My popped bench PSU.

[eccerr0r]

Re: My popped bench PSU.

What ended up being the last problem before it started working again?

[roadrash]

Re: My popped bench PSU.

Finally ive got it working again but I will have to get it better protected and better heatsink on the transistor.
Thank you everyone who helped...

[eccerr0r]

Re: My popped bench PSU.

But even if you do have a huge chunky heatsink, that still won't save your transistor from short-circuit, especially when you have large differential between the input and output voltages.

Yeah, I sort of made that call when I used a TIP42 as the pass transistor. This transistor is quite marginal to even help out being a pass transistor as its SOA is probably not much larger than the LM317T by itself. Now whether to go ahead and parallel a couple TIP42s or go find a TO-3 ... or do nothing until my spare parts box gets some extra TIP42s or a TO-3 to show up in it...

[momaka]

Re: My popped bench PSU.

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.

But even if you do have a huge chunky heatsink, that still won't save your transistor from short-circuit, especially when you have large differential between the input and output voltages. Reason being is that you will quickly exceed the SOA of the transistor way before the LM317 has time to react. This is why good quality bench PSUs use either many parallel transistors or one huge one. The idea is to keep the transistor(s) within the SOA, even if a short-circuit occurs when the PSU is outputting maximum current.

[momaka]

Re: My popped bench PSU.

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.

Ooops.
Well, I have to say that I had no idea that you'd use a "bench" power supply like that instead of a 5V power adapter, as I suggested. Not that it would have made a difference. That Dell PSU should have been connected through a series bulb for protection / primary current limiting purposes (always do that when working with faulty switchmode PSUs). If you had it connected correctly, then most likely it had something shorted on its output, and that caused your bench PSU to die, thus possibly dumping 30-odd volts back into the Dell PSU.

That said, looking at the pictures of that home-made bench PSU...
I really am speechless. If someone sold me that, I'd find where they live and smack them on the head with it.

Seriously... no fuse? Veryboard hanging around loose? Output transistor mounted to the case as a heatsink? This thing is a fire hazard!
To be honest, I don't think you should have any remorse about it dying. In fact, maybe it was for the better that it did - at least better in front of you than when you're away, possibly setting your bench on fire or who knows what else.

Also, I'm sure you could do a much better job at building your own bench PSU, even if you say you don't know much electronics. The guy who did this couldn't have known more than you did. But anyways, I'll stop with the bashing here and move onto getting this bench PSU working.

As a first step, I would say gut everything the guy did, minus the transformer. Before doing anything else, first add a line fuse in series with the transformer. A fuse rated for 1-2 Amps should be okay here. Make sure it's also rated for 250V or more.

Next, forget about the pass transistor. Yes, a pass transistor is good to have when you need larger currents. But what good is that if there's no way to protect it and the PSU self-destructs every time you encounter a short-circuit. For a _bench_ PSU, you need something more robust that can safely shut down in case of an accident. So going solely with the LM317 regulator would be wiser.

Open this datasheet:
https://cdn.badcaps-static.com/pdfs/...f670e845db.pdf
And built the circuit shown on page 10. That circuit has reverse-current protection diodes, so your bench PSU won't die if you accidentally happen to connect it to something with power.

Then simply, take the circuit you just built and connect it to the *rectified* and *filtered* *DC* output - i.e. a circuit you build with that transformer, the bridge rectifier, and capacitor that came with the original "bench" PSU.

As everyone else noted here: make sure the LM317 is attached to a nice big heatsink.

Contrary to what people often say, the LM317 is NOT just capable of 1.5 Amps of current - it can do more or less than that, depending on the input-output voltage differential. Many older LM317 datasheets tend to omit this, but the LM317 has more or less a limited power dissipation of 20 Watts, as shown at the top of page 6, 7, and 8 of that datasheet (where it says Pmax = 20 Watts). And on page 9 of the datasheet, you can see a graph of "Output current vs. input-output differential voltage". This essentially tells you how much current you pull from the LM317 regulator.

With a low input-output differential voltage, (meaning when your bench PSU is outputting close to its maximum voltage), you can get over 2 Amps of current output. But at a high differential input-output voltage (with your transformer, that means when the PSU is outputting 5V or less), you're limited to around 500 mA.

So it's understandable why many people put a pass transistor - it really amplifies the current you can draw through the LM317 - especially at low output voltages. But the lack of protections is a big no-no for a bench PSU, especially for "learning" or "experiments".

[eccerr0r]

Re: My popped bench PSU.

The transistor is used as the series pass device in emitter follower mode. Most of the power draw that leaves your PSU is going through the transistor instead of the LM317T. It's called "emitter follower" because the emitter follows the voltage at the base, which is connected to the LM317T output. The LM317T output is (theoretically) stable and thus the emitter voltage is also stable.

But the LM317T won't work and can be destroyed if that bulk capacitor, that 2200uF 40V capacitor, is actually connected to the output instead of the input of it. There looks like a little loop of red wire near that capacitor, if that red wire is actually connected to the positive of the capacitor and the capacitor only has one leg actually soldered into the board (!!!) then that might be correct.

[budm]

Re: My popped bench PSU.

I think you need to verify if the (+) of the rectifier is connected directly to the (+) of the that large blue cap or not, I cannot tell from looking at the bottom side of the perf board.
That filter cap should be connected to the output of the rectifier first to provide clean DC to feed the regulator circuit, if the cap is not present at that location then you will be feeding 100Hz AC into the regulator circuit.
Look at the diagram as how that big blue filter cap should be connected to the rectifier.
https://www.badcaps.net/forum/attach...3&d=1518994414

[roadrash]

Re: My popped bench PSU.

I am still not understanding the reason for the transistor. Sorry I'm no expert I'm still learning. If I have 23VDC out of the rectifier and a the input pin on the LM317T then cannot that alone regulate the voltage output according to the position of the variable resistor. Anyway I cannot read any output voltage from the LM317T yet I can measure a change in resistance between Gnd and the ADJ pin when the variable resistor is turned.

[roadrash]

Re: My popped bench PSU.

The TO-220 has three pins on one end and a metal thing sticking out the top (that has the hole in it). This tab is NOT isolated from the IC, it's connected to the middle pin, which is the output. If you screw this onto a grounded chassis, you'll short the output to GND. If it has its own isolated heatsink, that should be fine, just make sure it doesn't touch something that's not also connected to the output of the device.

Thank you for that info, I didn't realise that. Its OK though as I only attached the heatsink I used to the LM317T itself as fixing it to the circuit board would have been to awkward so it has no connection with the circuit board.

[eccerr0r]

Re: My popped bench PSU.

My take is that if you're trying to run 5A, you may need up to 500mA or so through the LM317. At this current, depending on voltage, yes indeed you would need a heatsink... but the 2N3055 would melt down first with a poor heatsink solution, and you then won't really need the heatsink on the LM317 after all

[R_J]

Re: My popped bench PSU.

Here is a picture, I dont think there is any real need to heatsink the lm317 if you are using it with the 2n3055, The 2n3055 should have a proper heatsink as it handles most of the current

[eccerr0r]

Re: My popped bench PSU.

The TO-220 has three pins on one end and a metal thing sticking out the top (that has the hole in it). This tab is NOT isolated from the IC, it's connected to the middle pin, which is the output. If you screw this onto a grounded chassis, you'll short the output to GND. If it has its own isolated heatsink, that should be fine, just make sure it doesn't touch something that's not also connected to the output of the device.

[roadrash]

Re: My popped bench PSU.

Did you also verify that the case of that power transistor is not shorted to the chassis?

yes

f you relocate the LM317T, be careful, its metal tab is physically the same as the middle pin - output, so don't short it against something.

what tab? its just a TO220 3 pin IC

its measuring around 23v at the input leg of LM317T

[budm]

Re: My popped bench PSU.

Did you also verify that the case of that power transistor is not shorted to the chassis?

[eccerr0r]

Re: My popped bench PSU.

Also if you relocate the LM317T, be careful, its metal tab is physically the same as the middle pin - output, so don't short it against something.

Ok, good info. 30V (AC RMS?) from the transformer.
Having just 25V out after the rectifier sounds too low if the transformer is 30V RMS.
Since the display doesn't illuminate, that's also good information. The wires are unfortunately a big jumble, not sure how this is connected to power...

Having a heatsink on the LM317T is "gravy" unless you want to get rid of the main transistor. The main transistor is the device that needs the heatsink. It's hooked up real weird as a emitter follower to amplify the output of the LM317T.

You should be getting an adjustable voltage out at the metal tab of the LM317T (since it's the output).

And looking at the pictures again... holy crap, is the input not filtered and the big capacitor is actually on the OUTPUT of the LM317T? This will kill the LM317T's...

You may need to redesign this thing if the 2200uF 40V cap is truly hooked up to the output... that's horrible. How much do you want to invest in this thing? Now I believe you that this thing is pretty much garbage if my eyes aren't deceiving me on how this thing is hooked up. Surprised there's no capacitor on the output either though they should be smaller than the input capacitor.

[R_J]

Re: My popped bench PSU.

what is the voltage on the input and output of the lm317, what is the voltage on the control. (adjust pin of the regulator)
The regulator ic (LM317) can only output about 1 amp. so you need the transistor to be able to supply more current, The lm317 provides the regulation and adjustability, and the transistor 2n3055 handles the current, The full supply voltage is applied to the collector, and the output voltage at the emitter is controlled by its base, which is connected to the lm317's output pin.

You say you mounted the lm317 to a heatsink? is the tab isolated?? or is it connected to ground, that tab is the output and can not be grounded

[roadrash]

Re: My popped bench PSU.

I cant seem to get this working. Its got a output of 30v from th tranformer and then a output of around 25v after the rectifier. But there is nothing output from the transistor. I replaced all the semiconductors including the LM317T which I relocated and added a heatsink to as suggested. I don't understand what the purpose of the transistor is so cant understand how this works. The Potentiometer has two of the three tabs bridged but that should be like that. But the problem must be with the circuit around the LM317T. Now everything is new can someone suggest a few tests to find whats wrong. The digital display does not illuminate either but I don't know if its working now either. I traced all wires & tracks on the PCB against the photos I took & the schematic given and it all seems to be correct. I'm stumped.

[eccerr0r]

Re: My popped bench PSU.

I notice there is a 47uf Electro cap between R1 & the variable resistor but mine doesn't have this. Can you tell me what the purpose of this cap is and should I put one on my psu?

This cap filters the output but really don't think it affects operation that much. It keeps base voltage more constant but it shouldn't be that bad to begin with. Adding one will not save the PSU from the apparent overload it sustained.

Again the larger heatsink is the top need, nothing else matters before this.

[roadrash]

Re: My popped bench PSU.

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

I notice there is a 47uf Electro cap between R1 & the variable resistor but mine doesn't have this. Can you tell me what the purpose of this cap is and should I put one on my psu?

[roadrash]

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.

Yes I see that now when removing the 2N3055. I need to open my eyes a bit more. thanks