That was supposed to be the minimal rating but there was differences from a manufacturer to another, FI RCA ones where said to support roughly 80V VCE, while the aluminium cased Motorola could eventually take more than 60V they also had the advantage of a 2.5MHz Ft instead of the usually minimaly specified 0.8MHz.
Those are odd indeed. They're large enough to not affect operation at steady-state, so I wonder if they do something during startup or shutdown. One could set up a simulation and find out.
For the B&O example the extra resistor could be to deal with the case where the bias pot wiper fails open. That doesn't seem to be case for the Wega though. The bias trimpots in the B&O were traditional carbon track, open frame types. I've replaced a couple of bias spreader transistors in those when I was in my teens.
Tom
At first look that seems to stabilise the quiescent current of the output stage.
Since the VAS current vary a lot with supply voltage due to the boostrap resistive load the voltage across the VBM will increase if supply voltage increase, in this case this resistor add more biaising to the VBEM base when the supply volltage and inversely.
Hi wahab,
The CCS tail current for a diff pair was well known as a best practice even back then. It comes down to the fact that a resistor is less expensive than all the parts used to make a CCS. The cost of a signal transistor wasn't killer back then unless you are talking about a cheap product.
Once manufacturers began looking at performance, late 70's and early 80's, you saw circuit design get serious. Otherwise it is merely a cost issue, including the cost of stuffing the PCB.
The CCS tail current for a diff pair was well known as a best practice even back then. It comes down to the fact that a resistor is less expensive than all the parts used to make a CCS. The cost of a signal transistor wasn't killer back then unless you are talking about a cheap product.
Once manufacturers began looking at performance, late 70's and early 80's, you saw circuit design get serious. Otherwise it is merely a cost issue, including the cost of stuffing the PCB.
Bootstraps were everywhere because a good PNP “VAS transistor” was often difficult to get - lots of good choices for NPN. Especially for a scavenger looking for video output transistors to use. That forced you into a PNP input stage but at least there was the 2N5087. 2SA720 or 872 were often used, but the latter was one of those types that should be avoided/replaced due to long term reliability issues. Biggest issue with a bootstrap is the quiescent current coming down as supply drops under load. That will force your diff pair out of balance, assuming it was balanced for full rail (Back in those days current mirrors were unheard of). Singleton input stages tend to sound better, when used with a bootstraped VAS.
Supply voltage limit of 80V was also imposed by the 2N3055. Only problem with BD139 was it was “bigger than it needs to be” for a VAS. If you wanted to get around both limitations you went for a video output type. They STILL call the C3503 a video output transistor - as well as its PNP. Those are the go-to VAS‘s for high performance amps. They take 300 volts and they’re not bigger than they need to be, which means less of your Cdom is stuck inside the transistor and dependent on VCE.
Well, the 2SA872A has zero reliability issues, and it is quiet. 2SA970 is also a good choice along with a few others.
Agree on the 2SC3503 and 2SA1381. I still have some originals. Often you could go with a TO-92M package, but the TO-126 packages are better for replacement when they run them hot in a design.
Agree on the 2SC3503 and 2SA1381. I still have some originals. Often you could go with a TO-92M package, but the TO-126 packages are better for replacement when they run them hot in a design.
Ahh, 2SC872 or 2SC872A?
I never did stock the 2SC872, but ordered literally hundreds of 2SA872A at a time, along with other signal transistors of the day. I never did see a noisy 2SC872, but many other earlier numbers tended to go that way. Generally in circuits where Vce was close to the rated on that part.
I think if you push the Vce rating on any low noise transistor, you're going to end up with more going noisy. Harmon Kardon was bad for doing that, Pioneer as well.
I never did stock the 2SC872, but ordered literally hundreds of 2SA872A at a time, along with other signal transistors of the day. I never did see a noisy 2SC872, but many other earlier numbers tended to go that way. Generally in circuits where Vce was close to the rated on that part.
I think if you push the Vce rating on any low noise transistor, you're going to end up with more going noisy. Harmon Kardon was bad for doing that, Pioneer as well.
A872 (were talking about PNP input stages). I still have a few pulls laying around. I wonder how noisy they test now, all these years later. They may be the pair I pulled out of Tony’s old 70 WPC Realistic amp he was DJing with back in 81. When that diff pair went crazy in the middle of a gig it damn near launched the woofer cones across the room. Thunder and lightning at full volume. Ran across several more doing the same thing but never quite that loud.
Hi wg_ski,
lol!
Not wanting to get too far off-topic ... but! Yeah, when the diff pair gets sick, and there isn't any protection for the speakers, you have real trouble.
I have seen fires in speakers back then (and in amp cases).
It would be really interesting to see how they test. Processes have come a long way, as well as chemistry and methods. We are sooo spoiled today (compared to what we had in the 80's - and that was way better than the 60's). I made it a habit to discard anything questionable. If you don't it will end up somewhere that you have to troubleshoot all over again. I'm talking personal projects, not customer units of course.
lol!
Not wanting to get too far off-topic ... but! Yeah, when the diff pair gets sick, and there isn't any protection for the speakers, you have real trouble.
I have seen fires in speakers back then (and in amp cases).
It would be really interesting to see how they test. Processes have come a long way, as well as chemistry and methods. We are sooo spoiled today (compared to what we had in the 80's - and that was way better than the 60's). I made it a habit to discard anything questionable. If you don't it will end up somewhere that you have to troubleshoot all over again. I'm talking personal projects, not customer units of course.
Next were for us Europeans the BF469/470/471/472 family of CRT driver transistors, and, in another case, the BF869 etc.
Previously there had been video transistors yet, of course (BF457/458/459 e.g.), but, as you said, only NPN ones.
Best regards!
Those transistors came quite lately in the DIY market, AFAIR best devices were the BC141/BC161 pair wich is unfortunately limited to 60V but is much better than the BD139/140 as VAS, the BCs had a 100V version under the form of the BD529/BD530 (6-12pF) wich allowed using +-40V supplies safely, for high power quasi complementary output stages 2N3055 were replaced by BDY56 wich is 150 V/15 A/115 W/10 MHz.