As a triple perhaps ? I think SOA is the real doubt . As a low frequency class D is nice . Does anyone have a circuit ? I remember one with NE 555 years ago that was OK for a sub woofer .
There must be loads of transistors that have unusual second lives . For example many TO 126 have great Rbb' . SMD also . I seem to remember BD 135/6 does ?
Nigel,
What you just put up was exactly what I was thinking would be a great format for listing the information that individuals are providing. If we could create a posting like this where information is added by cutting and pasting this information we could generate a very useful comparative chart of devices all in one place.
Thanks for that post.
What you just put up was exactly what I was thinking would be a great format for listing the information that individuals are providing. If we could create a posting like this where information is added by cutting and pasting this information we could generate a very useful comparative chart of devices all in one place.
Thanks for that post.
Nigel, i seldom use long tail pairs. My speciality is low noise, high speed anyway so i more or less focus on phonostages and linestages, also buffers and active crossovers.
Well, that is not 100% true. I just designed an MC input i called Abraxas. That is a parallel-symmetric-differential design so it uses complimentary long tail pairs with a twist.
My typical design is parallel symmetric, currentmirrors, transimpedance RIAA, buffer.
BJt and JFet have both their advantages. BJTs are lower noise in the bass because they have a lower 1/f frequency then J-Fets. They need only 1/10th the idle then J-Fets for the same nV/qHz spec. Problem can be how to generate the bias. It should not add aditional noise and it it difficult to DC couple ( in a parralel symmetric design, no problem with double diffential but with a 3dB noise disadvantage becaue the trannies in the long tail pair are in series with the noise ).
The base current is an error source but i found ways to " recycle" it.
J-Fets on the other hand are self conducting and high impedance at the input so bias and DC coupling is a breeze. They also "sound" a bit warmer and fuller then BJts but PRAT factor has to be watched or they sound less dynamic. They also have less problems with radio ingress. BJTs can have that proble because you could see the base-emitter diode as a detector....
Well, that is not 100% true. I just designed an MC input i called Abraxas. That is a parallel-symmetric-differential design so it uses complimentary long tail pairs with a twist.
My typical design is parallel symmetric, currentmirrors, transimpedance RIAA, buffer.
BJt and JFet have both their advantages. BJTs are lower noise in the bass because they have a lower 1/f frequency then J-Fets. They need only 1/10th the idle then J-Fets for the same nV/qHz spec. Problem can be how to generate the bias. It should not add aditional noise and it it difficult to DC couple ( in a parralel symmetric design, no problem with double diffential but with a 3dB noise disadvantage becaue the trannies in the long tail pair are in series with the noise ).
The base current is an error source but i found ways to " recycle" it.
J-Fets on the other hand are self conducting and high impedance at the input so bias and DC coupling is a breeze. They also "sound" a bit warmer and fuller then BJts but PRAT factor has to be watched or they sound less dynamic. They also have less problems with radio ingress. BJTs can have that proble because you could see the base-emitter diode as a detector....
Last edited:
I bought a pair of 1980 Maplin 50 watt amplifiers with transistors missing.
What I couldnt get at RS or Farnell I surprisingly found on ebay at sensible prices.
What I couldnt get at RS or Farnell I surprisingly found on ebay at sensible prices.
Hi Steven . I was almost overwhelmed 5 minutes into the project ! It is a good point . I wonder as Bob would be doing it anyway would he occasionally list the progress and possibly reject one or two if they are just NOS on eBay for examples found ? A NOS list might also be useful .
An odd twist to the story . I fixed a mans Dynaco amp years ago , Charles Thompson . We became friends . He built it with the help of his wife Jane whilst at Harvard on the kitchen table ( ST 70 ) . A year or so later he asked me to help his students do a project on the US second line suppliers of tubes . Sylvania was the number one of that group . I reported to Charles much like Dr Beeching reported to the British government about our railways in 1965 . Like Dr B I said two possibilities . No hope and Bob Hope . I had no idea what it was about . My report apparently closed the industry . I had no idea who Charles was , Harvard business school . Dr B was my grandfathers cousin . Reading about him recently he probably saved the railways . The reality would have become worse if nothing had been done , far worse was planned .
Actually, Nige, I do believe an off topic excursion can be very valuable here and there, kind of add some background to the basic thread topic. The trick is to keep it within reason.
I have never used parts from General Semiconductor, but they apparently make many "discontinued" devices. Anyone have any experience with their parts?
The best ever was a friend who must remain nameless saying any part could be cloned by her friends at 1 euro a piece . I said who are they ? Military . When I said but who is really paying the answer was the EEC . I was tempted . No idea if they really could do that ? It sounds like it would cost XXXXXXXXX ?
Dvv . Tube or transistors we don't want them to go .
Dvv . Tube or transistors we don't want them to go .
God forbid that anything should go, too much has gone already.
I'd like to comment on a pair of transistors somehow "in the shade". These would be BF 720/721, manufactured by Siemens, Philips and probably someone else as well.
Let me start from the key turnoff - unfortunately, they are available only in SOT-223 package, which is a drag, but not a critical one. With some care, all should be well, but you would do well to remember its package when planning to use it at some power level.
In my experience, it is an almost ideal cascode transistor - all it needs is a nice pass transistor "above" it and you have a great cascode stage. A particularly good match are the venerable Motorola MPSA 56/06- the end result is a clean and clear stage, which works very well indeed over a wide range of currents and voltages.
I have also used with it many of the classics - BC 550B/560B, BC 546B/556B, 2N 5551/5401, and so forth, and it always worked well. I may be imagining things, but I feel it works best with a pass transistor capable of above average current (for its packaging) - another match made in Heaven would using it with BC 639/640.
I often make SMD into a usable package . Solder on legs and heat shrink with 6.3 mm 3:1 .
It is surprising how well it works and stops it breaking .Easier than potting , squeeze it immediately for an almost potted look and cut to make it look neat . I sometimes use LM 285 as a voltage reference using the same method . It can be a 20 PPM device which as far as I know has no legged equivalent .
BC 639/640 I always suspected was BD139/140 repackaged .
It is surprising how well it works and stops it breaking .Easier than potting , squeeze it immediately for an almost potted look and cut to make it look neat . I sometimes use LM 285 as a voltage reference using the same method . It can be a 20 PPM device which as far as I know has no legged equivalent .
BC 639/640 I always suspected was BD139/140 repackaged .
- Status
- Not open for further replies.