A simple stereo Amp aroud a 2N3773

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm pretty sure the RCA device came first. And of course the 3055 before 3773 - they issued the part numbers in sequence until complementary pairs messed this all up.
In that time, various types was divided into groups resp. families by RCA (see also PDF attachement):
1) 2N3771 Family (related audio type = 40411)
2) 2N3773 Family
3) 2N3879 Family and for small/medium signal applications
4) 2N4036 Family

In this case this threads could be also of interest:

http://music-electronics-forum.com/attachments/10662d1281274151-40408.pdf
http://www.diyaudio.com/forums/parts/32140-old-rca-transistors-what-they.html
Popular Electronics February 1967
http://www.diyaudio.com/forums/solid-state/140988-vintage-brute-70-40411-clone.html
http://www.diyaudio.com/forums/solid-state/120226-best-transistor-substitutes-heathkit-ar-15-a.html

To find detailled information concerning the exactly line up of history all folders and databooks from RCA between 1963 and 1975 are necessary. Later released Follower devices was the BD550 series and the BD750 series; from this series I have various papers and the same I want to have from 2N3055 and 2N3773 and the associated types
 

Attachments

  • RCA 2N3771 Family - 2N3773 Family.pdf
    48 KB · Views: 502
There's a very interesting article in the Bob Pease book Troubleshooting Analog Circuits ( pages 84-85) regarding the 2N3055 Hometaxial, Epitaxial and the new Planar manufactured devices.

The Original 60's 3055 was a Hometaxial part and and was single diffused, however this process took up to 20 hours, which made it very expensive though the parts were the most rugged on SOA. In the late 70's and 80's 2N3055's made this way would have an H after the part number.
Epitaxial parts are faster cheaper to manufacture and have better characteristics except for ruggedness, and the Planar moves the game on another step. Most 3055's are likely to be made with the Epi process, the faster and newer parts such as 3281's etc with Planar and multiple emitter technology.

So the old school 3055's and 3771's were tougher than modern counterparts due to their construction however they were also slower.

If you want to get some of the original Hometaxial devices from this era then I'd hunt down the old aluminium base Motorola parts, or the RCA 1Axx and 1Bxx Audio specific transistors from around 1970.
As far as I know no one has faked these as there isn't the demand for them.
 
There's a very interesting article in the Bob Pease book Troubleshooting Analog Circuits ( pages 84-85) regarding the 2N3055 Hometaxial, Epitaxial and the new Planar manufactured devices.

The Original 60's 3055 was a Hometaxial part and and was single diffused, however this process took up to 20 hours, which made it very expensive though the parts were the most rugged on SOA. In the late 70's and 80's 2N3055's made this way would have an H after the part number.
Epitaxial parts are faster cheaper to manufacture and have better characteristics except for ruggedness, and the Planar moves the game on another step. Most 3055's are likely to be made with the Epi process, the faster and newer parts such as 3281's etc with Planar and multiple emitter technology.

So the old school 3055's and 3771's were tougher than modern counterparts due to their construction however they were also slower.

If you want to get some of the original Hometaxial devices from this era then I'd hunt down the old aluminium base Motorola parts, or the RCA 1Axx and 1Bxx Audio specific transistors from around 1970.
As far as I know no one has faked these as there isn't the demand for them.

Thank you for this advices.
mean you this book from follow URL?
Troubleshooting Analog Circuits - Robert A. Pease - Google Books
 
Ah but.... the 2N3773 family (these were introduced by RCA BTW) have a poor frequency response. ALthough the 40411 (a 2N3772 based device) is quoted as "800kHz typical" (which is minimum for the 2N3055) the worst case is 200kHz for the old hometaxial device.

Does not make a hifi amp.

The MJ21194/MJ21193 can do almost the same SOAR (these are about the best around) but have 4 MHz fT's ... much better.

John
 
Just like the 2N3055, there are different versions of the 3773. The most advanced had a Ft of 3Mhz (or 4, depending on the actual manufacturer) and still guaranteed very decent SOA (it had to since the originals had that spec). MJ21193/4 actually are several generations removed descendants of the 3773.
The 'faster' 3773 is actually quite capable as far as being robust within it's spec, but suffers quite large beta droop with current rise, something later transistors and technologies tried hard to remedy, ultimately resulting in LAPT/3-diffused types we use today. The 3773 was used en mass in motor control applications, and also was really the first truly usable audio output transistor. Tons were screened to higher voltages and used in many PA amps!
Regarding the 'Toshiba' 2N3773, there are two versions. Early devices are re-labelled original Toshiba parts, probably one of the 2SD5xx series. I've come across many of them in older PA amps and they work perfectly fine to this day. It is highly probable they were re-labelled by the company (or it's distributors) for a large order of second-source parts. At one time, I am told, these were actually quite sought after due to their good performance (this would be late 70s - early 80s timeframe), which eventually led to other distributors re-labelling inferior parts the same way. I actually have a few of both kind and the 'newer' and useless type is very similar but only until you look at it closer. It does have the familiar Toshiba convex cap (original 2N3773 have a standard flat cap) but thiner base (the 'original' toshiba part has a thicker base) and an inferior finish which tends to dull or corrode and become gray. The actual crystal in the 'original' is very similar (even slightly larger) but not the same as the 2N3773. It also tests slightly better on a curve tracer WRT breakdown voltage, on average (although genuine 2N3773 are very often over spec WRT breakdown voltage). The later 'fake' has a much smaller crystal. And, to make things worse, there are tons of 2N3055 re-labelled as 3773 and later MJ parts around too :(
 
Last edited:
Ah but.... the 2N3773 family (these were introduced by RCA BTW) have a poor frequency response. ALthough the 40411 (a 2N3772 based device) is quoted as "800kHz typical" (which is minimum for the 2N3055) the worst case is 200kHz for the old hometaxial device.

Does not make a hifi amp.

The MJ21194/MJ21193 can do almost the same SOAR (these are about the best around) but have 4 MHz fT's ... much better.

John
I don't agree. If that were true, were all amplifiers from those days, e. g. those from
http://www.diyaudio.com/forums/soli...-models-quasi-complementary-power-output.html
not suited for true hifi.
Of course - if one use such 200 KHz transistors for the output stages in topologies, developed for 70 MHz ft devices in ring emitter structure like 2SC2681 from NEC, you are right.
 
Last edited:
Hi

I used to think that the 2N3055 (hometaxial) made good hifi amplifiers too. And I will say they could produce fairly good sounding amplifiers, perfectly OK for some. They were "listenable to" if you were not critical. My first two or three stereo sets were based on this transistor. However, comparing to Arthur Bailey's complementary amplifier using MJ481/MJ491 I was greatly impressed by the openness - and largely this was down to the frequency response (in the region of 200kHz) as it is possible to virtually eliminate crossover distortion even for quasi-complementary amplifiers as long as you do not use Miller capacitors! Most 2N3055 amplifiers could stretch to 50 kHz, but 40411-based amplifiers could hardly manage 20 kHz at full power.

The new 2N3055 (epi) base transistors are better than the old RCA devices, I have to say, and can double the frequency response of most amplifier designs (with changes to some capacitors perhaps). I have heard a clear distinction between even a Quad 303 design using hometaxial's (30kHz response if I recall correctly) and ("Quad Mod") using epi's (70kHz at least on the bench).

This does not mean to say that a good amplifier could not have been designed for hifi performance using slow transistors. It just happens that most circuits based on typical component values that I have seen have had inadequate margins to be able to drive the output transistors once the gain falls at high frequencies. For example, take the 2N3772 transistor. With a minimum fT of 200kHz, to design an amplifier with 20 kHz response means having to provide one-tenth the output current (at least) even if the static gain is higher (20 or so), and providing a means of extracting base current in reverse. This can be done, BTW, using, say a low value resistor on the base (e.g. 10 ohms) but this again increases the demands on the driver, and most circuits used 100 ohms typically (Quad excepted).

It is possible that the 40411 had a higher frequency response in its selection criteria, but I suspect not.

John
 
I'd be very wary of those 300 W per channel ratings, I'd expect a pair of 3773's would be comfortable with around 150 a channel. Push them much past this and I'd expect the 3773's to become unreliable or fail.
It also appears the drivers are not on the heatsink, expect considerable crossover distortion as it will have been underbiased to prevent thermal runaway.
Looks like another board designed by someone with little audio knowledge.
 
The amp is mostly hyped. The 2N3773 MAY be OK today if it is built using an epitaxial technology, but the original would not be ... only 200kHz min. fT.
Apart from that, the power supply rails are only 25-35V, which are not going to give much power output but probably won't be a problem for a single pair of 2N3773's.
150W into 4 ohms needs 35V - so at best it would just squeeze 150W per channel at 35v where it would be clipping- so the 300W means both channels at full power into 4 ohms, and it does not say what the distortion is at 300W. Clipped power will be around 10% or more.
Without seeing detailed specs this seems more hot air than hot sound.

John
 
I have "full power" tested the CE1704 into both 8r0 test load and 4r0 test load.
The CE1704 is a quasi complementary single pair output stage using toshiba 2n3773.
It measured 110W into 8r0 and 200W into 4r0, when powered by a 35+35Vac transformer on a UK mains supply.
 
Thanks experts!

I had the same feeling. So which amplifier kit would you suggest me to go for among the entire lot? Power Amplifier

I have the Focal K2 Power 165 KR2 speakers which I'm intending to connect in series to give a 4 ohms impedance. 2 way speakers kit 165 KR2 - Focal Car Audio drivers

Please let me know if there are any other suitable amps as well? How would you rate the 'honey badger amp'?

Nitin
 
Hi Andrew
I agree that if you use 35V AC transformer then the power output should be achieved, but I still would be concerned with a single pair of 2N3773's - and AFAIK TOshiba no longer manufacture this, but devices from ON semi still seem to state 200kHz min, ft, even though they may be built on a 4 MHz process. I'd at least measure the ft before using a 3773 in an amp, but there again, it woudl be worth considering MJ21193/4 instead.

John
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.