LM317 variable power supply question.

[Doc38343]

On another front... I think this little bugger may be just what i need to isolate a 5V supply for my LED VM. I saw a circuit to modulate using a 555 then It may have used a power transformer but it did end up on a 7805 regulator... my led meter says from 5v to 30v is ok. I also have a 9v regulator i found so I was thinking if I add it and use a 12V as the supply I would have less overall dissipation due the the decreased voltage difference. also if i'm not mistaken I can use the 5V to feed the 555 and transistor and the peak to peak result would be 10v or close..

see what you think of this transformer... I need to build me a frequency modulator next so I can test these buggers!!

[tautech]

Looks like a common core mains filtering choke, the core might not be suitable for transformer use. 

[Ian.M]

All they are doing is acting as current sources (actually constant current sinks) to develop a bias voltage across D1,D2, and D3,D4 respectively to permit the output voltage to be adjusted right down to zero and also the current limit the same way.



A JFET's Idss is not a well defined parameter and its probable that the current would vary over something like a 2:1 ratio between different FETS with the same part number. 

e.g:
A 2N3822's Idss is expected to be between 2mA and 10mA
A 2N5640's Idss is expected to be between 5ma and 15mA

Substitution by any other sort of current sink should work, as long as there is enough current to get the diodes  above them conducting.   I would avoid OPAMP based current sinks as you may run into stability or bandwidth problems.

A -12V bias supply gives you 2V extra headroom  so a high gain NPN transistor with its base to a 2.7V Zener holding its base above the -ve bias supply with a pullup to 0V and an emitter resistor to set the current should work nicely.   Both sources can share the same Zener and pullup resistor.  Try 390R for the Q1 replacement's emitter resistor for just over 5mA current and 180R for the Q2 replacement for just over 11mA. 

Performance may actually be better than with the JFETs, but be prepared to fiddle with the emitter resistor values to get zero current and zero voltage right at the end of the control pots' rotations.   Maybe even use a 1K preset and a 470R preset.  What you are doing if you make them adjustable, is setting the current through the two diodes associated with each of the original JFETs so that their total Vf matches the associated LM350's reference voltage.   Some experimentation will be required.

200mA is a lot of current for meters - you probably should have gone LCD.   Scratch-building a reliable isolated DC-DC converter for 200mA out at 5V is not so simple.  I suggest either using the guts of a mains to 5V USB phone charger if you've got room to mount it safely or getting an off-the-shelf potted module, e.g. http://uk.farnell.com/tracopower/tmh-1205s/converter-dc-dc-2w-5v-0-4a/dp/1007558

[Doc38343]

Looks like a common core mains filtering choke, the core might not be suitable for transformer use. 

Aye aye.... I've literally got a box full of those little transformers but I agree with Ian and yourself... I will just power it from a wall wart i've got 3 boxes full of em. I have and embarrassingly large collection of ... junk? parts? huh old crap people bring to my shop for me to "dispose" of.. LOL

Before I started watching EEVBLOG on youtube and made up my mind to get back to doing actual repair, I was just a parts changer. Also I probably threw away 5 or 6  big 24" monitors not to mention the 20 or 30 various smaller sizes, that i know had a couple of vented caps and could have been easily repaired. (I was not thinking clearly)

Here is a couple of pics welcome to my world, and this is only the immediate surroundings in my work mess.. organized chaos! 

I also have a wall filled with old computers that i call the "Wall Of Shame"

[Doc38343]

UPDATE:
  I don't know what i did but the ATX supply shuts off under any load greater than a 555 or inputting the LM350's. I can hook up a LED or something really small current and it stays on but I tried hooking a standard 12v computer fan and it trips off instantly. even if it is the only thing I connect. my fluke shows only 5.4 mA when i have the 555 timer connected and the LED as well as the digital voltmeter. I was thinking wow that is low can't be right or that I was reading the meter wrong but I did it with another borrowed Fluke with the same reading. (I just got the 87V a few days ago and am still learning it) I used a cheap micronta radio shack before that and a really CHEAP Chem-tec $9 piece of crap so I am not sure if I'm understanding the new Fluke auto ranging and all.

What a good learning experience and total waste of time and resources converting this ATX  piece of crap. If I do another I will build a connector into the variable box and just plug it in so i can change them easy. However to heck with this crap I have some really good transformers and rectifiers so I am now changing gears to get me something that will smoke! LOL You guys helped me find some great schematics for build from scratch.... Phase 2 begins.... 

[Ian.M]

That's the trouble with ATX PC PSUs.  The rail voltages aren't that convenient and its a PITA to lift the secondary side ground.  Also if its used, the standby supply is usually nearly shagged as it runs all the time and if that fails it wont power on.

You can make a decent +ve only adjustable bench supply by hanging linear regulators off a laptop or printer power brick, but unless you can find one with a two wire mains lead, 0V will be grounded so you cant stack them for more voltage or to get a negative rail.

[Doc38343]

OK here is the "Beginner" noob-out... I got to thinking WHY? this is supposed to be a 550W PS and I know it is used but I also had it since new... I did repair it and it ran great but I didn't trust my repair fully so i replaced it and used it for this ... I know it would carry a load before I cut the wires and connected them so I back tracked and tested all wires are OK no shorts or anything until I tested the 5V rail.. and I haven't connected it to the terminals yet or anything they were inside heat shrink but my test showed .36 OHM! So I did have a ceramic 5ohm 5W load resistor I had installed per lots of instructions but I know it showed 4.7 OHM's on my new Fluke so how can it be .36 WAALA I learned something... This supply is more modern but has reverse compatibility to work with a wider range of PC's... It has a built in load stabilization! Just to be sure I cracked the lid and there it was in plain sight a ceramic load resistor. I disconnected mine and just like that it is working correctly. I was thinking perhaps the problem was due to the isolation i had done but I am very glad that it was the resistor. Now it may not be the best but it will certainly serve me on some levels and for experimentation. NOT a waste of time and resources after all... and an even better learning tool than i had hoped since my primary objective for electronics is to be able to repair things and understand the circuits at a workable level.. THANKS for the assist also.
Still going to do a scratch also... gotta get some components first.. freaken JFET's are hard to find salvage.