My eyes tell me Q3 is a goner.

I know it's not in the voltage path drawn, but the base-emitter voltage is about 11.9V too large to be a good transistor. That and the discoloration (brownish) in both photos.

Q3 is in the -15v circuit, which is working perfectly.

Good catch, both Q3 and Q1 have the appearance of too much heat dissipation compared to the other devices on that board, they don't shine anymore.

Lefty

Appreciate your observations guys, however.......
I just went and took a second look at em.
They look just fine.
What you're seeing is the way the light hit them when the pics were taken.

 

Taking another look at it, the right end of R7 (Q2 side) doesn't look like it's soldered very well either; nor does either side of R8.

I have to get going (actually, to your neck of the woods) - there's an airshow in Titusville today.

 

Sarge;
Yes, you could swing by, but, you've probably left already. I live right off Hwy#50 and Barna.
We'll definitely have to go fishing sometime. I have a nice little flats boat I get out in and tear em up with.
I certainly owe you for your help ! If you hadn't of helped me, I'd of already thrown this thing in the garbage and had the new/better $6.00 one installed by now LOL.
I am learning great things here, especially about reverse engineering power supplies LOL.
OK, on to business.......
---------------------------------------------------
Continuing from yesterday evening........
Confirmed that there indeed was no continuity from R7 to LM317/adj.
Right at base of adj pin, as Sarge pointed out, cold solder joint, that I put there.
In my defense, I beg to offer the obversation that this was the second LM317 I installed with both offering the same results.
I had soldered it in from the bottom side as most all the solder points seem to of been done.
I suspect the board was passed over a solder wave machine in mass production style, etc.
That leaves the topside not too well soldered, and I now have learned about soldering something from the bottom when the copper trace is located on the topside........
Anyway, fixed it.
Applied power, and ??????
Change, but no fix ! ! !
Now, instead of having fixed voltage of 18.5V, I have a new "fixed" voltage of 13.2V !
Still no variable pot.
Have checked all copper traces for continuity topside and bottomside.
Re-checked all voltages, and drops, noted in new pic, old values and new values.
Note--> Can check voltages one time, come in here and do computer stuff, return 30 min later and re-check = find values can change as much as 2.5V.
Example = LM317/adj was 13.6v, now is 11.8v 30 min later.
Doesn't help having digital multimeter jumping all around !
(My Kingdom! My Kingdom! for my old Simpson 260 back ! ! )
D3 varied from 1.7v to 5.5v
D4 varied from3.7v to 7.5v
I believe they're supposed to drop around 1.7v each ????
OK, that's my update for this morning.
Sure wish I could get this thing put to bed and move on !
Have a nice day.....
Oxbo

 

Started out this morning double checking all values, updating circuit diagram, etc.
Replacd Q5 with 2N3904 = not much change at all.
Replaced Q6 with 2N3906, output went from 13.16V to 7.2V, still, no varable.
Have updated the circuit diagram with these "last" values, see attached.
Well guys/Sarge, I believe it's about time to weigh the worth of this endeavour with the worth of our time that we could be moving on to something else, etc.
Will be ordering that new Pwr supply tonight.
I hate it when I can't figure something out.
Been a long time since I messed with circuitry, don't have a lot of patience these days, etc.
So, with this last message, as was with that infernal mono 555 timer Bernard and I were working on, I must realize an end to the project and settle for a compromise ending (I did get the little timer down to 12 sec, but, only one shot at a time).
I have literally spent days from first light to well past last light messing with this thing = burnt out !
Sarge, I appreciate all your help, but, lets us move on to other things now, and to heck with wasting any more time over a $6.00 pwr supply, etc......
Have a nice day,
Oxbo

 

Check all small signal transistors, any that have more than around 0.6V between Base and Emitter have issues, Q5 in particular, but there may be others.

 

Check all small signal transistors, any that have more than around 0.6V between Base and Emitter have issues, Q5 in particular, but there may be others.

OK, let me ask you a couple of questions.....
The -15V circuit works perfectly fine.
Q3 has a b/e voltage of 15.4v, so, what's with that?
My continuity meter continuously beeps (direct short) when I place it
on Q6 b/e, but placing it on Q4 b/e, also sounds a direct short, however, Q4 circuit works just fine.
Also, Q3 b/c sounds direct short. It also is in the good circuit.
Frankly, I'm lost..........

 

Sorry, I had a very long (yet fun) day today; left the house right after I posted, and didn't get back until after 11PM.

The airshow was great. I'd never seen an F-104 Starfighter fly before. Saw an F-4U Corsair flying that was featured in the 70's TV show "Black Sheep Squadron".

Saw the Shuttle launch; quite impressive. Had a great view from a "secret fishin' hole" place.

The trip back was murder. It only took around a half-hour to get to Titusville in the AM, but over three hours to get back home. Seemed like everyone on the planet wanted to see the launch.

Anyway, way off topic. I'll take a look at your results tomorrow.

 

OK,
I'm still on it, been on it since 5am. (is almost 3pm now).
--------------------
Removed R5, measured = 2.16K
Replaced R5, measured = 2.19K
Ck'd continuity, voltages = no change.
--------------------------------------
Mesaured VR1 voltages = pin 1= 1.3v, pin #2= -14.2v
Mesaured VR2 voltages = pin 1= 6.2v, pin #2= 6.2v.
Accidentally checked continuity of VR1/VR2 with power on.
Found VR1 = no continuity
Found VR2 = dead short, however, turning pwr off,it was same as VR1
Disconnected both (one at a time) from circuit checked values,
Both check same and OK.(entered values on circuit diagrams)
---------------------------------------
Discovered circuit error on my part.
Need to reverse these wires (to as shown), don't think it'll make a diff, etc..
See attached.....
------------------------

 

Finally ! ! !
I fixed it ! ! !
Have to re-check values and update drawings, etc, etc.
Will post in the morning.......

 

I'll bet that it has something to do with R5's connections.

R5 is supposed to pull the junction of R5/D4 to a negative potential. This allows adjustment of the output of the LM317 down to 0V. With +15Adj turned down as far as it will go, VR2 should be adjusted so that Vout is 0v or that the voltage on the ADJ pin is -1.26v.

Also, if you look really carefully, I'm pretty certain that you'll discover that pots VR2 and +15Adj are connected in series rather than in parallel.

 

Alrighty now ! !
Sorry for being so late.....
Having gone over the circuit so many times doing continuity checks and voltage checks, and just looking and looking at it, decided only one thing to do, start all over, re-replace everything !
I started out looking at Q5 & Q6.
To my utter amazement, having double/tripple checked them, I discover that some moronic idiot has put them in exactly/perfectly "BACKWARDS" ! !
Replaced them, and the dad-gummed thing worked ! !
Have spent the rest of the time documenting compared voltages.
Everything works fine now but there are a couple of differences (noted on drawings)
Would appreciate your analysis as to why these differences exist, etc.....
Sorry for being so stupid.......

 

When a transistor is bad/backwards/etc, voltages everywhere can change, even in seemingly unrelated areas. This is due to direct coupling and biasing changing everywhere in the circuit, making no voltages read normal.

This is especially true when the transistor is near the early stages, acting as a current source for biasing, or a current sensor for current limiting.

Below is a quick overview of how I look at a dead widget, of low to medium complexity "malfunction unknown". The areas related to electrolytics do not apply in your case.

Most often, a switching power supply is involved. Different circuits get different checks, based on common failure points, and the symptoms. First, remove dust, but do not use any solvents (water/alcohol/etc) that will erase clues/evidence. That is done once the initial lookover is completed.

When looking at a broken widget, when fuses are good, wires aren't broken, and I do not have a schematic (mostly never do). I start with anything that looks/smells burned in the case of transistors/resistors, or leaky/exploded in the case of caps.

No visible issues, I use an ESR or LCRZ meter on the Electrolytic caps and low value resistors, they will have the lowest resistance of anything else around them (usually under 2Ω), so measuring in circuit is easy and fast: only a couple minutes for an entire board. 20 years ago, most failures were from a transistor that died somewhere. Recently, electrolytic caps are more often than not involved in failures. Exploding electrolytic caps spray aluminum flakes everywhere that cause never-ending problems if not completely cleaned up before applying power

If nothing abnormal is found, I power up the widget, using an isolation transformer if item is line powered, and work from the power supply input "outwards" to the various sub circuits. Any time a transistor is found to be bad, odds are good that a resistor in the neighborhood is also toasted, and vice versa. Sometimes the components get hot enough to melt solder and form bridges. I touch the platic cases for anything warm, a cheap IR thermometer is useful for this to avoid shocks.

Overheated components often have their cases take on a brownish or greyish tint, both the black plastic, and the component writing. I compare with other similar componets to see if the color is "ok". Some manufacturers write in tan ink, which makes this less reliable. In addition, the solder mask (Usually green) will have a darker tint in an overheated area.

Most of the tests above are done in circuit, without desoldering, mostly at voltages too low to forward bias any PN junctions, usually 0.1-0.2VAC for ESR/LCR meters, and 0.2VDC for a Multimeter Resistance. With more and more surface mount components, it goes to intuition and black majic, as not all 3 terminal SOT-23 devices are transistors, and even finding a schematic often doesn't help much, as many items aren't labeled, especially at digital camera/cell phone circuit density and component sizes.

An ESR/LCR meter isn't a "Must Have Item", a scope of any quality is before that, as an adapter can be made to make a scope into an ESR meter.

 

thatoneguy;
Thanks for that.
This is a somewhat whole old/new ballgame for me.
There was a time (40 yrs ago) I fixed aircraft radios in the air force. Was not too difficult, having every kind of test equipment available, and especially TO's (Technical Orders).
These were the publications provided by the gov from the radio manufacters with all the detailed schematics and volumes of trouble shooting literature, etc.
But once getting out of the service I drifted on and never did get back into electronics. About 20 yrs ago I put together a little switching power supply, bought at Radio Shack, but I made a 555 varable pulse generator and one-shot to be included in the box, etc. Not too challenging, then bout 10 yrs ago did my little mail box door alerter (actually caught a kid yesterday opening the mailbox, quickly ran out there and caught his azz.) Now, I'll probably have my mailbox blown up LOL.
And now, decided to make me a boat burglar alarm, thus, back in it again.
I really enjoy messing with it, but try not to get too complicated.
Was about ready to toss it all again, till, finally, yesterday, managed to fix this thing, etc.
I certainly appreciate everybody's help. Feel pretty good today. Got my new power supply in, decided after all the time I put into fixing this one I'm going to keep it and make a new box for the new one.
I got it all put back together, installed a banana plug for extra lead, and a little switch where I can switch the -15V meter from the -15V supply to being able to utilize it on the breadboard, etc.
OK, supper's on,
Have a nice day,
Oxbo