two dead inverters

[shovenose]

1. Xpower 1000w

This one fan twitches briefly then fault light no output.

2. Go power electric inc. 1000w

This one is completely dead.

I used PSU for 12v input. Is that not enough?

[shovenose]

Re: two dead inverters

Since I can't edit anymore....
Photos! The black one is the XPower and the yellow is the Go Power

[shovenose]

Re: two dead inverters

So now that I'm home I've got something better to power them than a crappy "400w" raidmax unit, an older 750w PC Power & Cooling unit with a single 60A 12V rail. Unfortunately the yellow Go Power unit is still completely dead while the black Xantrex XPower's fan spins for a few seconds longer but then still shuts down.

I was told both units were broken so I'm not surprised they aren't working - but I really just don't know where to begin on Troubleshooting these. The Go Power one seems like it's a better unit since the Xantrex XPower is from like 2003...

Anyway hope somebody can help me and it's not too hard of a repair because I've love to have at least one of these for when I get a car. If I get both working I'll either sell it to my friend or put it in mom's car.

Question: Do inverters work like power supplies? In which case they will only draw the wattage needed of them like a computer PSU? If I put a 1000W inverter in an Accord 95 (mom's) but only use it to power a laptop (for example) it will only pull that much power from the car?

[Pentium4]

Re: two dead inverters

the Go Power looks much better. Is the fuse blown? I would check to see if you are getting any input first, check the bridge rectifier, and I like to check the soldering side to see if I can see any visible burn marks on it. That thing probably has a lot of rectifiers on that heat sink, one of them might have burned out or shorted

[cheapie]

Re: two dead inverters

Question: Do inverters work like power supplies? In which case they will only draw the wattage needed of them like a computer PSU? If I put a 1000W inverter in an Accord 95 (mom's) but only use it to power a laptop (for example) it will only pull that much power from the car?

Correct. Take a switching wall wart, remove the input rectifier, adjust the output higher, and attach the primary side of another one (with no rectifier) to that and attach the load where the transformer would be. With a few more component tweaks, you'd essentially have what these inverters are.

[shovenose]

Re: two dead inverters

the Go Power looks much better. Is the fuse blown? I would check to see if you are getting any input first, check the bridge rectifier, and I like to check the soldering side to see if I can see any visible burn marks on it. That thing probably has a lot of rectifiers on that heat sink, one of them might have burned out or shorted

Thank you for the logical path of troubleshooting. But first off:
-can I check the 40A fuses in circuit with my DMM?
-How do I "check to see if I'm getting input"?? I know the computer's PSU outputs 12V.

[kaboom]

Re: two dead inverters

There are two power supplies in those units.

Let's go over the black one first. There's a bank of MOSFETs driving those two transformers. The FETs chop the DC up and apply it to the transformers, which step up the voltage. There are TO-220 diodes rectifying the resulting HV AC. A DC bus of 160-170V is thus obtained.

Now, the interesting part. An H-bridge comprised of 8 MOSFETs, 4 banks of two in parallel, fed from the DC bus and operating at 60Hz, chops the HV DC into 120V RMS. If you trace the ckt, you'll find that neither side of the AC receptacle goes to GND. Only the EGC does. Because of the stepped nature of the output waveform, only a true-RMS meter will read correctly.

Usually, one or more of the "12V" MOSFETs shorts. Sometimes, the DC bus MOSFETs do, too. These are relatively easy to work on. Just remove the H-bridge MOSFETs and get the 12V-170V DC converter working.

For testing, you should only put a 10A fuse in there. If one of the 12V MOSFETs has shorted, the associated transformer will get hot, as it's now become an autotransformer feeding a short. If it draws more than about 1/2 amp without a load, there's something wrong.

Now on the yellow one, I can't tell you much. I'm not sure if the secondary side MOSFETs are driven with sec DC, or if the incoming 12V power drives them. Either way, it will have the same 12-170V booster as the black one does. It's not as easy to see, at least from that angle and board layout. As it's also a stepped-square inverter, the same principles apply, though.

There's usually a 78L05 for reference and/or opamp power somewhere, too. That's in addition to the 5VRef from the '494, if it uses one.

BTW, you can check those fuses, like any other, with your DMM.

Also, keep a voltmeter connected to the power supply you're using to test with. You'll probably see the voltage getting pulled down.

Hope it helps...

[shovenose]

Re: two dead inverters

Wow, that is a lot of information.

How do I identify what is what? There are a lot of things in various sizes on the big heat sink.

Btw I don't think those fuses are removable. Is there a chance they went bad for no reason? And replacing them would fix it? Could I put sockets instead? Thanks.

Oh and last thing, I am under the assumption yellow one is pure sin wave and black one is cheaper design?

[kaboom]

Re: two dead inverters

Wow, that is a lot of information.

How do I identify what is what? There are a lot of things in various sizes on the big heat sink.

I have edited one of your pics.

Towards the top are the MOSFETs that switch the incoming 12V.

Lower left is the bridge made of four individual TO-220 rectifiers.
Lower right is the H-bridge.

Btw I don't think those fuses are removable. Is there a chance they went bad for no reason?

They're soldered in. I was thinking of other equipment, with clips, when I suggested you temporarily replace them with smaller ones.

And replacing them would fix it? Could I put sockets instead? Thanks.

They blew for a reason- that unit is fused at 120A! If some of them began to blow, it was over the combined total (4*30A) for some time. Or, some MOSFETs have shorted.

Mouser has some clips specifically made for PC boards that accept "blade" fuses. These are the same ones commonly found on car amplifier PCBs. I don't have the P/N right now...

Oh and last thing, I am under the assumption yellow one is pure sin wave and black one is cheaper design?

You know, I thought the same thing, since I saw that choke wound on the double-toroid. I actually was going to include a description of how a true-sine worked, but didn't for this reason:

http://gpelectric.com/products/1000-...-wave-inverter
http://www.defender.com/product.jsp?...3738&id=611046

Hmm.. two inverters with the same model number???

Take a pic of the yellow one's label.

I'm leaning towards it being a true sine, that double-toroid inductor almost looks like the filter L true sines always have. There should be a film cap nearby and connected to it.

[shovenose]

Re: two dead inverters

Derp... the Go Power says "Pure Sine Wave" on the bottom

[shovenose]

Re: two dead inverters

Towards the top are the MOSFETs that switch the incoming 12V.

Lower left is the bridge made of four individual TO-220 rectifiers.
Lower right is the H-bridge.



They blew for a reason- that unit is fused at 120A! If some of them began to blow, it was over the combined total (4*30A) for some time. Or, some MOSFETs have shorted.

Alright... what do I need to check on these? For shorts? Can I do this in circuit or do I have to remove it?

Do you think I have a better chance fixing the black one since it at least lights up Fault?

[shovenose]

Re: two dead inverters

Whoa...
I thikn I just noticed something in the black one.
Every single cap in there is made by Su'scon.
Is it possible the two purple ones which are only 85C rated are bulged? It's hard to tell with the black plastic things on top. Could this cause Fault?

[kaboom]

Re: two dead inverters

Alright... what do I need to check on these? For shorts? Can I do this in circuit or do I have to remove it?

Do you think I have a better chance fixing the black one since it at least lights up Fault?

That helps a little.

Sometimes, there's a low value shunt resistor that current from the DC bus passes through. The voltage across it is relative to the current through it, V*R.

There would be an opamp (one of the 324s) monitoring this. Upon exceeding a certain threshold, it would signal the main (12V) PWM to shut down, and triggers whatever monitors overall status, lighting the "fault" LED. It sounds like this unit does that.

That would suggest a short in the H-bridge rather than the MOSFETs on the 12V side, as those generally burn up/in half, blow fuses, and even burn traces.

It's getting late, so later today, I'll make a neat little diagram and explain where to check that H-bridge. Also, there are some nice MOSFETs available now, should you want to upgrade the bridge.

Whoa...
I thikn I just noticed something in the black one.
Every single cap in there is made by Su'scon.
Is it possible the two purple ones which are only 85C rated are bulged? It's hard to tell with the black plastic things on top. Could this cause Fault?

Those two caps with the "caps" are the DC bus caps. They'll have the 160-170VDC when the unit operates.

I noticed that one of the bridge MOSFETs, the fifth one in from the left side, appears to have a discolored tab. Running hot?

You can do a quick check. Put your DMM in cont. check (beep) and measure from drain to source on those MOSFETs. You've got three leads on each one. Looking at the pic and considering the lower row of 8 MOSFETs, the left pin is the source, center and tab is drain, and rightmost (with resistor) is the gate.

At least one of those will show as shorted. It's also possible that they'll only break down at full voltage.

[shovenose]

Re: two dead inverters

That helps a little.

Sometimes, there's a low value shunt resistor that current from the DC bus passes through. The voltage across it is relative to the current through it, V*R.

There would be an opamp (one of the 324s) monitoring this. Upon exceeding a certain threshold, it would signal the main (12V) PWM to shut down, and triggers whatever monitors overall status, lighting the "fault" LED. It sounds like this unit does that.

That would suggest a short in the H-bridge rather than the MOSFETs on the 12V side, as those generally burn up/in half, blow fuses, and even burn traces.

It's getting late, so later today, I'll make a neat little diagram and explain where to check that H-bridge. Also, there are some nice MOSFETs available now, should you want to upgrade the bridge.

Thanks a bunch... I know I have a lot of posts here but it doesn't mean I know exactly what you're talking about. You're already explaining a lot but what is a 324??

Are you saying simply a little resistor is bad and nothing is shorted (on the black one, let's forget about yellow one right now).

[kaboom]

Re: two dead inverters

Thanks a bunch... I know I have a lot of posts here but it doesn't mean I know exactly what you're talking about. You're already explaining a lot but what is a 324??

A common quad opamp.

Are you saying simply a little resistor is bad and nothing is shorted (on the black one, let's forget about yellow one right now).

No. And besides, if a resistor went bad, we'd have to figure out why, even in that case!

Right next to the two 470u 200V caps are 4 resistors. Notice the silver band- they're low value shunts, [/i]and[/i] they're in parallel. The more current drawn thru them, the greater the voltage across them. An opamp monitors that voltage, and shuts things down if it gets too high.

Too much voltage across resistors=too much current drawn. Normally this is from too much load plugged into the inverter, but is just as effective should a bridge MOSFET short.

I actually have an inverter with the very same over-current monitoring. I did some serious upgrades to it, but that's for another thread.

Derp... the Go Power says "Pure Sine Wave" on the bottom

Thought so. I think the black item with the reddish pink pen mark on it is the cap for its LC filter.

[shovenose]

Re: two dead inverters

A common quad opamp.



No. And besides, if a resistor went bad, we'd have to figure out why, even in that case!

Right next to the two 470u 200V caps are 4 resistors. Notice the silver band- they're low value shunts, [/i]and[/i] they're in parallel. The more current drawn thru them, the greater the voltage across them. An opamp monitors that voltage, and shuts things down if it gets too high.

Too much voltage across resistors=too much current drawn. Normally this is from too much load plugged into the inverter, but is just as effective should a bridge MOSFET short.

I actually have an inverter with the very same over-current monitoring. I did some serious upgrades to it, but that's for another thread.



Thought so. I think the black item with the reddish pink pen mark on it is the cap for its LC filter.

I made a video of the way I'm testing it with a PSU and it Fault-ing. Would it be helpful?

[kaboom]

Re: two dead inverters

I made a video of the way I'm testing it with a PSU and it Fault-ing. Would it be helpful?

As long as you can do the following too:

Green wire/green arrow should be connected to chassis of inverter.

Make sure it is, then measure voltage from yellow arrow (eyelet) to chassis. That's your 170V DC bus. Make sure your meter is set to measure voltage! Don't "forget" to change it after you've been measuring for shorts with it set to resistance!

You should have good bus voltage for the short time it operates. After the time delay when it dies, it may or may not drop, so beware of that.

I'd like to see that vid, please...

[shovenose]

Re: two dead inverters

I used Google to figure out what the op amp was - now I understand a little more of what you're trying to tell me.

You want me to to test this while the unit is on?

[shovenose]

Re: two dead inverters

OK, I measured from the "eyelet" on the board to chassis... never gets any voltage whether it's off or on.

[kaboom]

Re: two dead inverters

OK, I measured from the "eyelet" on the board to chassis... never gets any voltage whether it's off or on.

And the power supply you're feeding it with never shuts down?

If you're never getting the DC bus voltage, it's possible the 12V MOSFETs aren't being driven.

First, check between that eyelet and the chassis for shorts. Shorted rectifiers will keep the DC bus from coming up.

At this point, you're gonna have to pull the board and check things. You'll probably be doing a few "hot tests" too.

Don't forget to look for that 7805/78L05, and make sure you're getting +5 out of it. Its 12V feed should come from the "power" switch. That's right- the entire 12V side of that thing always has power. The switch just supplies control power to the 12V PWM and opamp(s).

Have "fun."