@anatech - I see what you are saying and really can't argue with your sentiments as applied to the markets you speak of. When I heard "industry" my mind automatically went to those that I am more familiar with, primarily defense related. There, for the most part, the reliability and durability is there - and paid for very, very well. A business in the telecom industry, even the larger corporations, to my mind still falls mostly under the "consumer" umbrella - they reflect the market that they serve.. I really should have thought that through a little more and I would have realized the difference in emphasis.
I will agree with you on the general quality of most available goods, electronic or not, - and the country itself for that matter - has gone to absolute sh!t. And I fully feel that greed is, to put it in one word, what had gotten us to this place - and that we all bear responsibility in a personal way; I don't want to but if I am honest with my self it is true - after all, no single raindrop truly feels it is responsible for the flood. Didn't have to buy that iPhone but I did. Enough of that...
While I say metrology is my hobby that is only as a side effect in truth. What I actually do is restore vintage test equipment, primarily HP. So yes, I am indeed very familiar with something that is truly designed with elegance, skill and a thought to outlast the designer. I learn so much from studying the way the past masters have worked their craft and am constantly in awe of the genius and personal pride that went into most of anything made by someone like HP or Tek.
Hal
I will agree with you on the general quality of most available goods, electronic or not, - and the country itself for that matter - has gone to absolute sh!t. And I fully feel that greed is, to put it in one word, what had gotten us to this place - and that we all bear responsibility in a personal way; I don't want to but if I am honest with my self it is true - after all, no single raindrop truly feels it is responsible for the flood. Didn't have to buy that iPhone but I did. Enough of that...
While I say metrology is my hobby that is only as a side effect in truth. What I actually do is restore vintage test equipment, primarily HP. So yes, I am indeed very familiar with something that is truly designed with elegance, skill and a thought to outlast the designer. I learn so much from studying the way the past masters have worked their craft and am constantly in awe of the genius and personal pride that went into most of anything made by someone like HP or Tek.
Hal
In Spock’s voice:
”I am endeavoring, madam, to construct a high fidelity audio amplifier, using stone knives and bear skins.”
”I am endeavoring, madam, to construct a high fidelity audio amplifier, using stone knives and bear skins.”
This is my take... you may disagree with me completely...
You will never run 2A through each 2N3442, the Vce will be much higher than 4V DC. As a result, the real-life current gain-bandwith will be much higher than 80kHz. I suppose what I'm saying here is to give it a go with 3442's and then listen (and measure the response if you really have to, for peace of mind). Later when you get an itch, replace them all with something better. But be warned... using a much faster output device will force you to start upgrading the rest of your system... every single component 🙂 This might be good, might be bad... depending on what you want to achieve and how you are planning to get there (at what cost, and how quickly...)
I believe that in the ordinary run-of-the-mill systems, the 3442s may sound really nice.
Waver: I am not saying that your system is an ordinary run-of-the-mill system 🙂
5200/1943 pairs, good quality are in the 150 Rupees range here, less than $2.
Populated stereo board with 4 transistors and heat sink, and other parts, is about 550 Rupees / $7, so I feel there is no need to use multiple cases and heat sinks, use a tried and trusted, and possibly better performing option.
Populated stereo board with 4 transistors and heat sink, and other parts, is about 550 Rupees / $7, so I feel there is no need to use multiple cases and heat sinks, use a tried and trusted, and possibly better performing option.
Around here any place where you got $2 C5200’s you’d get fakes. 2N3055 (or worse) inside.
At one time that was the regular price. From Digikey.
At one time that was the regular price. From Digikey.
CDIL makes them, and maybe my dealer has NOS from the original makers.
My dealer's shop is old, he is the third generation to run the business, and can get good equivalents, and is honest enough to tel me if good or bad stuff.
Much of the manufacturing is in the Far East, and shipping to the USA, and particularly handling are expensive in high income countries.
That is part of the reason parts are expensive for you in small quantities.
My dealer's shop is old, he is the third generation to run the business, and can get good equivalents, and is honest enough to tel me if good or bad stuff.
Much of the manufacturing is in the Far East, and shipping to the USA, and particularly handling are expensive in high income countries.
That is part of the reason parts are expensive for you in small quantities.
Last edited:
I'm not agree with you. Yes those aren't original 2sc5200 but quite close to a genuine one & definitely much better than venerable 2N3055. Here in india most PA operators uses such transistors in their 'low cost' amplifiers with satisfactory performance. I'm attaching photo of such c5200 which costs less than $2 a pair if you buy in large quantities.
'Extreme', the quality is good if he says so.
CDIL has been around since 1964, they make military grade parts also, just because they are not well known in your area does not make them poor quality. That is just an example. No ties to CDIL.
The volumes in some components far exceed the volumes made by the original maker, and in some cases, like the 2030, they are not made any more by the original maker, so I have to use good copies.
CDIL has been around since 1964, they make military grade parts also, just because they are not well known in your area does not make them poor quality. That is just an example. No ties to CDIL.
The volumes in some components far exceed the volumes made by the original maker, and in some cases, like the 2030, they are not made any more by the original maker, so I have to use good copies.
Frankly speaking i don't like CDIL made transistors, instead i'll use KEC or UTC made components. CDIL is good if you don't have other options. In semiconductor manufacturing india is still taking baby step despite pioneering on complex technologies. I hope in near future situation will change.
Thank you very much for your encouragement, Extreme_Boky 😊!
What frontend would you suggest me to apply? Should I go for something more recent, say Valery's (R.I.P) Simpelstark or similar? Or should I stay with a contemporary RCA Power Transistor Manual design from 1970? If so, I'd be quite puzzled: There's a complementary amplifier with the CA3140A opamp at the input, and a quasi comp one with the CA3100A. The latter one also is shown in my BD550 data and application sheet from 1978, using 18 (!) power devices. Apart from that I've got the impression that the CA3100A always has been as unobtainable as the BD550's, at least here, why might RCA have made that decision? What's the difference between those two opamps? As it happens, there are some 3140's laying in a box here 😉 ...
Best regards!
See post #306... I was perfectly happy with that amp for quite a while, at a very young age... here it is again:
Then I discovered a Bible... "Audio Power Amplifiers Design Handbook" by Douglas Self. I was reading this book, thinking about it, reflecting, and going back to read it again and again... for quite some time.... around one whole year. In the end, I built the blameless amp and it sounded very clean, very accurate. I loved it.
Then, the Hiraga 30W amp came along... beautiful:
Then I experimented... along the lines of... less is more... and started reducing the number of gain stages in an amp... and decided to find an amp that can provide all the required gain with only 2 transistor gain stages - and I settled at Aleph J. To be honest, Aleph J2 is the true, real 2-gain-stage amplifier deal... this is why I am planning to build it as my next amp.
Good luck with whatever you decide to do... but if you haven't read the Douglas Self book... well you can finish the sentence 🙂
Then I discovered a Bible... "Audio Power Amplifiers Design Handbook" by Douglas Self. I was reading this book, thinking about it, reflecting, and going back to read it again and again... for quite some time.... around one whole year. In the end, I built the blameless amp and it sounded very clean, very accurate. I loved it.
Then, the Hiraga 30W amp came along... beautiful:
Then I experimented... along the lines of... less is more... and started reducing the number of gain stages in an amp... and decided to find an amp that can provide all the required gain with only 2 transistor gain stages - and I settled at Aleph J. To be honest, Aleph J2 is the true, real 2-gain-stage amplifier deal... this is why I am planning to build it as my next amp.
Good luck with whatever you decide to do... but if you haven't read the Douglas Self book... well you can finish the sentence 🙂
Last edited:
As already said, I've also built about a dozen of those Siemens amplifiers in #306 😉. There are two complains, though, why I wouldn't chose this design again with my 2N3442's: Siemens suggests BD139/140's as VAS and drivers, which indeed are good choices, but would limit me to rail voltages of just ±40 Vdc, where four pairs of power devices are well overengineered. Then, what's the purpose of that 8V2 zener in the LTP's tail? Having been literally clueless in my youth, I copied the design as it was. But nowadays I would increase the tail resistor value considerably and tie it directly to the rail, or better opt for a CCS.
Best regards!
Best regards!
I am not sure if you are asking what's the purpose of that 8.2V Zenner...? If you are -> It sets the current through input differential pair - it's like a voltage regulator that sets the voltage at the junction of R6&R7 ... You could decouple that junction point back to the positive voltage rail (Ub) with something like a 4.7uF capacitor.... It also sets the Vbe voltage for that overheat-sensing (via NTC1) & protection T1 transistor. The amplifier that I built did not have T1... I wouldn't use it anyway.
Last edited:
Yes, I asked exactly this one. Anyway, the LTP precision/balance would be increased dramatically with an increased common emitter resistor value. I definitively prefer a CCS in this role (but have to consider that back in 1971 even small signal transistors were rather expensive...).
Btw, I also omitted the temperature dependant clamp transistor right at the input in all of my builds.
In the meantime I found severe diferences between the CA3100A and CA3140A opamps: While the 3140 has CMOS inputs, the 3100 is almost nine times as fast (38 Mhz vs. 4.5 MHz gbw product). But does this explain why RCA chose the 3100 in the quasicomp design and the 3140 in the complementary one?
Best regards!
Btw, I also omitted the temperature dependant clamp transistor right at the input in all of my builds.
In the meantime I found severe diferences between the CA3100A and CA3140A opamps: While the 3140 has CMOS inputs, the 3100 is almost nine times as fast (38 Mhz vs. 4.5 MHz gbw product). But does this explain why RCA chose the 3100 in the quasicomp design and the 3140 in the complementary one?
Best regards!
For a front end with those 3442’s I wouldn’t do anything “special”. A PNP diff pair with CCS, followed by a video amplifier transistor VAS. Personally, I prefer CCS loads on those to a bootstrap. 6 mA. Then a triple QC following the BD 550 application note. +/-70V, 4 or 5 outputs in parallel. Faster drivers recommended to suck charge out of the output bank during crossover. To get lower distortion, make the “input” stage a single, apply local feedback from the output to emitter to give a gain of about 60, and use an LF353 op amp front end. That makes the Crown/Peavey circuit out of it. You’d be surprised how well it works, even with the slowest output transistors.
Why the LF353? It is IMMUNE to blowout problems if/when you throw DC to the output. As long as your feedback resistor limits fault current to 10 mA. I like to limit to 4mA for a margin of safety. I’ve tried a LOT of “better” op amps, often with disastrous results during heavy clipping or output faults.
Why the LF353? It is IMMUNE to blowout problems if/when you throw DC to the output. As long as your feedback resistor limits fault current to 10 mA. I like to limit to 4mA for a margin of safety. I’ve tried a LOT of “better” op amps, often with disastrous results during heavy clipping or output faults.
If it SAYS Toshiba on it, and it’s NOT, it’s a fake. And I wouldn’t trust it. Made by CDIL or UTC and marked as such is different.
I agree completely. If the part is not what it is marked as, you have no idea what it really is. In that case, the reality will not be good or they would leave it marked as it should be.
Anything that is not as it is marked, I destroy and discard. It may cost me how much I paid for them, but they will cost a lot more if I use them! I still get fakes from "trusted" sources occasionally.
With JIS standards, each part number is made by one company only. Unlike in North America (JEDEC) where several companies can make a part number. However, parts with a certain part number must adhere to the set standard. You got to know what to stay away from. Remember the TI TO-220 parts? Junk, and they looked like it.
Anything that is not as it is marked, I destroy and discard. It may cost me how much I paid for them, but they will cost a lot more if I use them! I still get fakes from "trusted" sources occasionally.
With JIS standards, each part number is made by one company only. Unlike in North America (JEDEC) where several companies can make a part number. However, parts with a certain part number must adhere to the set standard. You got to know what to stay away from. Remember the TI TO-220 parts? Junk, and they looked like it.