I'm building Carlosfm's AD815 preamp, and having some trouble. I built a +/-17v unregulated power supply for it. Unloaded, it tests perfect, within .01V. It is simple, build on a circuit board and uses a large transformer. I have little doubts about it.
I've built a seperate PCB for the regulators/DC nulling/op amp, and I think ts wacky. When I connect only the regulating circuit (set to +/-12.7V), I get +12.99V and -12.68V. The negative looks fine on a meter, but the positive is unstable - it settles slowly from 13.5V to about 13V, with some play remaining.
Then if I connect the DC nulling circuit and/or the opamp cirucit, the thing goes real off. The positive will only rise to about 6V, the negative drops to about -22V. I watch it on the scope, and turn the circuit off quickly. I havn't measured any amperages yet.
Does this look like a short somewhere? I made my own PCBs with Eagle, and I was very careful to give good clearance, and while soldering, not to create solder bridges (it would be hard, there is about a mm clearance).
I think something is up with the +ve regulator circuit. Here is Carlos' original circuit. I've changed nothing except change the 3k3 resistors to 2k2 to adjust the regulator output voltage. I used 1n4148 diodes around the regs, is that okay?
http://www.diyaudio.com/forums/attachment.php?s=&postid=765880&stamp=1131902668
I've built a seperate PCB for the regulators/DC nulling/op amp, and I think ts wacky. When I connect only the regulating circuit (set to +/-12.7V), I get +12.99V and -12.68V. The negative looks fine on a meter, but the positive is unstable - it settles slowly from 13.5V to about 13V, with some play remaining.
Then if I connect the DC nulling circuit and/or the opamp cirucit, the thing goes real off. The positive will only rise to about 6V, the negative drops to about -22V. I watch it on the scope, and turn the circuit off quickly. I havn't measured any amperages yet.
Does this look like a short somewhere? I made my own PCBs with Eagle, and I was very careful to give good clearance, and while soldering, not to create solder bridges (it would be hard, there is about a mm clearance).
I think something is up with the +ve regulator circuit. Here is Carlos' original circuit. I've changed nothing except change the 3k3 resistors to 2k2 to adjust the regulator output voltage. I used 1n4148 diodes around the regs, is that okay?
http://www.diyaudio.com/forums/attachment.php?s=&postid=765880&stamp=1131902668
Uh oh, I didn't use a LM317 positive voltage reg - for some reason I used a 'ld1086v33'.
http://www.mouser.com/search/productdetail.aspx?R=LD1086V33virtualkey51120000virtualkey511-LD1086V33
I though it was adjustable, but I think this specific item is 3.3V fixed!
More soldering at hand...
http://www.mouser.com/search/productdetail.aspx?R=LD1086V33virtualkey51120000virtualkey511-LD1086V33
I though it was adjustable, but I think this specific item is 3.3V fixed!
More soldering at hand...
okay, now that I've gotten a real adjustable voltage reg in, I get good unloaded voltages everywhere.
BUT, I still get the strange drop - with the DC null circuit in, and the opamp circuit connected, when I switch on the power, both + and - voltages jump up to about 13V, but immediately the +ve supply begins to sag, and after about 10seconds, it is at 4V and still slowly dropping...
Any ideas? A short?
BUT, I still get the strange drop - with the DC null circuit in, and the opamp circuit connected, when I switch on the power, both + and - voltages jump up to about 13V, but immediately the +ve supply begins to sag, and after about 10seconds, it is at 4V and still slowly dropping...
Any ideas? A short?
I isolated the problem to the null offset transistor. well, as it turns out, the pin out on my 79L09 transistor is not assigned correctly in Eagle. So I set that straight.
Unfortunately, with the transistor in with the proper orientation, the problem continues. I think I blew the transistor. Is that possible? Connecting the output to +15V, the input to ground, and the ground to out?
I bet so...
Any transistor buffs want to chime in?
As it stands, the circuit is done. Horray! I just need another transistor..........
Unfortunately, with the transistor in with the proper orientation, the problem continues. I think I blew the transistor. Is that possible? Connecting the output to +15V, the input to ground, and the ground to out?
I bet so...
Any transistor buffs want to chime in?
As it stands, the circuit is done. Horray! I just need another transistor..........
Check the transistor with a DVM having a diode check position- that would be most of 'em. If you get a diode drop between base and emitter, and between base and collector, it's almost certainly good.
FWIW, almost every layout problem I've ever had is because of bad models. There's some kind of unwritten rule that voltage regulator and transistor pinouts will be wrong, and if by luck it's right, the hole will be too small for the leads. This may sound crazy, but the first thing I do now with any layout program is throw out all the models that came with it, and make my own- taking care to have the physical part, the data sheet, and a pair of verniers at hand. Since I use AP Circuits for my boards, I be sure to create models that take advantage of their free hole sizes, unless something absolutely positively has to have its own specific size.
FWIW, almost every layout problem I've ever had is because of bad models. There's some kind of unwritten rule that voltage regulator and transistor pinouts will be wrong, and if by luck it's right, the hole will be too small for the leads. This may sound crazy, but the first thing I do now with any layout program is throw out all the models that came with it, and make my own- taking care to have the physical part, the data sheet, and a pair of verniers at hand. Since I use AP Circuits for my boards, I be sure to create models that take advantage of their free hole sizes, unless something absolutely positively has to have its own specific size.
Conrad Hoffman said:
probably a good suggestion. I have a vernier caliper, and use it often. It is very handy.
now, the part in question isn't actually a transistor, its a voltage reg in a to92 case. The data sheet says its 'virtually indestructable'. I have a DMM with diode and transistor testing, would your suggestion still work?
Here is the datasheet:
http://www.st.com/stonline/books/pdf/docs/2511.pdf
- Status
- Not open for further replies.