Dead Inverter: Repaired

[HighDesertOffgrid]

I am attempting to resuscitate a doa Reliable inverter with a short on one of the two DC inputs. I think heat got it. I removed FETs until the short was gone, but I can't find a cross-reference or data sheet on these. I can find replacements on ali baba but would like to possibly upgrade to a better piece while I have it open. Can anyone identify this and link me to datasheet and possible a cross-reference?

 

[scott c]

I am attempting to resuscitate a doa Reliable inverter with a short on one of the two DC inputs. I think heat got it. I removed FETs until the short was gone, but I can't find a cross-reference or data sheet on these. I can find replacements on ali baba but would like to possibly upgrade to a better piece while I have it open. Can anyone identify this and link me to datasheet and possible a cross-reference?
View attachment 240897

https://www.amazon.com/-/es/095R0NT-095RONT-SVGO95RONT-SVG095RONT-SVG095R0NT/dp/B0C3QGRJ17

DiscoverEE : Silan SVG095R0NT is a Silicon MOSFET rated at V, naA, Ω Rdson(typ) in TO-220-3L | DiscoverEE

Power, Mosfet, FET, Manufacturer: Silan, Part Number: SVG095R0NT, Package Name: TO-220-3L, Configuration: Single, N channel, THL

www.fet.discoveree.io

SVG096R5NT MOSFET Datasheet pdf - Equivalent. Cross Reference Search

alltransistors.com

 

[HighDesertOffgrid]

https://www.amazon.com/-/es/095R0NT-095RONT-SVGO95RONT-SVG095RONT-SVG095R0NT/dp/B0C3QGRJ17

I am trying to avoid reinstalling inferior parts. Thanks for the reply.

 

[Quattrohead]

It appears to be a 120A/90V TO-220

 

[SeaGal]

It's also possible that the "095" designation relates to its Rds - i.e. 0.095Ohms.

So a MOSFET with >120A, >90V and <0.095R Rds would be worth trying as an alternative.

 

[peufeu]

Chinese datasheet says 90V, 5.2 mOhms, 68nC gate charge

Best match:

IPP050N10NF2SAKMA1 | DigiKey Electronics

Order today, ships today. IPP050N10NF2SAKMA1 – N-Channel 100 V 19.4A (Ta), 110A (Tc) 3.8W (Ta), 150W (Tc) Through Hole PG-TO220-3 from Infineon Technologies. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

I'd go with this one because the price in qty 10 is cheaper, and you have to change all the paralleled FET for the same type

Runners up, maybe too much gate charge, depends on how fast it's supposed to switch:

FDP4D5N10C | DigiKey Electronics

Order today, ships today. FDP4D5N10C – N-Channel 100 V 128A (Tc) 2.4W (Ta), 150W (Tc) Through Hole TO-220-3 from onsemi. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

AOT66916L | DigiKey Electronics

Order today, ships today. AOT66916L – N-Channel 100 V 35.5A (Ta), 120A (Tc) 8.3W (Ta), 277W (Tc) Through Hole TO-220 from Alpha & Omega Semiconductor Inc.. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

 

[HighDesertOffgrid]

Chinese datasheet says 90V, 5.2 mOhms, 68nC gate charge

Best match:

IPP050N10NF2SAKMA1 | DigiKey Electronics

Order today, ships today. IPP050N10NF2SAKMA1 – N-Channel 100 V 19.4A (Ta), 110A (Tc) 3.8W (Ta), 150W (Tc) Through Hole PG-TO220-3 from Infineon Technologies. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

I'd go with this one because the price in qty 10 is cheaper, and you have to change all the paralleled FET for the same type

Runners up, maybe too much gate charge, depends on how fast it's supposed to switch:

FDP4D5N10C | DigiKey Electronics

Order today, ships today. FDP4D5N10C – N-Channel 100 V 128A (Tc) 2.4W (Ta), 150W (Tc) Through Hole TO-220-3 from onsemi. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

AOT66916L | DigiKey Electronics

Order today, ships today. AOT66916L – N-Channel 100 V 35.5A (Ta), 120A (Tc) 8.3W (Ta), 277W (Tc) Through Hole TO-220 from Alpha & Omega Semiconductor Inc.. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

Not sure how you found that, but I'm glad you did. My direct query on digikey yielded no results. Thanks for the link!!

 

[peufeu]

Hmm, that's weird! I did something like 90-150V, TO220, RdsON, Qg, sort by price

 

[HighDesertOffgrid]

It's ALIVE!
Thanks for everyone's assistance!
Board level troubleshooting/repair is really beyond my level of knowledge and I don't often get to put the smoke back in. But even a blind squirrel gets a nut every so often.
I have determined desoldering transistors (heat sinks) and cleaning out the holes sucks big time and thus only replaced the 8 FETs on the damaged channel. Mainly because I was unsure if there were any other damaged components.
My dc power supply is limited to 10 amps so no stress test today, but she's making ac power.

 

[SeaGal]

Result! but put us out of suspense... what MOSFET's did you end up using?

 

[solar8484]

I have determined desoldering transistors (heat sinks) and cleaning out the holes sucks big time and thus only replaced the 8 FETs on the damaged channel.

View attachment 252897

Get used to it as you will probably need to do it again with those inverters

 

[SeaGal]

A good de-soldering tool works well

 

[HighDesertOffgrid]

Result! but put us out of suspense... what MOSFET's did you end up using?

This one.

IPP050N10NF2SAKMA1 | DigiKey Electronics

Order today, ships today. IPP050N10NF2SAKMA1 – N-Channel 100 V 19.4A (Ta), 110A (Tc) 3.8W (Ta), 150W (Tc) Through Hole PG-TO220-3 from Infineon Technologies. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

I bought enough to replace the entire DC side, but the lack of a fancy electric suction tool required several beers for the first half.
Beyond that, I realized that an smd 10 ohm resistor was missing. After fruitlessly looking through all of the failed electronics in my junk box that I wasn't (haven't yet been) able to repair I had to get innovative.
There is a strong chance I inadvertantly desoldered that resistor and it slipped into another quantum realm.


I have it connected to my main bank and a 1400 wattish space heater running now. So far, so good.

 

[SparkyJJO]

I have a soldapult sucker. Manual but it works well.
I also use a stainless steel needle to help clear out holes when needed.

 

[HighDesertOffgrid]

Get used to it as you will probably need to do it again with those inverters

Final test. 3528 watts for 30 seconds. No shutdown.

 

[peufeu]

I bought enough to replace the entire DC side, but the lack of a fancy electric suction tool required several beers for the first half.

Ghetto desoldering method: grab a bit of copper wire, say 1.5mm thick, cut to length, put it across the pads you want to desolder, heat it with the soldering iron, add a bit of fresh solder, the copper wire will melt all the solder joints at the same time, pull the component out.

 

[Quattrohead]

I tried this with single core, but it didn't work

 

[peufeu]

It works really well if the soldering iron has enough power to heat all the pins at the same time. Otherwise it just makes a blob.

 

[altenergy21]

Mad respect, elite level

 

[Brett V]

If you don't have to save the component, clip the leads and remove the component first. This is easy with a pair of fine needle nose pliers and sufficient heat. Then clear the hole with solder wick.

 

[ScrotpusGobbleBottom]

Couldn't you just chop them off the legs, and then simply remove the pins?

Oh I guess I'm a minute late to mention this lol

 

[altenergy21]

I'm guessing for troubleshooting one starts with general problems with inverter - that they generate heat.
Then look at that huge heat sink
Turn it on, walk away
Then point the thermometer at that heat sink, if it's cold (oops somebody is not alive, stone cold dead)
Smell the transistors
Poke the transistors with a meter to see if there's any voltage/resistance etc.

There should be a small fan there or the heat sink should have more fins to better dissipate heat.
It could happen again.

Youtube has many videos on how to troubleshoot dead electronics. Unfortunately, I'm too dumb to follow them.

 

[matttelz]

Congrats on the repair! I'm discovering so many of the components in the inverter are sub par quality and some no name Chinese brand. Replacing the main components with quality parts is imperative for reliability ironically...

Trying to decide between these two mosfets. One is the one you purchased the other has a slightly higher amperage and wattage: Ultra mega long digikey link

How does the MOSFETS work? It looks like they're rated upto like 150w and even if that's the case, there's 16 of them... that'd mean the inverter could only supply 2400w...am I missing something?

 

[AntronX]

How does the MOSFETS work? It looks like they're rated upto like 150w and even if that's the case, there's 16 of them... that'd mean the inverter could only supply 2400w...am I missing something?

That's maximum heat dissipation from the mosfet itself, not the power that it passes through. When operated correctly they are over 99.5% efficient meaning only 0.5% gets wasted as heat. Another cheaper option is IRFB4410ZPBF at $0.94 ea for 10.

 

[HighDesertOffgrid]

I'm guessing for troubleshooting one starts with general problems with inverter - that they generate heat.
Then look at that huge heat sink
Turn it on, walk away
Then point the thermometer at that heat sink, if it's cold (oops somebody is not alive, stone cold dead)
Smell the transistors
Poke the transistors with a meter to see if there's any voltage/resistance etc.

I have no idea what the proper troubleshooting steps are. This was a working unit under extremely low load but it was in a black metal box in the desert sun. I replaced the inverter with a new one and painted the box gloss white. That seems to have helped.
The first step was a visual inspection looking for any signs of damage; burn marks, puffy caps, etc.
The inverter has 4- 50 amp fuses and the two on the right bank were blown but there was no other physical sign of damage.
If I had a thermal camera, I could have introduced a controlled current beyond the fuses and looked for hot spots. I have-seen others do this. Maybe I'll buy one some day.
The good news for me was the left bank seemed unaffected so I started probing for continuity differences between the two. All 8 right-bank mosfets read shorted, but they were also tied together with traces. So I started desoldering one at a time until the short was gone.
I pulled 5 of the 8, testing after each one to remove the short and 3 of those 5 tested bad.
This unit has two fans that function at variable speeds, but a larger heat sink would probably help a bunch.