Grafit repair. One IGBT exploded. Others damaged.

[Emulsion]

Hi,

Hope someone can help me.

I have an old Broncolor Grafit studio power pack/generator that I would like to repair myself. (Note: I have experience with flash electronics, high voltages and lethal capacitors).

The previous owner advised me that after storage the pack was tested and popping sounds were heard. I have opened the unit and discovered that one IGBT has exploded and others are damaged. Fortunately the IGBTs are a common type and are available from Mouser/Digikey.

The display and controls work OK however the caps dont charge and the display shows avail power of '0'.

Whilst I can simply replace the IGBTs I am unsure what may have caused the IGBT failures and am seeking tips from anyone who has repaired one of these units before.
I have seen another Grafit pack with similar IGBT damage so assume that this failure mode may be well known to techs.

The capacitors look physically ok however I will check the ESR of each cap during my repair. After repair I will carefully and slowly reform the caps.
(Of course I keep the caps shorted when not being tested or reformed).

As the unit isnt charging the capacitors I suspect that the charging circuits may also be damaged.

Can anyone offer service tips?

(I am a non-commercial/art photographer so cannot afford to have this unit serviced by an authorised repairer. My only hope is that I can repair this myself).

Many thanks
Emulsion.

 

[Emulsion]

Attached is an image showing the damaged IGBT at the top right. Several other IGBTs have small pinprick holes where they have failed also.

 

[mshchem]

What's an IGBT???

 

[Nicholas Lindan]

What is the purpose of the IGBT's? From the photograph it looks like they are there to short the capacitor bank to quench the flashtube (!) (?).

If they are across the bank then they could have failed to a short, and thus the capacitors can not be charged. It would also explain why they are a high failure rate component. I'd recommend replacing them with the highest amperage devices you can find that will fit, and make sure you have a minimum of 25% voltage margin (i.e., 400V max circuit, at least 500V IGBT).

The control circuit could have been damaged by such a catastrophic failure (or, it may be the control circuit malfunctioned to cause the failure).

If the failure is just the IGBTs then the problem is most likely one of component rating.

 

[Nicholas Lindan]

What's an IGBT???

Insulated Gate Bipolar Transistor. Sort of a cross between a plain-jane transistor (application transistor radio) and a MOSFET (Metal Oxide Semiconductor Field Effect Transistor) (application computer). IGBTs are commonly used in power electronics, as are large power MOSFETs. In power electronics both are used as on-off switches that can switch in a millionth of a second (if not faster).

 

[mshchem]

I've heard MOSFET didn't really know what it meant.

I will observe from the sidelines

 

[Warm Winter]

Since you have IGBT damage, they need to go first. Order extras just in case given they're usually cheap. Check through components with a multimeter set to continuity while everything is unplugged. You're going to want to look for stuff passing power that isn't supposed to be (resistor that's gone out of tolerance, capacitor that's gone to a short). Check resistances with it set to ohms, make sure your resistors are all in spec.
It'll be harder without the schematic or a repair guide, so try to find a service manual if you can.

To do a wild guess, IGBT and chunks of the epoxy missing sounds like too much power went through it, so you could just go for one with the same trigger rating but supporting a higher current capacity. The caps are likely fine.

Or if you like gambling, replace all the IGBTs in the pack with new equivalents (go by part number) and some spares and hope it doesn't go bang again, and if it does, it's something causing the IGBTs to explode.

Pretty much just poke with a meter, check things are in spec, replace things out of spec and damaged parts assuming it's a relatively easy enough issue to track down.

 

[JensH]

I've heard MOSFET didn't really know what it meant.

I will observe from the sidelines

You are not alone.

 

[AgX]

You're going to want to look for ....resistor that's gone out of tolerance, ... Check resistances..., make sure your resistors are all in spec.

First time I hear of resistors going out of tolerance.
I know of resistors burning through, like a fuse. But then it was caused by overload. This is seeemingly not what you mean. Can you explain?

 

[AgX]

The problem described in this case here seems common to us tinkerers: one has a defective appliance, finds a part that is obviously defective. But unless it is a part known to fail, already here one may wonder if it actually is the cause. If one finds heat damage at the leads around one may wonder even more on the cause. Without really understanding the circuit and if it is a possible harmful/dangerous device I myself typically stop at this stage. Thus it is interesting for us with limited electronics knowledge to learn if there really is a chance to trace down the actual culprit.

Your comments already indicate that the actual culprit may be impossible to find by visual inspection. For "metering down" the culprit though one needs enough understanding of the circuit to know how far/which strand to go with metering.

 

[Nicholas Lindan]

What is the purpose of the IGBT's? From the photograph it looks like they are there to short the capacitor bank to quench the flashtube (!) (?).

Scratch that. It can't be their purpose, circuits for stopping a flash discharge early by disconnecting the flash tube have been around for 50 years. I have little idea what that bank of IGBTs is for. Are they paralleled, all in series or is each connected to a separate flash-head outlet?

 

[Nicholas Lindan]

First time I hear of resistors going out of tolerance. Can you explain?

Old 'carbon composition' resistors - the dark brown ones with flat ends - will sometimes change value when they age, especially if they are subject to high stress levels or get hot in operation. They were made of carbon mixed with clay. As the resistor expands and contracts with heat and current and voltage surges the carbon/clay starts to crumble and the resistance changes.

More modern 'carbon film' resistors from the mid 70's onwards (usually light tan with a rounded shape) are more stable. 'Metal film' resistors (light blue with a rounded shape or rounded with numbers showing their value instead of colored bands) are very stable, as are surface mount resistors. But any component that is over stressed can go out of tolerance.

Electrolytic capacitors from the late 80's and early 90's have a deserved bad reputation. The Chinese sold very small and very cheap electrolytics that would bulge and leak - think cheap batteries. The Japanese got into the act by making Japanese copies of those low cost Chinese copies of Japanese capacitors - with predictable bad results. Electrolytic capacitors in transistorized electronics from the 60's and 70's tend to be fine. However, electrolytics used in vacuum tube electronics from the 50's and 60's are often dead.

 

[AgX]

Thank you.
I knew of the various kinds of resistors, but thought the different designs were just aimed at the production resp. load bearing af these and had no effect on longevity. Especially the classic one of a carbon layer on a ceramic base by their way of production seemed to me of being completely stable unless overloaded.

 

[Chan Tran]

I could be a simple bad IGBT. I work on corona treater power supply which has IGBT to invert DC to AC. The IGBT at times locked up and cause short circuit. It could be the firing circuit. I think in most inverter if both IGBT are fired at the same time that would happen.

 

[Nicholas Lindan]

the classic one of a carbon layer on a ceramic base by their way of production seemed to me of being completely stable unless overloaded.

That's called a 'carbon film' resistor - and, yes, you are right they are quite stable unless overloaded. The axial ones with leads are made by coating a ceramic tube with carbon, attaching leads/caps to make contact with the carbon and then cutting a spiral groove through the carbon film while measuring the resistance to create a resistor of the right value.

It is the older composition types that can be a problem.

Surface mount plain-ole resistors, called 'thick film', are made in much the same way except the film is planar and made from metal oxides. An 'L' shaped groove is made with a laser to trim the resistance.

 

[Emulsion]

I could be a simple bad IGBT. I work on corona treater power supply which has IGBT to invert DC to AC. The IGBT at times locked up and cause short circuit. It could be the firing circuit. I think in most inverter if both IGBT are fired at the same time that would happen.

Thanks, I hope it is just the IGBTs. I will test many of the surrounding components whilst I am waiting for the replacement IGBTs to arrive.

 

[Warm Winter]

First time I hear of resistors going out of tolerance.
I know of resistors burning through, like a fuse. But then it was caused by overload. This is seeemingly not what you mean. Can you explain?

Carbon comp and carbon film resistors can go out of tolerance, carbon comp can be sensitive to environmental humidity and just change over time.

Some others can go out if overloaded or overheated (say, a wirewound one getting a short between some adjacent windings ebcause it arced over or overheated and damaged the insulation).

In terms of capacitors, it varies. They should be checked but the electrolytics might be okay. The capacitor plague affected mostly smaller electronics, and to an extent well made major brand electrolytics will be fine and may just need reforming if they've sat for a long time.
Tantalums can die in annoying ways and go to a short and cause issues. Your flash isn't old enough for the old bumblebee capacitors (which should be assumed bad always).

Are the voltages from the power supply in spec? It'll probably have a few different voltages for the electronics (low voltage stuff) and for the capacitor bank.

 

[Emulsion]

Carbon comp and carbon film resistors can go out of tolerance, carbon comp can be sensitive to environmental humidity and just change over time.

Some others can go out if overloaded or overheated (say, a wirewound one getting a short between some adjacent windings ebcause it arced over or overheated and damaged the insulation).

In terms of capacitors, it varies. They should be checked but the electrolytics might be okay. The capacitor plague affected mostly smaller electronics, and to an extent well made major brand electrolytics will be fine and may just need reforming if they've sat for a long time.
Tantalums can die in annoying ways and go to a short and cause issues. Your flash isn't old enough for the old bumblebee capacitors (which should be assumed bad always).

Are the voltages from the power supply in spec? It'll probably have a few different voltages for the electronics (low voltage stuff) and for the capacitor bank.

Hi,
Many thanks for your reply.
It is difficult to know what voltages to check for as service manuals and schematics aren't available.
I seem to be limited to try and figure out parts of the circuit myself ...this makes repair more complex.
Kind regards
Emulsion

 

[Emulsion]

I saw another Grafit on eBay with exploded IGBTs.
Please see the attached photo.
Does anyone know what causes this fault?
Feel free to send me a DM.

 

[AgX]

This damage is impressive. I wonder what people at Bron would say about this.

 

[Emulsion]

This damage is impressive. I wonder what people at Bron would say about this.

I don't know how to make contact with any engineering people at Bron.
If anyone has any contact info for them please send me a DM!
Many thanks.