Im repairing an old Magnavox console for a friends mother. Her husband had bought it for her brand new in 1970 I believe it was. It’s been sitting for quite sometime broken and she would like to use it. She powered up and one channel didn’t work.
This early silicon did not hold up. Both PNP output transistors measured as germanium as they had about 1 mA of Ic leakage. Nearly all of the transistors in the bad channel were toast, measuring as either single diodes or with a gain of 2-3. So I’m going through and making replacements and I being what this is I want to put some lower noise transistors in this preamp section.
There are what I pulled out.
I believe these are rated for 800 mA of current (not sure if that’s max or continuous). The Sam’s manual gives a list of other transistors that would have n been used or can be used in those spots. Being this old there really isn’t much info on many of the options. One pair of transistors on the list had an Ic as low as 200 mA (again, not sure if max). If there is even far less current id like to go with something like the KSC1845/KSA992. The original NPNs have a gain of about 400 and the PNP about 130 (on DCA75)
Other options are BC550/560c, 2N4401/4403, 2N5401/5551, any other suggestions?
I want to put the lowest noise in as I can, the phono player will be used frequently. Will the 1845/992 be okay in this circuit?
Thank you,
Dan
This early silicon did not hold up. Both PNP output transistors measured as germanium as they had about 1 mA of Ic leakage. Nearly all of the transistors in the bad channel were toast, measuring as either single diodes or with a gain of 2-3. So I’m going through and making replacements and I being what this is I want to put some lower noise transistors in this preamp section.
There are what I pulled out.
I believe these are rated for 800 mA of current (not sure if that’s max or continuous). The Sam’s manual gives a list of other transistors that would have n been used or can be used in those spots. Being this old there really isn’t much info on many of the options. One pair of transistors on the list had an Ic as low as 200 mA (again, not sure if max). If there is even far less current id like to go with something like the KSC1845/KSA992. The original NPNs have a gain of about 400 and the PNP about 130 (on DCA75)
Other options are BC550/560c, 2N4401/4403, 2N5401/5551, any other suggestions?
I want to put the lowest noise in as I can, the phono player will be used frequently. Will the 1845/992 be okay in this circuit?
Thank you,
Dan
I think Q203's emitter current is approximately (8.69V - 5.02V) / 47K
I think Q202's emitter current is approximately [ (8.69V / 4700R) minus Q203's emitter current ]
But: I might be mistaken.
I think Q202's emitter current is approximately [ (8.69V / 4700R) minus Q203's emitter current ]
But: I might be mistaken.
The currents are only a couple of mA. Chances are any modern high gain transistor will be better than the originals. Silicon from those days was pretty bad, hence the PNP Ge output stage. Hopefully those are still good because they are harder to substitute. I did one of those consoles not too long ago, where The whole front end was pretty toasted but the PNP Ge output stage was fine. Usually they are transformer coupled at the input so it doesn’t cascade fail. 1 mA leakage is par for the course, so that’s no issue.
Well actually the output wasn’t supposed to be Ge, it was measuring that way. Looking at all of the suggested replacement devices, being Si, it didn’t make sense. Why would they use a silicon NPN and pair it to a germanium PNP. These are silicon, but have become leaky to where my tester thinks they’re germanium.
I have quite a large selection of outputs and wasn’t sure what to use. I ended up using MJ15003 and MJ15004. With some light testing they seem to work just fine. Think those are decent replacements?
Dan
Silicon NPN paired to Germanium PNP has been done before too, so it is a valid combination. Silicon power PNPs cost as much as a house back then, and blew up easy. They would look and feel sort of like a regular EF output stage, but biasing was different.
I liked the regular 2 PNP circlotron-like totem pole stage with the driver transformer better.
I liked the regular 2 PNP circlotron-like totem pole stage with the driver transformer better.
One CAN use a silicon device there in one of these Si/Ge pairs, but the bias circuit would need attention. As stock, it would be underbiased. That can be corrected if you know what you are doing.
Hi saabracer23,
So are you saying all original parts were silicon?
I recommend you post the entire schematic and parts list if possible.
Good luck.
So are you saying all original parts were silicon?
I recommend you post the entire schematic and parts list if possible.
Good luck.
I doubt it. That Motorola TO-3 with the red lettering probably IS Ge. MJ2955 hadn’t been invented yet.
I had no idea that they paired those together. I wonder why the Sam’s listed all of the alternative parts as silicon.
Well with the 4 silicon devices in I’m measuring the voltage drop across the emitter resistors, which are 0.5 ohms and I’m thinking I’m certainly under biased.
For the PNPs I’m seeing 0.06 mV across the resistors and I’m seeing 1.4 mV across the NPN emitter resistors. So looks like I need to make some resistor changes. It sounds perfectly fine, but I’m glad you said something. One of the NPN outputs is bad and then the two PNPs measuring as that much leakage I swapped. I’d prefer not to swap just the one bad NPN, do them as matched pairs.
Dan
Sams listed the alternate parts as silicon because the Ge’s were drying up, and Si could be made to work. That doesnt work in a totem pole stage without totally revamping the bias network, though.
The “bias” current with a Ge in there may very well all be due to leakage, and may not be any less with vbe=0. If you think about it, it may not be all bad. The leakage would provide a minimum bias for the NPN Si, turning the PNP into a current dumper. That’s better than notch distortion with no treatment whatsoever.
The “bias” current with a Ge in there may very well all be due to leakage, and may not be any less with vbe=0. If you think about it, it may not be all bad. The leakage would provide a minimum bias for the NPN Si, turning the PNP into a current dumper. That’s better than notch distortion with no treatment whatsoever.
It might be me, but I only see Q201-Q204 in the schematic. Those transistors are certainly not TO3. Where does the TO3 fit in?
I see the markings on the transistors you pulled out. How did you make a type number from that marking? Anyway, transistors rated for 800 mA AND a Hfe of 400 were pretty scarse back then. What is the type number you found and looked up in SAM's?
Like Mark Johnson wrote, emitter current thru Q203 is about 0.08 mA. (80 uA) And thru Q204 about 1.9 mA. And no, you don't have to really subtract the Ie of Q203 from that value. It is neglectable.
Anyway, the originals were silicon. You see that in the schematic because the CE voltage is 0.6V in both cases. And any transistor like BC557/8 and complements will fit perfectly.
I see the markings on the transistors you pulled out. How did you make a type number from that marking? Anyway, transistors rated for 800 mA AND a Hfe of 400 were pretty scarse back then. What is the type number you found and looked up in SAM's?
Like Mark Johnson wrote, emitter current thru Q203 is about 0.08 mA. (80 uA) And thru Q204 about 1.9 mA. And no, you don't have to really subtract the Ie of Q203 from that value. It is neglectable.
Anyway, the originals were silicon. You see that in the schematic because the CE voltage is 0.6V in both cases. And any transistor like BC557/8 and complements will fit perfectly.
Those were the preamp stage -all Si. No one ever made Ge in anything other than metal cans (Incompatible with the TO-92 molding process). The TO-3 came out of some place else.
Going by what wg_ski is saying I’m guessing not. Everything is silicon though, and the alternative PNPs in the Sam’s are silicon.
Here are some images from the manual. Hopefully they are readable. Here is the semiconductor list
Here is the amplifier section
Preamp
Power supply and such.
Dan
So what would your recommendation be? Should I just put the original Ge PNPs back in and only replace the NPNs leaving the MJ15003s in? Or would it be better to change resistor values to get bias up?
Dan
If I am being brutally candid, IMHO this is a bit of a "sow's ear." How fervently do you want a silk purse?
I think I'm advising a bit of objective planning. The stereo consoles I recall from my parents era (circa '70) had appealing cabinets, but not so much exemplary electronics. An option I remember considering was retrofitting a more modern receiver into the cabinetry. I never did follow through. 🙁
That said, improving an exiting electronic platform can be an interesting challenge. Time and money...
Best.
I think I'm advising a bit of objective planning. The stereo consoles I recall from my parents era (circa '70) had appealing cabinets, but not so much exemplary electronics. An option I remember considering was retrofitting a more modern receiver into the cabinetry. I never did follow through. 🙁
That said, improving an exiting electronic platform can be an interesting challenge. Time and money...
Best.
Did you notice something strange in the parts list and the schematic
So a PNP and NPN with the same type number is impossible.
I don't know what's going on. But when you measure the 152P1B is shows as a perfectly good PNP GE power transistor with a decent Hfe and leakage current.
Can you show what your tester says when you test the other transistor in the output? I don't dare to give you a designation or type. But I would like to see it on the tester.
| Designation | Schematic | Parts List |
| Q17 | PNP | 152P1B |
| Q22 | NPN | 152P1B |
So a PNP and NPN with the same type number is impossible.
I don't know what's going on. But when you measure the 152P1B is shows as a perfectly good PNP GE power transistor with a decent Hfe and leakage current.
Can you show what your tester says when you test the other transistor in the output? I don't dare to give you a designation or type. But I would like to see it on the tester.
I get it, before coming here I was asking myself how shiny can I get this turd by polishing it. Putting in new electronics really isn’t an option, it’s more of a nostalgia thing for my friends mother. She asked me if I could get it so that it’s reliable for the next 10 years or so, the kids really have no interest in it so after she passes, It’ll go onto another home.
But yeah, electronics are not the best, especially this earlier silicon stuff. I started this thread because I wanted to put the low noise transistors into the preamp, I get caught up in stuff like that. So I make sure the preamp is as good as possible all while it’ll be playing into some cheap speakers housed in styrofoam cabinets with a forward facing 12 and a cheapo diagonal firing horn.
So clearly fidelity will not be the highest priority. Though my friend is somewhat considering putting new speakers in. Guess I’ll figure that out once I hear it fully working.
Dan
I did not notice that, they must have Q21 and Q22 on the schematic reversed. Both of the PNPs have the collector tied to ground. They have it correct in the board layout. The two closest to the inside of the board the PNP.
Here is the NPN on the tester.
Oddly the “bad” NPN is measuring okay now…. When I pulled it and measured it the gain was 7. I measured it on like 5 different occasions during the day, because I had mixed it up with the good one and I wanted to determine the one with the low gain. Every time it came up as 6 or 7. Well I just tested both and the “bad” is now measuring with a gain of 86. I don’t think I should trust it. I don’t want my friend hauling this 100 pound cabinet back here.
Dan
For the purpose, I‘d probably do the minimum required to get it reliable.
If it were mine and I wanted to keep it, I’d convert the amp to a proper EF2, using Si devices. 2N5879/5881 outputs, whatever sustained beta drivers I could find that aren’t earmarked for something else, diode stack bias. Change Re to 0.22 ohm. The original circuit is just an EF1, with the driver/VAS running smoking hot. Drop the current in that stage by a factor of 10, but keep the darlington. Revisit the global feedback and compensation for good small signal square wave. Basically turning it into the classic Philips 15/25 watt circuit. And ditch the 470 uF output cap for a bigger one. Use one that actually works with the driver resonance to lower the effective corner frequency.
That would be nice, but I wouldn’t be able to achieve that without someone holding my hand. What is an example of a sustained beta driver?
Currently I have the MJ15003/4, driver is TIP41c which matched well to the part numbers and the other transistors in circuit are 2N5401/5551.
I can order 2N5879/5881. What makes them better, the lower voltage and current ratings? I have others like 2N6284/6287 and hundreds of pulls with lower ratings as well.
Or I could put the two PNPs back in.
I just want the bias to be up higher where it’s supposed to be, or rebuild the circuit as you suggested. From what I’ve heard her mother will be using this a lot and I want to return it in the best possible state I can.
Thanks all,
Dan