Inverter tech

The friendliest place on the web for anyone who enjoys boating.
If you have answers, please help by responding to the unanswered posts.
The charring on the AC wire connection points seems odd, and is probably a clue of some sort.


Also beware measuring any component when it is in the circuit as opposed to isolated. When in the circuit (installed) you are measuring the whole connected circuit, not just the component.


As an example, the DC input will have large filter caps, so if you measure across the DC pos and Neg, it will likely appear to be a short.
 
The charring on the AC wire connection points seems odd, and is probably a clue of some sort.


Also beware measuring any component when it is in the circuit as opposed to isolated. When in the circuit (installed) you are measuring the whole connected circuit, not just the component.


As an example, the DC input will have large filter caps, so if you measure across the DC pos and Neg, it will likely appear to be a short.



I was measuring every leg where it was soldered to the board from below.
 
I was measuring every leg where it was soldered to the board from below.


And that is measuring everything connected in any way to those legs. So if there are circuit board traces that connect one of those legs to capacitors, chokes, transformers, etc, you are measuring the impedance of all that stuff, not just the impedance of the FET.


Now this doesn't mean that you can't discern anything from the measurements. In fact, one of the tests after a circuit board is assembled is a so-called "bed of nails" test where probes contact many different points on the circuit board and measure the expected impedance. As I recall this is done without powering up the board, just as you are testing. The expected values are all calculated from the combination of the circuit design and the components used so the tester knows what to look for. This is done by the software used to design circuit boards, not with a pencil and paper. There are hundreds, if not thousands of test points.


You mostly need to understand that the in-circuit measurements can/will be different from what you would measure if the FET were removed from the board, and can therefore be misleading if you aren't aware of this.


Over the years I have attempted repairs of a variety of electronic equipment, almost always unsuccessfully. And I've got a degree in electrical engineering. The last attempt was an inverter/charger where I diagnosed the problem to the board level, and bought a replacement board. Put everything back together, applied battery power, it turned on which was great progress, then a few seconds later went Pop and burst into flames. I was then out the cost of the inverter/charger/ plus the cost of the replacement board which was not cheap. This is typical of failures where there are multiple things wrong. You find and fix one, but when you power it up the other faulty components cause the newly replaced components to fail once again and you are back to square one.



A few people here seem to have more hands-on repair experience than me, and certainly more success than me, so I think you are getting some good guidance. That said, I think the likelihood of success is very low. Even if you clearly identify bad components, removing them will require a high quality soldering iron and a good solder sucker, and even then it's really easy to burn or damage a thruhole. Many of the components are big so easy to work with, but there are also a bunch of surface mount components on those boards, and they are much harder to work with. Some even require specialized equipment. And component part numbers are often obscured, or not present at all on smaller devices, so figuring out what replacement parts to buy can be a real challenge. Only you can decide how to best spend your time and your money, but keep this in mind as you weigh the two against each other.
 
Steve noticed Mastervolt was not on your list. Last year the P.O. replaced his Victron (which failed) with a Master Volt. Is this a component I’d need to be concerned about?


Mastervolt has an excellent reputation, as does Magnum. After this thread about Xantrex, I've really lowered my thoughts about them, certainly not user friendly. And the searches and references I got in the past steered me away from them.



I could argue to make a call to the support side of a manufacturer prior to purchase, if you can't even talk to them, that says a lot. That's one of the criteria I used when considering Xantrex and Volvo engines and decided that my money was better spent elsewhere.


====


As for the OP, there's an argument to just bit the bullet and buy a new one, but certainly NOT a Xantrex. There are many quality ones out their in the ~$1500 range, which could easily be spent trying repair without success.


Yes, it sucks....
 
Let's get a bit more data on how the mishap occurred. I understand there was a socket rattling around in the connection area. Several scenarios seem possible.

a) AC in shorted (to ground or neutral) - should trip AC breaker but possibly too late
b) AC out shorted - unlikely since you did not have batteries connected
c) AC in shorted to AC out.
d) AC in shorted to DC in

Scenarios a) - c) would seem more benign. The output section deals with 120V AC, so caps and FETs are rated for that.
Scenario d) would seem to be the most likely. Shorting AC to the D12V DC input section would produce major mayhem. The input section is only expected to deal with 12V and FETs and caps are not rated to deal with more than 40-60V. Is scenario d) physically possible? Usually these terminals are not next to each other.

When you measure resistance on wires with large caps, you need to leave your leads connected for 30 seconds or more and watch your meter. It likely indicates a short for a few seconds and then slowly shows increased resistance as the caps get charged by your meter in resistance mode.
 
Well, based on the replies, I am getting to the point where I should declare this inverter as a loss. Let's hope someone will buy it for parts. I'll post it soon.

This troubleshooting was a good experiment to me and I have learned few things. Obviously, I am not qualified to fix this type of equipment, even if the issue was to be found.

I appreciate everyone's input and assistance. I will shut down this repair idea and move on. My new Mastervolt inverter should arrive today. I still have my Magnum inverter for backup, so I will be fine.
I hate to loose this fairly expensive Xantrex, since it has never been used, but that is life. You make a mistake and you will pay for it. Lesson learned. I enjoyed the troubleshooting, though.
Thanks everyone!
 
...
Now this doesn't mean that you can't discern anything from the measurements. In fact, one of the tests after a circuit board is assembled is a so-called "bed of nails" test where probes contact many different points on the circuit board and measure the expected impedance. As I recall this is done without powering up the board, just as you are testing. The expected values are all calculated from the combination of the circuit design and the components used so the tester knows what to look for. This is done by the software used to design circuit boards, not with a pencil and paper. There are hundreds, if not thousands of test points.
...

There are two distinct tests of this kind, one ensures the circuit board meets the spec, and that is done before the components are loaded, testing that each circuit conducts from points A,B,C, etc. as designed. A second type measures test points with the completed circuit in operation, sometimes with special test firmware, other times with the shipping code (assuming it's computerized, which almost everything is now).
 
I appreciate everyone's input and assistance. I will shut down this repair idea and move on. My new Mastervolt inverter should arrive today. I still have my Magnum inverter for backup, so I will be fine.
I hate to loose this fairly expensive Xantrex, since it has never been used, but that is life. You make a mistake and you will pay for it. Lesson learned. I enjoyed the troubleshooting, though.
Thanks everyone!

Good choice. These things can be fixed as long as the circuit board is not burned up. But if you have to pay someone it would not be cost effective. If you enjoy a challenge it could be fun.
However, not knowing what failed, the next steps would have required some live testing starting with a current limiting power supply. This would expose potentially deadly 160V DC even if the input current is limited. I'd not have recommended going down down this path without power electronics experience.
 
There are two distinct tests of this kind, one ensures the circuit board meets the spec, and that is done before the components are loaded, testing that each circuit conducts from points A,B,C, etc. as designed. A second type measures test points with the completed circuit in operation, sometimes with special test firmware, other times with the shipping code (assuming it's computerized, which almost everything is now).


The type we did was in-circuit after the board was assembled. Here's a good overview of the process.


https://en.wikipedia.org/wiki/In-circuit_test


All ours included boundary scan in the larger chips that had such a feature.


The test fixtures are very expensive, but will typically pinpoint any problem down to the component. This is one reason why repair centers typically only diagnose to the board level, and swap boards - it's just not practical to figure out what's wrong with a board without the manufacturing test setup. This varies with the complexity and density of the board, but trends continue to move towards more complex and dense boards. So boards get swapped and return to the manufacturer for test and repair using the same testers that are used in production.
 
The type we did was in-circuit after the board was assembled. Here's a good overview of the process.


https://en.wikipedia.org/wiki/In-circuit_test


All ours included boundary scan in the larger chips that had such a feature.


The test fixtures are very expensive, but will typically pinpoint any problem down to the component. This is one reason why repair centers typically only diagnose to the board level, and swap boards - it's just not practical to figure out what's wrong with a board without the manufacturing test setup. This varies with the complexity and density of the board, but trends continue to move towards more complex and dense boards. So boards get swapped and return to the manufacturer for test and repair using the same testers that are used in production.

Yes I've written firmware for bed of nails type of test fixtures, it's a ton of work to develop the testing and fixtures.

The board check tests performed pre-population are typically 'flying probe' and the tests are derived from the PCB design files. Robotic arms connect test probes to various points and take measurements. You can now do in circuit testing with flying probes as well, the setup is a bit more involved because you have to take the position/height and clearances of the components into account.
 
On a previous boat I had a failure on a xantrex inverter, in talking with Xantrex Tech they told me that inverter was unrepairable. But they could sell me a reconditioned one.......
Their reputation in the industry as well no many people steer clear of their products
 
On a previous boat I had a failure on a xantrex inverter, in talking with Xantrex Tech they told me that inverter was unrepairable. But they could sell me a reconditioned one.......
Their reputation in the industry as well no many people steer clear of their products


But wait, if the inverter was unrepairable, how did they recondition one? Seems like a self-incriminating statement.
 
But wait, if the inverter was unrepairable, how did they recondition one? Seems like a self-incriminating statement.

One of my first tech jobs was a printer repair technician for a dot matrix printer company, I was the 'authorized service rep' for my area for that brand. The repair manuals basically went like this (I'm paraphrasing):

  1. Do the lights come on? yes/no
  2. Yes: replace mechanism
  3. No: Replace motherboard

They only swapped major components in/out until it worked. To 'recondition' was the same process, anything external (covers, cases) was replaced with new if scratched up or damaged and the same test was applied yes/no to replace the internal parts in large chunks.

They probably only recondition the magnums at the factory where they have parts available and don't allow it to be done in the field and do not sell parts or provide service manuals.
 

Latest posts

Back
Top Bottom