Survival's 20A Diode Test Load

[OVNI]

I bought Survival's 20A Diode Test Load and discovered only Diode-1 (out of 8) was working properly. The other seven Diode's anodes appeared to be shorted to their respective cathodes.

So I sent it back for them to test/verify and immediately bought another one thinking I just had a defective unit and didn't want to wait for the first one to come back. Well the second one arrived and I had the same problem so I started looking at other things and discovered these test loads are populated with 20ETS08 diodes in a TO-220AC package. It turns out for the 20ETS08, its cathode is electrically connected to its tab (pictured below).

For Survival's 20A Diode Test Load, Diode-1's Cathode is connected to Diode-2's Anode, Diode-2's Cathode is connected to Diode-3's Anode, and so on. And all the tabs are electrically connected to each other due to the single heatsink thus effectively leaving only Diode-1 working. I removed the heatsink and it tests fine now but I'm not sure it can handle the heavy 20A loads anymore.

I looked for another part but did not find any*. Does anyone have one of Survival's 20A Diode Test Loads but with a different part where you could post the diode part number?

And does anyone know if I snip the factory soldered connection between pins 1 and 2 (visible on the bottom of the TO-220 package above) if the tab will become isolated? From the schematic in their spec sheet (pictured above) I would guess the answer is no but I don't want to try unless Survival says they'll still take it back if it doesn't fix it.

Thanks.

*EDIT: I found some on DigiKey with isolated tabs. Vishay has their VS-20ETS08FPPBF.

 

[Cyparagon]

OOPS

You can still fix it by adding sil pads and washers to electrically isolate the diodes from the heat sink while maintaining adequate thermal conductivity.

And does anyone know if I snip the factory soldered connection between pins 1 and 2 (visible on the bottom of the TO-220 package above) if the tab will become isolated?

Probably not, but it wouldn't hurt to try.

 

[OVNI]

OOPS

You can still fix it by adding sil pads and washers to electrically isolate the diodes from the heat sink while maintaining adequate thermal conductivity.



Probably not, but it wouldn't hurt to try.

Perfect. Thanks!

 

[Benm]

Cutting the center lead will probably do nothing: virtually all diodes in this package style have the tab connect to the cathode, even if the center pin is completely absent. It has to do with getting the lowest thermal resistance from junction to case, sacrificing electrical insulation in the process.

This can be worthwhile in many applications where each diode has its own heasink or only one diode is required (say reverse polarity protection). In bridge rectifiers 2 can share a heatsink which is often quite practical.

So in this case: get some of the fully isolated diodes.

DSEI60 would be good and cheap (i've seen them under $1), can take 60 amp forward, but sadly have 300 mil pin spacing where this board is designed for 300.

Isolation pads and plastic bolts are always an option, though not all fare that well under really hot conditions. Metal bolts with insulating washers are available too, which a probably better when running these things at a decent load.

 

[Garoq]

Thanks for the heads up and the fix suggestion, I've contacted our supplier (who is also an LPF member) to alert him of the issue. :beer:

BTW the 20A load has worked normally in my testing of drivers ranging from well under 1A to 5A, and the results have correlated closely to those observed when using our standard 3A test load.

 

[paul1598419]

I use 10 amp axial lead silicon rectifiers that are $4.89 for a lot of fifty. This site: www.ebay.com/itm/271863164068. Not 20 amps, but enough for what you will probably run up against. I use these with a 0.1 ohm 5 watt resistor that has a tolerance of 1%.

Edit : this item has since been removed, but I'm sure you can find some at the same price. Sorry.

 

[Benm]

For a test load it would be best to use heatsink-mounted ones though.

The problem is not only that overheating kills the diodes eventually, but also that their forward voltage drops quite a bit as they heat up. The diodes will be fine operating at say 120 celcius, but the forward voltage will have dropped 200 mV (each!) compared to room temperature at that point.

Laser diodes do have a similar correlation between temperature and forward voltage, but you'd not want to run them nearly that hot.

As for this specific test load: Using uninsulated TO-220 package diodes on a common heatsink is a bit of a design fail. Virtually all bare metal TO-220 package diodes have the cathode connected to the heatsink tab, i've yet to see one where it's isolated.

 

[Cyparagon]

overheating kills the diodes eventually

Yes. However, the full 10A of a 10A10 diode for example would give a temperature rise of only ~90C (110C assuming 20C ambient), which is well below its 150C max.

The Vf drops with higher heat, but it also increases with higher current. These two factors can somewhat offset each other.

The test leads, the board traces, the connection points, the jumpers, and the current sense resistor all offer Vf variation that is a bigger problem anyway.

No one operates these things longer than a few seconds. Most of them would melt if they were operated for more than a minute or two. The 20A load in question for example dissipates upwards of 160W with all diodes at 20A. The heat sink they're fastened to could not take that for more than a minute I think.

Screwing around with your own DIY stuff is one thing. For a company to offer a "professional" product like this is another.

i've yet to see one where it's isolated.

Is TO-220fp close enough? :shhh:

 

[OVNI]

Thanks for the heads up and the fix suggestion, I've contacted our supplier (who is also an LPF member) to alert him of the issue. :beer:

BTW the 20A load has worked normally in my testing of drivers ranging from well under 1A to 5A, and the results have correlated closely to those observed when using our standard 3A test load.

Can you tell us what the diode part number is on the ones that work?

 

[Garoq]

Can you tell us what the diode part number is on the ones that work?

Well, the one that worked for me is the second one I sent you.

I can ask the vendor for the diode part number, but they are probably the same on all the units.

 

[OVNI]

...DSEI60 would be good and cheap (i've seen them under $1), can take 60 amp forward, but sadly have 300 mil pin spacing where this board is designed for 300...

I found the DSE160 comes in a TO-247AD package. And the board is designed for a TO-220AC package. So did you mean "DSEI60 ... have 430 mil pin spacing where this board is designed for 200"?

EDIT: NEW QUESTION
The 20ETS08 diodes on the Survival Diode Test Load can withstand a VRRM of 800/1200V. Is there any reason to have VRRM that high? Anyone know what drives the VRRM spec for purposes of testing drivers used for LDs?

 

[Cyparagon]

The only reason is the voltage rating is so high is because an otherwise equal diode at 200V costs about the same - so "why not?"

 

[OVNI]

If I go down to around 50V, they're about half as much. And I'll probably get 10 for the price break (instead of just 8) so I thought I'd save around $10 too by getting the lower voltage versions.

 

[Cyparagon]

Which part is that? 50V is not standard. Unless they're schottky diodes, which have a much lower forward voltage (not good for test loads).

 

[OVNI]

Which part is that? 50V is not standard. Unless they're schottky diodes, which have a much lower forward voltage (not good for test loads).

You are right, it is a schottky diode. I thought I had screened out the low Vf parts but missed this one. Thanks.

 

[Benm]

@OVNI, i meant that it would not fit the 200 mil distance on the PCB indeed - it is too wide to fit on the board by any chance.


[qoute]
Is TO-220fp close enough?
[/quote]

As a practical replacement on this board, sure!

But the vendor picture clearly shows diodes with uninsulated packages mounted on the same heatsink, so they would have been regular TO-220 packages with insulated anode and cathode from the tab in order to work.

Perhaps such diodes exist, but i've never seen one among lots of to-220 packaged diodes. If you replace them with a generic diode you will short stuff out through the heatsink regardless.