts. AND OF COURSE -> My "lot" has no bearing on the next lot, those that came before and after are probably great. No reason for them not to be

* other than the hand-populated from junk parts pcb C O U G H

so
If I were Quality Engineer over at DALY I would clean up my mosfet supply chain. YEA, you could put the test at system level but it would just result in endless rework. Of course they can test the boards while they are still unpotted, but - why.

* Pedigree Parts

American Buyers have to drive that or it wont happen.

GRRRR!!


-methods

-methods

 

[Allex]

Ever tried Ant BMS ?
They come in different flavors and all with bluetooth.
Personally I like them better but this is just because the balancing function works better on them. And they are usually cheaper than Daly

 

[methods]

I will buy one right now.
Can I get them off Amazon Prime or do I have to go to someone's website?

-methods

 

[methods]

I found some that go up to 16s
I found some that go up to 20s and down to 8s

I found someone private websites and on eBay.

.... I think I need to take my own advice and pay a little more. To be fair I'm looking at "one's" pricing and it may be a better price when I buy 50 pcs.

* Raising my budget by $20

I don't even know what my budget is... I have to set a metric

* $24 too cheap for 16S 40A
* $100 too expensive either way

* $80 barf
* $60 ack
* $42 yes

My budget is $42
Maybe that's gross maybe that's shipped, maybe that's for one maybe that's for 100pcs

I am going to pay $42 US dollars for a 16S life BMS that will survive short circuit current and give me 20 amps all day without complaining. It will balance at any rate, I don't care if it has Bluetooth or not.

If it doesn't have margin in the form of programmability, then it has to pass margin testing. This will be full loud slamming

* Slamming LVC at full load
* Slamming LVC at burst load
* Slamming LVC at short circuit

* Slamming overtemp at all loads
* Slamming HVC at all loads

Sigh... Guessing it will be a 60A rated part to give me 20A service.

-methods

 

[Bbbbrass]

Thanks for the reply! I have micro solar, and many batteries for back up power, camper van, golf cart, and just random projects. Anything with more than 2 cells gets a BMS (except for a 1 off 3S powered airhorn for a prank...).

I really like the JBD (overkill, RadioB), and would love your take on it. I work for no one, but here are links I have used.

$22 (plus $8 shipping), ordered but haven't tried out yet.
https://batteryhookup.com/collections/accessories/products/li-ion-lifepo4-3s-20s-30a-bms-with-balance?variant=42283333845154

$62 for double capacity, bluetooth, and silicon leads. I've used these from other vendors and they have been great. Balance is still a bit weak but it's fine if you IR test and use a pack builder like repackr or cell saviors.

 

[Allex]

I will buy one right now.
Can I get them off Amazon Prime or do I have to go to someone's website?

-methods

I get them of Ali. ICgogogo store there is legit
Up to 24S and up to 450A
They come in many different shapes also, just search for "ant bms" on Aliexpress
"Big ones"
https://www.aliexpress.us/item/2251...00022995957694!sea&curPageLogUid=bHmdeHbcu9h2

"Small ones"
https://www.aliexpress.com/item/330...b516589548143101961e33d6!12000022368141650!im

Would be interesting to see what you think of them.

 

[Bbbbrass]

My budget is $42
Maybe that's gross maybe that's shipped, maybe that's for one maybe that's for 100pcs

Here are some photos of the $22 10S "30A" BMS


I got from BH[/url]https://batteryhookup.com/collections/accessories/products/li-ion-lifepo4-3s-20s-30a-bms-with-balance?variant=42283333845154[/url].

They are Xiaoxiang units, which I believe are also JBD.
Some MOSFET closeups as best I could.

 

[lnanek]

Looks like 120A rated MOSFETs, similar to ANT:
https://datasheet.lcsc.com/lcsc/1912111437_CRMICRO-CRST055N08N_C410925.pdf

Nice they included so much headroom on the components.

 

[Bbbbrass]

This just arrived from Amazon. $20
https://www.amazon.com/gp/product/B0B25FXXFP/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

I have low expectations, but am wondering if I can add a temperature sensor.

Clearest photos I can get of the internals.

From what I can find and tell, not quite as good a FET as the last BMS http://www.chipsourcetek.com/Uploads/file/PE58120P.pdf

 

[lnanek]

As long as we're ripping apart random BMS now, here's a 60A BMS from BestWay:




I do like how they had a temperature sensor right in there with the MOSFETs, the heat sink was bolted down with through bolts, and the MOSFETs had easy to find datasheets and lot numbers.

 

[methods]

Keep posting
More is better

I'm almost done with the review on the unit I got from battery hookup. It's looking awesome! I got past the testing that was killing the DALY

* Force thermal shutdown during 20 amp charge off of noisy charger, SPF 5000 ES

So far I'm completely and totally stoked on the battery hookup unit. I finished all the LVC and over temp characterization. Now I am 10 amp hours into the charge testing.

* On board thermal protection works perfect
* External thermal protection works perfect
* Cell level LVC works perfect
* Low voltage lockout works perfect
* MOSFETs have a data sheet
* Board has the diodes!!!!!!!!!!
* Conformal coated
* Really clean parts machine assembly
* Nice copper bus bar
* Room for growth

If I were going to modify a BMS... This would be the one. Really excited because they are super super cheap. I went all in and bought 20 (gulp) but so far it's looking like I got hooked up and not ripped off.

SWEET

I will process all the pictures and video when the testing is complete. Keep posting pictures and videos of random BMS tear downs.

-methods

 

[methods]

* units are 1mOhm per bank and produce no heat
* Thermal pads are sticky and gooey
* Unit tears apart with only four screws
* You can visually inspect everything no tricks
* It doesn't look like they cut corners or cut cost

All the power numbers add a perfectly to one watt and 1mOhm. The MOSFETs are 5.5mOhms each, 5 in parallel, then of course two of those banks in series for common mode.

It's a little fussy.... But so long as you're not banging off of LVC constantly it will work perfect. I had to jump start mine to get it going because I had cells resting below 200 millivolts.

* I collected all the data
* I will post the characteristics tonight or tomorrow

-methods