Northern Electric preamp modules

Hello,

A friend has had these preamp modules lying around for years and hasn't been able to use them because they don't have a power supply. He's a fantastic engineer, so I thought I'd build him the supply so he could keep recording music that I like.

Here they are:



Top

And... (Promise to keep CJ away!!)

Transformer

Here are the insides:

Insides

From what I can gather, Northern Electric was Canada's branch of Western Electric. These modules are apparently out of some old console, but I can't find any information about them anywhere. The module is #R6409A. Does anyone have any clue as to where these came from?

Better yet, does anyone have a schematic?

I've been staring at these things for a while now and have managed to figure out the following:



I can't yet guarantee it's perfect, but I checked it a couple of times and I'm pretty sure it's right.

I haven't figured out the input transformer yet.

There is a component in there that I do not recognize, and before I desolder it and play around with it, I figured I'd ask if anyone here knew what it was. It's a trimmer of some sort, but it's not labeled on the chassis. It appears to be in the (first) feedback loop, and it's the question-mark in the schematic. In the photo it's the white thing in the center:

weird part

And what's with having TWO feedback loops? I've never seen that before.

Another weird thing is the heater wiring. Only half the 12AX7 is used. (Pins 1, 2, and 3 are grounded). So they only bothered to heat the other triode, using 6 volts from pin 5 to 9. This surprised me, not only because I hadn't seen it done, but also because I would have thought the inside temperature of the glass envelope would be dependent on having both heaters powered, and that conduction would suffer with less heat.

It's a bit tricky to figure out exactly what's going on because all external connections are made through an 18 pin Amphenol connector, and obviously I don't know what they're connected to.

One of the coolest things is the input potentiometer, which is also a switch that opens when the pot is fully clockwise. It seems the input can be switched to line level once the switch is open, since the pot becomes a series resistor with no reference to ground.

If anyone has any comments, suggestions, etc. I'd love to hear them. What voltage do you think I should make the B+?

Thanks,

E.

> Only half the 12AX7 is used. (Pins 1, 2, and 3 are grounded). So they only bothered to heat the other triode, using 6 volts from pin 5 to 9. This surprised me, not only because I hadn't seen it done, but also because I would have thought the inside temperature of the glass envelope would be dependent on having both heaters powered, and that conduction would suffer with less heat.

That is odd. Not for the reason you cite (cathode temperature isn't much affected by total heat in the bottle) but because there is no standard for which heater goes with which triode. You could plug in a different 12AX7, and the not-used side might get hot, while the used-side stayed cold and dead.

> a component in there that I do not recognize

Trimmer cap. Interleaved metal and mica with a screw to press the leaves together to increase capacity. Big. Might be 1,000pFd.

Two feedback loops because at high frequency the feedback from the transformer vanishes. Although I looks like the trim-cap gives positive feedback at high frequency. Leave it alone. When you have it working, and decided on a load (it may be fussy), and have good test set-up, trim that cap for best response at the top of the audio band.

Run 250V B+. That's what most such modules ate, and I don't see anything to say this is different. Run it in the dark and be sure the 6BH6 plate does NOT glow the least bit red.

Output: take any two of the 150Ω windings in series as a 600Ω output.

Input: connect a mike to any combination of A-B-C-D. You will find two ways that work right, corresponding to two 37Ω windings, Connect them in series for a 150Ω input.

When series-ing windings, if you get zero output, reverse the phase of one winding.

I very much doubt you should put a Line Input into this. The click-switch would be in the all the way up position for all normal work. That input pot clobbers signal/noise ratio. You would only take it off the click for VERY loud sources. Probably an afterthought when rock-n-roll was invented.

Yes, my Ma Bell expert confirms that Northern Electric was the manufacturing branch of Canadian Bell. US Bell got out of the PA/Radio racket in the 1920s, but as you see things was different in Canada.

The blue-caps scream "1970". So do the Teflon wires. The hand-stamps under the front lip may have a date-code in them. If true (not a repair), this must be some of the last broadcast tube-modules ever made.

{moderator hat ON} I'm going to link-in some of those pix so your post does not overwhelm folks with slow modems. They can click to see more if they have time.

[quote author="PRR"]there is no standard for which heater goes with which triode. You could plug in a different 12AX7, and the not-used side might get hot, while the used-side stayed cold and dead.[/quote]
Wow... that is certainly not what I would have expected. The base diagram for a 9A tube is pretty specific, and it definitely implies that pin 4 is the heater for the first triode and 5 for the second.

Thanks for the help, PRR. I'll try 250 volts and report back in a couple of days.

PRR I have not seen a 12a*7 have different heaters hookups for the triode halfs the times I wired a 12a*7type, 4,9 was for the 1,2,3 side.

I am interested in what tubes were not wired this way. .

what did you use to draft that schematic? it looks great. it looks like a professional-drafted schematic from the pre-computer days.

[quote author="solder_city"]what did you use to draft that schematic?[/quote]

I just used Adobe Illustrator. Lately I've been trying to be more organized, so I've been drawing schematics on the computer so that I have better records to look back at later. Since I have all the various components drawn already, it's pretty quick to draw a different circuit from scratch just by copying and pasting...

Ideally I would use some Mac based schematic capture program and generate net and part lists for Osmond when I need to make PCBs, but unfortunately I haven't found a program that I like.

Thanks for the nice comments!

By the way, has anyone ever seen a source for that bad-ass cloth-covered shielded wire? That stuff looks great.

I agree, that schematic looks great. Adobe Illustator, huh? I'll have to look into that. I use OrCad most of the time, which does the job but has that "drafted on a computer" look which I dislike. Yours looks like a real drafted-on-paper schematic, which is appealing.

Did you create the symbols yourself, or lift them from existing schematics? Would you be willing to share your symbol library with others? :green:

As for cloth-covered wire, you can get it from a few sources. Here are three:

http://www.newsensor.com
(New York homeboys, of course I had to put 'em at the top of the list :wink: ).

http://www.tubesandmore.com/
(Be careful with ordering through this website; I recall that Ethan was saying that they didn't implement SSL properly. Perhaps better to pick up the phone...).

http://www.mojomusicalsupply.com

[quote author="NewYorkDave"]Did you create the symbols yourself, or lift them from existing schematics? Would you be willing to share your symbol library with others?[/quote]

Most of the symbols I drew myself, but I have redrawn schematics from time to time when the original is barely readable, so in that case I will often trace the original.

I'd be happy to share my symbols. Be forewarned though, Illustrator is a powerful drawing program and is probably totally overkill for drawing schematics. I use it because I've used it for years for my day job and I'm pretty fast already.

I went through a couple of schematics and copied random chunks onto a blank template, incase anyone wants them. Hopefully someone will find them useful.

For Mac folks there is a stuffed version here.

For PC folks, there is a zipped version here.

I currently use Illustrator CS (version 11). If someone has an older version and wants me to back-save to that version, PM me and I'll send it along.

Here is a GIF of the chunks I included in the file:



Thanks for the cloth-covered wire links. I actually do have some of the kind Antique Electronic Supply sells, but I haven't seen shielded paired wire yet. I love the non-stripping thing... just pull back the cloth. It doesn't twist nicely, however.

Cheers...

What a coincidence, I got my pair up and running only a few weeks ago ...

[quote author="PRR"]The click-switch would be in the all the way up position for all normal work. That input pot clobbers signal/noise ratio. You would only take it off the click for VERY loud sources. Probably an afterthought when rock-n-roll was invented[/quote]

... thx for that info !

The guy I got mine from said I was paying for the iron, basically said the same thing about the front-end pot and the topology not being that great overall (maybe over-inflating himself I dunno). I tested both with the input tranny windings tied in parallel on an Audix I5 and they both sound ok, maybe a little metallic (bright?). I haven't tried them on my other mics yet. The 600ohm connected output is going straight into a DAW, I assume it should be terminated with a 620ohm resistor.

I'm curious though, if you wire in the extra 150 ohm winding you get 1.5 times more signal out, no ?! And what should the terminating resistor be in that case ?

I hand-drew the layout and schematic from my units about ten years ago - I'm not 100% sure they're completely accurate.

NE-R6409A Schematic
NE-R6409A Layout

Tried several times to power them up but always got a loud hum at the outputs when running the filaments at 12vac - mine were wired in series. It's only recently that I wired the filaments in parallel and ran them on 6.1vdc that I got the units running nice and quiet. I used a modded PowerOne running at 210vdc for B+

If I understand correctly the unit is always running at full gain, and therefore at full noise too. I was thinking of fixing the S/N thing and also giving the unit more max gain by taking the unused half of the dual triode and wiring in a variable-load Mu-stage together with the first triode straight off the input transformer - the gain/volume pot would load down the mu-stage gain node and not the transformer signal. In effect the noise would drop with gain keeping the S/N more constant. I'm hoping to achieve enough gain for my Nady Ribbon.

Is this too risky or the wrong approach ? Or maybe I should use the iron for a better Ribbon mic pre circuit ?!

I was looking at some Northern Electric material in my files, and I remembered this old thread. I don't have any documentation on your preamp, but here's an info sheet on a similar model:

R5409CA

Sorry I missed this before.

> got a loud hum at the outputs when running the filaments at 12vac - mine were wired in series.

It isn't the series thing. This is one of the very few classics that relies on affordable DC heater power. Look at the heater wiring in your excellent sketch: this is all wrong for AC heat. Even Fender would not have botched it so bad. Dave's schematic clearly specifies "d-c" for the heater supply.

You could have used a 12V DC supply. Not the nastiest one on your bench: that wiring is so unbalanced and intertwined with audio that it needs to be quite clean.

Nothing wrong with 6VDC power, as long as you don't have to swap modules in a broadcast console where everything better be exactly interchangable.

> I was thinking of fixing the S/N thing and also giving the unit more max gain by taking the unused half of the dual triode and wiring in a variable-load Mu-stage together with the first triode straight off the input transformer - the gain/volume pot would load down the mu-stage gain node and not the transformer signal. ... I'm hoping to achieve enough gain for my Nady Ribbon.

The "Mu-stage" was unknown in 1964. I also fear you would upset the feedback.

The classic way would be to leave these two stages alone, add a 3rd plain common-cathode stage in front, pot between. Break V1 pin 2 from the input transformer. Wire-up a basic 12AX7 volt-amp on V1 pins 6 7 8. Plate resistor 100K or 220K, fed from C3.1. Cathode resistor 1K or 1K5. Grid to input transformer. Plate to 0.05uFd 400V cap, to pot, to ground. Pot wiper to grid V1 pin 2, plus a 1 Meg to ground.

That gives about 32-35dB gain on top of the existing 40dB gain. In any normal situation you would want an audio-taper pot and keep it half-way or lower: about 54dB gain or less.

Full-up you get 74dB gain. You also get tube noise around 5mV-10mV or about -40dBm, total dynamic range about 55dB, input overload about 1mV. This is really Output overload: too much gain and only +15dBm max out. The input stage overload is more like 100mV, and by trimming the pot you can optimize the output level so it doesn't freak-out.

You could try reducing the output stage gain by fiddling the feedback. But that will upset the response. And they already talk of trimming C1 to compensate other errors.

In Dave's plan, the input transformer is wired for a very-low nominal 30Ω. Most modern mikes will prefer the windings in series for nominal 120Ω.

... thanks PRR

I understand about sticking an extra Common-Cathode stage in front of the whole thing, but the circuit already seems a tad noisy as is - tempted on using a lower gain 12au7 in that case ...

I'm afraid I don't understand the relevance of the 15dbm statement and the output freaking out - is this a transient response thing ?

... in your estimate (or anybody's) how would this augmented version compare in overall performance to what's considered very good for a Ribbon Mic pre ?

... jcm

Well, I made the supply and now my pal who owns these things decided he wants to butcher them and build something else with the iron.

So I have a nice pair of 1:4 input transformers and a pair of 7:1 output transformers. He wants a nice mic pre. Hmmmmmmmmmm...........

Suggestions welcome.