I have been reworking a germanium amplifier, and I understand the circuit below is an emitter follower (besides just the schematic saying so ), but there are some oddities about this circuit that I do not understand. I don't know anything about the transistor beyond it is germanium, obviously PNP. A pox on Magnavox for using internal numbering, and I don't have the photofact for the Magnavox A531 to look up an equivalent or whether it is even listed. Questions below . . .
Questions:
I understand R17 and R16 set the -7.2V bias voltage, but why are R15, C10 needed? Input impedance? Would not the emitter follow the base without them?
Why are R18 and C31 necessary? I'm guessing this transistor maybe cannot handle -19V on the collector, so R18 reduces the voltage drop across the transistor while C31 maintains an AC ground? But C31 (0.001uF) seems awfully small to be a bypass capacitor- maybe it is involved in a filter from the collector side somehow?
Thanks for your help!
Questions:
I understand R17 and R16 set the -7.2V bias voltage, but why are R15, C10 needed? Input impedance? Would not the emitter follow the base without them?
Why are R18 and C31 necessary? I'm guessing this transistor maybe cannot handle -19V on the collector, so R18 reduces the voltage drop across the transistor while C31 maintains an AC ground? But C31 (0.001uF) seems awfully small to be a bypass capacitor- maybe it is involved in a filter from the collector side somehow?
Thanks for your help!
Hi. R15 usually defines input resistance. Again , usually voltage divider lije R16 R17 sets the dc voltage point , and a capacitor is placed from there to ground. In that case AC input resistance would be equal R15. But here we have another end of C10 not grounded, but connected to output, which has same phase signal, so this reduce current through capacitor , that's equal to increase input resistance. Small capacitor and resistor at collector is just rc filter i think.
Adding R15 and C10 allows the base bias to be “bootstrapped”, greatly increasing the input impedance at AC frequencies. I think your guess re purpose of R18 and C31 is probably correct and helps reduce dissipation in the transistor.
If the transistor has failed, a silicon replacement should work as a substitute with no problems.
Edit: +1 for @ximikas comments. I still suspect R18 is for dissipation and C31 to suppress Miller capacitance. Filter corner frequency is too high for hum suppression and emitter follower should have good rejection of supply anyway.
If the transistor has failed, a silicon replacement should work as a substitute with no problems.
Edit: +1 for @ximikas comments. I still suspect R18 is for dissipation and C31 to suppress Miller capacitance. Filter corner frequency is too high for hum suppression and emitter follower should have good rejection of supply anyway.
Last edited:
Thank you! Ximikas- I totally see that now- I never considered how C10 would radically increase the impedance- I figured maybe an additional form of negative feedback given some limitation of the germanium transistor. I know Ge transistors have a low Beta, so perhaps the base was hard enough to drive to warrant this extra treatment. I'm also guessing the source impedance from the upstream radio/phono chassis must have really sucked. Now that I am converting this to a stand-alone amplifier, and will drive the input from a modern low impedance source, I can consider whether I want to keep this detail. Do you think the very high input impedance would be a liability for induced noise?
Thank you! BSST- You're right- the F3db for 0.001 and 22k (7.2Khz) does not seem to make much sense as a filter, unless there was a high frequency interference coming from the radio/phono chassis that they were worried about. These consoles were not very high fidelity, especially when it came to high frequency- I can see a 7 Khz low pass judged as being "good enough" for the horns/tweeters they were using.
I see now how the opposite phase voltage at the collector incurred by the 22K could decrease the gain at higher frequencies through the miller cap. I think I can see how the 0.001uF cap would then come into play to limit the amplifude at the collector to counteract this. If that is the case, why not use a larger capacitor? Kill it dead? I might just remove the cap and see what happens with the scope, if I could even detect it.
Thank you! BSST- You're right- the F3db for 0.001 and 22k (7.2Khz) does not seem to make much sense as a filter, unless there was a high frequency interference coming from the radio/phono chassis that they were worried about. These consoles were not very high fidelity, especially when it came to high frequency- I can see a 7 Khz low pass judged as being "good enough" for the horns/tweeters they were using.
I see now how the opposite phase voltage at the collector incurred by the 22K could decrease the gain at higher frequencies through the miller cap. I think I can see how the 0.001uF cap would then come into play to limit the amplifude at the collector to counteract this. If that is the case, why not use a larger capacitor? Kill it dead? I might just remove the cap and see what happens with the scope, if I could even detect it.
Probably you will not see difference if that 1nF capacitor is removed. Emitter follower repeats input signal, not supply voltage, this capacitor is just power supply filter. Capacitor bypasses to gnd just to prevent ac floating collector.
Thank you for everyone's help. I love this forum.
From here, I'm assuming 0.001uF and 22K (forming a 7.2kHz low pass from supply to collector) is to suppress the higher frequencies at the collector that would interact with the Miller capacitor and reduce the gain. It's still a mystery why 22K is there, other than to reduce dissipation, and why C31 is not larger to just overkill the problem if it were one. Knowing which dang transistor it was (again, a pox on Magnavox) would help a lot. Wish I could jump in a time machine to 1961 and ask these guys. I'd have a lot of questions about the rest of the design.
From here, I'm assuming 0.001uF and 22K (forming a 7.2kHz low pass from supply to collector) is to suppress the higher frequencies at the collector that would interact with the Miller capacitor and reduce the gain. It's still a mystery why 22K is there, other than to reduce dissipation, and why C31 is not larger to just overkill the problem if it were one. Knowing which dang transistor it was (again, a pox on Magnavox) would help a lot. Wish I could jump in a time machine to 1961 and ask these guys. I'd have a lot of questions about the rest of the design.
19-14.5V = 4.5V is across R18 (22k) which gives a collector current of 0.204mA. The emitter current however has 7.1V across R19, witch is also 22k, giving and emitter current of 0.323mA. That suggests that the base current is 0.323 - 0.204 mA = 0.119mA = 0.12mA.
The base is biased by the potential divider R16, R17, which sets the junction at around -6V. That leave 1.2V across the 10k resistor R15, giving a base current of 0.12mA, entirely consistent with the above.
That suggests that the transistor current, gain IC/IB is 0.2/0.12 ~ 1.7
That is truly lousy even for a germanium transistor. But that is what the design voltages and resistor values suggest.
The base is biased by the potential divider R16, R17, which sets the junction at around -6V. That leave 1.2V across the 10k resistor R15, giving a base current of 0.12mA, entirely consistent with the above.
That suggests that the transistor current, gain IC/IB is 0.2/0.12 ~ 1.7
That is truly lousy even for a germanium transistor. But that is what the design voltages and resistor values suggest.
Thank you, Sawyers, I appreciate that. It has been humorous understanding the effort undertaken to deal with a high source impedance, and the frequency and gain limitations of sucky transistors.
This amp has date codes of 1962, and being in the belly of a lower end Magnavox console (the bigger higher end models of the same year had better tube amps) It is obviously an economy design using cheap transistors and laughably undersized power transformer. It would be all too easy to upgrade the predriver section, (or the whole amp, really) to silicon transistors and have something better. I have cleaned and dressed this unit up a bit while preserving it's original character, and despite my temptations I think I will leave it alone, limitations and all. It does work quite well, so long as you only care about frequencies from about 60Hz up to 15kHz at best, as that is the limitation of this interstage transformer, so upgrades around it would be pointless.
This amp has date codes of 1962, and being in the belly of a lower end Magnavox console (the bigger higher end models of the same year had better tube amps) It is obviously an economy design using cheap transistors and laughably undersized power transformer. It would be all too easy to upgrade the predriver section, (or the whole amp, really) to silicon transistors and have something better. I have cleaned and dressed this unit up a bit while preserving it's original character, and despite my temptations I think I will leave it alone, limitations and all. It does work quite well, so long as you only care about frequencies from about 60Hz up to 15kHz at best, as that is the limitation of this interstage transformer, so upgrades around it would be pointless.
I think all is said, but here is similar schema from a book "GE Transistor manual" (7th edition year 1964) there is also earlier editions available online as pdf. Highly recommended data for germanium amps.
I have original book from Finnish distributor Yleiselektroniikka.
I have original book from Finnish distributor Yleiselektroniikka.
Well, when you don’t have transistors with hFE rank 400 to 630, and the ones you do have 100uA ICBO, you do what you’ve got to do. You still see that with silicon transistors when you want a couple hundred k input impedance.
Fascinating. I really appreciate the pictures HJH- I never considered temperature stability as a concern. In the first circuit on the left of the figure 4.8, I am surprised to see an identical R4 in both the collector and emitter side- I understand how the amplifier is still unity gain when taking output from the collector (Rc/Re = 1) but I am still not sure why the output is taken from the collector and not the emitter (is it to avoid loading the feedback driving C2? and results in higher still input impedance?). In the schematic for the amplifier that motivated this thread shown below, I see an identical 22K in the collector and emitter of Q1, and given my mindset of "emitter follower" with output at the emitter it made sense, especially in providing the bias voltage for Q2 for free, except for why the 22K was in the collector- I assumed this was to drop the voltage at the collector due to voltage limits of the device, but now I see it is following the design on the left side of figure 4.8, although they do not take the output from the collector. Further questions as to why the 0.001uF cap is there?
BTW I did find the GE transistor manual online at the World Radio Forum- the various versions are here. I need to do some reading . . .
https://www.worldradiohistory.com/BOOKSHELF-ARH/Bookshelf-GE.htm
BTW I did find the GE transistor manual online at the World Radio Forum- the various versions are here. I need to do some reading . . .
https://www.worldradiohistory.com/BOOKSHELF-ARH/Bookshelf-GE.htm