RCA MI-3523 mic pre for a project

[Jim Zuehsow]

I put some pics over on Coppermine of the RCA MI-3523 mic pre. Looks like it could be a good project since it doesn't require exotic transformers like the Altec and Langevin do. Specs at 44.5 db gain and +24 out. It seems this unit has been overlooked as you always see schematics for the old RCA tube and early transistor preamps. Pictures include what it looks like, the schematic, and the pc board layout. Check it out.
Here is shortcut to the pictures:

http://www.twin-x.com/groupdiy/index.php?cat=10094

 

[mikka]

Could you link please ..... for infidels?

 

[Jim Zuehsow]

Here is the link:


http://www.twin-x.com/groupdiy/index.php?cat=10094

 

[emrr]

Any more frame of reference? Where'd you find this? What was it in? etc, etc. That's a new one to me.

The Altec 9470A transformers are no more specialized than these would be.

 

[Jim Zuehsow]

I have a rack full of the Altec 9470s and 4 of these racked up with pads, phantom and phase reverse They sound very good. What I meant about the "exotic" transformers is that the Altec are "astaticly" wound, with very tight balance between windings as far as capacitance goes, at least that's what the old Altec literature says. Not something single-ended transformers are known for. You could use an old A series UTC and it would work fine. I did build a couple of these using Hammond transformers, and they sound identical to the RCA units with the Microtran iron. By the way, the output transformer is 2400 ohms CT to 600 ohms.
These were meant as a replacement upgrade for older RCA amps in consoles and also for custom installations. I believe the rack would hold 8 of them plus a power supply. This info came from an old AES preprint I had laying around.

 

[emrr]

They look like direct plug-in replacements for BA-71 or BA-72 amps, but the time frame is curious to me. They have the 71 type open frame, but I'd expect a package like the later 72 enclosed box. I'd guess they have to be at least 1972 or later. First RCA piece I've seen with Microtran iron also.

There's another mid '70's RCA custom console series that uses the API 2520 opamp with a UTC A-24 as input transformer.

 

[skipwave]

This one looks unique to me. Fully differential, all PNP, interesting biasing arrangement for the output pair.

Do you happen to have one of those wonderfully thorough circuit descriptions that RCA did for each of their models?

Thanks for sharing.

 

[NewYorkDave]

all PNP

Very common back in the bad old days of germanium, even though these are silicon. Maybe these were intended as a plug-in upgrade to replace older amps that used the noisy leaky germaniums (germania? :wink.

interesting biasing arrangement for the output pair.

The outputs are Darlington-connected, and direct-coupled to the input transistors, nothing unusual there. The biasing of the input transistors (taken off the output emitters via series resistors) is also conventional. Then you have another feedback loop from output collectors to input emitters. It's all one big direct-coupled circle jerk.

 

[skipwave]

[quote author="NewYorkDave"]

interesting biasing arrangement for the output pair.

The outputs are Darlington-connected, and direct-coupled to the input transistors, nothing unusual there. [/quote]

Ah ha! Oh man, shameful that I missed the darlington connection. I'm used to schemes with the modern, single package darlington shown as a single symbol.

So how does this design (and your Dynair lineamp) get away with no coupling cap between the collectors and the output transformer?

[quote author="NewYorkDave"]It's all one big direct-coupled circle jerk.[/quote]

Now that's just dirty.

 

[NewYorkDave]

The direct current in the output xfmr primary is balanced, it's a standard push-pull arrangement.

Note the big-assed collector resistors on Q1 and Q4. This was normal back when they were making direct-coupled amps using all transistors of the same polarity. The collector voltage at that point had to be fairly low to provide appropriate bias to the base of the next transistor, and the easiest (and least wasteful of power) way to do this was to make the resistor a high value. The low collector voltage limits the dynamic range of that first transistor, of course, and the high resistance contributes noise, but now I'm just nitpicking :wink:.

 

[emrr]

It is a direct plug-in replacement for the BA-71 / 72 amps.

RCA kept making the BA-73 program amp for the later SS consoles, and it was always all germanium, through at least five revisions. The BA-71A was all germ, and got changed to 1 germ by 1962. The 1969ish BA-72A revision retained the 1 germ. 72 and 73 were made right up to the end (1976-77?). That's what's curious about this thing to me; where it falls in the product line. I don't know that the 72 was ever outright replaced by anything else.

RCA stopped using UTC audio iron on the standard series at some point, and switched first to Kenyon, then to SEI audio iron on the last production I've seen. Microtran widens the mystery.

I've never seen a noisy BA-71A, and I've had six with original transistors. I've seen the original caps cause noise. Never seen a noisy BA-71B or BA-72A and I've had about 30 of them in recent years. So the noisy germanium speculation doesn't seem likely to me. Love to see more more documentation, and a catalog that has this listed with a year reference.

 

[Jim Zuehsow]

[quote author="NewYorkDave"]The direct current in the output xfmr primary is balanced, it's a standard push-pull arrangement.

Note the big-assed collector resistors on Q1 and Q4. This was normal back when they were making direct-coupled amps using all transistors of the same polarity. The collector voltage at that point had to be fairly low to provide appropriate bias to the base of the next transistor, and the easiest (and least wasteful of power) way to do this was to make the resistor a high value. The low collector voltage limits the dynamic range of that first transistor, of course, and the high resistance contributes noise, but now I'm just nitpicking :wink:.[/quote]

I just posted the RCA description of this circuit and they claim the 1 meg resistors increase the gain and lower the noise of the input stage. I tend to believe their claim since I have 4 of these and their noise and dynamic range is as good as all the other preamps I own.

 

[emrr]

Ah, looks like 1967 date codes on the transformers. Maybe they tried these and no one cared. :shock:

 

[Jim Zuehsow]

By the way, this was presented in an AES preprint at the 1968 convention, so there is the timeframe for this unit. I think they finally went into production shortly after that. They claim that the output transformer was specially designed to have very low distortion at 20 Hz.

 

[NewYorkDave]

I tend to believe their claim

You can believe both of us, since what I said doesn't contradict the claims you just paraphrased. (I haven't read the document yet).

A higher input stage gain does result in lower noise, but a larger resistor is noisier than a smaller one. That's basic physics and hard to dispute. I said the large resistor contributes some noise, I never said it contributes too much or that it outweighs the increase in gain.

A transistor running at a low collector voltage has less headroom (range between cutoff and saturation) than one running at a higher collector voltage. Again, this is fundamental and I never said the headroom in that stage was too low, just that it was lower than it would be with a higher collector voltage.

noise and dynamic range is as good as all the other preamps I own.

Cool. Will you post some measurements?

 

[hitchhiker]

Jim! Thanks for showing us this one.I'd like to mock one up. What is the input trans ratio?

Hey Doug! I recently got to fixing my ba33b amps. A couple years ago you were kind enough to email me a schematic and I thank you again. I agree that the ba-33 type are possibly the best of the RCA transistor stuff. The PWR SUP 5W 20r is running very hot .(180 degrees C) Is that normal or are my 1500 uF filter caps bad? The sound of these makes me very curious about the mi-3523 Jim has found.

cheers!
Lance

 

[emrr]

Lance,

Sounds like you need some filter caps, at least in principle. Have you checked voltages against the schematic yet? Shouldn't be that hot. I'd do the output cap also. Those are frequently bad. I definitely like the gain structure of the 33 best of all the RCA SS program amps. Yet another that is higher gain and partially germanium; haven't seen a noisy germ transistor in one yet. I think it depends on the case style of the type selected.

I bet Jim's beasts sound very different from the normal RCA SS stuff. Damnit, another one to find.

Jim,

Is there a BA-xx number assigned, or only the MI- (master index) number?

 

[Jim Zuehsow]

I believe I measured the input transformer as 10K CT. I built a couple of these using transformers from Hammonds 500 and 800 series broadcast transformers. Not quite as clean as the Microtran units RCA used.
These are very nice sounding units and compare pretty close to the Altec modules except they seem to have better high and low end body. I'm working a swing shift up at the earth station tonight, but I will post some of the waveforms and distortion plots from the article when I get home.

 

[Jim Zuehsow]

I bet Jim's beasts sound very different from the normal RCA SS stuff. Damnit, another one to find.

Jim,

Is there a BA-xx number assigned, or only the MI- (master index) number?[/quote]

It does indeed sound different, and cleaner than the "crunchy" sound I associate with some of the other RCA transistor preamps.
As far as I know, there is only the MI number with this unit. I have never seen one go on ebay, so maybe I have the only 4 ever made!
I wouldn't mind getting some more of these, if I could find them.

 

[Jim Zuehsow]

I added 2 more pages of info from the AES preprint about the preamp:

http://twin-x.com/groupdiy/displayimage.php?pos=-734

Incidently, I have the schematic pack for the ISA430 and the Focusrite 6 if anyone is interested or has some room to host them

JZ