I have a power amp board (old acoustic control corp brand amp) with an old PNP can transistor that's labelled RCA 40410 in the Q7 position. Power wise I believe I have one that will cross (2N4918), but the bias must be off because it's dropping the voltage at the mid point between C10 and C11 (supposed to be 36V) to 14V.
Can I re-bias this stage in order to get the transistor I have on hand to work? It looks like it may have a problem on Q4 if I mess with the bias on Q7. I don't have a great understanding of those two stages unfortunately. Thanks in advance! I'd love to get this old amp working.
I was reading here:
Replace 40409, and 40410 with what?
about substitutes and someone (post #4) mentions a Miller Cap may need to be added if going to a plastic case transistor. Could this be my problem? Pardon my SS electronics ignorance!!
Can I re-bias this stage in order to get the transistor I have on hand to work? It looks like it may have a problem on Q4 if I mess with the bias on Q7. I don't have a great understanding of those two stages unfortunately. Thanks in advance! I'd love to get this old amp working.
I was reading here:
Replace 40409, and 40410 with what?
about substitutes and someone (post #4) mentions a Miller Cap may need to be added if going to a plastic case transistor. Could this be my problem? Pardon my SS electronics ignorance!!
Attachments
Last edited:
There is a difference in the bias between the 40410 and the 2N4918
also big difference in the Ft and 50 volts difference in max. working voltage .
Its not the ideal replacement.
Rail is 76 volts and the 2N4918 is 40 volts maximum whereas the 40410 is 90 volts maximum.
also big difference in the Ft and 50 volts difference in max. working voltage .
Its not the ideal replacement.
Rail is 76 volts and the 2N4918 is 40 volts maximum whereas the 40410 is 90 volts maximum.
40410 was a 100 mhz ft transistor. So I doubt if substituting 30 mhz modern transistors like MJE15029 is going to be a problem. Big problem is modern pinout, usually bce. 40410 was ebc.
40410 was a 40406 with factory installed heat sink. Surprise, you can buy 40406 from surplussales.com of nebraska. I bought some 1/27. RCA logo, date 98 something. You'll have to add your own TO5 heatsink from arvid or somebody.
Biggest potential problem 40411 was homotaxial 2n3055 selected for high voltage. Homotaxials have .15 ohm b-e resistance. Modern 2n3055 or any epitaxial transistor does not. So if you substituted the output transistors with modern parts, you need to add emitter resistors at least. Or else suffer thermal runaway and toasted output transistors if you turn the volume up and warm them.
OTOH, surplus sales alleges he has 2n3055h in the bargain category. He doesn't test his parts, I have no idea if they are really homotaxial parts or not. But you can give them a try if you want. They are not RCA.
without homotaxial output transistors , you need to install emitter resistors on them. .15 ohm minimum, I'd use .33 ohms myself. 3 watt or 5 watt wirewound resistors.
Another possible cause of RF oscillation of epitaxial output transistors is high Ft - 2 mhz instead of 400 khz of old 1975 parts. People get around this by putting 10 ohm 1 w resistors series the base of the output transistors.
So check if your cause of wrong center point voltage is really oscillation. Also might be something else blown up. I doubt if 2n4918 would throw it that far off, unless the gain was 5 or something instead of 25-30 of 40410.
Edit, duncan may be right, 2n4918 is leaking due to low Vceo.
40410 was a 40406 with factory installed heat sink. Surprise, you can buy 40406 from surplussales.com of nebraska. I bought some 1/27. RCA logo, date 98 something. You'll have to add your own TO5 heatsink from arvid or somebody.
Biggest potential problem 40411 was homotaxial 2n3055 selected for high voltage. Homotaxials have .15 ohm b-e resistance. Modern 2n3055 or any epitaxial transistor does not. So if you substituted the output transistors with modern parts, you need to add emitter resistors at least. Or else suffer thermal runaway and toasted output transistors if you turn the volume up and warm them.
OTOH, surplus sales alleges he has 2n3055h in the bargain category. He doesn't test his parts, I have no idea if they are really homotaxial parts or not. But you can give them a try if you want. They are not RCA.
without homotaxial output transistors , you need to install emitter resistors on them. .15 ohm minimum, I'd use .33 ohms myself. 3 watt or 5 watt wirewound resistors.
Another possible cause of RF oscillation of epitaxial output transistors is high Ft - 2 mhz instead of 400 khz of old 1975 parts. People get around this by putting 10 ohm 1 w resistors series the base of the output transistors.
So check if your cause of wrong center point voltage is really oscillation. Also might be something else blown up. I doubt if 2n4918 would throw it that far off, unless the gain was 5 or something instead of 25-30 of 40410.
Edit, duncan may be right, 2n4918 is leaking due to low Vceo.
Last edited:
Thank you Indiana. Looks like some good options. I will read through your post a few times before I order a transistor for this thing. I don't mind giving the Q7 transistor its own small perf-board, heatsink, and some jumper wires. There's plenty of room in this chassis. It's like a 4 cylinder VW engine in a 1973 Cadilac's engine compartment in this thing. lol.
The OP tranistors are original and still test good thankfully.
I tried again with a bit of an experiment and I'm getting the correct voltage (36V at the in between output transistors point marked in the schematic) IF I do not have the base of my substituted 2n4918 connected to the circuit. I also measure 36 volts at the wire that feeds the base of the transistor in question (Q7). So I guess it's the transistor being full on at idle causing the voltage drop.
Duncan2: I agree with you on your data, and yes the rail is 76 volts, but there is 36V across the 40V rated transistor in the way the A/B stages are laid out. Not much margin for error there, but I just wanted to try what I had to see if it was worth investing in proper parts for this thing (I absolutely think it will be but I want to make sure it's salvageable). A 40v transistor should survive 36v, or am I overlooking something?
Last edited:
The 2N4918 is a 40 volt part the 2N4920 is 80 volt. This part should work as it is below the breakdown. It will need to have a heatsink to work.
Duke Aguiar, just like the signed schematic.
Duke Aguiar, just like the signed schematic.
You will need a 76 volt or higher part in there. It may have 36 volts on it in quiescent, but it will swing almost the entire supply voltage. 80 volt or better parts are needed. Something like a BD139/140 pair. ECG128/129 is probably as close as you can get to the originals, but they are very likely to be fakes. If they are fakes, you’ll blow it up. I wouldn’t worry if the fT is lower than the 100 MHz originals, since the 390 pF collector-base capacitors slow them down anyway. 100 MHz is not needed, 10 or 20 will do.
If you use more modern epitaxial outputs, you do want emitter resistors. R26 sort of functions as one, but adding one in the emitter leg of the upper transistor can’t hurt.
If the lower driver transistor is “leaky” due to insufficient voltage handling, you won’t get a DC offset like you’re observing. You’ll get too much quiescent current or just blow the output stage. Something else is causing the “center point” to be at 14 volts instead of 36. Look toward the front end, where DC feedback is applied, to diagnose. The problem may be “off screen” of the partial schematic you show.
If you use more modern epitaxial outputs, you do want emitter resistors. R26 sort of functions as one, but adding one in the emitter leg of the upper transistor can’t hurt.
If the lower driver transistor is “leaky” due to insufficient voltage handling, you won’t get a DC offset like you’re observing. You’ll get too much quiescent current or just blow the output stage. Something else is causing the “center point” to be at 14 volts instead of 36. Look toward the front end, where DC feedback is applied, to diagnose. The problem may be “off screen” of the partial schematic you show.
And just because a transistor is “rated” at only 40 volts doesn’t necessarily mean that’s all it will take. Especially if they are part of a “family” with increasing voltage ratings. They are sample tested and binned. If too may in the sample don’t pass the spec, they just drop whole lot down to the next bin. You can always select for higher voltage if you need to - just like they did with the 40411. It is nominally a “60 volt” 2N3055. You would want to test Vceo and Iceo *at temperature* when doing so.
You may very well have gotten lucky and had a 4918 that would take 80 volts, or even 100. I’ve had TIP 42’s with no voltage grade letter take the full 100 volts of the “C” part. You MUST test them if you are going to do this.
You may very well have gotten lucky and had a 4918 that would take 80 volts, or even 100. I’ve had TIP 42’s with no voltage grade letter take the full 100 volts of the “C” part. You MUST test them if you are going to do this.
Awesome! Great work on the acoustic amps. I bought this one dead for $75 hoping to get it working (I have some background in tubes but an absolute rookie in SS), I played through an acoustic owned by a friend for many years and always loved it! I had to get one of my own.
Interesting. I didn't know it worked like this. I got a lucky one then! It's still functioning. It does get a little warm (slightly) when I have the base connected to the rest of the circuit. The wire going to the base has 35V on it (not connected to the transistor to test if this is the problem and it is). When the base isn't connected to the circuit, the voltage point (marked 36v) measures ok. So it is this transistor dropping the voltage.
I have pigtails going from the PCB to my 2n4918 transistor (heatsinked).
It's not liking what's on the base since as soon as I connect the base to the 2n4818, the voltage drops to half of what it should be.
Here's more of the schematic btw.
Attachments
Last edited:
You don’t know if you’re “lucky” unless you get it to full output power. It may be ok quiescent and go bang when you start putting out power.
To test a transistor to se if it can be used at higher voltage, put it in series with a 100k resistor and apply full voltage. Measure the voltage across it and the current through it. If it breaks down it will act like a zener diode. The resistor will limit the current through it to a safe value. If the “zener” voltage is high enough, it won’t let any current flow at all. That’s what you want to see. Leakage under a micro amp or so.
Put a soldering iron to it and raise its temp until you can’t stand to put your finger on it, and make sure is still doesn’t “leak”. You cannot always do this to select high power transistors, because there is also second breakdown to consider, and thus won’t address it. But for drivers or VAS, it almost always works. The reason they could select 40411 that way is that it is known to be very rugged by design.
To test a transistor to se if it can be used at higher voltage, put it in series with a 100k resistor and apply full voltage. Measure the voltage across it and the current through it. If it breaks down it will act like a zener diode. The resistor will limit the current through it to a safe value. If the “zener” voltage is high enough, it won’t let any current flow at all. That’s what you want to see. Leakage under a micro amp or so.
Put a soldering iron to it and raise its temp until you can’t stand to put your finger on it, and make sure is still doesn’t “leak”. You cannot always do this to select high power transistors, because there is also second breakdown to consider, and thus won’t address it. But for drivers or VAS, it almost always works. The reason they could select 40411 that way is that it is known to be very rugged by design.
Measure the voltage across R26, and see if that’s changing when you connect the driver transistor. And you can try shorting Q4 collector to emitter to intentionally remove output stage bias, and see if the voltage on R26 goes to zero.
I like the idea of this test. So what I know of zeners is that they'll clamp a voltage (almost like a regulator). If the transistor is acting like a zener, what should I expect to measure across the 100k and the transistor with voltage applied across the two in series? I assume I don't put a voltage on the base of the transistor?
wg_ski
Take a ½ wave ac voltage source (100v peak or ?) with a current limiter (20k or ?) and place a scope E to C (B open) and see the waveform clip & sometimes you may see a burst of oscillation.
EBay has some real RCA 40490 & 40410 that could save you some time.
Duke
Take a ½ wave ac voltage source (100v peak or ?) with a current limiter (20k or ?) and place a scope E to C (B open) and see the waveform clip & sometimes you may see a burst of oscillation.
EBay has some real RCA 40490 & 40410 that could save you some time.
Duke
I found an ECG replacement today that's pretty close. I like to experiment though for the sake of learning. I ordered some modern options to play with along side the actual parts. I'm ready for smoke to be let out 😀. I know nothing about epitaxial etc... I have a LOT to learn. I'm tempted to stick with 12ax7s and 6L6s, but I really want to get at least some grip on transistor amplifier basics.
I don't know who ECG is, but if they are like NTE, or Central, or New Jersey semi, they leave all the specs out. Like Ft - They will give you Vcesat, which is great if you are making a switch. An amplifier is not a switch. I had NTE49/50 as drivers, they sounded like mud on high freq material.
I had TIP31c/32c as drivers that had the Vceo I needed for 70 v rail, and it sounded like **** compared to the surviving 1970 40406 40407 on the other channel. Took MJE15028/29 drivers to put the sparkle in the top octave piano & tinkly bells. Ft 30 mhz. The TIP31c/32c are ft 3 mhz.
Reason I don't respect BD139/140. the ones we can buy in the US, fairchild, have no Ft spec. Besides the pinout being wrong to replace 40406 40407.
The 40406 40407 driving 40411 quasi complementary output, that is what came in my ST120, as you had 40409 40410 in your amp. Dynaco bought their driver heatsinks from somebody besides RCA. When the volume was loud enough to get the idle bias current over 15 ma (they had to charge up a cap to get there) they sounded great.
I had TIP31c/32c as drivers that had the Vceo I needed for 70 v rail, and it sounded like **** compared to the surviving 1970 40406 40407 on the other channel. Took MJE15028/29 drivers to put the sparkle in the top octave piano & tinkly bells. Ft 30 mhz. The TIP31c/32c are ft 3 mhz.
Reason I don't respect BD139/140. the ones we can buy in the US, fairchild, have no Ft spec. Besides the pinout being wrong to replace 40406 40407.
The 40406 40407 driving 40411 quasi complementary output, that is what came in my ST120, as you had 40409 40410 in your amp. Dynaco bought their driver heatsinks from somebody besides RCA. When the volume was loud enough to get the idle bias current over 15 ma (they had to charge up a cap to get there) they sounded great.
Last edited:
I don't know exactly what wgski meant in post 10.
I do a Iceo test, that is current with base open. I use 47 kohm & 12 or 24 v. I put ma scale of a dvm between transistor & gnd. NPN plus & 47k go on collector, PNP plus & 47 k go on emitter. Current is microamps, the transistor is good. Current =sourcevoltage/47k, the transistor is leaky.
If you have 55 or 100 v supply you could also do this Iceo test, maybe with 100 k instead of 47. But be careful, 100 v from one hand to the other can stop your heart.
Just had another scan of 1972 RCA power transistor databook. 2n5321 & 5323 are complimentary npn/pnp TO5 transistors at Vceo 75. 2n5320 & 5322 are complimentary TO5 transistors at Vceo 100v. Both have 50 mhz Ft. both are EBC. 40406 is pnp with Vceo 55 v. Surplus sales didn't have any but 40406, but he did have Nat Semi 2n2270, TO5, npn, 60 v Vceo. Had gold plate pins, unlikely to be from land of counterfeiters like the ebay vendors.
I do a Iceo test, that is current with base open. I use 47 kohm & 12 or 24 v. I put ma scale of a dvm between transistor & gnd. NPN plus & 47k go on collector, PNP plus & 47 k go on emitter. Current is microamps, the transistor is good. Current =sourcevoltage/47k, the transistor is leaky.
If you have 55 or 100 v supply you could also do this Iceo test, maybe with 100 k instead of 47. But be careful, 100 v from one hand to the other can stop your heart.
Just had another scan of 1972 RCA power transistor databook. 2n5321 & 5323 are complimentary npn/pnp TO5 transistors at Vceo 75. 2n5320 & 5322 are complimentary TO5 transistors at Vceo 100v. Both have 50 mhz Ft. both are EBC. 40406 is pnp with Vceo 55 v. Surplus sales didn't have any but 40406, but he did have Nat Semi 2n2270, TO5, npn, 60 v Vceo. Had gold plate pins, unlikely to be from land of counterfeiters like the ebay vendors.
Last edited:
I remember once grabbing a chassis of a tube amp build as it was falling to the floor. I somehow ended up with the 460VDC B+ on one hand to the other holding the chassis. Goes without saying that I hope to not repeat that! lol. Solid state chassis sit flat to a bench a little nicer than tube amps thankfully.
Sounds like a good diagnostic for leaks. I'll set that up on a bread board with a meter using clip leads so all I have to do is flip the switch.
C E O means collector to emitter with base open. That’s how you connect it, just like it sounds.
You have it backwards anyway. Emitter to collector breakdown will be about 15 volts. And reverse beta is pretty bad, too. So don’t hook up transistors backwards - they don’t like it.