• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

WE121A, Please critique me

So I've just about finished a restoration of a pair of Western Electric Mic Pres. Everybody's told me I shoulda parted out the iron for big clams on evilBay but that's completely beside the point. So anyway I'd greatly appreciate some feedback on the mods.

Here's the Original Schematic

Here is my Current Working Schematic

What I have done:
1. Replaced all the tubes. 6J7's replaced with 6SJ7's.
2. Replace all caps. Electros w/ canisters, couplers with Solens.
3. R3 changed from a fixed resistor to a pot for 2nd stage gain control.
4. Added input and output pads. They're juicy old Daven 500/500 T-pads.
5. Added output meters.
6. Enclosure with remote power supply. I wired the interconnecting power cable so that if it becomes unplugged, the power supply shuts down.

I want to add phantom mic power to the inputs as well.
Here's a Picture
 

EC8010

Ex-Moderator
2003-01-18 7:57 am
Near London. UK
I hope those pads are properly balanced, otherwise, they'll destroy the input transformer's common-mode noise rejection.

The best way of adding phantom power is to take a 3k4 resistor from the +48V supply to the centre-tap of the input transformer. But you can't do that with your pad in the way. Why not move the pad to after the transformer?
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
> I hope those pads are properly balanced, otherwise, they'll destroy the input transformer's common-mode noise rejection.

Not if the input winding is floating.

At low frequency, CMRR will be unaffected by the pad. At high frequency, stray capacitance matters. It may not be a real concern in a small quiet studio. I worked for years with unbalanced low-Z dynamics and rarely had a real buzz.

Also the very existence of an input pad suggests large mike levels, enough to drown out any buzz.

(In another time and place we discussed reducing the high total gain, a WE-approved mod, but it didn't happen.)

> Why not move the pad to after the transformer?

Because he has these lovely low-Z pads, and putting ANY extra stuff on the high-Z side of the input transformer will severely screw-up the response. Big step-up there, very high impedance, transformer is wound for Open Grid, minimal wiring, no resistance.

Also the input transformer may saturate. Though this may not be bad. And the tranny is proportioned for higher levels: a WE-mod omits one stage of gain.

> The best way of adding phantom power is to take a 3k4 resistor from the +48V supply to the centre-tap of the input transformer.

Lowest noise, but I have never seen a Phantom mike so weak that such added noise was an issue.

Oh, and the Phantom will work with those constant-R pads: it adds a few hundred unbalanced ohms but phantom values are not that critical.

Capacitor coupling the Phantom is probably a bad idea.
 

EC8010

Ex-Moderator
2003-01-18 7:57 am
Near London. UK
PRR said:
> I hope those pads are properly balanced, otherwise, they'll destroy the input transformer's common-mode noise rejection.

Not if the input winding is floating.

I'm afraid it will. If the pad is an unbalanced attenuator, then looking back into the pad from the transformer side we will no longer see the source as balanced impedances to ground. The WE input transformer will have been wound to have balanced stray capacitances and leakage inductances to ground from either input terminal so that common-mode noise on the input signal causes equal currents to flow from either leg, resulting in no voltage difference between the transformer's input terminals. If the source has unbalanced impedances to ground, then the noise voltage drop across those impedances will be different and they will be converted from common-mode noise to differential-mode, to which the transformer is sensitive.

Ah, I wasn't aware that the transformer had a large step-up ratio - keep the pad on the input of the transformer, and make sure it's balanced.

The reason for adding the phantom power at the centre-tap of the transformer is that it preserves input balance (it's much easier to achieve transformer balance than resistor balance). If you use a 6k8 to each leg, any imbalance in the resistor value unbalances the input, reducing noise rejection.

Lest my concern with balance be thought to be theoretical whingeing, I've worked in TV studios where microphone cables and cables for (dimmed) floor lighting were unavoidably mixed, and also sent baseband analogue audio all over the country. We had to be very careful about balance if we didn't want dimmer spikes and telegraph (long while ago) getting all over our precious programme.
 
OK, I read up on T Pads and these are not balanced, but they are floating. The old studio that these relics came from was centered around a home-made console of these pads.

As for the WE approved mod of removing the first stage, I did try this on one of the amps. I had hoped it would be as simple as to shatter an old octal tube (something that's pletiful in my camp) and make a bypass jumper. Before I got all gungho, I realized that this would put serious DC on the secondary of the input transformer. It could be done with a SPDT switch, set to toggle the source between the grids of V1 & V2. On the down side, eliminating a stage throws the output out of phase. In all honesty, the previous owner's mod of changing R3 to a pot makes a huge difference. The units get much quieter (I'm talking about noise floor here, not the obvious gain reduction) plus this allows me to balance the 2 channels.

The whole input padding issue came about after testing the units with a a pair of AT condensor mics set up to record a piano. Even with the mic's internal pads set at -20dB, I was still getting distortion. My typical enviroment is much louder than an acoustic piano. Output padding came about after I blasted the inputs of a Sound Blaster sound card. I wanted to have the output VU meters be adjustable so that the unit can be callibrated to match whatever it is connected to.

I'm not in any big rush on the mic power issue since I do have a pair of MXL power supplies I can use for the time being, but it just seems silly to have a stereo mic pre w/o mic power.

Thanks!
-Richard
 

EC8010

Ex-Moderator
2003-01-18 7:57 am
Near London. UK
Hmmm. Your problem is much more involved than simply adding input pads. The WE microphone amplifier is clearly a bit short of headroom with modern microphones. Plus, your soundcard is probably too sensitive. You need to sort out your gain structure...

It's usual to need to apply 60dB of gain to a condenser microphone directed at a formal conversational level speaking presenter. Real piano is both louder and has higher peak to mean ratio than speech, so you will need less gain. Further, your microphone position is likely to be only 6" away from the strings, and you have a big plane reflector above the microphone (I'm assuming a grand piano here). You probably need only 35dB of gain to produce peaks of +8dBu.

Your soundcard might not expect peaks of +8dBu. In fact, it might only expect fiddly little peaks of 250mV. The WE amplifier was designed for old (insensitive) microphones and professional (high) signal levels.
 
For the moment I'm going to stand by the padding solution namely because the old fella I got this collection from basically told me this was the way to go. He ran a successful studio from the late 40's into the mid 70's. The few hours I spent with him were quite humbling. I enjoy being around smart people even though it makes me regret all the brain cells I roasted in my youth.

I wasn't very clear in my previous descriptions. When I said the piano audio was distorting, it was very minor distortion, so yes there is an input headroom issue. As for cooking the SB inputs, that's what happens when you put 100 milivolts in with 70dB of gain. That's something in the neighborhood of 300 volts fed into the sound blaster. :hot:

I'm going to button up a few loose ends on it tonight and do some proper test recordings this weekend.
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
> That's something in the neighborhood of 300 volts fed into the sound blaster.

Output iron is 3:1 or 4:1 stepdown. Even if you are outputing 250V square-waves, lightly loaded, you only get 60 or 80 volts out. It is less harmful than plugging your sound-baster input to a wall-socket.

Still: You NEED a fixed output pad of -12dB. Else that 3-WATT output is going to smoke chips. AND those irreplaceable Daven faders.

I forget the rating, but the 6V6 can put 3 clean watts which in 600Ω is 43 volts. Chips running on +/-15V rails may smoke at 11V input. If the amp were low gain, maybe you would never get near 43V output; but if you are "looking for flavor" or using hotter sources and mikes than intended, you will be pushing the amp hard. You could push the amp to 43V without smoking chips if you had a 11V/43V= -12dB pad. Even when you run the tube "clean", at say +10dBm peaks, your sound card can pull-up the -2dB peaks to near nominal digital levels.

And it was quite routine to throw a -6dB pad on the output of almost everything, to force the impedances to 600Ω no matter what. The output of this amp is lower than 600Ω, which will confound gain calculations (and those Davens if not loaded); but after a -6dB 600Ω pad it will BE 600Ω near enuff. And your "600Ω" load could be 500Ω, or 300Ω if someone patches two loads at once, or who-knows-what; with the 6dB pad the amp will always see ~600Ω load.

Since you are sure you want the 500Ω vari-pad, but don't want to waste the top 6 steps (or risk turning-up too far and smoking chips), and exact 600Ω "matching" is moot, this will work:
 

Attachments

  • we-pad.gif
    we-pad.gif
    3 KB · Views: 111
Wow PRR! Thank you for the design work. I will certainly implement this for I seldomly intend to fry stuff. That's just a by-product of my "bull in a china shop" bench work.

Currently I have 6F6's installed in the 3rd stage because that's what I have a matched pair of. I could easily round up a pair of 6V6's if you think this would be better. One of the last cosmetic things I intend to do is get a complete matching set of RCA tubes (and replace that ugly black canister cap). The previous owner modified it to use 6SJ7's instead of 6J7's. I was going to step these up to JAN5693's.

Again many thanks!
-Richard
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
> I have 6F6's installed

6F6 is similar but not-the-same as 6V6. Have you checked that the idle current is not excessive for those priceless WE transformers, and idle power is not too much for the 6F6's lower(?) ratings?

6F6 should have less "flavor" than 6V6. Though I have a feeling that the two types tended to converge in later years. When demand for 6F6 slowed, it may have been expedient to adapt 6V6 parts rather than keep all the 6F6 tooling hot.

One odd source for matched pairs of top-grade 6F6 is the older H-P 200AB audio oscillator. Ones with attached power cord were mostly two 6F6, and rarely replaced (the 200AB asked for <1Watt from an 8-watt pair). But later ones (roughly when they switched to detachable power cables) used the dull little 6AQ5s.

Make the 1K2 resistor switched. With it, the amp sees essentially 600Ω as intended. Without it, the amp sees a higher load, which may reduce distortion and extend or even bump-up the top octave. You could go so far as to try a 250Ω 5 Watt resistor there, to make the amp "grunt", go 2nd-harmonic, and droop the treble a bit.

> a complete matching set of RCA tubes

Of course what it really wants is WE-made tubes. RCA's industrial tubes were very good, and some of WE's tubes are "common types", but when WE made tubes they made them for their own use: they knew they would be using these tubes for many-many years, didn't want down-time. Some of the best tubes ever made.