Author Topic: 300W inverter repair - problem with H-bridge  (Read 1240 times)

0 Members and 1 Guest are viewing this topic.

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
300W inverter repair - problem with H-bridge
« on: October 19, 2024, 05:25:51 pm »
I have been trying to repair this 300w inverter. When I got it everything in H-bridge was blow. I replaced all the components from the optocouplers to the Mosfets. I wasn’t able to identify the gate protection diodes (D4&D6 in the schematic) but there were remains of glass package so I put two 1N4148s.Everything ran fine for a while (maybe less than e minute without load) and then the whole top part to the bridge including the BJTs blew. I thought the cause were 1N4148s so I replied everything and put two Schottky diodes instead. In the schematic D4 and D6 they are marked as 1N4007 but I believe some faster diodes are need for gate protection. After a while the same thing happened this time blowing 3 mosfets and the BJTs. The mosfets don’t heat up at all before they blow. Also I have looked at the Gate pulses of the top mosfets - on one of them the impulses were clean, on the other one it looked like there is some oscillation but didn’t have enough time to check the measurement. Unfortunately I don’t have a differential probe to see multiple signals. I have looked at the schematic and started wondering what goes bad first - the BJTs or the mosfets? Interesting thing is that the optocouplers are still good so that leads me to think that mosfets blow first. Mosfets and the BJTs are bought from Aliexpress so there is some probability of being fake, but at this point I want to be sure there is no other issue before go out and buy some genuine mosfets. I attach a photo of the mosfet if somebody has experience with that brand, a schematic of just the H-bridge and an old blog post explaining the whole device: https://cschirp.de/index.php/elektronik/wechselrichter-reverse-engineering.html

 

Offline BennoG

  • Regular Contributor
  • *
  • Posts: 214
  • Country: nl
Re: 300W inverter repair - problem with H-bridge
« Reply #1 on: October 19, 2024, 07:09:47 pm »
put the 300V part on a 100V current limited psu (or the highest you have)
connect the output of the BCP 817 to ground and see what the current is doing.

step2 connect a scope to the outputs of both BCP 817 and see if they are not ringing or doing hf stuff.

Benno
 
The following users thanked this post: bborisov567

Offline Microcheap

  • Frequent Contributor
  • **
  • Posts: 266
  • Country: 00
Re: 300W inverter repair - problem with H-bridge
« Reply #2 on: October 19, 2024, 07:17:50 pm »
... Mosfets and the BJTs are bought from Aliexpress ... [/url]

That's your problem right there. I wouldn't waste another second trying to debug the thing before replacing the parts for ones from a reliable source.

I have already learned the lesson: NEVER trust components ordered from Aliexpress or eBay, especially transistors and op-amps. Always assume they are fake junk, because most of the time they are.  :horse:

There's no point, you can get good parts from LCSC for nearly the same prices and some times cheaper.

 
The following users thanked this post: shakalnokturn, bborisov567

Offline BennoG

  • Regular Contributor
  • *
  • Posts: 214
  • Country: nl
Re: 300W inverter repair - problem with H-bridge
« Reply #3 on: October 19, 2024, 07:31:55 pm »
I missed the Aliexpress part, yes replace the from a respectful source Farnell / digi-key / mouser / RS components / tme / lcsc.

Benno
 
The following users thanked this post: bborisov567

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
Re: 300W inverter repair - problem with H-bridge
« Reply #4 on: October 19, 2024, 08:22:14 pm »
I guess will get some genuine parts on monday. What about the gate protection diodes for the reverse voltage on the gates of the mosfet? Should i put back 1N4148s or leave the schottky diodes?
 

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
Re: 300W inverter repair - problem with H-bridge
« Reply #5 on: October 21, 2024, 04:36:25 pm »
Update on the topic: i put genuine IRF840s and for the BJTs i used FZT458, for the protection diodes still have the schottkys from before. I ran the inverter only once to get the oscillograms. Both gate impulses on the top drivers look same. It appears that there is some oscillation when the the mosfets are on(picture 1 and 2). Also occasionally there are big spikes at the end of the impuse(picture 3). I am not sure if these spikes are because of the oscilloscope probe? Otherwise if there are really present the mosfet should have blown by now - peak is 80V! I am a little bit afraid to turn on the inverter again because the previous two times everything blew when turning on, not when working continuously. What do you think about the oscillations?
 

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: 300W inverter repair - problem with H-bridge
« Reply #6 on: October 21, 2024, 06:11:04 pm »
sure it shouldn't be zeners?
 

Offline Chris56000

  • Frequent Contributor
  • **
  • Posts: 913
  • Country: gb
Re: 300W inverter repair - problem with H-bridge
« Reply #7 on: October 21, 2024, 06:32:41 pm »
. . .It's easy to determine if the broken small diodes D4 and D6 are zener diodes or not – simply see if the marked (cathode or + end) of the diodes connects to the gate (left–hand leg) of each MOSFET or not – if they do, either directly or via a low–value resistor, then you want 15 V 0.5 W zener diodes to replace them!

Chris Williams
It's an enigma that's what it is!! This thing's not fixed because it doesn't want to be fixed!!
 

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: 300W inverter repair - problem with H-bridge
« Reply #8 on: October 21, 2024, 07:05:39 pm »
. . .It's easy to determine if the broken small diodes D4 and D6 are zener diodes or not – simply see if the marked (cathode or + end) of the diodes connects to the gate (left–hand leg) of each MOSFET or not – if they do, either directly or via a low–value resistor, then you want 15 V 0.5 W zener diodes to replace them!

Chris Williams

zener or not, they would be in the same direction ...
 

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
Re: 300W inverter repair - problem with H-bridge
« Reply #9 on: October 22, 2024, 05:21:28 am »
Currently i am using NRVTS2H60ESFT1G schottky diodes so they should clamp the big spikes at in one direction, i am staring to think it is more because of the measurement. Just to be sure there is protection in both directions i will replace them with 18 V zeners (thinking about bzxt55c18). What about the other oscillations when the mosfet is on?
 

Offline Harry_22

  • Frequent Contributor
  • **
  • Posts: 603
  • Country: 00
Re: 300W inverter repair - problem with H-bridge
« Reply #10 on: October 22, 2024, 02:25:43 pm »
Hi!

I can’t understand why you have -13.2 volts on the trace (trigger mark in the second picture)? Please clarify where the probe and ground were connected.
 

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
Re: 300W inverter repair - problem with H-bridge
« Reply #11 on: October 22, 2024, 03:24:32 pm »
Probe is connected between source and gate on the top mosfet. I should have inverted the trace. These -13.2 v are actually 13.2 V of Vgs.
 

Offline Harry_22

  • Frequent Contributor
  • **
  • Posts: 603
  • Country: 00
Re: 300W inverter repair - problem with H-bridge
« Reply #12 on: October 23, 2024, 08:02:20 am »
I tried to simulate the circuit in Microcap but the upper T6 opens too slowly. I had to increase the period from 25 µs to 100 µs so that the peak reaches 240V.
I don't see any problems with the diodes, but it looks like the circuit is slightly different.
« Last Edit: October 23, 2024, 08:16:34 am by Harry_22 »
 
The following users thanked this post: bborisov567

Offline bborisov567Topic starter

  • Regular Contributor
  • *
  • Posts: 133
  • Country: bg
Re: 300W inverter repair - problem with H-bridge
« Reply #13 on: October 24, 2024, 05:41:05 am »
I also simulated the circuit on LTSpice and results are looking really good. Yesterday i had some more time to test and it turns out that if connect the probe gate-source the impulses are looking really clean 12 v p-p, but if i connect it source to gate (gnd to gate and active probe end to source) then i observe the oscillation shown before. I believe there might be some issue with the oscilloscope grounding to the grid that causes the bad measurement.
 

Offline BennoG

  • Regular Contributor
  • *
  • Posts: 214
  • Country: nl
Re: 300W inverter repair - problem with H-bridge
« Reply #14 on: October 24, 2024, 05:50:05 am »
I also simulated the circuit on LTSpice and results are looking really good. Yesterday i had some more time to test and it turns out that if connect the probe gate-source the impulses are looking really clean 12 v p-p, but if i connect it source to gate (gnd to gate and active probe end to source) then i observe the oscillation shown before. I believe there might be some issue with the oscilloscope grounding to the grid that causes the bad measurement.
Never put the gnd clip on anything else then the gnd of your DUI.
Even when you have everything doucopled.
Use 2 probes and substract the value in the scope.
Or use a differential probe.
 
The following users thanked this post: bborisov567

Offline fzabkar

  • Super Contributor
  • ***
  • Posts: 2804
  • Country: au
Re: 300W inverter repair - problem with H-bridge
« Reply #15 on: October 24, 2024, 07:16:22 am »
Here is a similar circuit:

https://elektrotanya.com/vector_vec034d_300w_power_inverter.pdf/download.html#dl

The diodes are 1N4007 and 1N4148.
 
The following users thanked this post: shakalnokturn, bborisov567, Harry_22

Offline shakalnokturn

  • Super Contributor
  • ***
  • Posts: 2218
  • Country: fr
Re: 300W inverter repair - problem with H-bridge
« Reply #16 on: October 25, 2024, 11:42:25 pm »
Thanks fzabkar, I don't see need for any zener here, the bootstrap is supplied from 12V so negative voltage protection.

As for the oscillations and probing method, are you battery powering the inverter for these tests or using a SMPS?
Depending on your setup the oscillations may be the scope's or test SMPS's leakage current and in some cases a good way to pop IRF840's...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf