WE 407A Push-Pull Mic Preamp
- Thread starter Larrchild
- Start date
Help Support GroupDIY Audio Forum:
Larrchild
Well-known member
ioaudio
Well-known member
hey, very nice project!
i see there´s still some place in the box for rectified side chain amp
regarding c2 & c3 : isn´t 100nF on the small side?
regards, max
i see there´s still some place in the box for rectified side chain amp
regarding c2 & c3 : isn´t 100nF on the small side?
regards, max
Larrchild
Well-known member
In a normal amp circuit, I'd say go for big there. (coupling caps between 1st and second stage) But since I want to apply dynamic gain control to the first stage, The huge DC shift of the 1st stage during gain control will gag the second stage until things catch up ( see: Altec) So I wanted to minimize DC shift to the second stage.
Plus it's a 40k Rp tube, so .1 is not egregious.
Plus it's a 40k Rp tube, so .1 is not egregious.
enthalpystudios
Well-known member
I love this thread. I really need to finish all the clone junk I've amassed and move on to something interesting like this! Awesome work you guys, truly cool
Besides, I feel like doing something tube/p2p would almost be easier at this point for me. PCB's kind of drive me nuts. Especially when I want to hear a different cap, or what have you.
Anyway, very cool, can't wait to hear more samples (maybe some good ol' strat->blackface stuff?)
regards
billy
Besides, I feel like doing something tube/p2p would almost be easier at this point for me. PCB's kind of drive me nuts. Especially when I want to hear a different cap, or what have you.
Anyway, very cool, can't wait to hear more samples (maybe some good ol' strat->blackface stuff?)
regards
billy
Larrchild
Well-known member
As you approach cutoff, yes the THD climbs fast. But within the first 10 volts or so, (10-15 DB), it was suprisingly consistent.
While watching the real-time FFT of the harmonic signature, as I cranked the DC controlled gain pot, the dual sections of the tube would come in and out of balance slightly and THD would either come up or drop. But it was all fairly minor and never became more than .1% or .2% in non feedback mode (more revealing).
The conclusion was that although the sonics change slightly, it was never out of my self-prescribed boundries of .3%
The tracking of the dual section tubes is the crux of a happy outcome in this project. In places where the THD suddenly dips to 3 digits past the decimal, you can tell the load lines of both tubes are in happy convergence, thats for sure.
That said, balancing the cathodes and choosing tubes carefully make this a somewhat fussy project to embark upon, but I sure like the sound it makes. NYD sent me a 6BQ7 to try, because I want to make this a Lab Project with better avail on the tubes. But the pinouts were so different, that I would have needed to hack this up more than I want to now in order to try.
So I decided to make another module from scratch using non-bastard transformers that can be bought. Since this was a junkbox project, I used what I had. Now with Dave's tube, I will build something more reproducable. Also looking at a 6FQ7 as an output instead of another 407A.
The DC gain control is only a mic trim. I'm not trying to get 25db out of it. Just 10-15, so I'm still in a somewhat linear portion of the curve it seems.
The front end pad and feedback controls set the rest of the gain structure.
And output control too.
Being fully-differential, I can't add a hi-z 1/4" input jack sadly..like most pre's have, but I'll find something else here with one of those, I'm sure.
This started out as a home hifi line section. And then I realized that anything other than Lundahls or Jensens would sound too colored. Then I realized I had all these average sounding mic pre's here and decided to do a P-P mic pre instead. So far, it's been pretty smooth to get this going. I need to regulate the phantom supply and give it soft-start before it and the schematic are finished and I post the schematic. This week. Work permitting. But other than that, for my first ever mic pre, I'm quite happy.
I can see running out of gain reduction range if you had it manually cranked down, so I'm not sure if the DC control will stay, if and when I put a sidechain on this. Or, I may fall in love with DC controlled gain as a sound and want to control it during mixing with my Protools 002R using a motorized fader set to midi with a midi to cv converter on it. Automated Tube Mixer Baby! I'm sure it's duty as a line amp could also be optimized.
In the final analysis, the consistency of 6BQ7's will determine the validity of this project=)
>Besides, I feel like doing something tube/p2p would almost be easier at this point for me. PCB's kind of drive me nuts. Especially when I want to hear a different cap, or what have you.
Sometimes just wiring something up is faster. Now if you need 5 of them, it's different. But in the world of proto, you can't always wait for boards!
While watching the real-time FFT of the harmonic signature, as I cranked the DC controlled gain pot, the dual sections of the tube would come in and out of balance slightly and THD would either come up or drop. But it was all fairly minor and never became more than .1% or .2% in non feedback mode (more revealing).
The conclusion was that although the sonics change slightly, it was never out of my self-prescribed boundries of .3%
The tracking of the dual section tubes is the crux of a happy outcome in this project. In places where the THD suddenly dips to 3 digits past the decimal, you can tell the load lines of both tubes are in happy convergence, thats for sure.
That said, balancing the cathodes and choosing tubes carefully make this a somewhat fussy project to embark upon, but I sure like the sound it makes. NYD sent me a 6BQ7 to try, because I want to make this a Lab Project with better avail on the tubes. But the pinouts were so different, that I would have needed to hack this up more than I want to now in order to try.
So I decided to make another module from scratch using non-bastard transformers that can be bought. Since this was a junkbox project, I used what I had. Now with Dave's tube, I will build something more reproducable. Also looking at a 6FQ7 as an output instead of another 407A.
The DC gain control is only a mic trim. I'm not trying to get 25db out of it. Just 10-15, so I'm still in a somewhat linear portion of the curve it seems.
The front end pad and feedback controls set the rest of the gain structure.
And output control too.
Being fully-differential, I can't add a hi-z 1/4" input jack sadly..like most pre's have, but I'll find something else here with one of those, I'm sure.
This started out as a home hifi line section. And then I realized that anything other than Lundahls or Jensens would sound too colored. Then I realized I had all these average sounding mic pre's here and decided to do a P-P mic pre instead. So far, it's been pretty smooth to get this going. I need to regulate the phantom supply and give it soft-start before it and the schematic are finished and I post the schematic. This week. Work permitting. But other than that, for my first ever mic pre, I'm quite happy.
I can see running out of gain reduction range if you had it manually cranked down, so I'm not sure if the DC control will stay, if and when I put a sidechain on this. Or, I may fall in love with DC controlled gain as a sound and want to control it during mixing with my Protools 002R using a motorized fader set to midi with a midi to cv converter on it. Automated Tube Mixer Baby! I'm sure it's duty as a line amp could also be optimized.
In the final analysis, the consistency of 6BQ7's will determine the validity of this project=)
>Besides, I feel like doing something tube/p2p would almost be easier at this point for me. PCB's kind of drive me nuts. Especially when I want to hear a different cap, or what have you.
Sometimes just wiring something up is faster. Now if you need 5 of them, it's different. But in the world of proto, you can't always wait for boards!
bcarso
Well-known member
Have you looked at IM yet?
Larrchild
Well-known member
no, but i will. I hope I'm not shocked=)
Larrchild
Well-known member
Ok, here's some numbers:
with tertiary feedback:
DC gain full= .12% IMD
DC gain -10db = .32% IMD
Feedback OFF=
DC gain full = .6% IMD
DC gain -10db = .23% IMD
IMD signal = 1000 + 60hz 4:1 ratio.
with tertiary feedback:
DC gain full= .12% IMD
DC gain -10db = .32% IMD
Feedback OFF=
DC gain full = .6% IMD
DC gain -10db = .23% IMD
IMD signal = 1000 + 60hz 4:1 ratio.
NewYorkDave
Well-known member
I'd just use a floating input jack for an instrument DI input, selectable with a switch. There's no law that says the low side of the guitar pickup or whatever has to be grounded.
Naturally, you'll need coupling caps if you're applying a DC control voltage to the grids.
Naturally, you'll need coupling caps if you're applying a DC control voltage to the grids.
Recieved 75EI laminations but no bobbins! :twisted:
Also found a few more books on transformers, which is slowing down things due to updated design tweaks.
Hang in there, this will be a good transformer.
cj
Also found a few more books on transformers, which is slowing down things due to updated design tweaks.
Hang in there, this will be a good transformer.
cj
[quote author="Larrchild"]NYD sent me a 6BQ7 to try, because I want to make this a Lab Project with better avail on the tubes.[/quote]
I haven't read through the entire thread, so it may have been mentioned.
The 5670/2C51W is the same tube with a "normal" 6V heater, and very cheap. You can get the Russian/Chinese 6N3p also. Or the 7861 if you like 12V heaters... And there are many other versions.
You can see my linestage with the same tube here: http://stiftsbogtrykkeriet.dk/~mcs/RemTPre_small.jpg
Best regards,
Mikkel C. Simonsen
I haven't read through the entire thread, so it may have been mentioned.
The 5670/2C51W is the same tube with a "normal" 6V heater, and very cheap. You can get the Russian/Chinese 6N3p also. Or the 7861 if you like 12V heaters... And there are many other versions.
You can see my linestage with the same tube here: http://stiftsbogtrykkeriet.dk/~mcs/RemTPre_small.jpg
Best regards,
Mikkel C. Simonsen
Larrchild
Well-known member
>The 5670/2C51W is the same tube with a "normal" 6V heater, and very cheap.
I owe you a beer, Mikkel. 5670 it is. :grin: I even have some 5670's. It's a ruggedized version of 6BQ7, so microphonics may improve. Thank you for that info. Nice line-section,sir.
It's remote cutoff zone is similar to the 407a too. That matters more in this design because of the DC control aspect.
The 5670 is what Manley was using in the varimu when the world ran out of 6386's.
I looked at its plate curves and it starts cutting off steeply in the 10-20v zone.
I'm sure if I add a peak limiter/sidechain to this box, it will sound aggresive with this cutoff-curve, but that might be fun to have on a switch.
I've decided I like the 200 ohm input vs 600 ohm much better when using my AT condenser. beefier. I still need to try this pre on my TLM-103 Neumann. If it can de-burr that I'll eat my hat.
CJ, broe, I'm just wrapping up all the other aspects of this rig, so I can Lab-it. :thumb: No Hurry! I am already starting another chassis for these substitute tubes.
I owe you a beer, Mikkel. 5670 it is. :grin: I even have some 5670's. It's a ruggedized version of 6BQ7, so microphonics may improve. Thank you for that info. Nice line-section,sir.
It's remote cutoff zone is similar to the 407a too. That matters more in this design because of the DC control aspect.
The 5670 is what Manley was using in the varimu when the world ran out of 6386's.
I looked at its plate curves and it starts cutting off steeply in the 10-20v zone.
I'm sure if I add a peak limiter/sidechain to this box, it will sound aggresive with this cutoff-curve, but that might be fun to have on a switch.
I've decided I like the 200 ohm input vs 600 ohm much better when using my AT condenser. beefier. I still need to try this pre on my TLM-103 Neumann. If it can de-burr that I'll eat my hat.
CJ, broe, I'm just wrapping up all the other aspects of this rig, so I can Lab-it. :thumb: No Hurry! I am already starting another chassis for these substitute tubes.
Remember this?
Of course you do!
Anyway, back at it after a couple of false starts, a trip to the Hole which lead to other calamities we won't discuss, and a bobbin mixup, I finally recieved the non square bobbins for the 75 EI core (Neve output type lams), so I am getting this going again. We were after more inductance, thus, the larger non sq core.
Here are the 5 bobbins, 3/4 square, 3/4 by 7/8, 3/4 by 1. by 1 1/4, by 1 1/2.
Right to left, doh!
I wound 1000 t each on the five bobbins to do a test study on inductance.
Stripped the final candidate already.
3/4 by 1 1/2 > Here is what the xfmr will approximate after winding:
Here are the inductances from the test:
75 EI - 29ga M6 Measured on B-K Precision 875 B
1000 t # 30
3/4 Sq: 11.9 Henries
7/8 stack: 12.43 H
1 inch stack: 13.29 H
1 1/4 stack: 15.68 H
1 1/2 stack: 18.69 H
Some considerations to ponder:
1) the xfmr sits inside a feedback loop. This means two things:
a) the impedance of the transformerand therefor the turns ratio and turns, might be able to be lowered. This will help lower lumped C's.
b) we have to watch for instability. If the circuit and xfmr has phase shift, the transformer will become unstable when the feedback reaches 180 degrees (positive feedback-yikes!). This can happen if there is too much capacitance, a real threat in a transformer of this ratio.
If this happens in the upper the upper pass band, we could be in trouble.
Some unstable xfmr's have been known to pop tweeters becuase of this.
Ersel B Harrison comes to mind. (Peerless, one of his models was a screamer!-forgot which)
I believe in Larry's circuit, since we are direct coupled and only have one stage, instability will not be a problem.
Maybe the reason why the Langevin AM 16 is direct coupled. Not caps to cause phase shift in that circuit which features xfmr feedback.
Anyway, now we have our Al numbers for the 5 cores.
Useful stuff for future projects.
The transformer seen was laced Lap 3, this is the way the final version will be.
This will cut down on any non balanced DC causing problems with the core.
So now we can get our turns and henries for pri and sec.
10 henries is acceptable for a 600 ohm winding.
Some folks use 20, which is even nicer, especially nowdays, with all this sub wolf stuff, but this is a mic pre, and Barry White has passed, so we don't need to worry too much, I guess.
Capacitance is the main obstacle here. Especially with the big core>longer wire length.
So we will use every trick in the book in order to deal with it.
One trick is to wind two pri's, side by side, ala UTC A-10 (sec) .
What this does is lower the "plate" area of the winding by two.
If you connect the two "pies" in series, you lower the C by 1/4 th, which is why UTC and others used this on ratios in the uper integers.
We will wind a bi fi primary to provide balanced DCR and capacitance.
Since this bi-filar winding is going to be used as a continous primary, and not a pri-sec bi fi, like the API, we do not need to worry about pri-sec capacitance.
Actually, the API uses this capacitance to it's bandwidthal favor.
A xfmr can be viewed as a transmission line. Calculate the L's and C's in this line and you can get an idea of what the performance will be like.
Some transformers can be modeled as a two pole filter, (like a 600:600, as all the L's and C's are the same on both sides, pri and sec)
Others can be modeled as a three pole filter network.
Pick the ripple you want in db along the pass band, and you can predict the rolloff in the stop band.
This is why the graphs I did on unterminated transformers resemble rolloff curves off some of the better known Bessel and Tolyschchenov type filters. (sp?)
Inter layering, while decreasing leakage, increases capacitance, so we do not want too many layers. Plus, the bi fi pri will greatly reduce leakage, but we have to watch voltage gradients.
Pri Sec Pri is a popular choice.
Aslo, we need to know a little about the end user's setup in order to do the proper hookup.
Larry, is one end of the secondary going to be grounded, or un grounded (balanced) ?
This will help me pick the best hokup for balanced capacitance.
Thanks!
Of course you do!
Anyway, back at it after a couple of false starts, a trip to the Hole which lead to other calamities we won't discuss, and a bobbin mixup, I finally recieved the non square bobbins for the 75 EI core (Neve output type lams), so I am getting this going again. We were after more inductance, thus, the larger non sq core.
Here are the 5 bobbins, 3/4 square, 3/4 by 7/8, 3/4 by 1. by 1 1/4, by 1 1/2.
Right to left, doh!
I wound 1000 t each on the five bobbins to do a test study on inductance.
Stripped the final candidate already.
3/4 by 1 1/2 > Here is what the xfmr will approximate after winding:
Here are the inductances from the test:
75 EI - 29ga M6 Measured on B-K Precision 875 B
1000 t # 30
3/4 Sq: 11.9 Henries
7/8 stack: 12.43 H
1 inch stack: 13.29 H
1 1/4 stack: 15.68 H
1 1/2 stack: 18.69 H
Some considerations to ponder:
1) the xfmr sits inside a feedback loop. This means two things:
a) the impedance of the transformerand therefor the turns ratio and turns, might be able to be lowered. This will help lower lumped C's.
b) we have to watch for instability. If the circuit and xfmr has phase shift, the transformer will become unstable when the feedback reaches 180 degrees (positive feedback-yikes!). This can happen if there is too much capacitance, a real threat in a transformer of this ratio.
If this happens in the upper the upper pass band, we could be in trouble.
Some unstable xfmr's have been known to pop tweeters becuase of this.
Ersel B Harrison comes to mind. (Peerless, one of his models was a screamer!-forgot which)
I believe in Larry's circuit, since we are direct coupled and only have one stage, instability will not be a problem.
Maybe the reason why the Langevin AM 16 is direct coupled. Not caps to cause phase shift in that circuit which features xfmr feedback.
Anyway, now we have our Al numbers for the 5 cores.
Useful stuff for future projects.
The transformer seen was laced Lap 3, this is the way the final version will be.
This will cut down on any non balanced DC causing problems with the core.
So now we can get our turns and henries for pri and sec.
10 henries is acceptable for a 600 ohm winding.
Some folks use 20, which is even nicer, especially nowdays, with all this sub wolf stuff, but this is a mic pre, and Barry White has passed, so we don't need to worry too much, I guess.
Capacitance is the main obstacle here. Especially with the big core>longer wire length.
So we will use every trick in the book in order to deal with it.
One trick is to wind two pri's, side by side, ala UTC A-10 (sec) .
What this does is lower the "plate" area of the winding by two.
If you connect the two "pies" in series, you lower the C by 1/4 th, which is why UTC and others used this on ratios in the uper integers.
We will wind a bi fi primary to provide balanced DCR and capacitance.
Since this bi-filar winding is going to be used as a continous primary, and not a pri-sec bi fi, like the API, we do not need to worry about pri-sec capacitance.
Actually, the API uses this capacitance to it's bandwidthal favor.
A xfmr can be viewed as a transmission line. Calculate the L's and C's in this line and you can get an idea of what the performance will be like.
Some transformers can be modeled as a two pole filter, (like a 600:600, as all the L's and C's are the same on both sides, pri and sec)
Others can be modeled as a three pole filter network.
Pick the ripple you want in db along the pass band, and you can predict the rolloff in the stop band.
This is why the graphs I did on unterminated transformers resemble rolloff curves off some of the better known Bessel and Tolyschchenov type filters. (sp?)
Inter layering, while decreasing leakage, increases capacitance, so we do not want too many layers. Plus, the bi fi pri will greatly reduce leakage, but we have to watch voltage gradients.
Pri Sec Pri is a popular choice.
Aslo, we need to know a little about the end user's setup in order to do the proper hookup.
Larry, is one end of the secondary going to be grounded, or un grounded (balanced) ?
This will help me pick the best hokup for balanced capacitance.
Thanks!
Larrchild
Well-known member
is That it?
Beautiful!!! "El Grande Hierro"
No, I got some solid gold bells on the way.
:grin:
Still experimenting, so......
Hey, ungrounded or grounded sec?
Thanks!
Also, if you can relax your 600 ohm input, to say 1 K, I can save some turns and capacitance.
This is called "cheating the circuit" a bit, a little imp mismaqtch can save a lot oc C.
:grin:
Still experimenting, so......
Hey, ungrounded or grounded sec?
Thanks!
Also, if you can relax your 600 ohm input, to say 1 K, I can save some turns and capacitance.
This is called "cheating the circuit" a bit, a little imp mismaqtch can save a lot oc C.
Larrchild
Well-known member
Not always. I'm digging the "global feedback off" setting a lot, too.
When I add too much global FB, it can get squirrely because of the phase-shift issue in the iron. Also, loading on the main secondary will cause unknown conditions to occur. If this becomes an obstacle to primary performance, lets give it a back burner and for a tertiary winding,
just wrap 10% of primary t's around the whole thing " loosely coupled" and call it a day.
1K source Z might be ok, if yer againt the wall. Not my first choice. Should be ok.
Output is balanced, not grounded. But one can expect one side to be grounded in real world. So plan for that.
API's use 30v. This uses 200, Good! Let's not kill anyone.
That feels the most zen for this, somehow.
Like DC imbalance? Cause I'd hate to have this giant beautiful core lose LF due to offset. You decide. I might have a little creep in.
Didn't he invent the MIG? Stalin loved him.
I have a feeling that the phase shift in an API tertiary is lower due to lower ratio and that I cant compare a hi-ratio xfmr in that application here.
Smells like ultrasonic bursts of trouble.
All those oldtyme Langevin schematics with cap coupling for GFB are probably for that reason. Putting the xfmr in the loop has a certain charm too, but not if it oscillates.
Thanks!
rafafredd
Well-known member
Tell us all about it at the brewery!
Will you make separate feedback winding, or will you take feedback AND output from the same secondary?
