diy solar

diy solar

problem with capacity tester?

My tester has two diodes (Dula diode common Cathode) to the left of the MOSFET (IRFP260).

Same sort of failures showed up in aerospace applications.

This power MOSFET spec'd to dissipate up to 280W, suitably heatsinked.
Safe Operating Area (SOA) shown for various pulse widths, but graph does not include lines for DC as data sheets used to show.
Device is constructed with trenches for extra area and behaves like many MOSFETs in parallel. Problem is thermal runaway where one area gets hotter, carries more than its share of current, gets hotter, eventually fails.

iPRF260 SOA.jpg





Figure 6 of NASA paper shows additional lines sloping downward, cutting off at some Vds each of the pulsewidth lines.
I saw reports that even where modified SOA said it should be OK, thermal runaway occurred.
MOSFETs either have to be kept quite cool or be operated well within, perhaps orders of magnitude within SOA.

Sounds like these battery testers are being rated 150W or 180W with a 260W transistor (rating assumes case held at 25 degrees C)
With 150 degree C max junction temperature, 260W indicates 2.2 W/degree C.

"Linear Derating Factor 2.2 W/°C"


This simply explains that power dissipation decreases with ambient temperature, because not as much power is needed at higher ambient temperature to reach maximum junction temperature. Same analysis I always do for any electronic component. It does not address thermal runaway during linear operation.


"Some (usually expensive) MOSFETs are specified for operation in the linear region and include DC SOA diagrams, e.g. IXYS IXTK8N150L."


When you use your battery tester for higher voltage packs instead of single cells, more likely to violate the modified safe operating area.
The Vishay part identified doesn't document a safe DC operating point.
You could try repairing your failed boards with a transistor that does specify DC operation. Also allow more margin especially at higher voltage.
 
so that burnt piece is a MOSFET? (been reading the post behind this one too/ while below par on understanding electronic boards) ... Does anyone think on could replace that burnt MOSFET piece, and possibly get a the Ah Tester working again?
Fortunately i had a 150W spare tester(a dial up one rather than push button setting). My capacity testing is at about an end...i'm on to tweaking using my inverter/charger now at higher real life amperages and i dont expect to use the working tester overly much in the future.
 
so that burnt piece is a MOSFET? (been reading the post behind this one too/ while below par on understanding electronic boards) ... Does anyone think on could replace that burnt MOSFET piece, and possibly get a the Ah Tester working again?
I replaced the MOSFET on mine, I have the same MOSFET that I use for other project which I order from DIGIKEY, you just need good 40W soldering iron. My failed MOSFET has resistance reading of about 40 Ohms between Gate and Drain, good MOSFET should show infinite resistance.
 
Maybe one of these (available from DigiKey) would work as an upgraded replacement MOSFET.
I haven't looked into gate voltage, just Vds, Id, SOA. They have "DC" included in the SOA graph.



higher current:


Looks like 40A, 12V DC, 150A at lower voltage (single cell)

That's assuming your heatsink can keep case at 25 degrees C. Water line and single-pass cooling system? Or stick it in the fridge?
 
Those are not covered by the heatsink in mine, and seem to be the source of heat that discolors the PCB. I suspect these would last much longer if those had some sort of heat dissipation.
So knowing what you know about these cheapo 150W/180W testers and the components they use, would you have a recommendation for ‘safe’ current levels to run when testing an 8S string?

Or said another way, the temperature read out on these boards seems pretty reliable - is there a temperature you would recommend staying under?
 
So knowing what you know about these cheapo 150W/180W testers and the components they use, would you have a recommendation for ‘safe’ current levels to run when testing an 8S string?

Or said another way, the temperature read out on these boards seems pretty reliable - is there a temperature you would recommend stayis ng under?
Turn the amp knobs to zero. When you turn up the knobs turn them slowly and try not to go over the rated watts.
I never had a problem using as many amps possible as long as I stayed under the max watts.
 
Turn the amp knobs to zero. When you turn up the knobs turn them slowly and try not to go over the rated watts.
I never had a problem using as many amps possible as long as I stayed under the max watts.
I think the point of this thread is that many people have burned out their units under the rated Watts, especially when driving maximum current through them testing a single cell (so only 66W or 73W depending on whether your unit maxes out at 20A or 22.2A).

So running maximum current through these appears to be a weakness (possibly because of the diodes not having any heatsink).

The units have a maximum W setting and seem to do a good job shutting down as soon as Watts exceed that setting.

What they do not have is any kind of maximum current protection nor any kind of maximum temperature protection.
 
I'll just point out that i used mine several times right on the maximum edge...took the amps up slowly, and for 6-7hours at at time....guess in the end i just over stressed it. Probably about a 1/2 doz uses before it gave up. A little kindness and it will probably do fine :ROFLMAO:

I wonder if you holes drill in the board around the mosfet and a cpu fan mounted underneath the board, if it might be happy then...that might give a better passage of air past it????
 
I'll just point out that i used mine several times right on the maximum edge...took the amps up slowly, and for 6-7hours at at time....guess in the end i just over stressed it. Probably about a 1/2 doz uses before it gave up. A little kindness and it will probably do fine :ROFLMAO:

I wonder if you holes drill in the board around the mosfet and a cpu fan mounted underneath the board, if it might be happy then...that might give a better passage of air past it????
We’re you taking Watts ‘right to the maximum edge’ or current?

When testing individual cells, you are generally maxing out on current while Watts are modest, while when testing a series string of cells as a battery, you are generally maxing out Watts with modest currents...
 
Turn the amp knobs to zero. When you turn up the knobs turn them slowly and try not to go over the rated watts.
I never had a problem using as many amps possible as long as I stayed under the max watts.
Exactly correct, move amps slowly. The MOSFET will develop hotspots internally and die (I'm told, by people who know). Usually they die as a short, that would be very bad. This failure is different and usually dies as an open. It is the thermal shock from rapid power changes.
 
So knowing what you know about these cheapo 150W/180W testers and the components they use, would you have a recommendation for ‘safe’ current levels to run when testing an 8S string?

Or said another way, the temperature read out on these boards seems pretty reliable - is there a temperature you would recommend staying under?
Yes and most likely yes. I am less concerned about the MOSFET blowing if you adjust current slowly then I am about the very hot diodes next to it that are not covered by the heatsink and cause my PCB discoloration. I just received more fuses so I can run more tests myself, I highly recommend having a fuse just in case the MOSFET dies as a short. I am using #10 Guage copper wire, 40 amp automotive fuse, and 45 amp anderson connectors, now I have enough to run 4 simultaneously instead of just one. Using a Riden 12 amp supply to charge, it has a nice battery charge mode, set at 3.65v and when amps drop below .1, it will shut itself down (and tell you how much it put into the cell). Very accurate and even has a temperature probe, 3.65v on the Riden is 3.654v according to my Fluke. I can take power out of cells faster than I can put it into a single cell. Waiting patiently for the upgrade board to make it an 18 amp supply (and a second complete 18 amp kit).
 
Mine went up to 24A. I pointed a fan at it.
I ran mine through 3 full cells at 22 or more amps. The MOSFET is no problem, however the diodes right next to it gets too hot to touch and discolor the PCB. Had to drop it to 10 amps (display said 45 degrees C) to bring that to a level I was comfortable with. The temperature displayed is of the MOSFET (pretty sure), It would not worry me to run that to 70 or 80 degrees Celsius, but those diodes get crazy hot if you do.
 
I ran mine through 3 full cells at 22 or more amps. The MOSFET is no problem, however the diodes right next to it gets too hot to touch and discolor the PCB. Had to drop it to 10 amps (display said 45 degrees C) to bring that to a level I was comfortable with. The temperature displayed is of the MOSFET (pretty sure), It would not worry me to run that to 70 or 80 degrees Celsius, but those diodes get crazy hot if you do.
Sounds like the lack of any heatsink near those diodes is the design flaw in these cheapo capacity testers.

I’m not sure what the design used is, but in general, the voltage across a diode when forward-biased is relatively fixed independent of current, meaning power being dissipated through a diode is roughly proportional to current (where power through a power transistor is proportional to current times voltage).

What that means is that at low-voltage, high-current testing of single cells, those diodes are going to be generating a great deal more heat than they are when testing battery strings at higher voltage and lower current.

Overlooking any kind of heatsink in the diodes sounds like the root flaw, but it’s less likely to be a problem at lower currents than higher currents.

So for now, I’m going to stick to my 10A limit on my 150W tester. Among other things, 10A is the calibration point, so testing at 10A is likely to provide the most accurate measurement (after calibration).
 
Sounds like the lack of any heatsink near those diodes is the design flaw in these cheapo capacity testers.

I’m not sure what the design used is, but in general, the voltage across a diode when forward-biased is relatively fixed independent of current, meaning power being dissipated through a diode is roughly proportional to current (where power through a power transistor is proportional to current times voltage).

What that means is that at low-voltage, high-current testing of single cells, those diodes are going to be generating a great deal more heat than they are when testing battery strings at higher voltage and lower current.

Overlooking any kind of heatsink in the diodes sounds like the root flaw, but it’s less likely to be a problem at lower currents than higher currents.

So for now, I’m going to stick to my 10A limit on my 150W tester. Among other things, 10A is the calibration point, so testing at 10A is likely to provide the most accurate measurement (after calibration).
Sounds like pretty good logic to me. I am testing a more robust (supposedly) design called the ET5410 next. So far except for the "designed by a drunken donkey" menu interface, it looks promising. Still limited to 40 amps or 400 watts. The drunken donkey comment is a translation from a Russian review. I agree with it, terrible menu interface.
 
Sounds like pretty good logic to me. I am testing a more robust (supposedly) design called the ET5410 next. So far except for the "designed by a drunken donkey" menu interface, it looks promising. Still limited to 40 amps or 400 watts. The drunken donkey comment is a translation from a Russian review. I agree with it, terrible menu interface.
What did that model cost you?
 
We’re you taking Watts ‘right to the maximum edge’ or current?

When testing individual cells, you are generally maxing out on current while Watts are modest, while when testing a series string of cells as a battery, you are generally maxing out Watts with modest currents...
I was watching the watts as i moved the current up and took it to 179W. I'll just add, that it was when i was doing it on my whole pack, not an individual cell. The secret seems to be, move it up slowly. I maybe got a little impatient.
 
I'll just point out that i used mine several times right on the maximum edge...took the amps up slowly, and for 6-7hours at at time....guess in the end i just over stressed it. Probably about a 1/2 doz uses before it gave up. A little kindness and it will probably do fine :ROFLMAO:

I wonder if you holes drill in the board around the mosfet and a cpu fan mounted underneath the board, if it might be happy then...that might give a better passage of air past it????

I would guess they don't know that NASA learned at the school of hard knocks.
That particular transistor someone identified doesn't even spec DC operation.
You probably need to either derate further or modify the design.
For an 8s pack, use two testers each draining 4s, half the voltage and watts.

Or put in a better transistor.
Or add a resistor in series, so some of the voltage and watts are dissipated elsewhere (tune for a specific V/I operating range)

NASA's publication showed curved areas where they observed thermal runaway of MOSFET even though within data sheet SOA

NASA Thermal Runaway.jpg

NASA added lines sloping more steeply down on the SOA graph to show where Operating Area that was actually Safe from thermal runaway.

NASA SOA with thermal instability.jpg

Although within published watt limits, higher voltage required lower current to avoid thermal runaway.

This would relate to failing shorted. Some members report the device failing open, which sounds like wirebonds burning out. With a battery behind it, could be shorted transistor resulted in the wirebonds blowing.

It also sounds like diodes need upgrading (or heat sinking) as well. Perhaps copper wire coming off the leads would serve as heatsink, adding to whatever PCB copper area is present.
 
Back
Top