I don't know, as I haven't got any experience with them. Note that they were very expensive, much too expensive for me as a poor pupil 
.
German manufacturer Klein & Hummel used them as power devices in their ES-707 integrated amplifier from 1969, whose power amplifiers basically were RCA Power Transistor Manual designs. But I also don't know anything about the reliability of this ES-707. They were also used by Naim.
Best regards!
.German manufacturer Klein & Hummel used them as power devices in their ES-707 integrated amplifier from 1969, whose power amplifiers basically were RCA Power Transistor Manual designs. But I also don't know anything about the reliability of this ES-707. They were also used by Naim.
Best regards!
Cost components was an issue but there was often tricky designs, FI this Sonab P4000 has a lot of good things, CCSs for both the LTP and VAS, this latter being enhanced with an emitter follower, wich is a very good idea, but it s not perfectly designed with a hasardous compensation scheme among others, output stage use two 2N3055, because at the time manufacturers would sell you an NPN power device for say 2$ while they scamed you by asking 10$ for the PNP complementary.
Attachments
Stick a PNP in there and increase R714. Potential pot wiper problem solved, and no need to kluge the PCB doing it.
Keep the NPN output stage, but use MJ15003’s and you can 4-ohm it all you want to. Back in the day when the 3055 QCs we’re all over the place, there was a complementary pair available - 2N3716 and 2N3792. Yeah, the PNPs were pricey. But if you thought epitaxial 2N3055’s had a ruggedness problem you ain’t seen nothin yet till you tried 2N3716’s which did have an 80 V VCEO rating. Plug it in, turn up the volume and wait for the kaboom. Only thing that goes faster are germanium.
Are you sure? Seems to me the circuit would still benifit from a baxdal
Diode//resistor.
Maybe, but keep in mind we're talking about 2N3055 outputs. It's like putting makeup on a pig. Why?
The 2N3055 is an economy transistor used for the sole purpose of saving costs by the time better ones came out. Almost anything performed better, MJ15003 / MJ15004. 2N5631 / 2N6031 and of course the 2N3773 / 2N6609 being solid industry performers. If I ever saw a blown 2N3055, I always used something better for replacement. The original was a linear power supply pass transistor. That was the original design and one reason why originals had low fT (so they wouldn't oscillate mounted off the PCB on a heat sink). As semiconductors matured, we were treated to even better parts and higher amplifier performance as a result.
If I had an amplifier I was inclined (for whatever reason) to keep and use, that had 2N3055 outputs, I'd yank those suckers and install something a lot better. Then I would make slight improvements in the circuit. There is almost zero point in improving a circuit with those 2N3055's installed unless it is a really marginal design. Then, why? Use it as is, really improve it (new outputs + ??) or chuck it.
The 2N3055 is an economy transistor used for the sole purpose of saving costs by the time better ones came out. Almost anything performed better, MJ15003 / MJ15004. 2N5631 / 2N6031 and of course the 2N3773 / 2N6609 being solid industry performers. If I ever saw a blown 2N3055, I always used something better for replacement. The original was a linear power supply pass transistor. That was the original design and one reason why originals had low fT (so they wouldn't oscillate mounted off the PCB on a heat sink). As semiconductors matured, we were treated to even better parts and higher amplifier performance as a result.
If I had an amplifier I was inclined (for whatever reason) to keep and use, that had 2N3055 outputs, I'd yank those suckers and install something a lot better. Then I would make slight improvements in the circuit. There is almost zero point in improving a circuit with those 2N3055's installed unless it is a really marginal design. Then, why? Use it as is, really improve it (new outputs + ??) or chuck it.
And 2N3055 sure as hell doesnt save cost today. I just paid $2.32 a pop for genuine 2SA1943’s on the last Mouser order. PNP and NPN damn near the same price. The N package version is even less and fits TO-3 heat sinks/sockets. Believe it or not it (and the 5200) is more rugged and won’t blink an eye on +/-35V rails.
By the mid 70s original hometaxial 2N3055 were no more produced and were replaced by epitaxial versions that had the same characteristics voltage and current wise but also a higher Ft, Motorola specified its version at 2.5MHz, STMicro at 3MHz and RCA at 2.5Mhz, 2N3055A was 6MHz, so not sure that using slowish 2N3773 was that good of an idea unless there were high supply voltages.
FTR the known complementary is the MJ2955 but in Europe there was also the BDX18, and other derived devices that had good Ft and supported 100V like the BDW51C/52C pair wich i once used.
2955 didn’t exist right away. It came along when epitaxial 3055’s did since that was a complementary-capable process.
2N3773 wasnt slowish any more than any other similar device. Epitaxial versions were always 4 MHz*ish*. JEDEC said a “2N3773” needed to be 800 KHz or more. A good hometaxial will pass that and any epitaxial off the assembly line will. A lot of older motorola epitaxials were 1MHz minumum (ie 2N5630, MJ15001). There isn’t anything that makes them inherently slower than 4MHz 2N3773’s. They have a loose spec, that’s all. One manufacturers claim of 2.5 MHz, vs another’s at 3 or 4 as a minimum makes no difference in practice. They are ALL slower than a 2SC5200 or MJL3281. If you want a real boost in speed, start there - not comparing one old epi type to another.
2N3773 wasnt slowish any more than any other similar device. Epitaxial versions were always 4 MHz*ish*. JEDEC said a “2N3773” needed to be 800 KHz or more. A good hometaxial will pass that and any epitaxial off the assembly line will. A lot of older motorola epitaxials were 1MHz minumum (ie 2N5630, MJ15001). There isn’t anything that makes them inherently slower than 4MHz 2N3773’s. They have a loose spec, that’s all. One manufacturers claim of 2.5 MHz, vs another’s at 3 or 4 as a minimum makes no difference in practice. They are ALL slower than a 2SC5200 or MJL3281. If you want a real boost in speed, start there - not comparing one old epi type to another.
Actually you could do very good amps with minimalistic distorsion with a pair of such old devices but at the expense of complexity
although that was mainly more low power transistors, the advantage of the 2SA1302/2SC3281 wich were released in the late 70s was that you had better perfs with as simple designs as in the early 70s, and hence also low cost, you just have to look at schematics of mass consumers dedicated japanese amps of the era.
although that was mainly more low power transistors, the advantage of the 2SA1302/2SC3281 wich were released in the late 70s was that you had better perfs with as simple designs as in the early 70s, and hence also low cost, you just have to look at schematics of mass consumers dedicated japanese amps of the era.
The only real downfall in those old receivers from the late 70’s and 80’s was manufacturers pushing the SOA limits of the C3281’s and similar types. Manufacturers had traditionally used multiple old slow TO-3 pairs to combat beta droop, but with these new types you “didn’t need to“ anymore. So they push for 100 watts out of one pair at +/-55V, which was fine until you put a second pair of speakers on it. Some ran them as high as +/-70 volts pushing for 150+ watts, and this was downright dangerous. You dont know how many Sanken MT200’s I’ve had to change out on these marginal designs. Yeah, the transistors were good but they were expecting MIRACLES.
Hi wg_ski,
Yep. Completely agree with everything you said. RCA typically had much faster transistors and you often had to recompensate the amplifier.
Specs back then were pretty loosy-goosy. I had to test every single part before use, and I still do. One nice thing about the old days - no fakes. That started with TV horizontal outputs (I didn't fix TV's thank god!).
As for using one pair when two or more should have been used - YES! It's almost like they didn't have any clue what an SOA curve was. Nikko and Sansui were pretty bad, Pioneer also. I didn't fix the cheaper stuff normally.
Yep. Completely agree with everything you said. RCA typically had much faster transistors and you often had to recompensate the amplifier.
Specs back then were pretty loosy-goosy. I had to test every single part before use, and I still do. One nice thing about the old days - no fakes. That started with TV horizontal outputs (I didn't fix TV's thank god!).
As for using one pair when two or more should have been used - YES! It's almost like they didn't have any clue what an SOA curve was. Nikko and Sansui were pretty bad, Pioneer also. I didn't fix the cheaper stuff normally.
Here an additional vintage amp with excellent sonic performance: Bedini-25/25 - go to
https://marantzhallo-fi.blogspot.com/2015/05/bedini-2525-class-power-amplifier.html
in the attachment from post #29 under
https://www.diyaudio.com/community/threads/bedini-amp-repair-advice.305981/page-2
you will find the circuit diagram.
A Vbe multiplier with 2N3055 I have never see before - usual there is in use a BC549 or BC550, sometimes a BD139 in a TO126 outline for heat sink mounting.
https://marantzhallo-fi.blogspot.com/2015/05/bedini-2525-class-power-amplifier.html
in the attachment from post #29 under
https://www.diyaudio.com/community/threads/bedini-amp-repair-advice.305981/page-2
you will find the circuit diagram.
A Vbe multiplier with 2N3055 I have never see before - usual there is in use a BC549 or BC550, sometimes a BD139 in a TO126 outline for heat sink mounting.
Last edited:
What I see about axial (electrolytic) caps is that a lot of 2nd and 3rd tier brands show up as “audiophile” quality, usually high voltage stuff for use with tubes. And people will pay through the nose to get them. The “really good” ones tend to only be available in radials or snap-ins (which is also radial).
Vishay, CD & Kemet all still make a fairly good selection of axial caps with performance matching that of radial - albeit at a higher price. As for using 2nd or 3rd tier brands or, God forbid, "boutique" brands... that is just asking for a problem no matter what marketing BS terms they hang on them.
Also, I very much agree with @anatech that a distortion analyzer is pretty much required if you want to try changing things around effectively. There is a lot of equipment available that will do the job without breaking the bank - even a PC with a decent sound card will allow you to see what works and what does not.
Hal
Also, I very much agree with @anatech that a distortion analyzer is pretty much required if you want to try changing things around effectively. There is a lot of equipment available that will do the job without breaking the bank - even a PC with a decent sound card will allow you to see what works and what does not.
Hal
Hi wg_ski,
A snap-in is a type of radial, but the internal attachments to the foil allow for higher current and pulses. For example, using a normal radial in a Carver M1.5 (or similar) is asking for trouble. I use snap-ins on a PCB I copied from the factory replacement (I still have one of those). Note the PCB has very thick mounting pins as well, not wire leads. I wouldn't really compare those two types of capacitor except for the lead arrangement.
Audiophile caps in general only equal good industrial brands - at the very best. Then you pay extremely high money for a typically over-sized part that doesn't fit the mounting space. If anyone can think of a better reason to avoid these things, please speak up, but those reasons should be more than good enough. It is all hype, marketing stories, and my favorite these days "white papers". Now if you buy from Ebay or other similar sources, you're on your own. Buy a lottery ticket while you are at it.
Hi Hal,
Yes, I use those. They are getting dear and are being phased out. Too bad.
I clearly remember radial caps appearing when all we had were axial caps. It made servicing look messy with the negative lead looped over.
PC sound cards can be used BUT they are often 5 V peak input and higher voltages may well damage them. Plus you still need a real 8R dummy load, non-inductive if you ever hope to have comparative readings. Nothing you see is valid otherwise. Sound cards also have various input impedances and capacitances, so making a divider is a bit more complicated and you need to characterise the sound card inputs. Scaling? Forget it. Your voltmeter also must be accurate to 50 KHz on AC volts if you want any kind of accuracy. That and the meter must have a basic accuracy that is good enough. Most cheaper meters have an LSD (least significant digit) that doesn't mean anything at all. Then there is calibration. If you don't believe me, work out the error specs on your meter. If it doesn't have those specs, they are horrible and you need a real meter. The average human will read a digital scale and take that as accurate reading. Not hardly!
I receive a steady stream of equipment repaired elsewhere that is "clean" or operating properly that are not. It isn't uncommon to see 1/2 to 2% THD on equipment that passed final check using a listening test, and maybe even an oscilloscope. Anyone who believes they can hear distortion below 1% is delusional. You can hear the difference on an A-B, but not checking equipment in isolation that you just repaired.
A snap-in is a type of radial, but the internal attachments to the foil allow for higher current and pulses. For example, using a normal radial in a Carver M1.5 (or similar) is asking for trouble. I use snap-ins on a PCB I copied from the factory replacement (I still have one of those). Note the PCB has very thick mounting pins as well, not wire leads. I wouldn't really compare those two types of capacitor except for the lead arrangement.
Audiophile caps in general only equal good industrial brands - at the very best. Then you pay extremely high money for a typically over-sized part that doesn't fit the mounting space. If anyone can think of a better reason to avoid these things, please speak up, but those reasons should be more than good enough. It is all hype, marketing stories, and my favorite these days "white papers". Now if you buy from Ebay or other similar sources, you're on your own. Buy a lottery ticket while you are at it.
Hi Hal,
Yes, I use those. They are getting dear and are being phased out. Too bad.
I clearly remember radial caps appearing when all we had were axial caps. It made servicing look messy with the negative lead looped over.
PC sound cards can be used BUT they are often 5 V peak input and higher voltages may well damage them. Plus you still need a real 8R dummy load, non-inductive if you ever hope to have comparative readings. Nothing you see is valid otherwise. Sound cards also have various input impedances and capacitances, so making a divider is a bit more complicated and you need to characterise the sound card inputs. Scaling? Forget it. Your voltmeter also must be accurate to 50 KHz on AC volts if you want any kind of accuracy. That and the meter must have a basic accuracy that is good enough. Most cheaper meters have an LSD (least significant digit) that doesn't mean anything at all. Then there is calibration. If you don't believe me, work out the error specs on your meter. If it doesn't have those specs, they are horrible and you need a real meter. The average human will read a digital scale and take that as accurate reading. Not hardly!
I receive a steady stream of equipment repaired elsewhere that is "clean" or operating properly that are not. It isn't uncommon to see 1/2 to 2% THD on equipment that passed final check using a listening test, and maybe even an oscilloscope. Anyone who believes they can hear distortion below 1% is delusional. You can hear the difference on an A-B, but not checking equipment in isolation that you just repaired.