Amazon load testers, Magic Smoke

goatman

10 MW
Joined
Jun 23, 2019
Messages
3,183
Location
Surrey, B.C.
havent even used it for a week and it just went BOOM :D right after taking this picture

IMG_20210306_232023.jpg

dead

IMG_20210306_233133 (1).jpg

like these ones
https://www.amazon.com/Maxmartt-Adjustable-Constant-Electronic-Capacity/dp/B084GP95QF/ref=sr_1_97?dchild=1&keywords=180+watt+battery+load+tester&qid=1615102811&sr=8-97

i guess its time to open it up and see if its fixable

IMG_20210306_235500.jpg

IMG_20210306_235453.jpg

looks like a mosfet

IMG_20210306_235653.jpg

be nice if i could read the numbers

looks fixable

IMG_20210307_002718.jpg

wonder if it was bad paste

IMG_20210307_002944.jpg

cross reference for mosfet

https://alltransistors.com/mosfet/crsearch.php?&struct=MOSFET&polarity=N&pd=280&uds=200&ugsth=4&id=46&qg=230&rds=0.055&caps=TO3P
 
Well, assuming they didn't use a counterfeit part, that's an IRFP260
https://www.vishay.com/docs/91215/91215.pdf
basic specs:
VDS (V) 200
RDS(on) (Ω) 0.055 when VGS = 10 V
Continuous Drain Current when VGS = 10 V
46A when TC = 25 °C
29A when TC = 100 °C
Maximum Power Dissipation TC = 25 °C PD 280 W

So, if it's actually a real part, and being used "correctly" by the design, it should be able to easily handle what you were doing with it at the time.

My suspicion is that either it is counterfeit (and thus we have no idea what it's actual specs are), or it is not being driven correctly (problems with the gate drive perhaps putting it into it's "linear" region, rather than fully on, which makes it more resistive and a lot hotter), etc.

If they are not using it as a switch, but instead are using it linearly, as a resistor, I dont' know how well it should handle that (I'm not enough of an "engineer" for that).


Regarding the paste--there should *only* be paste in the imperfections between the two surfaces, so that as much metal is in contact as possible. The paste has a much lower thermal conductivity than direct metal-to-metal contact does, especially as it's solvents evaporate and it dries out, leaving airgaps, which has even worse conductivity. Unfortunately it is very common for manufacturers of all kinds of things to use relatively thick pads of paste of various kinds, simply because it's much easier to apply this way by machine (or by unskilled hand labor). Also because the surfaces are often not really flat, so there can't be much metal-to-metal contact.

"Overclockers" and those looking for best thermal performance may "lap" and polish their heatsinks and devices specifically so they are as perfectly flat as possible, with as few imperfections as possible, so that as much metal as possible will be in direct contact. Not worth doing for most applications, but for very high power dissipation it is often worth it, when there is limited surface area for heat transfer, as well as for limited heatsinking ability. (sometimes both apply, sometimes just one)
 
im sending it back for a refund and get another one

thought about just fixing it but what if something else fried

the thermal paste shouldnt be dry and crusty, right?

if i have blown mosfets in my controller the thermal paste gets all over my hands
my thinking was get another one, pop the fan off and put good thermal paste on it
 
goatman said:
the thermal paste shouldnt be dry and crusty, right?

if i have blown mosfets in my controller the thermal paste gets all over my hands
my thinking was get another one, pop the fan off and put good thermal paste on it
There are many kinds of thermal paste. Some of them have no significant solvent in them, so they feel dry and crusty but are supposed to be that way. The most common ones like that are a gray waxyish substance that flakes easily at room temperature. It is typically a better paste than the very basic titanium dioxide + silicone oil compound that is the white slimy stuff you find on most ebike controllers.

The white stuff is the cheapest possible stuff they can use...which is why they use it. Some of them skip it entirely for the cost of it and the labor to apply it. When the oils evaporate, leach out, or "crawl away" from surface tension (after varying levels of time), it is basically an airgap at that point, and actively harmful to the heat transfer process, because it is nearly always applied thick enough to see it all over the surface, which means it is preventing metal-to-metal contact (which should never be the case for best thermal conductivity).


The actual thermal conductivity difference between most thermal pastes (when properly applied!) is negligible; you can look up many tests at overclocking / cpu forums and websites, with all sorts of graphs and charts.

So the important qualities to look for are either solvents that will not evaporate (or leach out, etc), or ones that have minimal solvents, and mostly just thermally-conductive material, and in as finely-ground a particle size as possible.


A separate problem in controllers is taht the typical ebike controller uses an extruded casing that is *not flat* !!! on the inside where the FET mounting bar screws down. It has ripples and waves in the surface, whcih create huge airgaps between that bar and the heatsink of the case itself. Filling this will thermal paste is better than the air gap, but only until the solvent is gone. :( The mounting bar itself is rarely flat, either, multiplying the problem. The screw holes in both bar and case are a complication because they are typically deformed at the hole edges such that they have a ring of higher material there, preventing the screws from being able to pull what flat surfaces there actually are together.

There's no simple easy quick fix for these problems, but running a good flat file (that is long enough to completely go across the whole area at the same time) over the whole surface till it's as flat as it can make it will help. Then thermal paste just enough to fill in all the scratches the file made, sliding the heatsink bar around on the casing while applying pressure to push any excess paste out of the way, so that you can actually feel the grittyness of metal to metal contact as much as possible. It's almost certainly going to be better than the way it started out. ;)
 
i went searching and found this one, its 4-wire, hooks to a lap top or cell phone for about the same price. go read the comments on how some have modded them for heat

https://www.amazon.com/Electronic-electronic-adjustment-constant-discharge/dp/B08MQ1G57D/ref=sr_1_2?dchild=1&keywords=180+watt+load+tester&qid=1615187791&sr=8-2

and a video showing the fake 260, i linked to the video at time stamp with english subtitles

https://youtu.be/CxCRc8gjHJo?t=860
 
ordered a irfp260pbf-nd from digi-key
200v 46A to247-3

IMG_20210311_165526.jpg

next to old one

IMG_20210311_165605.jpg

scavanged some paste

IMG_20210311_172835.jpg

turned on and still had the memory from when it blew

IMG_20210311_173416.jpg

connected a battery and tried to trip the safety shut off

IMG_20210311_173446.jpg

seems to be working fine

IMG_20210311_173645.jpg
 
FWIW, that'[s WAAAAAAAYYYYY too much thermal paste, and will help insulate the heat from passing out of the FET.

You really should read what I wrote previously about that. ;) Or if you don't believe me, you can do instrument everything with temperature sensors, and do testing yourself on temperatures of the FET and the heatsink with varying amounts of paste.
 
goatman said:
i did, the white stuff from blown bike controllers is garbage, so, load it up :lol: :lol: :lol:
The worse the compound is, the LESS of it you should use. Otherwise you are just insulating your parts, not helping them shed heat. ;)
 
I've built several moderately overlocked (air cooled) PCs lasting way beyond their tech cycle and on copper heatsinks, you only need a rice grain sized dot of arctic silver and maybe a 2 rice grain sized glob on aluminum. As Amberwolf pointed out, polishing the heatsink yields much better results.
 
amberwolf said:
FWIW, that'[s WAAAAAAAYYYYY too much thermal paste, and will help insulate the heat from passing out of the FET.
It makes no practical difference (to thermal performance) how much thermal paste you use, as long as:
1. It is actually thermal paste.
2. There is enough to cover the surface of the component (or hot area).
3. The component is correctly clamped against the heatsink (preferably to datasheet specs).
4. Any "excess" doesn't interfere with other circuit operations (for example, if the paste is conductive).

If the thermal paste isn't allowing the component to adequately contact the heatsink, then it's not a problem of excess paste, but rather of inadequate fixing between the component and heatsink.

If the fixing is adequate, then any excess just gets squeezed out.

Whilst it's true that in some cases all you need is a "grain of rice", I would say that more importantly, "excess" paste is better than not enough.

Don't take my word for it. Here's some random internet reference:
https://www.gamersnexus.net/guides/3346-thermal-paste-application-benchmark-too-much-thermal-paste
 
While that is true...the problem is usually that people putting excess paste on don't always have sufficient pressure to push the excess out...then depending on the kind of solvent or oil the paste uses, it can either migrate out from between the surfaces, or it can evaporate; in both cases this leaves an airgap, which is a great insulator. (the thicker paste itself isn't directly that bad of an insulator and it is a great deal better than air...but when the paste goes away that airgap can be a serious problem).

This is all from many years of fixing all sorts of electronics (mostly for other people) with heatsinks and hot parts, where this particular problem was the root cause of the failure (or for those that didn't outright fail, the performance problem).

The percentage of times this causes a problem is likely very low...but if it isn't allowed to happen at all, the problem it causes won't arise. ;)

That's why I try to get people to do it "right" initially (they're unlikely to go back and fix it later, unless it blows up again).



Side note: Heatsinks and components are not infrequently clamped together incorrectly, whcih leads to the same kind of problem whether or not there was excess paste--but this is a separate problem. (can't count the number of computers with performace problems or even actual damaged CPUs or GPUs I dealt with back when I did computer repair, because one or more of the heatsink fastening points wasn't secured correctly or at all. :( )
 
I have the same load tester and I smoked it few weeks back, never do battery stuff before your 2nd cup of coffee, ask me why. :mrgreen:

will this MOSFET work as a replacement

https://www.amazon.com/Bridgold-IRFP260NPBF-IRFP260N-IRFP260-Transistor/dp/B07R49H4HZ

generic 260n mosfet from amazon

Vishay from digi is 6.5/per piece with flat shipping of $8
 
Check out the following heatsink mod (similar to Maxmartt 150W) on the ATORCH DL24P Load Tester at ... https://www.amazon.com/Electronic-electronDL24Pic-adjustment-constant-discharge/dp/B08MQ1G57D/ref=sr_1_28?keywords=torch+load+tester&qid=1645036171&s=industrial&sr=1-28 ... scroll down to: Top Reviews From United States to read the heatsink mod review by John E titled ... It works, and is accurate, but will burn out sooner or later.

That's the same upgraded ATORCH DL24P model i purchased (thanks to you) and it's still working AOK. I don't discharge at more than 10 amp rating (runs cool), and more often closer to 5 amp rating, so hopefully it may outlast me. Truly like it ... glad you replied to my post last year on your "desulfate" thread ... so decided to order this upgraded 180W DL24P model instead of the 150W Maxmartt or MakerHawk model.

Took a little playing with to learn how to use its functions, but now its a Fun experience. Couldn't get along without it ... like for 30Q cell/pack experimental use. Still no success with noticeable "regeneration" so have given up on attempting to significantly reduce unacceptable self-discharge on those nine 30Q cells (after 125 c/d pack cycles in 2020) suffering from higher than normal self-discharge.

7134aTOxLNL._CR200,0,1200,1200_UX175.jpg


This is one of 4 photos showing heatsink mod at bottom of the Amazon DL24P review by John E. Click on that heatsink mod photo review posted by John E to enlarge it. The following is a portion of his rather lengthy, but helpful review ...
The diodes (not the MOSFET) on this get VERY hot running at 20 amps, so I thought I'd show how hot, and how I measured. Please note that somewhere between 10 and 15 amps is a safe current to run this ...
Also this review on page 2 by Casey Fenfox ...
Accuracy seems quite good and the capabilities allow for a lot of potential uses. I did notice that the included cooler is grossly undersized for the rated 180W capacity but thankfully it uses standard LGA 115x intel mounting holes and can easily be upgraded. The fan is 3-pin 12v but a 4-pin fan will usually work as long as it isn't maglev. I upgraded mine to a Noctua cooler with new thermal grizzly kryonaut thermal paste that can actually handle the heat load and it stays nice and cool even up at 200W.
Hope these offer some hope with resurrecting your Load Tester :thumb:
 
Back
Top