Circuit explanation

[Rhodesplyr]

I'm mainly a tube amp tech--and largely self-taught in electronics--but, lately, I've been working on some solid state amps for clients, including a 1974 Peavey Musician Series 400 head. This one was brought to me by a regular client of mine who often buys things in non-working condition out of curiosity. With a blown power supply cap replaced and a power supply short traced to a power transistor replaced without the insulator (!), it's up and running again.

I was checking out the front end to make sure there were no problems--it had more hiss than I'd like--and I noticed on the input transistors of both channels that their base and emitter voltages were exactly the same, no 0.7V drop. Yet, both channels appear to work normally. Just to continue my self-education in electronics, if someone could tell me what this method of biasing a transistor is called, I can familiarize myself with it. I've attached the schematic.

I had the amp plugged into an old Radio Shack two-way monitor speaker just for testing purposes (no great loss if blown), but when I plugged in a regular 12" musical instrument amp speaker, the hiss became much less apparent. I'm guessing Peavey figured no one would hear it in a typical application, but I wonder if swapping low-noise metal films for the few high-value base-biasing resistors would lower it.

[teemuk]

...I noticed on the input transistors of both channels that their base and emitter voltages were exactly the same, no 0.7V drop. ...if someone could tell me what this method of biasing a transistor is called, I can familiarize myself with it.

There is no name for it because such biasing scheme simply doesn't exist. A working transistor circuit will have about 0.6V (silicon) or 0.3 (germanium) voltage drop between base and emitter. If you ain't reading this voltage drop then either the transistor or your measurement device isn't working properly.

[R.G.]

With a blown power supply cap replaced and a power supply short traced to a power transistor replaced without the insulator (!), it's up and running again.

Good save.

I noticed on the input transistors of both channels that their base and emitter voltages were exactly the same, no 0.7V drop. Yet, both channels appear to work normally.

Teemuk's right - bipolar NPNs cannot work with a base-emitter voltage of zero. I'm guessing that you may have made one of the same mistakes in measuring that I do all the time.

Or the input transistors may be blown and the thing is still working on signal leaking into the following stages. Follow the signal with a scope and you can tell.

but when I plugged in a regular 12" musical instrument amp speaker, the hiss became much less apparent.

Musical instrument speakers have a much poorer high frequency response than anything with an explicit tweeter. Probably the speaker just can't do it.

I wonder if swapping low-noise metal films for the few high-value base-biasing resistors would lower it.

Maybe. Based on having revived several amps like this, I'd do this:
1. Measure the DC voltage to ground on all the transistors in the preamps. Write them down, and then think. For an NPN to be working as an amplifier, the emitter has to have the lowest DC voltage, the base has to be 0.5V to 0.7V above it, and the collector needs to be above the base. If that's not true *it's not amplifying*. For PNPs, the reverse is true: emitter highest, base 0.5 to 0.7V lower, and collector lower than that. Devices with the base and emitter at the same voltage or emitter and collector at the same voltage are either dead or something else is holding them off.
2. If it were me, I'd simply replace all the low power transistors in the preamps. I'd also replace 100% of the electrolytic caps in the preamps. Electros that old get leaky and that can not only affect their operation, they can make noise too. As you know, small signal transistors are about $0.05 each, and you can replace almost all small signal NPNs with a 2N5088 and PNPs with a 2N5086 or 5087.
3. Once you have that straightened out, start looking for hiss. In many old circuits, especially like this one, the base-emitter of the input transistors can be reverse biased by capacitors on the base or emitter during power up or power down. Reverse breaking a base-emitter like this does not prevent the device from working (in the short run at least) but causes permanent noise degradation. So replacing the front end transistors may fix the hiss. If it does not, then start putting in metal films. Do this from the very front first. The first stage of an amplifier is amplified by all succeeding stages, so the first stage dominates the noise performance of an amplifier.

Oh, yeah. Find out if your friend will pay for making it work right, not just barely work.

[Enzo]

GUitar amps are not designed to be hifi amps, by 5kHz we are generally pretty well rolled off. You never see a tweeter on a guitar amp, other than the special case of something like an "acoustic" amp. It isn't about PV not thinking you'd notice, it is about the speaker wasn;t going to reproduce the high freq noise anyway. That is why guitar usually sounds thin and crappy through stereo speakers. Plug your guitar into the PA mixer and see how lifeless it sounds.

There is normal hiss and then ther is excess hiss. Excess hiss is usually, in my experience, some semiconductor.

Iagree with RG, get rid of all those e-caps, those little 2uf and 25uf coupling caps are just itching to cause you trouble.

[Rhodesplyr]

Thanks for the help. I'd already planned to do what R.G. suggested and check all the DC voltages at idle, especially since they're not indicated on the schematic. One important thing to the diagnosis is to point out that there's an error on the schematic: in the Normal Channel, the PNP transistor is wired the same way as it is in the Effects Channel. Its collector connects only to the 47k resistor to ground for current flow and not to the emitter of the following transistor.

On the preamp board, someone had already replaced all the 100uF power supply decoupling electrolytics with Vishay/BC Components 016 AS series capacitors, so it was probably done in the last decade; my sense is that someone tried to fix this amp and gave up. They left all the 2uF caps, and I'd already ordered replacements. They should arrive today from Mouser. All the electrolytics on the output board are slated for replacement, too. They are original, and one of them is showing electrolyte leakage around its seal.

I'm trying to get better at thinking in solid state, so here's my theory about what's wrong. The curious thing is that whatever is wrong should be the same in both channels because I find the same voltages/operating points in both: no b-e drop at the first transistor. Or, to be more accurate, I'm finding only a ~0.05V b-e drop. The 2nd and 3rd transistors show the expected 0.7V b-e forward voltage. The fact that the problem is the same in both channels was what lead me to ask whether or not this was an amplification stage design I wasn't familiar with.

Since the PNP 2N4249 transistor is switched on, it must be getting base current from somewhere. I'm thinking that the first transistor at the input must have failed in such a way that it's leaking some current even though the base-emitter junction isn't on. The only other possibility, which I think less likely, is that both the 39pF capacitors have shorted. In that case, the second transistor's collector voltage would be controlling the following (in both senses of the word) NPN transistor's base and emitter voltage, causing the emitter voltage of the third transistor--and thus, the first, since their emitters are linked by a voltage divider--to rise high enough that no base-emitter current can flow in the first transistor.

I agree with R.G. that the input signal is leaking through/around the input transistors, which is why the preamp section is still "working." If I read the schematic correctly, the PNP transistor is doing most of the voltage amplification in the input section. The collector voltage of the first transistor is held at 0.7V below the supply voltage by the base voltage of the second transistor.

I think I'll swap out the first transistors when I shotgun the electrolytics and see what happens. Peavey lists the MPS-A18 as a replacement, and I think I have some of those around here somewhere. In the end, I guess the design question is: did this circuit exceed the operating limits of the input transistors, or did both channels get abused?

David

[R.G.]

Abused? Bad design? Old age? Very hard to tell. I'm many years past it, but I did do a detailed analysis of that circuit at one time in the past; possibly 20 years ago. It didn't strike me as all that unusual, but then I was much greener then. Tell you what. I'll go dump the front end preamp into a circuit simulator and see what the math machine thinks the voltages should be.
(... muttering and smoke emitted from the back room...)

OK. Here's what the circuits machine thinks happens:
Number the transistors from left to right. Q1=NPN, Q2=PNP, Q3=NPN all rest NPN.

Q  B      C      E
1  15.1  21.6  14.6
2  21.6  14.9  22.2
3  14.9  22.2  14.3
4  0.98  13.7  0.42
5  13.7  22.9  13.1
6  0.61  11.0  0
7  11.0  24.0  10.4

I modelled the power supply drops between sections as well. Resistors were nominal value, NPN = MPSA18, PNP = 2N5087 (I didn't have a 2N4249 in the simulator library).

In my experience, the actual circuit measurements will probably be within 10% of those values if I typed in the right numbers. I did not put in any of the capacitors, as a capacitor is a DC open, only looked at DC conditions. Anywhere there is a cap, the DC conditions are broken. If they're not, the cap is bad.

Added: I did the normal channel, and yes, the schematic should be as you mentioned, Q2 collector NOT shorting the base-emitter of Q3.

[Tonewood]

metal film

If it does not, then start putting in metal films. Do this from the very front first. The first stage of an amplifier is amplified by all succeeding stages, so the first stage dominates the noise performance of an amplifier.

Would you replace all first stage resistors with metal film or is there a select few that would most likely be the problem - as in tube designs?

My Roland JC-120 is very hissy and this thread has motivated me into looking at it. I have been avoiding the learning curve of solid state.

Thanks

[Rhodesplyr]

R.G., Wow, you really went beyond the call of duty here.

Here are my actual measurements of the Normal Channel, using the same format.

Q B C E
1 15.4 22.6 15.3
2 22.6 23.3 15.7
3 15.7 23.3 15.0
4 0.96 13.4 0.40
5 13.4 22.9 12.9
6 0.60 12.0 0
7 12.0 22.9 11.4

I have noticed a few more discrepancies between the schematic and the actual circuit. One of these is apparently the order of the groups of transistors served by one supply capacitor. On the board, the first section is only one 180 Ohm resistor removed from the 24V supply.

In any case, I'm amazed by how accurately the software modeled this circuit. As you can see, the only major difference is the emitter of Q1.

Of course, at this point, the only thing left to do is to start replacing parts to see what cures it. My suspicion is that Q1 has some sort of partial E-C short.

[Rhodesplyr]

Tone,

Again, I'm no solid state expert, but one of the nice things about tubes is that their grid-leak resistors require no DC current bias. The base bias resistors required for a BJT do pass a constant DC current. If these resistors produce noise under current flow (sometimes called "popcorn noise"), this noise enters the signal chain. It's most noticeable if the signal being amplified is low amplitude and needs a lot of gain, and the worst offenders are usually carbon composition resistors. The higher the resistor value, the greater the potential for noise. Carbon films are quieter, metal films, quieter still.

For example, in the first preamp stage of a Wurlitzer 140B Electric Piano, you'll find the transistor base biased by two 680k resistors. I've had good luck replacing the carbon comp originals with low-noise metal films. They don't change the tone, but a lot of the noise goes away. It won't solve a noisy transistor problem, it's an improvement. The resistors I like to use are the Vishay/Dale RN/CMF types, the ones with brownish tan bodies. How far you go in resistor replacement is up to you, but I concentrate on the early gain stages. Glenn Frey liked the results on the one I rebuilt for his studio. (It's a long story.)

Back in the early 1960s, Dynaco, in its tube preamps, used carbon film for cathode and plate resistors in phono preamp stages--same principle.

So, on the Peavey, when I see a 1Meg resistor biasing the base of the first gain stage transistor, it occurs to me to think about replacing it.

[R.G.]

Would you replace all first stage resistors with metal film or is there a select few that would most likely be the problem - as in tube designs?

There are a select few.

My Roland JC-120 is very hissy and this thread has motivated me into looking at it.

In the JC120 schemo you left, I'd suspect R1, R2, R5 and R6. If it were mine, I'd arbitrarily make those metal film. Any resistance that's directly in series with the input signal or in series to the input adds its thermal and excess noise directly to the signal.

Further, the JC120 comes close to but does not quite do a very noise free biasing setup. I would tack a 10uF in parallel with R3 to further reduce any noise that R3 and R7 produce, although with R3 at 6.8K, this will not be a huge thing. It might help to replace Q1, as input overloads degrade junctions, but JFETs are not as prone to this as bipolars.

Keep in mind that it is possible that something is oscillating at RF and the hissing you hear is the RF noise being rectified and mixed in with the audio. You might try my favorite abusive noise hunting trick of attaching a 10uF capacitor to ground at places along the signal path. This dramatically shunts all signal to ground unless you do it right after a follower of some sort. The hiss from earlier in the amplifier is also killed. If you shunt the input to the power amp and still hear hissing, it's in the power amp. If you shunt right after the input stage and the hissing goes away, it was in the input stage and the rest of the amp is relatively clean.

Rhodesplyr is entirely correct in his comments about resistors.

@Rhodesplyr: I'm always suspicious of simulators, but some things they do very well, and ferreting out DC conditions is one of them. Saves a lot of soldering iron work.

On the base-emitter of Q1, I've sometimes found that a high gain device will have such a high impedance bias network that the meter pulls down the base and causes an erroneously low reading of base voltage. When you suspect that, put the meter leads directly on the base and emitter leads, not base to ground. That tells you Vbe directly, and it's a low impedance measurement because the base-emitter diode itself is low impedance as is the emitter. But measuring base to ground can pull down the base and the emitter; you just are not measuring the emitter at the same time if it's your meter pulling it down. That's why a lot of tube schematics tell you that voltages are measured with a VTVM - VTVMs have a lot higher input resistance than digital meters.

[Tonewood]

R.G & Rhodesplyr thanks a lot for the great info. After I isolate the problem with the "abusive noise hunting trick", I'll order some new resistors and cap and report the results.

After my last post I noticed that if volume in both channels is turned completely down, hiss goes to acceptable level. Slightest increase in either volume knob and hiss comes in. Once the hiss is back in, it is doesn't change with further increases in the volume control.

[R.G.]

After my last post I noticed that if volume in both channels is turned completely down, hiss goes to acceptable level. Slightest increase in either volume knob and hiss comes in. Once the hiss is back in, it is doesn't change with further increases in the volume control.

That makes me very suspicious of RF oscillation. Oscillation grows to be as big as it can be as soon as there's enough gain. If you have an oscilloscope, use it. If not, it's going to be tougher.

[Rhodesplyr]

R.G.,

Never let it be said that I'm someone who can't admit that I was wrong. I did what you suggested last night, and I found that by measuring from Emitter to Base on the input transistors, I found 0.5V. So they are working. The strange thing is that I thought I'd made that measurement, but I guess I didn't. I may have been distracted by the fact that, at that point, the whole power supply was being held together via alligator clips. The power supply is now hard-wired and reassembled, and all smaller electrolytics on the output board are new. That got rid of some noise issues.

However, the overall noise drop made it more obvious that there was still some noise in the preamp circuits that was hard to localize, though it seemed to be coming from the input/high gain areas. What I traced it back to was the voltage adjustment pot associated with the 24V regulated supply to the preamp board. After a thorough spraying and cleaning, it's much better. Thus, the high-gain circuits were amplifying noise from the power supply. I should take my own advice: always troubleshoot the power supply first.

If it were mine, I think I might replace the stock 24V regulated supply with, perhaps, a voltage dropping resistor and a standard three-terminal 24V regulator IC. I used one of these in a copy of a Farfisa F/AR power supply I had to build for my Compact Duo, and it seems to have better regulation and lower noise than Peavey's 1974 design. I have a feeling that adjustment pot for the output voltage may start to cause trouble again.

But, on the whole, I think this has bumped up my skill level in terms of diagnosing transistor circuits. I've run into situations before where a meter would load down a high impedance circuit, but I wouldn't have predicted it would do it in this case.

Back to the 24V regulated supply, do you think there's any reason that Peavey chose what seems to me a rather low capacitance (50uF) at its output? Regulated supplies in combo organs I work on usually have 2000uF to 5000uF capacitors at the outputs. My instinct would be to bump it up to at least 220uF or so, but I don't know if there's a reason Peavey chose a lower value--other than cost.

David

[R.G.]

Never let it be said that I'm someone who can't admit that I was wrong. I did what you suggested last night, and I found that by measuring from Emitter to Base on the input transistors, I found 0.5V. So they are working. The strange thing is that I thought I'd made that measurement, but I guess I didn't.

I won't even get started on things I thought I did but really didn't. There is the whole day I spent debugging a new board from the bottom side where I could get at all the signals only to find I'd forgotten to put the ICs in the sockets, for instance. oops...

If it were mine, I think I might replace the stock 24V regulated supply with, perhaps, a voltage dropping resistor and a standard three-terminal 24V regulator IC. I used one of these in a copy of a Farfisa F/AR power supply I had to build for my Compact Duo, and it seems to have better regulation and lower noise than Peavey's 1974 design. I have a feeling that adjustment pot for the output voltage may start to cause trouble again.

I vote for this one. Three terminal regulators are simply too good, too cheap, and too easy to use to not use them. Even better, you can often use the low power "L" version in the TO-92 package and put a regulator on each circuit section that needs power and have separately regulated power supplies that *can't* interact then. As of this morning, you can get a 78L24 100ma regulator in a TO-92 package for $0.25 and a 7824 1A regulator in a TO-220 for $0.40. They self current limit, self thermal limit, and produce very accurate, solid power. About the only way to kill them is to put more than 40 V into them.

But, on the whole, I think this has bumped up my skill level in terms of diagnosing transistor circuits. I've run into situations before where a meter would load down a high impedance circuit, but I wouldn't have predicted it would do it in this case.

I wouldn't have thought of it but I recently ran into a situation where that was all that told me the truth. The scar was fresh.

Back to the 24V regulated supply, do you think there's any reason that Peavey chose what seems to me a rather low capacitance (50uF) at its output? Regulated supplies in combo organs I work on usually have 2000uF to 5000uF capacitors at the outputs. My instinct would be to bump it up to at least 220uF or so, but I don't know if there's a reason Peavey chose a lower value--other than cost.

Possibly stability. I haven't looked at the 24V regulator, but some active circuits have difficulty with driving a highly capacitive circuit. But cost entered into it. A good regulator (three terminal 78xx included) generally doesn't need huge capacitors at low frequencies, because their active output impedance is quite low at low frequencies. It's only at higher audio and above that they run out of internal gain to lower output impedance and need external caps to keep the supply impedance low. That's why good designers literally sprinkle 0.1uF caps around the board - it puts high frequency decoupling right on the spot where it's needed.

The 78xx series is in general not driven into oscillation by output capacitance, but it doesn't need much at low frequencies, and big caps won't help at high frequencies.

I just finished a set of PCBs that replace the entire internal workings of the Thomas Vox Beatle amplifier and dispense with 90% of the internal cabling that make these things a misery to work on. I put 78L24s and 78L18s on the several small, local PCBs that make this up. This let me ditch the row of power resistors and axial caps that eat up the bottom of the chassis enclosure for the preamp.

Yum... jelly bean voltage regulators!! Love 'em.

[Enzo]

The point of large filter caps in power supplies is to remove ripple. In this case the raw voltage for the 24v supply is already filtered, so ripple is not an issue. The little cap is all that is needed for stability and decoupling.