Vintage RCA audio preamp I'm trying to fix/learn from

[rp55]

I own a vintage RCA audio premp (BA-72A). I bought this as a module some time ago and it quit working on me for some reason. So it did work ok for quite a long time but not now. I get nothing out of it. Here are the schematics I have:

http://imgur.com/OOuqhNP
http://imgur.com/GIsxqAT

I'm trying to use this as a "6 foot deep end of the pool" learning experience as I'm studying electronics on my own and I've been trying very hard to understand the basics and stuff: transistors (depletion region... yep got it), caps (capacitance in farads which is work per volt) etc. I'm doing the work, I'm doing the math, and I'm pushing thru and hopefully getting it. But I seem to be completely lost when it comes to this spaghetti circuit?? I thought I could get some helpful guidance from some of you genius folk out there as I know there are very knowledgeable people here and I learn a lot every time I merely lurk over a post.

This RCA preamp is apparently a 4 transistor preamp (2 stage input and 2 stage output) which has (in addition) a transformer on the input and the output.

I've got a 700ohm resistor on the output transistor output to satisfy load requirements. I'm running the input using my computer soundcard with a 440hz wave or something. The power source (DC) is an actual RCA power supply used with this module and is giving a clean -30v-0v.

----------------------------
So now here's what it's doing: Zero. No output.

I've got the input ac audio signal coming in on pin E1 via the transformer (and e2 is ground for that signal). I've got a good -30v dc attached. I've pulled off one leg of a few of the resistors involved in the first stage to test them and they're ok. I've pulled the first transistor and it checks ok with simple multimeter test (inconclusive). Out of the collector of Q1 I get a little of the positive high peaks of the input signal.

Now I've "studied" the circuit (as much as I can with my limited knowledge) and I just don't see the input being biased up into the working range of the first transistor (Q1). So I'm guessing that it's only going to deal with the upper audio wave portion in the first transistor??.

I've tried modeling the first half stage with falstad's wonderful ciruit designer java applet:
(copy and paste/use the below link for the falstad model):

http://www.falstad.com/circuit/#$+1...125E-4+0+-1 o+26+64+0+35+40.0+7.8125E-4+1+-1


And I just don't get it. It's apparently not biasing the initial transistor. I guess I'm just not sure what I'm looking at. I mean... why so many connections?

Sorry so long guys, and thanks for any "duh" epiphanies you can give me. Hopefully I'm getting over some hurdles with this stuff with your help!

 

[rp55]

I'm not completely lost when it comes to typical inner circuit structures. For instance, I successfully modeled a circuit where I added a dc offset to an ac (audio) input:
http://www.falstad.com/circuit/#$+1...+-7.428951293706854 o+0+64+0+35+5.0+0.1+0+-1

So that way I can look for that sort of thing. I understand basic resistive voltage dividers (including transistors as one of the resistors). So I'm not completely in the dark there.

 

[Baluncore]

Bad bias is usually a cracked solder joint to a bias resistor, or a dead electrolytic capacitor.
Resolder the circuit board. Examine the caps for domed ends or extruded dielectric.

 

[meBigGuy]

You can use the multimeter to check all the transistors in circuit. Don't need to take them out. You probably already know this much, but one lead on the base, check to emitter and collector, then reverse leads and check again. One way higher impedance than the other.

The first stage (Q1) should show the voltages indicated with no input signal. It's pretty well DC isolated from the rest of the circuit. If not you need to determine why.

The drop across R17 should be small. If it isn't (and the transistors seem OK), then you probably have a leaky capacitor(s).

Generally when troubleshooting, one first looks for swollen capacitors, cold solder joints or board discoloring from heat, then blown transistors. Then you need to start doing detailed troubleshooting. Again, I'd start with the 1st transistors bias voltages.

My money is always on bad electrolytic capacitors. Look carefully for ANY swelling on the top of the capacitor. It's pretty common to just shotgun it and replace all the capacitors if the transistors ohm out OK.

 

[AlephZero]

Now I've "studied" the circuit (as much as I can with my limited knowledge) and I just don't see the input being biased up into the working range of the first transistor (Q1).

The base bias is from R2 and R3, which for a 30V supply would give about 30 x 91/(91 + 33) = 22V. The voltage gets onto the base of the transistor through R1. 22V is near enough the 20.5V on the diagram to make sense.

R1 and C3 are a filter circuit, presumably to roll off any high frequency noise on the input, since the corner frequency is about 100 kHz.

With luck you haven't caused any more damage by unsoldering components, but the right way to troubleshoot this is measure all the voltages given on the diagram, and if there are any large differences look at the circuit and figure out what might be causing it. Usually, the cause is a component that is open or short circuit. Don't start soldering till you know what the problem is!

if you tell us the voltage measurements you get, somebody might be able to help.

 

[Okefenokee]

If it helps I can share some of my bag of tricks. Carefully look at the board, preferably with a magnifying glass.

Check components. Do any have a cracked or crumpled surface? Do any look discolored as if they were getting really hot?

Do you have a good nose? smell the board. Burnt electrical components have a distinctive smell.

Check solder joints. Are they all nice and shiny? A bad solder joint will look scuffed.

Look at the traces. Are any raised from the surface? Is there a bulge in the board material around the solder joint? A bad solder job can make the traces too hot and cause them to separate from the board when it swells. The traces are weaker that way and can break apart. The break can be so fine that you barely see it.

Tarnished connectors. I find many problems are nothing more than dirty corroded contacts where the device connects to power or signals.

Can you remove all the transformsers and coils? The problem with troubleshooting components where coils are present is that the coil will give you DC continuity. Look at c13 in your diagram. If you tried to measure the capacitance of c13 you will fail because the coil in transformer 102 will act like a short.

Check continuity from component to component. Test from the radial of one component to the radial of another component. That way you include the solder joints and trace in your continuity test. Print your diagram and use a highliter to mark what you've checked.

Brute force! Take a solder gun and reflow every solder joint. Be patient and don't overheat the board.

By the way. The right way to remove components is to use a desoldering pump! It's quick so you don't apply heat for so long.

 

[rp55]

Fantastic tips and info all. Thanks!

Good news, I found the first problem via continuity tests. R17 had failed (cracked in half just trying to hold the probes to the sides of it). So I've only got a 1/2 watt 150ohm replacement (which just burned due to my next question...) hopefully as it's calling for a 130 that won't affect things too much.

So I was getting some signal on my 700ohm output load post the output transformer. A little clipped but I hadn't really measured my input level yet.


But now I tried something else. I was wondering previously about the E3 connector on the schematic (lower left). That is connected to the input transformer case (ground).
In another schematic I have for a similar preamp, they show it connecting to ground (http://imgur.com/RirnpOV) although this may not actually be E3 but it looks very similiar. Oddly, they say the com rail is +30 and the other rail is -30 but I don't know if that's right and it looks very similar to my unit. They probably just mean +30 in regards to the -30 (which is 0 volts)? No?

So anyways, I figured that connection could be connected to ground. When I did that, my newly inserted resistor started to burn. It just doesn't seem right to have e3 (which is essentially -30v) going to ground thru that R17 resistor. Not sure what I'm missing there... Isn't that current coming from (conventionally) common 0 volts down to the -30v rail and out that connection? Why would it need to be sent to ground?

Anyways thanks again for all the help. Very useful and getting some clue about which resistors were doing the biasing also really help confirm what I thought so I could proceed.

 

[rp55]

I just noticed that the schematic makes a note that all resistors are 1/2 watt (unless noted, and R17 isn't noted). So that E3 connection must not be right and was trying to run 6 watts (V^2/R so -30^2/150 = 6watts) thru that resistor. No wonder it burned. I don't understand that connection then?? It works without it. And it sounds as good as old now.

 

[Okefenokee]

You're schematic says that voltage are measured from E6. That would mean that if you measure -30V at E7 then it must be 0V (0V-30V = -30V). If that were true then connecting E3 to ground should do nothing to R17.

That note seems confusing. What they must have meant is that the voltage was measured from E6 and they got -60V. Then they did the math (-60V = E7 - E6 ; E7 = -30V) and they wrote that result on the schematic. That's just my guess. So when you grounded E3 you put 6 Watts on R17.

 

[rp55]

You're schematic says that voltage are measured from E6. That would mean that if you measure -30V at E7 then it must be 0V (0V-30V = -30V). If that were true then connecting E3 to ground should do nothing to R17.

That note seems confusing. What they must have meant is that the voltage was measured from E6 and they got -60V. Then they did the math (-60V = E7 - E6 ; E7 = -30V) and they wrote that result on the schematic. That's just my guess. So when you grounded E3 you put 6 Watts on R17.

It's definitely -30 between E7 and E6. Not 60v. And even then -30 would put 6 watts thru R17 (as it did in my case when it burned). Perhaps somewhere along the line one of the "customizers" of these pre's thought E3 should be grounded. That last schematic looks more modern (like someone made it in the past years), whereas the original schematic for my unit doesn't say what to actually do with E3 (even though it is tied to the xformer case on mine). Know what I mean? I don't see any logic in sending current thru r17 straight to ground at all, and it works without it. ?

 

[Baluncore]

Neither E3 nor E5 should be connected the 0V common rail. E3 and E5 are on a local preamp low voltage rail that will be slightly above -30V.

The total current through the preamp flows through the R17, 130 ohm resistor. R17 isolates noise on the external -30V rail from the preamp internal low voltage rail.

Since all current flows through R17 it gets hot and has died of old age. Replace it with a higher rated resistor combination with a similar value.

Failure of C11 as high leakage current could cause the death of R17.

 

[rp55]

I measured the voltages around the transistors to see how they match spec and here's what I got:
2nd column is schematic specified value. They are all a little high for some reason?? My psu is exactly -30v. Is that a big deal?

Q1
B -21.8v -20.5v spec
C -17.4 -16.2v spec
E -22.5 -21.0v spec

Q2
B -1.33v -1.5v spec
C -13.6v -10.4v spec
E -1.17v -1.35v spec

Q3
B -13.6v -10.4v spec
C -0.7v -0.8v spec
E -14.3v -11.1v spec

Q4
B -22.80v 22.2v spec (negative lacking on chart?)
C -14.3v -11.1v spec
E -23.50v -22.9v spec

This unit has been hot rodded a bit with orange sprague drop caps around C6 and C12. Not sure if that may be affecting things.

Perhaps just due to component age? Or perhaps the fact that I'm using a resistive 700ohm load? Or my 150ohm replacement R17 vs the spec 130ohm. I guess it's interesting to see the difference but it doesn't mean anything to me. Perhaps you guys can look at this stuff and just know where problems are.

 

[rp55]

Neither E3 nor E5 should be connected the 0V common rail. E3 and E5 are on a local preamp low voltage rail that will be slightly above -30V.

The total current through the preamp flows through the R17, 130 ohm resistor. R17 isolates noise on the external -30V rail from the preamp internal low voltage rail.

Since all current flows through R17 it gets hot and has died of old age. Replace it with a higher rated resistor combination with a similar value.

Failure of C11 as high leakage current could cause the death of R17.

Thanks for the reply. How does the R17 resistor isolate noise? Isn't that what C11 is for (It looks like a bypass cap, yes)? But I see what you mean that C11 leakage would cause R17 to slowly cook.

 

[Baluncore]

R17 and C11 together make a low pass filter from which the quieter low voltage rail is derived.
Noise on the -30V supply appears across R17.

 

[rp55]

R17 and C11 together make a low pass filter from which the quieter low voltage rail is derived.
Noise on the -30V supply appears across R17.

I guess I'll have to read up more on that. I didn't realize a low pass filter was used on a dc power rail. I figured just a bypass cap was enough on the front of it. I guess I don't understand why you'd want to pass any frequencies to a dc power rail (i.e. low ones). What does that do?

 

[Baluncore]

A quiet supply rail is needed. The -30V rail will probably have switching noise from the rectifier diodes present. A bypass cap alone would solve the problem if the -30V supply had a higher impedance, but it is very low impedance and is referenced to a chassis connection some distance away. Using a series resistor raises the supply impedance and so does not require such a large bypass capacitor to eliminate the supply noise audio component.

 

[Baluncore]

-30V DC is a low frequency. You turn it on one day, you turn it off the next. It must reach the bypass capacitor.

Bypass capacitors are used to isolate the inductance of the supply rail and so lower the local impedance of the supply. They prevent oscillation on the supply line that could result from the reflection of transients off the voltage regulator output. They prevent noise from one module from coupling through the supply to other modules.

 

[rp55]

A quiet supply rail is needed. The -30V rail will probably have switching noise from the rectifier diodes present. A bypass cap alone would solve the problem if the -30V supply had a higher impedance, but it is very low impedance and is referenced to a chassis connection some distance away. Using a series resistor raises the supply impedance and so does not require such a large bypass capacitor to eliminate the supply noise audio component.

-30V DC is a low frequency. You turn it on one day, you turn it off the next. It must reach the bypass capacitor.

Bypass capacitors are used to isolate the inductance of the supply rail and so lower the local impedance of the supply. They prevent oscillation on the supply line that could result from the reflection of transients off the voltage regulator output. They prevent noise from one module from coupling through the supply to other modules.

Thanks for the mind blown. Is there a more intuitive way to understand what you're saying? I've got inductance (in depth) on my list next to study. I understand the concepts but am still struggling with it in terms of impedances.

I guess in my mind, the typical linear psu has smoothing caps (which this RCA psu does) after the rectifier bridge. So noise would still get thru that? Perhaps they were expecting the psu to be far away from these preamps... is that some of what you are saying? I'm not quite sure what you mean by switching manually on and off as being a frequency (i.e. low)? And I thought bypass caps are just to take out high frequency interference noise. How does the psu "circuit" provide a low impedance voltage? Isn't voltage just voltage (i.e. the work that can be done per units of charge which I presume are available via the medium's conductivity)?

I guess I understand about needing to have a quiet rail now that I think about it as that is in direct contact (so to speak) with the analog signal going thru speakers etc. So somehow the resistor changes the impedance and settles down the transient echos (which I don't understand) of the current?? Sort of like breaking current waves/oscillations so they calm down or something? Is that it? If so, where are these transients tending to come from? RF? AC?

 

[Baluncore]

If you look at the (AC supply synchronous) noise waveform on the -30V supply you will see diode switching noise. The smoothing cap is really there to maintain voltage during the parts of the cycle when diodes are not conducting. Because the smoothing capacitors have some self inductance and resistance, they are not perfect capacitors.

The PSU voltage is from a transformer with low impedance. When turned on, the diode bridge also has very low impedance, so the imperfect smoothing caps will let some noise pass.

Transients come from the signals in the modules and from the rectifiers in the power supply. The power rails are transmission lines that can have reflections, (and so oscillations), if you don't block propagation with series resistance, series lumped inductance or bypass capacitors.

 

[rp55]

If you look at the (AC supply synchronous) noise waveform on the -30V supply you will see diode switching noise. The smoothing cap is really there to maintain voltage during the parts of the cycle when diodes are not conducting. Because the smoothing capacitors have some self inductance and resistance, they are not perfect capacitors.

The PSU voltage is from a transformer with low impedance. When turned on, the diode bridge also has very low impedance, so the imperfect smoothing caps will let some noise pass.

Transients come from the signals in the modules and from the rectifiers in the power supply. The power rails are transmission lines that can have reflections, (and so oscillations), if you don't block propagation with series resistance, series lumped inductance or bypass capacitors.

I think I see a bit of what you're talking about. Having fixed another speaker amplifier and having to troubleshoot the rectifier (smoothing caps had gone out) that the ac ripple is still a bit there. But you're talking about very tiny harmonics so to speak on those psuedo dc "waveforms". In other words, if I zoomed in on the scope I would see these artifacts more clearly normally occurring? Perhaps they are getting amplified somehow via the transmission distance? I'll keep in mind what you say as I get deeper in, thank you. Perhaps it will become more apparent as I toy around and build a simple amplifier and see if these problems manifest themselves so that I have a mother of invention scenario and these solutions become obvious.

 

[rp55]

If you look at the (AC supply synchronous) noise waveform on the -30V supply you will see diode switching noise. The smoothing cap is really there to maintain voltage during the parts of the cycle when diodes are not conducting. Because the smoothing capacitors have some self inductance and resistance, they are not perfect capacitors.

The PSU voltage is from a transformer with low impedance. When turned on, the diode bridge also has very low impedance, so the imperfect smoothing caps will let some noise pass.

Transients come from the signals in the modules and from the rectifiers in the power supply. The power rails are transmission lines that can have reflections, (and so oscillations), if you don't block propagation with series resistance, series lumped inductance or bypass capacitors.

One thing I just realized about this circuit is that they aren't even making differential use of the balanced input signal connections (i.e. it's not a balanced amplifier doing anything to cancel out the input noise).

 

[Baluncore]

The input and output are isolated using coupling transformers so the external connections are balanced and there should be no ground loop or common mode problems.

The circuit starts with an NPN followed by a PNP stage. Both will have some non-linearity but the curves will cancel to some extent. Since it is a preamp, and the signals are small compared with the supply voltages, the non-linearity of the preamp will not be significant.

 

[rp55]

The input and output are isolated using coupling transformers so the external connections are balanced and there should be no ground loop or common mode problems.

The circuit starts with an NPN followed by a PNP stage. Both will have some non-linearity but the curves will cancel to some extent. Since it is a preamp, and the signals are small compared with the supply voltages, the non-linearity of the preamp will not be significant.

Well I'm not comprehending (yet) much of what you're saying, but it does give me some more things to read about and hopefully learn from.

One really stupid question I have that I'm sure you can answer... since this is just a preamp the output is obviously going to be somewhere around 1 volt (with the input being obviously factors below that). Why does it need to amplify within the inner stages up to -20 volts if it's going to merely "collapse" it back down to 1 volt? I guess I'm just not finding these obvious details anywhere I'm reading about amplifier basics?

 

[rp55]

The input and output are isolated using coupling transformers so the external connections are balanced and there should be no ground loop or common mode problems.

I guess I don't see how this unit would be eliminating any interference injected noise on the input (via the distance of the signal cable) as it is not doing a phase inversion of one of the balanced connections (and summing) to eliminate it. Are you saying that the transformers won't pass the noise on input signal? I must be missing something.
I can see how it won't pass a ground hum since it's not varying and thus the transformers won't pass it.

 

[Baluncore]

Why high voltage supplies for small signals?

If you plot the input to output transfer function of a class A amplifier, over it's full output range, you will see that the response is slightly curved. A curved response causes asymmetrical distortion of the signal which generates even harmonics.

Distortion due to non-linearity is proportional to (Vsignal / Vsupply ) ^2. The curvature of the line has less effect on smaller signals, so higher voltage supplies with smaller signals reduces distortion.
If you increase the input signal amplitude by 1%, harmonic distortion rises by 2%.
If you double the input signal amplitude you will get 4 times the distortion.
If you halve the supply voltage you will get 4 times the distortion.

 

[Baluncore]

Why transformers? They are "baluns" used to convert between BALanced and UNbalanced signals.

An amplifier cannot tell the difference between signal and noise on it's input. It must amplify whatever is presented to the input.

By isolating the interface with a transformer, an external differential signal is interfaced to a single sided internal signal. Since the external signals are balanced, any noise picked up by the twisted pair cables should cancel.

 

[Baluncore]

I noticed some veteran terminology on the circuit diagram.
See note 1. “All capacitor values in MFD unless otherwise specified”. An MFD is a “Micro FaraD” = uF.
Did you notice C3 and C5? MMFD is a “Milli Micro FaraD” which is a nano farad = nF.

C3 has an unusual value being 430MMFD = 430nF = 0u43F. The preferred value of 43 is a very rare value for capacitors. It is in the (5%) E24 series.

 

[rp55]

I noticed some veteran terminology on the circuit diagram.
See note 1. “All capacitor values in MFD unless otherwise specified”. An MFD is a “Micro FaraD” = uF.
Did you notice C3 and C5? MMFD is a “Milli Micro FaraD” which is a nano farad = nF.

C3 has an unusual value being 430MMFD = 430nF = 0u43F. The preferred value of 43 is a very rare value for capacitors. It is in the (5%) E24 series.

I did actually notice that and figured out that meant 430pF. I was using a nomograph I downloaded to easily see the reactance of various caps in the circuit and this one just seems to be a way to keep out anything up high frequencies above audio range. Would it matter much what they used as long as they didn't come down too low into the upper ideal harmonics above 20k? Perhaps that was the popular make of the day?

I was also reading about what you said how transformers handle the balancing and I think I understand that now. Such that if the hot and cold were in-phase with each other normally, the transformer wouldn't induce anything at all. There would be no "reciprocal" field of charge for whatever signal charge at any moment (unless the cold was ground of course). So since the noise ends up on both it's like the same scenario I just stated: the moment the noise hits the transformer it is negated since there is no "reciprocal" energy charge for the noise. Is that basically correct? And assuming it is, doesn't that imply that a balanced connection has more voltage than non-balanced which just has a hot referenced to ground?

 

[Baluncore]

When you have a twisted pair feeding a differential input, (such as a transformer), one input is termed inverting, the other non-inverting. Signal that is present with the same phase and amplitude on both inputs is called “common mode”. A differential amplifier is designed to amplify the difference signal while significantly rejecting any common mode signal.

“hot and cold” usually refers to unbalanced systems such as chassis = ground = cold, and live = signal = hot.
For your use of the term "reciprocal" I think it better to use the term “differential”.

 

[rp55]

When you have a twisted pair feeding a differential input, (such as a transformer), one input is termed inverting, the other non-inverting. Signal that is present with the same phase and amplitude on both inputs is called “common mode”. A differential amplifier is designed to amplify the difference signal while significantly rejecting any common mode signal.

“hot and cold” usually refers to unbalanced systems such as chassis = ground = cold, and live = signal = hot.
For your use of the term "reciprocal" I think it better to use the term “differential”.

Right, thanks! That terminology makes perfect sense now. Thanks for the clarification.

 

[rp55]

BTW I simulated the entire preamp using falstad's circuit designer. Does it seem like it is a representative simulation? (full screen the window to see all of it)

http://www.falstad.com/circuit/#$+1...0.0+0.00625+1+-1 o+35+64+0+35+10.0+0.05+2+-1

 

[rp55]

In taking a guess at some things that are happening:

It appears C6 removes the dc bias of Q1. A new bias is put on Q2 via R11-R12. The manual says something about C8 bypassing Q2's emitter which causes more current out of its base and hence more voltage on R13. Then the last two stages look like a push-pull design for the output. That's about all I can see given my current knowledge. I guess those inner caps are some sort of feedback balancing between transistors as their reactances related to frequencies don't appear relevant (or perhaps I'm off base there... likely).

 

[rp55]

So I seem to get an antenna receiver (when in boost mode via connecting E4 and E5 together per spec) when E3 is not connected to the transformer per previous replies. The boost has worked in the past. I get radio white noise unless E3 was connected to the transformer body. Also get some very low hz oscillations happening when switching out of boost (E4 and E5 not connected). No Earth ground or 0v connected to the body otherwise. It is fixed by connecting (grounding?) E3 to the transformer body. I don't understand why and I'm unsure about ground reference differences between various nominal considerations of ground (if that makes sense).

Perhaps it's merely (after just thinking about it) just having a long wire hanging off of -30 (E3) which creates an antenna?

 

[tfr000]

Did you notice C3 and C5? MMFD is a “Milli Micro FaraD” which is a nano farad = nF.

MMFD is micromicrofarad = pF. Very old schematics always list caps as uF or uuF. nF are more recent.

 

[Baluncore]

You are correct, apologies. A senior moment I believe.