Repairing a GW Instek Triple Channel Power Supply that Struggles to Turn On

[The_PCB_Guy]

Although you say the manual I linked to in post #4 is different than your supply, if you follow the primary switch wiring on page 89 of the one I linked to, you will see that although they are drawn differently the wiring and switches are identical. The drawing you have in post #13 showing the primary wiring/switches has an error. There are 2 separate primary windings of 120 volts each with a tap at 110 volts. For low voltage (110/120) they are switched in parallel and for high voltage (220/240) they are switched in series. I've assumed where it looks like you are in the U.S. that you have the switches set correctly for 120 VAC or parallel primary connection.

If the two primaries are in parallel there is an extremely low chance that both are open and if you follow my marked up drawing of the one you had in post #13 you can see two complete and separate paths for the primary current through the two windings. Unless the supply was wired for 120 and connected to 240 to burn both windings (or a buried thermal fuse for each winding), you have a wiring error somewhere. As others have stated the resistance of the primary windings should be a very few ohms. Checking the resistance on J102/J103 should show low resistance and be identical on both connectors.

I understand your sketch and that makes sense. However I can assure you there is no connection (or at least the connection is very high-resistance - >1 Mohm) between the gray wire and the black wire, and the same goes for the orange wire when the lower switch is positioned to the right.

[ArthurDent]

Then there is only one other possibility. As I said, there probably was a thermal fuse inside the transformer and that has opened. The gray wire you say goes directly to the transformer isn't directly connected to any winding but goes through a thermal fuse to the black wire on the lower winding and that goes through the connector to the selector switch. This revised drawing would be what the wiring would be in this case.

If the fuse opened there is a fair chance the transformer windings are o.k.. Unless there are some shorted turns causing the transformer to overheat, there is a way to get the supply working. Instead of the hot lead going to the gray wire, disconnect it and run the hot lead to the switch as shown. Carefully check the current draw and temp rise on the transformer to make sure there it stays cool and if it does, you could epoxy a thermal fuse to the outside of the transformer to replace the blown internal one and maintain the thermal protection.

[edpalmer42]

The fact that there is an intermittent connection tells me that the transformer hasn't overheated.  Thermal fuses don't 'sort of' open.  Rather this looks like a simple bad connection i.e. a manufacturing fault that was good enough to get through final inspection.  Matt, what does the unit look like?  Well used or basically new?

Ed

[The_PCB_Guy]

The fact that there is an intermittent connection tells me that the transformer hasn't overheated.  Thermal fuses don't 'sort of' open.  Rather this looks like a simple bad connection i.e. a manufacturing fault that was good enough to get through final inspection.  Matt, what does the unit look like?  Well used or basically new?

Ed

"Intermittent" is a bit generous. It never fully powers up, the lights barely flicker and even that is rare. I tried powering it up again today after sitting for a day and there was absolutley nothing.

The unit is definitely used, and from what I gather was in a school setting for quite some time. This alone worries me - I've seen schoolchildren, and even high school/college students, wreak havoc on electronic equipment. That being said I'm not sure how they could possibly damage the transformer without opening the unit up. I'm sure there's protection to prevent over-current conditions from damaging the primaries/secondaries

I will snap a few photos of the front panel when I get a chance, to show the condition.

[edpalmer42]

Students - that is scary.  Those little monsters can bend the laws of probability!  A lifetime ago when I was an engineering student there was a story about how some years previously, the students in the power lab did something that didn't blow the very sensitive breakers on the lab, but somehow managed to blow the 15KV fuse on the high side of the transformer serving the lab!  The power company had to come and change it!

Did you give this supply the 'sniff test'?  Does it smell like anything got overheated?  The pictures don't show anything questionable.

Ed

[The_PCB_Guy]

Students - that is scary.  Those little monsters can bend the laws of probability!  A lifetime ago when I was an engineering student there was a story about how some years previously, the students in the power lab did something that didn't blow the very sensitive breakers on the lab, but somehow managed to blow the 15KV fuse on the high side of the transformer serving the lab!  The power company had to come and change it!

Did you give this supply the 'sniff test'?  Does it smell like anything got overheated?  The pictures don't show anything questionable.

Ed

The "sniff test" was the first one I performed when I opened up the unit. No burnt smells whatsoever.

[The_PCB_Guy]

I'd like you to connect the variac between the black or orange wires and the blue or red wires.  With the voltage selectors set to 115V that should put the variac output across the two windings in parallel just like they should be.

I'm a bit confused about this. If I have the variac connected to the black/orange and blue/red wires the AC selector switches are not in circuit. In order to get at these wires I need to have them unplugged from the switch board. Perhaps I'm misunderstanding you?

Thanks,
Matt

[The_PCB_Guy]

Ok, just for a quick test, looking at the schematic ArthurDent posted, I simply connected the Variac between the red and black wires only (directly across one of the primaries on the transformer. I turned on the Variac and slowly brought up the power and sure enough, the unit powered up! No adverse effects as far as I could tell. So this suggests to me I could jumper the brown stub to the black wire as Ed & Arthur were suggesting and the unit should work as expected, no?

Thanks again guys for the continued support!

Cheers,
Matt

[edpalmer42]

I'd like you to connect the variac between the black or orange wires and the blue or red wires.  With the voltage selectors set to 115V that should put the variac output across the two windings in parallel just like they should be.

I'm a bit confused about this. If I have the variac connected to the black/orange and blue/red wires the AC selector switches are not in circuit. In order to get at these wires I need to have them unplugged from the switch board. Perhaps I'm misunderstanding you?

I should have specified that the connectors were plugged into the voltage selector board and the exact method of making the variac connections was up to you.

Ok, just for a quick test, looking at the schematic ArthurDent posted, I simply connected the Variac between the red and black wires only (directly across one of the primaries on the transformer. I turned on the Variac and slowly brought up the power and sure enough, the unit powered up! No adverse effects as far as I could tell. So this suggests to me I could jumper the brown stub to the black wire as Ed & Arthur were suggesting and the unit should work as expected, no?

Sounds good to me!  Again, the exact method of making that connection is up to you.  You could just cut some insulation off the black wire and solder the tap or you could use a Scotchlok connector from your local automotive store.  Personally, I'd go the extra distance and 'restore' the circuit to something like it would have originally been by adding a new brown wire onto the stub and routing it inside the sheath with the 2 blue and 1 brown wire and terminating it back on the voltage selector board.  Yeah, it's moving into OCD territory, but I'm used to being there! 

Ed

[The_PCB_Guy]

Personally, I'd go the extra distance and 'restore' the circuit to something like it would have originally been by adding a new brown wire onto the stub and routing it inside the sheath with the 2 blue and 1 brown wire and terminating it back on the voltage selector board.  Yeah, it's moving into OCD territory, but I'm used to being there! 

Funny, this is exactly what I was planning to do

I have OCD too.

[The_PCB_Guy]

You, guys, have been awesome. I am proud to report I now have a fully-operational GPS-3303 power supply!

Thank you once again for all your help!

Matt

[edpalmer42]

Congratulations Matt!  Glad I could help.

It turned out to be a trickier problem than I expected.  I like those! 

Ed

[ArthurDent]

Sometimes trying to diagnose a problem like this where you can't actually have the problem in front of you is like a calling your doctor and asking him to explain why your leg hurts.    It happened in this case that a 'conference call' with multiple 'what ifs' and several back and forths produced a good outcome although the consultations probably aren't covered by your health insurance plan.

On another note, being as OCD as I am, I probably would have put a thermal fuse in the center of the transformer donut hole to really restore it to as close to original as possible. Enjoy your supply.