WE 407A Push-Pull Mic Preamp

[Larrchild]

Yeah, all distortion was measured at the 600 ohm output although I chose +10 dbm for that. At +30 it was pretty high but still looked like a sinewave.

I'd like No distortion in the iron as that would tend to be f2 and we hate that.

We want mild f3 distortion from the tubes.

 

[CJ]

OK, this 75 M6 EI is perfect.

Should get you 300 henries with 20 ma pri dc and 30 db output.
Turns look like about the same as the Langevin.
Wire is gonna be 38 or 39, which is a good ga to work with.

But dcr is gonna be about 1500 to 2000 ohms.
No big. Better than plate resistors, parafeed chokes and caps, right?

 

[PRR]

> distortion in the iron as that would tend to be f2...

No; with all DC balanced it would be mostly odd order.

> ...and we hate that.

Simple folk prefer sweet second and find third to be rough. Of course the generation raised on Fender has their own taste.

The key thing about iron distortion is that it is vanishingly small midrange, gets gross in deep bass, but can phatten-up deep bass for small speakers. The tubes of course don't care what frequency they distort (except for the varying load circuit reactance throw on the tubes).

> But dcr is gonna be about 1500 to 2000 ohms.

Foo. That's low. It's a 40K winding now, right? Any DCR under 10% of working impedance is just gilding the lily. Maybe you have to bump-up B+ a few volts to cover DC loss, NBD.

 

[Larrchild]

Man, I sure was getting a lot of (mostly) f2 when I hit this parafeed xfmr circuit too hard.
I'm looking forward to that changing with a dc bal xfmr.

I think it's safe to say that people who like SE amps are ok with f2 distortion and Push-Pull folk dig the f2 cancellations. Dog and Cat people.

 

[CJ]

OK, got the core figured, got the turns figured, now figured out winding structure.

Using single bobbin for 40k push pull means uneven dcr and lots of capacitance from the 13,000 sec turns, so going to go with Tele disk style, 4 discs of 3,250 turns, and two disks with 812 turns ea. And the fb in between.
This kills capacitance and gives even dcr all for the same price.

Building seperators for bobbin now.

But Def Tones tonite S/O show, get baked and be there!
I'll be in the pit. :razz:

 

[Larrchild]

That is a neat trick to cancel "C". 40k primary is a test of mettle in that arena.
I've been in deep work mode the last few days but I'm due to revisit this in a day or so.

Just as soon as I get it off of the heart-lung machine, I will hook it up in the living room and post an mp3 =)

Go listen to your Devil-Rock. And thanks!

 

[Larrchild]

http://www.globalnetvillage.com/design/preproto.jpg
I took a break and loaded up the nano with a few thousand songs and went to the Keys to chill.

Then, revitalized, I returned to finish the mic pre. I am working on the phantom supply, but the preamp is now off the respirator and self powered and looking Retro as we require.

CJ?[/url]

 

[NewYorkDave]

Looks great, Larr!

What's that forlon-looking TO-220 in the lower right corner of the frame?

 

[bcarso]

[quote author="NewYorkDave"]Looks great, Larr!

What's that forlon-looking TO-220 in the lower right corner of the frame?[/quote]

Wasn't that sacrificed to the Tube God at midnight, and the corpse left behind as a warning?

 

[CJ]

I have 10 pounds of 75 EI coming in next week, along with U brackets.
(Build your own 1166, 4 dollars a pound for steel, 25 cents for a bracket, and 25 cents for a bobbin, you supply wire, Ken - Ram Sales 1-818-997-8057- he will take a check, but you have to wait- no min order)

I have the transformer 99 percent designed, but the last 1 percent is killing me!

I have the core, I have the flux, I have the turns, I have the wire size, I have the leakage inductance down, but I can't seem to get the capacitance down to a tolerable level.

The problem is that I am fighting two variables here, the wire size, (remember, I need to feed 10 ma to the 707's, use 20 ma safety and low dcr) and turns. The bigger wire leads to more capacitance, (imagine winding with copper the size of a 2 by 4, compute your capacitance).
And the high impedance pri means many turns and thus capacitance. I found out that the disk method (Oliver type) does not do good with stray C at higher ratios, so I am not going to wind it like that.

You can not wind bi-tilar pri to get good balance because capacitance is to high with that many turns. You can cancel capacitance by putting half of winding under the cathode, but this reduces gain of output stage due to NFB.

Sheesh, no wonder nobody makes a 40K:600 at 10 ma balanced!

But I think I can get you a 3 bd rolloff at 40 k hz if that is adequate.
Low end will be fine due to many henries of un gapped core. (push pull)

cj

 

[Larrchild]

yeah -3db at 40khz is gonna be about right with all that skinny wire. Rock-on.
I'm just glad I got 'er boxed up and nearly useable.
(Dave, the TO-220 is one of the 2 regulators this gets, 20vdc and 48vdc)

 

[skipwave]

Larr - That really is beautiful! Thanks, guys, for the great discussion.

CJ - Can't wait to hear how the transformer turns out. If you make a few, I'll buy one. :thumb: The image of you moshing to the D-tones will have me giggling all morning. :green:

 

[CJ]

OK, I got it all figured out. I was thinkin a Ni core would be the ticket, as it would reduce turns by 1/3, but then we would need a lot of Ni, which could become easily magnetized with a slip of the voltmeter probe, etc. Plus, it's expensive and Larry wanted M6 for cool bass.

So I called up Ken and he told me he has non square bobbins for 75EI !!!
This means I can lace up a twice stack and reduce turns by 1/2 due to area going up by factor 2. DCR goes up due to longer mean turn distance, but wire ga can come down to compensate, since less turns means more room for thicker copper. Triad winds a lot of 1 by 2 cores on some classics, like the HS 50 and 52 outs. Effieciency of coil does not change much from square to rect core.

And, I forgot all about e shields!

So, with a fat stcak and a couple of shields, I think we can wind upo a +/- 1 db 5 to 30 k cps, what say>?

 

[Larrchild]

Ok, I wasn't ignoring the question . :grin: I was testing the preamp in the living room studio to get a feel for it, and answer the iron vs nickel question.

Short of a meter-buffer, it's working really well. And I will build one of those today. But first, had to perform sea-trials on this thing.

I have few condenser mic choices here. 2 Neumann TLM-103's and 2 AT 4040's. And they have always sounded too damn crispy. This preamp has solved that, and is a good compliment to crispy condensers. I was recording through a Mackie 1402, so that isn't a big pre to conquor, but here's an A/B:

Mackie 1402 Pre

WE 407A Preamp

Microphone was an AT 4040 at 3-4 inches from the instrument. It was recorded to tracks 3 and 4 of an Otari MX-70 1" 8 Track, and bounced into ProTools at 96k/24 bit.

I like my cheap 4040 much better now on the WE 407a. =)

So regarding Nickel on the output, Im indecisive. I don't want to go any more rounder-sounding than it is now. But more clarity might ruin the fun too, if ya know what I mean. I'm still leaning toward M6. Although, since I'm deriving feedback from the output transformer, if we got less phase shift winding with Ni, I'd go with that.

Here's some pics:

Maiden Voyage

Front
Inside
Inside
Back

 

[skipwave]

Hmmmm... You must've forgot to put something in. :? All that empty space in the center of the box. :wink:

Seriously, that is some super-clean wiring and layout. Hard to believe its a prototype. Great work! :thumb:

 

[CJ]

I got Ken to drop in a few rectangular bobbin samples for 75EI in the shipment of laminations, that way, I don't have to send him another check, wait for that to clear, etc. So I will get crackin as soon as that stuff gets here.
I am surprised it is not here yet.
Maybe today.

The nickel makes me nervous, sittin around the B+ like that. I mean some people worry about magnetizing there core with a DMM, what would 250 at 100 ma do if you had a short or something?

People do use it on outputs, but nearly always in a parafeed configuration, with no dc on the core. The 670 has direct coupled nickel core outputs in the Triad HS-52. The Pultec has direct coupled output nickel inthe Peerless S-217-D. Hmmm, those are a couple of nice boxes. Mayeb Ni would be the ticket.
I notice some added series resistance in the plate circuit of the 670. Ths might be to protect the iron just in case. And it serves a metering purpose also.

I have never seen anything bigger than 625 EI in Nickel.

If I can get Mag Met to sample me some 75 EI 50/50, we could give that a whirl. Maybe lace it 2 x 2 just in case.

Then you can do a shootout with two feedback models.

 

[Larrchild]

Skipwave Wrote:

Hmmmm... You must've forgot to put something in. Confused All that empty space in the center of the box. ... You must've forgot to put something in.

Well, since this is a Voltage Controlled Preamp..that big open area in the center is for a sidechain control amp (as PRR guessed 3 pages ago..)
to form a peak limiter addition within the pre itself. :thumb:

But that's later. Lemme turn this into a Lab Project first.

Also, the reason it looks like a box of air is to try to get the power transformers into a seperate zipcode.

Mayeb Ni would be the ticket.

I'm gonna find out, if I take one of my 50k Ni Fairchild outputs and wire that in.=)

Meanwhile, hone those calculations, and ask your engineering team what they think. Thanks.

 

[NewYorkDave]

I hear some weird artifacts in the low-end in both samples (encoding, I guess, or a guitar that badly needs a set-up); but even through that dark glass, the WE pre clearly smokes the Mackie. No surprises there.

This is the kind of stuff that keeps me from giving up on this place altogether. Thanks once again to Ethan for creating the Drawing Board, so interesting stuff like this has a chance to be read rather than disappearing quickly and quietly beneath an avalanche of "my SSL clone don't work" threads. Thumbs up to Larr and all the others who are driven by the desire to experiment and explore.

 

[Larrchild]

I know one thing, My AT4040 room mic on the drums never sounded like this.
http://www.globalnetvillage.com/design/predrums.mp3 647kb

I'm really happy with the fat-flavour of this thing. I think my mics are seeing a better load. Then there's that push-pull thing. This was worth the time. I now have a non-sucky mic pre at the cribb. :grin:

 

[mwkeene]

Man, I hate to rain on the parade, but I can't really hear a difference between those two guitar samples. I know I know, nobody will take me seriously after this comment, but I just gotta be honest. Now, if you put a clip of the drums with the Mackie and the tube pre, im sure I'd hear a difference. Drums are just so widely distributed across the frequency spectrum i feel like id have more sonic information to analyze. Were the clips compressed at all? Both clips sound like they were done in the same take...how did you split the signal from the mic? That noise is interesting, it almost sounds like someone is blowing on the mic a little around 1k ...

That thing looks really sweet, I admire the ability to so quickly get something together...
-mike