Here is the simplest design that I may use for my engine simulator. It doesnt get more basic than this.
https://makingcircuits.com/blog/simple-100-watt-amplifier-circuit-using-2n3055-transistors/
https://makingcircuits.com/blog/simple-100-watt-amplifier-circuit-using-2n3055-transistors/
The amp post 321 makingcircuits.com has 2 2n3055 as drivers which messes up the highs. power transistors don't make good drivers, too much output capacitance. You need 30 mhz Ft for drivers, to reproduce highs, I've proved by experiment. Also low input gain, not compatible with 1.4 v preamps. .
Same number of transistors, stacked modern 2n3055 out (to 70 v rail) couple of more resistors, AX6. Really great sound and reliable to 70 w/ch. I had mine put out 24.2 vac for 5 seconds on SP2-XT speakers on Rhianon Shut up & Drive. https://www.diyaudio.com/forums/solid-state/236256-retro-amp-50w-single-supply-42.html Doesn't require 4 ohme speakers like post 321
I used 2n5320/22 or MJE15028/29 as drivers and VAS, 2n5401 or MPSA56 as input transistor. I used MJ15003 as output transistors but modern 2n3055 or MJ15015 would have worked. Use heat sinks.
Same number of transistors, stacked modern 2n3055 out (to 70 v rail) couple of more resistors, AX6. Really great sound and reliable to 70 w/ch. I had mine put out 24.2 vac for 5 seconds on SP2-XT speakers on Rhianon Shut up & Drive. https://www.diyaudio.com/forums/solid-state/236256-retro-amp-50w-single-supply-42.html Doesn't require 4 ohme speakers like post 321
I used 2n5320/22 or MJE15028/29 as drivers and VAS, 2n5401 or MPSA56 as input transistor. I used MJ15003 as output transistors but modern 2n3055 or MJ15015 would have worked. Use heat sinks.
Last edited:
That circuit in post 321 is terrible, if I may say so. There are no current sharing resistors in the lower pair, and a redundant collector resistor in the second 2N3055.
AFAIK the 2N3904/6 drivers are only rated to 40V.
Asking for trouble.
Not even as good as some of the crude early circuits RCA published.
AFAIK the 2N3904/6 drivers are only rated to 40V.
Asking for trouble.
Not even as good as some of the crude early circuits RCA published.
Last edited:
The datasheets I have for RCA transistors show the 2N3055 hometaxial ft to be 800kHz minimum. The 2N3442 was 200kHz, as was the 2N3771/2/3.
I used the original 2N3055 in a number of amps. Wasn't the best for highs, I agree, but certainly the others with 200kHz were terrible. I got as far as simulating a 3442 70W circuit but the circuit could barely manage 10kHz so never used it for any audio. (You have to check the models for those older devices as SPICE simulation came out after they'd been manufactured the first time. Some models typically use the epi spec of 3MHz - and this epi 2N3055 is better. I suggest those who say "never use the 3055" are thinking of the old device originally sold for power supplies. )
Some manufactures of 3055 MAY have lower fT than 800kHz but that would not meet the original specification, let alone the newer epi one.
I used the original 2N3055 in a number of amps. Wasn't the best for highs, I agree, but certainly the others with 200kHz were terrible. I got as far as simulating a 3442 70W circuit but the circuit could barely manage 10kHz so never used it for any audio. (You have to check the models for those older devices as SPICE simulation came out after they'd been manufactured the first time. Some models typically use the epi spec of 3MHz - and this epi 2N3055 is better. I suggest those who say "never use the 3055" are thinking of the old device originally sold for power supplies. )
Some manufactures of 3055 MAY have lower fT than 800kHz but that would not meet the original specification, let alone the newer epi one.
Hi John,
You're right. The original 2N3055 was used by Lamda power supplies (they suggest they designed or spec'd it). It had a low fT because the pass was mounted on a heat sink remotely and they didn't want it to oscillate. Somewhere I have the original 1960's data sheet for it. But the Jedec spec is for the original with c-e breakdown of 60 VDC. That is what a real 2N3055 is. Also, the construction wasn't too linear with current vs beta either.
Anyway, why use an ancient design even if it's performance is ... "okay". Today's transistors are so superior in every way, why limit yourself (seriously limit) by using something that is long obsolete? Of course more modern designs will make the most of your parts too.
You're right. The original 2N3055 was used by Lamda power supplies (they suggest they designed or spec'd it). It had a low fT because the pass was mounted on a heat sink remotely and they didn't want it to oscillate. Somewhere I have the original 1960's data sheet for it. But the Jedec spec is for the original with c-e breakdown of 60 VDC. That is what a real 2N3055 is. Also, the construction wasn't too linear with current vs beta either.
Anyway, why use an ancient design even if it's performance is ... "okay". Today's transistors are so superior in every way, why limit yourself (seriously limit) by using something that is long obsolete? Of course more modern designs will make the most of your parts too.
I'm paying $4 for alleged military surplus. On Semi, any TO3 is $10 up. I bet Chinoy and NareshBird can do a lot better in India for 2n3055 with fabs there still making them.
I scored 40 On MJ15015 for $1.60 each 8 years ago. All hauled to the steel scrapyard by the burglar 9/20. In 4 SUVloads he didn't even make enough to ransom his Cherokee out of the tow lot. I saw him riding a bicycle by my house 3 days later.
I scored 40 On MJ15015 for $1.60 each 8 years ago. All hauled to the steel scrapyard by the burglar 9/20. In 4 SUVloads he didn't even make enough to ransom his Cherokee out of the tow lot. I saw him riding a bicycle by my house 3 days later.
“Alleged” military surplus. If you’ve still got a trusted surplus supplier I guess it still makes some sense. All three of the places I used to get obsolete TO-3’s cheap folded during covid. So any new parts come from Mouser or Digikey. I probably have a lifetime supply of TO-3’s now anyway.
Somebody in the US, looking to fix his amp or build a new one, doesn’t have a bunch of inventory, and is unwilling to take chances on what he gets, is stuck paying ten bucks for 3055’s or 21195’s. Which is a better use of the ten bucks?
Somebody in the US, looking to fix his amp or build a new one, doesn’t have a bunch of inventory, and is unwilling to take chances on what he gets, is stuck paying ten bucks for 3055’s or 21195’s. Which is a better use of the ten bucks?
I guess that despite the superiority of todays devices, the 2N3055 truly is a classic device of much virtue.
I have two questions for you all:
1. What was the most useful Darlington output device in old amplifiers?
2. What is the basic story of the SJ devices that were 'bared' from the consumer market?
I have two questions for you all:
1. What was the most useful Darlington output device in old amplifiers?
2. What is the basic story of the SJ devices that were 'bared' from the consumer market?
VCB0=100V, VCE0=100V, VCER=70V, according to the datasheet. So VCER lower than VCE0. Really?
That is the company spec...if you ask them questions, they may or may not reply.
It works as a replacement or new, and at about 50 cents each, I am happy. Available in enough quantity if required, can be sent from a larger center if needed.
The 'ST' marked versions are about 60 cents, retail price for single units at a shop.
BEL are sometimes fake, and like mentioned above there are more capable transistors.
My work in electronics is mostly repair for self and a few close friends, and I can get the parts I need.
Some seller in UK was slling CDIL TO-3 at 1.92 GBP each, about 4x Indian price, but they need to eat as well ...
It works as a replacement or new, and at about 50 cents each, I am happy. Available in enough quantity if required, can be sent from a larger center if needed.
The 'ST' marked versions are about 60 cents, retail price for single units at a shop.
BEL are sometimes fake, and like mentioned above there are more capable transistors.
My work in electronics is mostly repair for self and a few close friends, and I can get the parts I need.
Some seller in UK was slling CDIL TO-3 at 1.92 GBP each, about 4x Indian price, but they need to eat as well ...
My point about the 2N3055 is that many dismiss it on the grounds of having a too low fT for good audio, but that is really only true for the original device which is obsolete. The epi devices which are now manufactured have ft's around 2-3MHz, so all I am suggesting is that the 3055 is unfairly dismissed if people regard it as the being the old (obsolete) device. For example, other devices which have been suggested are the MJ802/4502 (also 2MHz but obsolete) MJ15003/4 which are also 2MHz, and so on. So it is possible to design a "good" amplifier with 2N3055/MJ2955 or equivalent (TIP3055/TIP2955, but the lower SOA may mean parallelling a couple).
Of course, the newer transistors are better in almost every respect - gain linearity and ft being two important parameters. The changeover point for me came when they also offered better SOA than the original 2N3055. The original could handle 1.95A at 60V, but modern devices like the MJL21193 can do 2.25A at 80V (single pulse 1 sec). Though for power handling I still prefer the TO-3 MJ21193/4 - they can do 250W. They were my "goto" devices for 100W now but the MJL1302/3821 (or 2SC5200 etc) are, though you need more in parallel.
And along with the newer plastic devices, TO-3 cans are being priced out of the market, it seems to me.
-> franzm, the original spec was Vceo60, Vcer70 and Vcbo 100. Some may be manufacturing devices in a similar way to the original 2N3055 which meet those specs, but Vceo100 is an error, I suggest.
Of course, the newer transistors are better in almost every respect - gain linearity and ft being two important parameters. The changeover point for me came when they also offered better SOA than the original 2N3055. The original could handle 1.95A at 60V, but modern devices like the MJL21193 can do 2.25A at 80V (single pulse 1 sec). Though for power handling I still prefer the TO-3 MJ21193/4 - they can do 250W. They were my "goto" devices for 100W now but the MJL1302/3821 (or 2SC5200 etc) are, though you need more in parallel.
And along with the newer plastic devices, TO-3 cans are being priced out of the market, it seems to me.
-> franzm, the original spec was Vceo60, Vcer70 and Vcbo 100. Some may be manufacturing devices in a similar way to the original 2N3055 which meet those specs, but Vceo100 is an error, I suggest.
This type and those version from Siemens (BD130) was good known in Germany only approximately 10 years later than in the United States.
Maybe one of the member can upload one of the first US advertisement of RCA's 2N3055 (early 60's). Thank you very much.
P.S.:
A very experienced technician once told me that there is no other type of transistor that has had and still has so many different internal designs under the same naming "2N3055".
Unity gain commercial power amplifiers in this topology I haven't found until now, unfortunately; usual power amplifier stages have gain factors between 20 and 60 (VAS+Buffer or LTP-VAS+Buffer). Most easy topology include voltage gain was published in the German magazine "Populäre Elektronik" in the 70's - the voltage gain stage consists only of one transistor and the whole stage operates in the inverted mode (one of my favorite design). Maybe one of the member can upload this article - maybe there exist the same article in english.
For me most interesting steps for such complementary and quasi complementary power amp stages for optimizing sound character are described under
http://www.renardson-audio.com/classbff.html
and
https://www.diyaudio.com/community/threads/new-cherry-ndfl-amp.148066/
file:///C:/Users/admin/Downloads/ndfl.pdf
https://www.diyaudio.com/community/...k-how-it-realy-works-and-some-examples.96634/
Concerning 2N3055 there are additional this threads here on diyaudio:
https://www.diyaudio.com/community/threads/2n3055-the-early-years.382690/
https://www.diyaudio.com/community/threads/amplifier-based-on-2n3055.10337/
https://www.diyaudio.com/community/threads/2n3055-amplifier.276300/
https://www.diyaudio.com/community/threads/120-watt-amp-in-2-ohm-loud-with-2n3055.240814/
anywhere is posted a schematic from Mr. Nelson Pass' favorite 2N3055 power amplifier design - as I recall right a "circlotron".
Predecessors of 2N3055
https://www.diyaudio.com/community/threads/vintage-transistors.167680/
P.S.: I have receive your PM; I am glad that my information was helpful for you
Last edited:
John, that is correct, but it iss still abysmally low for audio.
I don't know of any amp with such relatively slow devices that comes even close to a modern amp with much faster devices in terms of low distortion for instance. They may be better than amps with the original 3055, but still a lot worse that modern ones with the faster 30MHz devices.
Jan
By some short digging in my stashes and boxes, I just found something that I had built about 45 years ago. I was a young student then, much impressed on Vceo, PD and Ic ratings, but literally clueless about fT and SOAR:
These are four pairs of 2N3442's, intended as output devices, and a fifth one as drivers, bolted via back-to-back aluminium L shaped extrusions onto relatively massive heatsinks. I recall to have been inspired by a 1978 RCA BD550 data and application sheet, but didn't manage to get my hands on these devices. (Btw, did they ever see the market's light?)
I had pulled the 2N3442's from abandoned lab bench supplies at a company I was working during my holidays, and still have some hands full lying in another box. As you see, the amplifiers never came to fruition, as in the meantime I learned about transistion frequency and SOAR
.
I had pulled the 2N3442's from abandoned lab bench supplies at a company I was working during my holidays, and still have some hands full lying in another box. As you see, the amplifiers never came to fruition, as in the meantime I learned about transistion frequency and SOAR
.