Blew up Atorch DL24P battery tester

[Zambam]

I received my 180W DL24P a little over a month ago and It's been woking great testing LiFePo4 cells at cell levels. I went to test the capacity of a 22S 74V 20Ah LiFePo4 battery this morning. I heard a pop with a spark when I hit the Start button. I had cutoff set at 55V (2.5V/cell), discharge current set at 2.3A (below limit of 180W/74V=2.43A).

The spark was in the vicinity of the two R010's on the circuit board where it got blackened. Both R010's are open circuit. I suppose they are current shunts? I checked the 2 power devices, a MOSFET and a Schottky diode, both were shorted (MOSFET shorted between Drain and Source). The MOSFET and Schottky diode I was able to find on Aliexpress. What about the two R010's? What's the p/n? W/o being able to get two R010's, I will not attempt to repair it.

Has this happened to anyone else and were you able to fix it? I wonder if this failure is covered under warranty? I wrote to Atorch. Waiting for a response.

 

[amberwolf]

I received my 180W DL24P a little over a month ago and It's been woking great testing LiFePo4 cells at cell levels. I went to test the capacity of a 22S 74V 20Ah LiFePo4 battery this morning. I heard a pop with a spark when I hit the Start button. I had cutoff set at 55V (2.5V/cell), discharge current set at 2.3A (below limit of 180W/74V=2.43A).

The spark was in the vicinity of the two R010's on the circuit board where it got blackened. Both R010's are open circuit. I suppose they are current shunts? I checked the 2 power devices, a MOSFET and a Schottky diode, both were shorted (MOSFET shorted between Drain and Source). The MOSFET and Schottky diode I was able to find on Aliexpress. What about the two R010's? What's the p/n? W/o being able to get two R010's, I will not attempt to repair it.

Those are in the right place for shunts.

If the SD is downstream of the FET from the battery input, then the most likely sequence is that the SD failed from overvoltage, then the short circuit current thru the FET overloaded it and shorted it, then the resulting too-high current overloaded the resistors which vaporized their resistive surfaces and during the evaporation arced across the gap until it was too wide to sustain the arc, at which point the current flow stopped.

If the SD is upstream of the FET then the first two events are probably simply reversed.

The reason either part could fail from overvoltage could be a switching spike if they are switched, or if the parts are counterfeit (not uncommon), or simply underspec'd for the job.

If the FET isn't switched but instead is used linearly as a resistive element (not typical), it could easily have overheated whenever it's used and accumulated damage, and then failed at powerup this last time.

BTW, it's not uncommon for FETs to damage their gate drivers (external to the FET) when they blow up. This one doesn't appear to have actually had a plasma event (which would blow pieces off or crack it's case and let the smoke out like the resistors did), so that's less likely, but something you can check for by testing the gate pin for the modulation it should be getting from the MCU.

 

[Zambam]

Quick diagram of the parts involved. Unit is supposed to be rated from 2 - 200V @ 180W.

STPS41H100CG PDF

Part #: STPS41H100CG. Description: LOW DROP POWER SCHOTTKY RECTIFIER. File Size: 98.63 Kbytes. Manufacturer: STMicroelectronics.

www.alldatasheet.com

JCS50N20ABT-O-AB-N-B PDF

Part #: JCS50N20ABT-O-AB-N-B. Description: High frequency switch mode power supplies. File Size: 1129.78 Kbytes. Manufacturer: JILIN SINO-MICROELECTRONICS CO., LTD..

www.alldatasheet.com

Found below schematic from reply #63 in this thread Cheezeball DC Load: DL24P: Pump, or Dump ??? - Page 3

 

[amberwolf]

Quick diagram of the parts involved. Unit is supposed to be rated from 2 - 200V @ 180W.

STPS41H100CG PDF

Part #: STPS41H100CG. Description: LOW DROP POWER SCHOTTKY RECTIFIER. File Size: 98.63 Kbytes. Manufacturer: STMicroelectronics.

www.alldatasheet.com

JCS50N20ABT-O-AB-N-B PDF

Part #: JCS50N20ABT-O-AB-N-B. Description: High frequency switch mode power supplies. File Size: 1129.78 Kbytes. Manufacturer: JILIN SINO-MICROELECTRONICS CO., LTD..

www.alldatasheet.com

Based on the handdrawn diagram, then it could be using the FET in it's linear region as a controllable resistor (essentially like they're used in audio amplifiers).

You can verify that by testing the gate signal--if its a changing analog voltage, that's what it's doing.

If it's a PWM'd digital signal, then it's using the FET as a switch to short the battery out periodically (most likely a rapid pulse train).

In the former version, simple overheating of the FEt is the most likely failure cause of an operational failure, but shouldn't happen on powerup.

In the latter version, there's three easy possible causes:
--in operation (not at powerup) diode fails, and the pulses of current creating inductive spikes in the wiring and traces can now make negative-voltage spikes that exceed the reverse voltage of the FET.
--in operation or at powerup, the FET just internally fails (defect, not meeting spec, cumulative damage) shorted.
--at powerup, the MCU controlling the gate signals starts but crashes or otherwise fails to turn the FET gate *off* (or the gate driver itself fails stuck *on*), so the FET stays on, rapidly overheats and fails shorted...

Usually a FET that fails shorted is not as low a resistance as it would be when actually fully turned on, so it heats up even faster, and can vaporize enough internals to blow the plastic casing apart, leaving just vaporized metal deposition on the former die location that keeps it all shorted together (even the gate sometimes, which is where the gate drivers/etc get blown up too).

 

[Zambam]

Amberwolf, thanks for the detailed analysis! While I did not exceed its ratings, I'd think a device like this should have overload protection from destroying itself.

 

[Zambam]

Found this long 16 page eevblog discussion on the DL24P from 2020 Cheezeball DC Load: DL24P: Pump, or Dump ??? - Page 1

 

[docw009]

Thanks for the link. Something to do someday ... add a zener etc to the Mosfet.

I blew my first one up the first time I used it. I filed a dispute with Aliexpress, and was willing to take a partial credit, enough to buy the DL24p base w/o fan. That one has been working several years. Threw the old base away though, I don't have space/time to keep it,

 

[Zambam]

Thanks for the link. Something to do someday ... add a zener etc to the Mosfet.

I blew my first one up the first time I used it. I filed a dispute with Aliexpress, and was willing to take a partial credit, enough to buy the DL24p base w/o fan. That one has been working several years. Threw the old base away though, I don't have space/time to keep it,

What did you do to blow the first one up? w/o a fan what kind of heat sink did you add? What's the highest voltage/ wattage you ran through it?

 

[Zambam]

What about the two R010's? What's the p/n? W/o being able to get two R010's, I will not attempt to repair it.

Reply #23 says "The shunts are 5w, R010 or 10 mOhms". That's 0.01 ohms.

Cheezeball DC Load: DL24P: Pump, or Dump ??? - Page 1

Cheezeball DC Load: DL24P: Pump, or Dump ??? - Page 1

www.eevblog.com

edit: from the same thread found this 10W 0.01 ohm shunt

https://www.mouser.com/ProductDetail/71-WSHP2818R0100FEA

 

[Zambam]

I got the below reply (second one) from Atorch which suggests I damaged the unit by shorting + to - and exceeded the 180W limit. I wrote back that I did not do either and was very careful.

"Hello friend, when using high voltage, be very careful. When connecting the positive and negative poles, a short circuit may cause damage to the product due to high voltage. According to your description, the board is currently severely damaged, and the 74V high voltage cannot be directly set above 180W, which is also one of the reasons for damage!If you need to test high voltage above 74V, it is recommended that you replace the DL24MP product with 4 discharge tubesThank you for your inquiryBest wishes!"

Anyone know what "discharge tubes" are?

 

[Zambam]

Below was their first (hilarious reply), which basically says I need to protect my reputation by not "applying for a return process", otherwise I will be banned from any future purchase. Since they are blaming me for damaging the unit. I have no choice but to file a dispute.

"Dear customer friends, our company is a legitimate brand company, and every product is shipped after strict testing. In terms of quality, please have absolute trust. Before purchasing, please be cautious whether our products meet your needs. After purchasing, if you have any questions when receiving and using the goods, please leave a message to contact us. We will actively solve your questions. In order to protect your reputation and have the opportunity to purchase our products again in the future, we suggest that you do not apply for a return process unless absolutely necessary. Thank you!Working Hours:Monday-Saturday: 8:30-18:30"

 

[docw009]

What did you do to blow the first one up? w/o a fan what kind of heat sink did you add? What's the highest voltage/ wattage you ran through it?

I used the fan with LED's from the first unit that blew up. The credit was just enough to buy a new board with shipping.

I did nothing to blow up the first one, just connected it a battery with a 1 or 2 amp current, and Poof.

 

[Zambam]

A Russian thread on the Atorch with lot's of pics

Atorch DL24, недорогая электронная нагрузка с четырехпроводным подключением

Atorch DL24, недорогая электронная нагрузка с четырехпроводным подключением на техническом блоге. Электронное измерительное оборудование. Технические обзоры. Обзоры измерительных приборов и

www.kirich.blog

 

[amberwolf]

Amberwolf, thanks for the detailed analysis! While I did not exceed its ratings, I'd think a device like this should have overload protection from destroying itself.

You may not have exceeded the device's ratings as advertised, but it's always possible to exceed the actual real ratings of the device as those are not even known.

(maybe by the engineers that designed it, but that info isn't going to make it to the actual users...the ad-writers will, based on stuff I've seen for assorted devices, look at the highest spec on any part and call that the limit, even though other parts have lower limits, and even though the device as a whole couldn't do that even if all the parts had the same limits, because of conditions that occur during operation....

Unfortunately without reverse engineering the device and determining all the factors in it's design and usage that create lower limits than the part specs, or simply testing a bunch of them to destruction in various ways, the end user can't really know what the limits really are.

That's true of most of the stuff out there, not just testers like this or other ebike/etc related stuff.

Once upon a time (long long ago in a century far away ) most manufacturers would at least publish or have available the real specs of what they made, even if the ad copy got a little...loose. And they would be using consistent parts and have QC steps that helped ensure most devices met those specs, rejecting and either scrapping them or repairing them and retesting. Not everyone did this to the same level, and unscrupulous things happened now and then, but overall there was a higher quality standard.

But that all costs money and time, and these days many (not all) companies skip some or all of this for that reason. Making stuff so much cheaper doesn't happen without some cost somewhere along the line, so this is how the end user pays for that--by getting less reliable, often untested, devices, often claiming to do more than they really reliably can.

With *really* cheap electronics, counterfeit parts (that sometiems are not even "real") and qc-rejected parts (that should have been destroyed and recycled) supplied by truly unscrupulous people make these situations even worse. (both of these are problems also extending into batteries and cells, which are a likely cause of various fires)



Another way to make things cheaper is to leave out all the parts that aren't actually required for it to function at it's most basic level, which includes protection components (how many things do you have with no fuse?) both for itself and the end-user. So, that happens a lot in the really cheap stuff.

 

[amberwolf]

Anyone know what "discharge tubes" are?

Usually that means FETs. Probably a holdover translation from way back before transistors.

 

[amberwolf]

Below was their first (hilarious reply), which basically says I need to protect my reputation by not "applying for a return process", otherwise I will be banned from any future purchase. Since they are blaming me for damaging the unit. I have no choice but to file a dispute.

"Dear customer friends, our company is a legitimate brand company, and every product is shipped after strict testing. In terms of quality, please have absolute trust. Before purchasing, please be cautious whether our products meet your needs. After purchasing, if you have any questions when receiving and using the goods, please leave a message to contact us. We will actively solve your questions. In order to protect your reputation and have the opportunity to purchase our products again in the future, we suggest that you do not apply for a return process unless absolutely necessary. Thank you!Working Hours:Monday-Saturday: 8:30-18:30"

These days, any seller, of any kind from any culture or nation, that claims to test their products, especially fi they say they do 100% testing, is almost certainly lying and almost certainly does nothing but take a box out of the shipping container and slap a shipping label on it.

There probably are *some* that do some spot checks, and there might be a handful that actually test things, but I can guarantee you that virtually no one is doing 100% testing of 100% of their products--it would take a huge team of people with a lot of money tied up in test equipment, a calibration department with their own separate expensive not-often-used equipment, etc. And it would also mean that they would be throwing away a fair percentage of products they could have otherwise sold, because they failed one or more tests along the way.

No, they're not testing anything.

 

[pegaz666]

I got the below reply (second one) from Atorch which suggests I damaged the unit by shorting + to - and exceeded the 180W limit. I wrote back that I did not do either and was very careful.

"Hello friend, when using high voltage, be very careful. When connecting the positive and negative poles, a short circuit may cause damage to the product due to high voltage. According to your description, the board is currently severely damaged, and the 74V high voltage cannot be directly set above 180W, which is also one of the reasons for damage!If you need to test high voltage above 74V, it is recommended that you replace the DL24MP product with 4 discharge tubesThank you for your inquiryBest wishes!"

Anyone know what "discharge tubes" are?

Discharge tubes = Mosfets, DL24MP has 4 mosfets under the heatsink so the load is split between 4 not just on one.

“While I did not exceed its ratings… (180W)” - see aliexpress auctions - they all mention that you can use 180W limit but up to 36V, in the range of 36V-80V you can pull max 60W, above 80V-200V only 45W max is allowed. So technically speaking at your 72V you’ve exceeded the max power by 3 times…
Funny thing is they didn’t have this warning 1,5yrs ago otherwise I would not have bought mine… No idea why this voltage limitation exists though. I’ve damaged mine when I was playing with 14s (58.8V) LiIon battery. I’ve replaced the mosfet (genuine) and added a zener diode for protection but I’m a bit worried to test it again…

Does anyone know what’s the technical reason for this voltage to power limitation?


This screen shoot comes actually from Atorch’s pdf manual available here: http://en.atorch.cn/upload/20241224153104.pdf