Peavey Regulator on the fritz

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

CJ

Well-known member
GDIY Supporter
Joined
Jun 3, 2004
Messages
15,664
Location
California
replaced all the transistors and checked diodes, darn regulator won't regulate,

don't understand why reversed biased transistor (bottom) is in the circuit,

first version missing 10K resistor, so we added it to the pc board but still n o worky

maybe a one chip solution is the way to go, but we are right at the max input V for most reg chips,

any ideas?

 

Attachments

  • reg.jpg
    reg.jpg
    122 KB · Views: 54
CJ said:
replaced all the transistors and checked diodes, darn regulator won't regulate,

don't understand why reversed biased transistor (bottom) is in the circuit,
A reverse biased base-emitter junction will zener at around 6.8V. That is the reference voltage the regulator is regulating with  (schematic says 9V not sure I'd expect that high.)
first version missing 10K resistor, so we added it to the pc board but still n o worky
without that 10k the current source supplying the output base drive will not turn on, so no output voltage.
maybe a one chip solution is the way to go, but we are right at the max input V for most reg chips,

any ideas?
That is a sh__ simple regulator, with typical voltages on the schematic so see where it is off the rails.

top device is pass element with base current coming from 4249 current source.

3904 is doing actual regulation pulling base current from the pass element if output voltage too high.

3391 is simple current limiter.. (at roughly 100mA).

Whose initials claim credit for that design..? They deserve special recognition for using 5 different transistor part numbers in a 5 transistor design.  :eek: :eek: :eek: If he worked for me I'd smack him...

JR

PS: In fact Peavey had real zener diodes in the system that were machine inserted like resistors, before we got the machine insertion of transistors (in the 90s) so the labor to hand insert the transistor more than cancelled the savings of using the transistor as a zener... Some engineer just trying to be clever, instead of smart.  8)
 
CJ said:
Thanks JR! we are working on voltage readings,

looks like Jack checked it so it should be alright,
Damn that's "the" HDP who approved it.. ;) . Probably Jack was the design engineer, showing off for Hartley's benefit using a transistor as a zener.  ;D ;D I don't recognize the draftsman. That's a good ten years before I worked there...

I guess I wouldn't smack Jack Sondermeyer (RIP.  8)

Knowing Hartley they were probably trying to use up old transistors that were laying around so they didn't have to throw them away.

JR
 
The current limiter transistor PNP 2N3391 needs a base resistor to limit the current , say 1k or so else the current through the EB will rip out the junction.
John's suggestion to pull some current through the ref device (10k pullup) is needed for a better design.
Duke ;)
 
on the board i have there is a 4.7 ohm resistor on the base of the current limiter.

did a quick breadboard of the circuit and it works fine. even without the 10K, the pot moves the voltage from about 9 to 32 volts, double checked the pc board circuit and it matches the breadboard  circuit i did, last culprit is that the +40 has somehow been linked to the 24 volt output side,

this circuit is not super sensitive to transistor types, we used MPS8097 for the voltage ref and current limiter, and a PN2907a for the 2N4249 PNP transistor.


http://adrsoundsense.co.uk/technical/peavey/PV_Semi_Cross_Ref.pdf


 
Audio1Man said:
The current limiter transistor PNP 2N3391 needs a base resistor to limit the current , say 1k or so else the current through the EB will rip out the junction.
The current available is already limited by the pnp current source, but there is a 1k in series also...  I'm pretty sure the design works.
John's suggestion to pull some current through the ref device (10k pullup) is needed for a better design.
Duke ;)
I may have misunderstood CJ's original post...  If he was talking about the 10k only present in the second version of the schematic  that will improve regulation but circuit should still work without it. I thought he was talking about the10k connected to pnp current source that makes it work.

Looks like the later version only uses 4 different transistor types.  8)

JR
 
CJ said:
on the board i have there is a 4.7 ohm resistor on the base of the current limiter.
The resistor establishes the current limit... 100mA through 4.7 ohm develops 470mV that turns on the current limit transistor and sucks out base drive current from the pass transistor.
did a quick breadboard of the circuit and it works fine. even without the 10K, the pot moves the voltage from about 9 to 32 volts, double checked the pc board circuit and it matches the breadboard  circuit i did, last culprit is that the +40 has somehow been linked to the 24 volt output side,

this circuit is not super sensitive to transistor types, we used MPS8097 for the voltage ref and current limiter, and a PN2907a for the 2N4249 PNP transistor.


http://adrsoundsense.co.uk/technical/peavey/PV_Semi_Cross_Ref.pdf
Yup...  pretty conventional PS topology...  using the reverse biased b-e junction is a little too cute IMO (they are noisy used that way too), but other than that a solid design. The adjustment range seems a little large for a fixed rail... Prudent design would reduce the adjustment range, and also make it so a broken open-circuit pot (trimpot) defaults to some reasonable output voltage.

JR
 
ok we got er fixed,  turns out the regulator transistor 2N3904 does not like anything over 40 volts, this is a 1975 amp so we see 42.5 volts on the input rail which the transistor might see during  startup , Fairchild 3904 series not he best,

that 10K tends to keep the ref voltage stable while turning the pot to adj voltage by supplying more zener current,

we have drawn in the 4.7 resistor on the 3392 base.

we stuck a 100 volt part, MPSA06 (because that is what we had) for the reg transistor,
voltage from the pot rotation is still wide which could be fixed by using a lower hfe part, but things are clean and stable so what the heck, engineers always seem to go for too much control instead of too fine so if anything changes they do not need to switch pot values, (witness Fender bias pots)

small note of interest, in one schemo the collector of the ref is left open, on the other schemo it is grounded, does not seem to matter,

we are still puzzled as to where the current goes thru the 2N3904 as the rev bias ref probably draws very little current, maybe the current goes thru the BCjunction and thru the pot to ground?

here is the voltage chart for a working regulator thanks for all the help!  :D

we could have used a 7824 with a few diodes in the ground leg and held our breath but we are glad to have the whole Mary Ann in there like a stock amp.

 

Attachments

  • pv fixed.gif
    pv fixed.gif
    27.1 KB · Views: 15
CJ said:
ok we got er fixed,  turns out the regulator transistor 2N3904 does not like anything over 40 volts, this is a 1975 amp so we see 42.5 volts on the input rail which the transistor might see during  startup , Fairchild 3904 series not he best,
Actually during start up the current source has to charge up the capacitor so I would not expect turn on transient higher,,, 3904 is between base of pass element at 25V and zener connected transistor so all of 15v, unless trim pot turned to zero then it sees maybe full unregulated collector to base... BTW I still don't like that trimpot going full range (especially to 0V).  :eek:
that 10K tends to keep the ref voltage stable while turning the pot to adj voltage by supplying more zener current,
yes, stiffens it up, but only 1.5mA not huge in context of a couple mA down 3904. Does give a sharp cut off when pot turned below reference V (but pot shouldn't be turned that low). 
we have drawn in the 4.7 resistor on the 3392 base.
That should not be necessary if circuit working properly, but could protect the current limit transistor "if" PS output shorted to ground "and" pass element has failed hard on. May reduce repair cost/complexity (something Peavey thought about at least during my time there 10 years later).
we stuck a 100 volt part, MPSA06 (because that is what we had) for the reg transistor,
voltage from the pot rotation is still wide which could be fixed by using a lower hfe part, but things are clean and stable so what the heck, engineers always seem to go for too much control instead of too fine so if anything changes they do not need to switch pot values, (witness Fender bias pots)
transistor beta should not affect adjustment range. We're talking single digit mA at that circuit node. I still don't like the full range adjustment and no protection against pot going open wiper, a common failure for cheap trimpots. But adding three resistors may be undesirable in a circa 1970s design. 
small note of interest, in one schemo the collector of the ref is left open, on the other schemo it is grounded, does not seem to matter,
does not matter. The base collector diode grounded would only matter after device has failed at which point it doesn't matter very much. 
we are still puzzled as to where the current goes thru the 2N3904 as the rev bias ref probably draws very little current, maybe the current goes thru the BCjunction and thru the pot to ground?
no the reverse biased base emitter junction of 953 looks like a low impedance above zener voltage. Roughly 8 mA flowing into that node, and 3904 is sucking out all but the few mA the pass element needs to deliver power supply output. I am too lazy to look up beta of 430 but will speculate around 50x  so several mA dumped through 3904 into the zener. 
here is the voltage chart for a working regulator thanks for all the help!  :D

we could have used a 7824 with a few diodes in the ground leg and held our breath but we are glad to have the whole Mary Ann in there like a stock amp.
A simple resistor divider from output to ground, connected to base of 3904, with a resistor added in series with adjustment pot wiper will reduce the adjustment range and make it less concerned about trimpot failure especially if fixed resistor divider is setting nominally correct voltage.

I stand corrected about the zener voltage. I expected 6.8V. I have never done that in any of my designs. Back in the day I messed around on the bench trying to use one as a noise generator, but IIRC I was not happy with it for that either.  Note another silly discrete transistor trick is to instead of using just the reverse emitter to make the zener diode, connect backwards across  collector-emitter. That places the forward collector base diode in series with the backwards base-emitter junction.  This gives another roughly 0.5v drop and I have seen npn and pnp pairs of these used in clamps across mic inputs (low Z clamp at around 10V but high Z at mic levels)..  I think the added collector base diode delivers a different temp-co than just the base-emitter, but I've never used one as a reference.  It is can be useful to understand how backwards junctions behave when troubleshooting circuitry failure modes.

JR
 
took your advice on the trimmer,  in fact the original thumbwheel  smoked while i was turning it, kudged a spider pot in there but too coarse, went for a voltage divider which gets trimmed out for 24, that transistor ain't gonna go nowhere now, unless somebody plugs the amp into 240,

thanks for all the awesome help, that is info we can use elsewhere, where did you learn all that stuff?  :D
 

Attachments

  • pv 24.jpg
    pv 24.jpg
    173.6 KB · Views: 28
CJ said:
thanks for all the awesome help, that is info we can use elsewhere, where did you learn all that stuff?  :D
One of my early jobs in the 60's was as an electro-mechanical tech working at MIT instrumentation Lab. I got assigned to tech for a graduate engineer who was debugging a switching power supply (for a Navy rescue submarine DSRV).  Back then switching power supplies were not common. That job turned out to be the grad student (engineer?) blowing up the PS transistors, and then me repairing it, so he could blow it up again (repair and repeat). :eek:  After a few months I became proficient at troubleshooting semiconductor failures. I would grind the cover off the top of TO-3 power transistors and by looking at the die, determine what happened. Over-voltage punch-throughs were visible as a tiny sink of melted silicon, open base was literally a melted wire, etc .  A few months later when the grad student returned to classwork, they didn't give me a new engineer to train,  8) they just had me continue the project by myself.  I got drafted before we finished that Navy project but that is where I got started on discrete electronics troubleshooting.  I might have been able to finagle a draft exemption for working on a military project but received my draft notice with only 6 days to show up, so no time for paperwork through slow channels.

Later I was lucky enough to work with a few good engineers generous enough to answer my questions about what they were doing  and why...

JR

PS: The DSRVs "deep submergence rescue vehicle" were launched in early 70's and used to rescue submarine crews stranded on the bottom. They can go down 5000 feet and couple to a top hatch. They have even had brief cameos in a few movies. They are battery powered so PS efficiency was a valid concern.
 
i remember that thing, my mom and dad both worked for Lockheed, dad graduated from M IT and got out of the service because they wanted him to build stuff to kill people which he did not like, but hey,

he worked on the Polaris project, and did supersonic jets like the X-15 and related models, also did the wings for the B1 bomber and did a composite Lear jet, (carbon fiber)
 
CJ said:
i remember that thing, my mom and dad both worked for Lockheed, dad graduated from M IT and got out of the service because they wanted him to build stuff to kill people which he did not like, but hey,

he worked on the Polaris project, and did supersonic jets like the X-15 and related models, also did the wings for the B1 bomber and did a composite Lear jet, (carbon fiber)
MIT Instrumentation Lab did a lot of different projects (like Apollo moon mission software). By law when you get drafted, your employer is supposed to give you your old job back when you get out.  When I was drafted I was working on rescue submarines,  ;D when I got out they gave me a job working in a different group on inertial navigation for missile guidance systems.  :(  I was not enthusiastic about that work and I lasted there about 3 or 4 months before I got recruited by an engineer I worked with previously, to join a small company working on pitch shifting speech (so blind people could playback recorded talking books at double speed but with correct pitch).

While I never intentionally pursued working with audio, that is where I ended up after several incremental steps.

JR
 
CJ said:
we could have used a 7824 with a few diodes in the ground leg and held our breath but we are glad to have the whole Mary Ann in there like a stock amp.

An LM317 would make everyone happy an saved you a lot of troubleshooting time!

But good work CJ

Thanks
 
we get a ton of those old PV's in the shop because they refuse to die, plus we got a cool engineer who used to work there, so it was about going after the knowledge, now we can fix those broken regulators with a 10 cent transistor that we keep in stock,

how hot would a 317 get? this TIP31c sits above ground quite a ways, so it is probably going to run cool, but yes, the more modern amps would certainly use a simpler solution and maybe a transformer tap, but taps add to transformer cost,

 
> how hot would a 317

Same as your TIP. Waste is waste, whether you put it in boxes or bags. (I'm starting on a second 15 cubic yard trash container, so my mind is trashy.)

Ad-vantage of the IC: if it thinks it is too hot, it shuts-down. The TIP tries until it dies.
 
CJ said:
how hot would a 317 get?

You could even put an heatsink on an Lm317, it would not be hot enough to be a problem and it would work great n your scenario.
But if you sort your problem just by replacing 1 transistors, then "Bravo"

By the way CJ, I sent you email regarding a mic transformer I wanted to give you, you didnt replied
 
yeah we are goin from 40 to 24, same heat regardless of base elevation, got it, doh ,  :p

whoops are you the BBC guy? thought you  might be in the UK so why send transformers for 60 bucks shipping just for a hack, if in US then we should put hose on the butcher block immeduately,
 
CJ said:
whoops are you the BBC guy? thought you  might be in the UK so why send transformers for 60 bucks shipping just for a hack, if in US then we should put hose on the butcher block immeduately,

I'm not the BBC guy, don't know who is that guy.
I'm in Europe. Well I will send you the email again, you might have missed it.

I don't pay 60 bucks to ship a transformer, I pay $5 max, I dont mind if you want it.

 

Latest posts

Back
Top