My popped bench PSU.

[eccerr0r]

Re: My popped bench PSU.

There are a lot of possible improvements:
1. The current "heat sinking" of that 2n3055 is completely unacceptable. Need better heatsinking. If you add a heatsink, adding a fuse might be helpful. Without the heatsink on the 2n3055, the transistor will protect the fuse.
2. Removing the 2n3055 completely and bolting the LM317T to a large heat sink will allow the self protection of the LM317T to protect against overcurrent damage. However this will drop the current rating to 1.5A instead of the 3 to 5 amperes that it currently can handle (if heatsinked properly).

Other improvements:
Using a pnp transistor design (or eliminating the pass transistor) will help voltage regulation. The current design will change its voltage a few hundred millivolts depending on current draw. Using the pnp design will improve the regulation to tens of millivolts.

My homemade LM317T with pnp pass transistor has CC/CV control though it complicates things a bit. Despite current limiting, I still can destroy the pnp pass transistor if I overheat it.

[roadrash]

Re: My popped bench PSU.

Thanks for all that information and that very helpful Schematic. The red wire I was talking about is the one you can see in this picture connected to the earth tag on the chassis. Thats before I removed it and thought I better check first that it did come of there.
I checked the 2N3055 transistor and its shorted out so ive got some on order. What is the best way to add some protection to this supply so this cant happen again. It would be nice if I can access it from the outside. A reset button would be nice and easy if you can reccomend one, or is a simple low amperage fuse be better?

[eccerr0r]

Re: My popped bench PSU.

Speaking of upgrades I'd convert from NPN-only to PNP-amplified NPN or PNP pass transistor... much better regulation IMHO. However still would not protect against self destruction upon overload. Honestly if you find the 2N3055 to be dead, it may not be much worse than before if you bolted the LM317T (with appropriate insulation materials) to the chassis and use only the IC's internal pass transistor. I have another homemade PSU that someone else built that did this, it also works just fine.

Hmm. Maybe we should have a homemade "bench" PSU thread to showcase all our jury rigged and ghetto designs

[budm]

Re: My popped bench PSU.

That red wire is connected to the Collector of the 2n3055 which is NOT connected to the chassis. If you check the connection between that red wire, it should also be connected directly to the positive leg of that Blue filter cap as shown in RJ schematic.
I would at least add the filter cap to the output.

[eccerr0r]

Re: My popped bench PSU.

Based on the NPN-only pass topology, if one end of the wire is the LM317T pin 3 (rightmost pin) then that pin also connects to the collector of the NPN pass transistor, which is the metal outside of the transistor - which was my guess. In the NPN-only pass topology, the LM317T needs to pass a fraction of the full load to keep the pass transistor conducting and thus depending on the load, may need to dissipate some power.

Anyway the LM317T is also one of those virtually indestructible regulators as long as you don't go over voltage (its weakest point is if you go over voltage in reverse...); it has internal dissipation limiting, so it probably is fine though it's kind of hard to test via an ohmmeter. Checking the voltage on pin 3 when on with the 2n3055 disconnected is one way to check the viability of the LM317T.

However the transistor, not so much so, that probably would be the first thing to check. I have a PNP - pass LM317T homemade PSU and I've already had to replace that pass transistor once because I exceeded the heatsink's ability to remove heat away from it.

[R_J]

Re: My popped bench PSU.

The right leg on a LM317T is Vin, center is Vout, left is adjust

[roadrash]

Re: My popped bench PSU.

The red wire is connected to the right leg of the of the LM317T (looking from the front) and looks like it came away from a earth tag on the chassis. What is the right leg on this chip do and should it ne earthed?
There are only 3 semiconductors in the whole power supply and that is this LM347T (regulator I believe), a GBU608 which is a rectifier I believe and a 2N3055 NPN bipolar transistor.
Anyway which of these is most likely to have failed in a short between the 2 voltage outputs at 5V, if its the LM317T, how do I check it with a multi meter?

[eccerr0r]

Re: My popped bench PSU.

Definitely should make sure you understand what you're doing. Also make sure where you connected it to -- is also not shorted.

The problem with this particular PSU design (and many other homemade designs) is that it has no limiting and can very well self-destruct. The heatsinking itself also insufficient on this design - not enough contact and very thin metal.

Just to be right, this is the red wire connection that broke? If so, this is the collector of the 2N3055 and not actually ground, there are insulators there to prevent it from shorting to case/GND.

[roadrash]

Re: My popped bench PSU.

I had it set to output 5v to the Dell power supply so I could runthe supply without its 5v section (on a small daughter board) and diagnose and test its 12v section. There is a thread here that will explain what I was trying to do. https://www.badcaps.net/forum/showth...t=64063&page=3

I think fixing that bench PSU should be easy enough as there is very little in it to test. I was just unsure where that red wire had fallen off but it really looked like it came off that earth tag.
The real problem is going to be this big Dell PSU that I was trying to test and now quite possibly caused more damage to. Something in there got hot and stunk and a faint bit of smoke. I killed it quick but damage probably done now. You can see this one on the link above, it was a member "momaka" who suggested trying this but maybe I did something wrong but I was careful to make sure I had it connected right and only used 5v as told.

[budm]

Re: My popped bench PSU.

App notes:
https://cdn.badcaps-static.com/pdfs/...da6e24753a.pdf
Just one main filter cap, no filter cap and small bypass on the output side at all? I wonder if that thing oscillates like crazy with certain type of load or long output wires.
Probably from site like this:
https://forum.allaboutcircuits.com/t...ulator.107545/

[eccerr0r]

Re: My popped bench PSU.

If that loose red wire was connected to the screw tag that's on the 2n3055, it's isolated and connected to the case (collector) of the 2n3055. If you mean pin 3 of the LM317 (which is the input), then that sort of makes sense in the emitter follower type of series pass.

If you have the time I'd think it have better regulation if it were converted to a PNP-type pass transistor, either by using a helper small PNP or using a MJE2955. The circuit posted above with the raw 2N3055 with no pnp helper has poor regulation versus current draw as the LM317 has no exact input of the true output voltage; though designing it this way simplifies things greatly.

[budm]

Re: My popped bench PSU.

"Just popped this trying to test a Dell PSU" May I ask how you have this power supply hooked up to the DELL power supply? Connect it to the HOT side? Cold side?

[eccerr0r]

Re: My popped bench PSU.

Assuming the first two pictures aren't actually part of the design...

This looks like a typical series dissipative external pass bench PSU. Most of the dissipation is done by that 2N3055, and by the looks of it, it's very poorly heatsinked and likely what burned out. How many amperes/volts were you trying to get out of it, and what's the highest voltage you can get, just to get a good guess at how much power that 2N3055 needed to sink.

Could this be the reference design of your supply: http://www.bristolwatch.com/ele/lm317.htm

[roadrash]

Just popped this trying to test a Dell PSU I am trying repair and something obviously wasnt connected right. I bought this off ebay a few years ago. It was from a radio amateur who said he built it himself using a PC power supply. Ive just opened it up and its not exactly like what he said.
From what i can see its just a PC power supply enclosure he used and has fitted a Transformer with a Regulator and rectifier mounted on a bit of vero board It also has a Power transistor mounted at the front but (I dont know what that is for) and a Variable resistor and LCD display, When I took the lid off I


The vero board was just laying loose on top of a small sheet of plastic and was not held in place with anything and free to move around. It could have shorted out on something at any time. Not very well made!!
When I moved the vero board I found a loose (red) wire that looks like it came off a earth tag I think, but I am not 100% certain. It connects to the right hand leg of the LM317T.
So even the soldering is crap. Maybe this was a omen before it burst into flames at sometime. Well I am obviously going to get this fixed up properly when I get it repaired, He was also very naughty as I cannot see any inline fuse anywhere either. Its was miraculously all working fine until I hooked it up to the power supply I am trying to repair and it ddin't like it. There was no burning or smoke it just went dead.
Anyone suggest what to test first and maybe suggest any improvements apart from a inline fuse and securing that vero board. Anyway don't laugh, here are some pictures of it.
https://drive.google.com/drive/folde...4C?usp=sharing