RCA transistor app notes / amplifier design

[Lucien Nunes]

I'm trying to find an RCA application note or section in a handbook dating from the late 60's / early 70's that includes their original design for a quasi-complementary audio amp, based on the following transistors:
40408 pre driver
40595 & 40596 drivers ?
2x 40636 outputs
The circuit was used in various forms in many amps, especially the Harman Kardon Citation 12. A friend is repairing a fried amp in a Compton organ, which I am expecting to be a dead ringer for the original RCA design.

I'll happily reimburse anyone for a scan or photocopy of the schematic & words.

TIA
Lucien

[DangerMan]

Hi Lucien,
Try here, and scroll down:
http://www.passdiy.com/pdf/citation.pdf
Is that it? It has a differential input pair before the 40408 and looks a distinct possibility; but it's only the circuit diagram.
TTFN
Pete

[Lucien Nunes]

Thanks for the link Pete, it looks potentially similar, but the Compton example isn't currently here to compare. I know we have a few more transistors, I suspect for protection circuitry, what I'm not sure about is what the differences were between the Citation and the original RCA design. The aim of the comparison is to find out whether it is a slavish copy or whether perhaps there was some creative input.

The driver transistor numbers didn't quite sound right when I typed them from memory hence the question mark; as per the Citation schematic they are indeed 40594 & 40595

Cheers
Lucien

[ppppenguin]

I have an RCA power devices data book from 1976. Most of the types mentioned are not in the book. There are no app notes except for a series of abstracts. None of these specifically relate to audio amps.

You are welcome to borrow the book if you think it may be of any use.

[Lucien Nunes]

I'm surprised that the transistors were already discontinued by 1976 but the design would have been obsolescent. I think the venerable boxes full of 2N3055s or 2N3773s were giving way to fully complementary circuits by then.

Actually the 40636 might have been dropped from production as the Hometaxial process was phased out (anybody know when that was?) because it was a 2N3055 selected for high Vceo. The Hometaxials had 100% SOA and the early epitaxials probably couldn't meet the spec of the 40636 at high voltages.

There were other selected '3055s, including the ones used in Leak amps of the era. Many moons ago I was in the workshop at Gerry's, fixing a Leak Stereo 70, puzzling over the transistor numbers 39250, 39251 & 39252. Not being ordinary RCA 40xxx numbers I couldn't look them up in my data book. I was toying with the idea of 2N3773s while being a bit concerned about the rail voltage, when a chap came in and started looking around. The dialogue went something like this:

Visitor: A Leak!
Me: A dead Leak, wish I knew what these transistors were
Visitor: 39250 is a 2N4036, 39252 is a 2N2102, 39251 is a selected 2N3055
Me: That's jolly clever, you must be an audio amp expert then...
Visitor: Hello, I'm Ray Whitcombe
Me: Er, oh...

Lucien

[neonlamp]

Hi Lucien
For some reason best known to themselves, RCA re-named their range. The 40594 became an RCA-1A03, the 40595 the RCA-1A04. That was in their 1978 data book. My view, they had a well-known "brand" name with the 40,000 series that I think this was an error of judgement. The first 70W circuit using the lineup 40406, 40407, 40408, 40409, 40410 and a pair of 40411's was pretty dire. Luckily this was updated to the differential set using an input pair RCA-1A02 (40406) RCA-1A17 VAS/pre-driver (40408) and driver/outputs as you mentioned execpt the 40636 became the RCA-1B02. Two other trannys were used for protection, and several diodes for a bias string, protection, . Just like their 50W based on the 2N3055 only more so.
You could probably substitute a BD139-BD140 pair for the drivers, and MJ15003's for the output or use a complementary PNP MJ15004. But you might have to add a couple of 100pF caps to the driver bases and appropriate rail to stop sprogging as these are faster transistors than the 40636's

John

[Lucien Nunes]

Thanks for the insight, John, I've never had cause to investigate those late RCA numbers and hadn't realised they were exactly the same parts rebadged. However I now have in front of me an identical amplifier from one of my own organs (which works), apparently dated 1972. Although my friend will probably have repaired his amp by now I would still like to compare with the RCA design, so I'll draw and post the schematic. Incidentally the organ has four of these 70W channels, two for treble and two for bass, so even if it is not yet working they should be OK for the service on Sunday albeit at reduced volume.

The MJ15003 sounds like a good choice, I have used it before as a get-out-of-jail card for this sort of thing. However if it were my own amp under repair, I'd avoid the BD139/140 in favour of something in a TO-5 simply to maintain the original appearance. 40594 and 40595 appear to map onto 2N5320 and 2N5322 respectively - were they selected or just renumbered to fit in with the range?

Lucien

[jimmc101]

Now that starts to make sense, I have a 1968 'RCA Technical Presentation' titled
'Characteristics and Applications of RCA Silicon Power Transistors' which has a single circuit for 12, 25, 40 & 70W amplifiers.
The 70W version of the circuit uses 2N5320, 2N5322 and 2N3055s.

I will try to scan and post it tomorrow (Oops it's later today now).

I also have the '72 Power Transistor manual it has the 40594, 40595 & 40636 listed in the 'Special Audio Power Transistors' section.
The 2N5320 & 2N5322 are listed as 'general purpose'.
There are no amplifier circuits in the book.

Jim

[neonlamp]

Hi Lucien

Yes the 40594, 40595 appear to be selected versions of the 2N5320/2N5322. Probably these would work, but the selection was for 95V Vcer instead of 90V.

If the equipment uses the old RCA spec. 42V rails the 2N5320/22 will be OK (up to 45V) but I guess the additional margin could give lower distortion (better Early voltage).

John

[jimmc101]

Scan as promised. Sorry for missing edges, the bindings a bit tight.

Jim

[Lucien Nunes]

Excellent! Thanks Jim, that must be the design I was thinking of. I'll trace out the Compton one as soon as I've cleared up my house-bashing project for the night.

But first - can anybody spot the howler in the schematic?

Lucien

[Lucien Nunes]

OK, can anybody spot the two howlers in the schematic? Whoever drew these was having a bad hair day. There's also an error in fig. 24, although that gets right one of the errors in fig. 19.

Lucien

[neonlamp]

Hi all
Yes, this is the one. I recall Mainline selling 12, 25, 40 and 70 W amplifier kits in the 70's, which I take it are transistor and diode sets for these amps, but I never bought one.

The bootstrap and bias diode connections are nearly right. Just wrong.
I would have said that the drivers and pre-driver would be chosen from the 2N2102 series (40361, 40635 for the lower powers), because the 2N3053 is essentially a 40V transistor.
Some manufacturers still sell a 2N3019 (npn) and 2N4033 (pnp) which are TO-5 cans rated at 80V so may be OK for anyone wanting to build a version up to 50W.
MOdern 2N3055's from ST and ON Semi are faster than RCA originals, so the circuit may need sprog stoppers. On the positive side, because of the way in which epi transistors break down compared with the originals, it is likely that most 2N3055's will break down at well over 80V and be suitable for the 70W amp. Not because epi is inherently better, (it may not be) but because they have to make the transistor effectively a higher voltage to achieve the 100V BVcbo rating. Chances are fair, but check your trannys first.

[Lucien Nunes]

Hah, surprise, the Compton amp is almost identical, I was almost expecting to find the mistakes in it too! I have an additional 1nF from Q3's collector to ground, and different values at the input filter, otherwise it is a carbon copy.
It has its own minor mistake, at the PCB edge connector the tab for Q6's base is mislabelled 'Q6 collector' in the silkscreen.

For the limiter it uses MPS6530 & MPS6533, for the long tailed pair there are TO-105 packaged devices by SGS although the number means nothing to me: U19482/187045. All diodes are 1N3754s except D10 & D11 which are 1N3193.

Compton were prolific pipe organ builders who became deeply involved with electric organs, having gained early success with their Melotone electrostatic add-in tone-generator for cinema organs. Like most pipe organ builders they were extremely conservative, which in the case of electronics meant using existing designs long past their technical obsolescence. This version of amp replaced a 40W transformer-coupled job, and at the time they folded in the early 70s it was the most modern piece of technology in the instrument. Because they copied it from the RCA book!

Thanks to everyone who has helped out with this, Here are a few pics of the instrument and my amplifier chassis from a similar one.

Lucien

[alayn91]

Hello,

The HK Citation_12 service manual is here:
- http://manuals.harman.com/hk/Service...WELVE%20sm.pdf

Regards.
Alain.

[Lucien Nunes]

Thanks for the link - the HK circuit clearly stems from the same source but has concessions to its hifi application. It wouldn't do, surely, for a hifi amp not to have adjustable Iq, and they've shunned the limiter circuit because nobody would miswire or short out the speakers in their living room, would they? Unlike Compton, they felt the need to adapt the circuit to suit.

Lucien

[peetleetech]

I know that book. I built a few of it's examples. I am peetleetech (Droopy) and if I can find the notebook into which I copied a few drawing before passing the book on to my(now deceased dad) I can help you. It is a very straight forward quasi complmentry circuit. RCA does a sneaky thing sometimes, so beware of pins marked e and should be c.
Call you back when I've found it. I always liked the 40636, beautiful and honest. The 30 amp version, 40411 was better and interchangeable.

[kiewavalley]

Lucien

Is Ray Whitcombe on this forum?
How can I contact him via email?
The contact details I have for him date back to 1999.
At that time he was assisting me with the LEAK history.....

Regards
Steve Spicer

Originally Posted by Lucien Nunes

Visitor: A Leak!
Me: A dead Leak, wish I knew what these transistors were
Visitor: 39250 is a 2N4036, 39252 is a 2N2102, 39251 is a selected 2N3055
Me: That's jolly clever, you must be an audio amp expert then...
Visitor: Hello, I'm Ray Whitcombe
Me: Er, oh...

Lucien

[Lucien Nunes]

I don't know, I had thought so but can't find the post that made me think that. Perhaps the mods can help?

Lucien