Push-pull microphone preamp with UTC LS and Tamura transformers

[waltzingbear]

for those wishing to see the 428A schematic, its on Chris Preston's web site
http://www.vintagewindings.com/products/AltecDVD.html
look at the Altec 250 docs

cheers
Alan

 

[shabtek]

thanks for bringing up that altec design.
not p-p throughout and not off-the-shelf output trans.

I've lately been curious about tertiary windings in cathode circuit of p-p output stage
edit: I was looking at 428/429--thanks for contributing this info Alan

 

[Winston OBoogie]

thanks for bringing up that altec design.
not p-p throughout and not off-the-shelf output trans.

Not off the shelf transformers no, but both the 458 and 459 are def. push-pull all through.

That 459 is a mini beast!  40mA B+ with the two parallel 12BH7's on the output.

You could build that one with off the shelf transformers but you'd expect less than the 56dB gain, probably 46dB is realistic with what's available in IPT.

Combine the parallel 6072's cross neutralized with the parallel 12BH7's and it'd be killer I bet.

Thanks Alan for the link.

 

[aazaa]

Thanks Alan for the link.

Indeed Alan, also thanks. A lot of HQ Altec documentation over there!

Paul

 

[abbey road d enfer]

for those wishing to see the 428A schematic, its on Chris Preston's web site
http://www.vintagewindings.com/products/AltecDVD.html
look at the Altec 250 docs

cheers
Alan

thanks for teh link.
Interesting how they qualify a very low Z source of "unterminated".

 

[aazaa]

......    Regarding the high end roll-off.
I know you're happy with it, and you also added a compensation switch to flatten it out if need be......
......    An old, but easy, trick to ameliorate this somewhat is to add small pF value caps from each input anode to the opposing valve's grid.  Cross-coupled neutralizing.....

The Altec 458A is a good example of cross-coupled neutralizing......

First thought: 'a nice proposal' ... but after some reflection I see that what is excellent in the Altec 458A is not going to work in my preamp. 'Cross-coupled neutralising' done this way would be disturbed by the +-12dB global NFB in the circuit.

The solution drawn in grey in the circuit pictured in the first post, works perfectly to compensate for the high roll-off. When switched 'on', there is slightly less global NFB for the highs. (Note that this global NFB is also cross-coupled.)

As the values of the capacitor and resistor in the compensation network need to be different when using other transformers, I prefer not to specify them in my drawing.


Paul

 

[abbey road d enfer]

'Cross-coupled neutralising' done this way would be disturbed by the +-12dB global NFB in the circuit.

I see that global NFB is voltage-to-current. Ain't that supposed to reduce considerably the input Z?

 

[aazaa]

I see that global NFB is voltage-to-current. Ain't that supposed to reduce considerably the input Z?

Interesting comment! Yes, you are right, the global NFB is reducing the input Z.

The input impedance on the 500 Ohms connection of the UTC LS12 is measuring slightly higher than 190 Ohms. For high frequencies the impedance is a bit higher (>200 Ohms) as there is less feedback for the highs.

The input impedance seems to be ok for my microphones. I mainly use Neumann M49 and Sennheiser MD422U with this preamp, both are 200 Ohms.



Paul

 

[ruffrecords]

I cannot believe it. I have missed another great tube mic pre thread. Must be old age creeping up on me.

Anyway, very interesting design. Not only dc coupling but closing the loop down to dc too.

For even better performance have you considered using an 6922 instead of the ECC82? The ECC82 is well known for having poor intrinsic distortion which combined with relatively low mu means it is rarely an ideal choice. The 6922 has twice the mu and about 10dB less intrinsic distortion so it could make a worthwhile improvement to an already excellent design.

Also, do you have any thought on a wide ranging gain control for this design?

Cheers

Ian

 

[aazaa]

I cannot believe it. I have missed another great tube mic pre thread. Must be old age creeping up on me.
Anyway, very interesting design. Not only dc coupling but closing the loop down to dc too.

Old age??? No, Ian or... maybe, ... but if so, then only in the mirror, not here!

For even better performance have you considered using an 6922 instead of the ECC82? The ECC82 is well known for having poor intrinsic distortion which combined with relatively low mu means it is rarely an ideal choice. The 6922 has twice the mu and about 10dB less intrinsic distortion so it could make a worthwhile improvement to an already excellent design.

Nice, your suggestion for a better performing tube. Can be useful for people who want to build a variation on my PP-preamp.

From the very start I knew the possibility of using more appealing tubes. This circuit can be easely adapted for different tubes/transformers. (I used the - excellent sounding - old Philips tubes laying around.)
The heaters of the 4 tubes in my preamp are all connected in series. Only 150mA charge for a very relaxed LM317HVT. There is a limiter for a soft start of max. 165mA. (The heaters of the proposed 6922 are 300mA...)

Max. Vkf of the 6922 tube is 60V. For safety it is best to connect the heaters of this tube to at least +60V.
Max. Vkf of ECC82 is 180V.

Low mu seems to me an advantage for a tube connected to a PP output transformer. The gain of DC unbalance is lower.

-----------------------------------------------------------------------------------------------------------------------------------------
[edit]

Also, do you have any thought on a wide ranging gain control for this design?

The preamp has a fixed and a variable output. Stepless volume control at the variable output.

Proposal for gain control in the final version: posts 130-135
-----------------------------------------------------------------------------------------------------------------------------------------
I've chosen to draw only the basic circuit. It is easy to read, easy to understand.
On the site here are many excellent examples of volume control, filters, phantom etc... including yours.
It is a pleasure for diy-people to choose their own tubes, to adjust the circuit to their taste, to add bells and whistles themselves.

And... thanks for the nice reaction, Ian!


Paul

 

[Winston OBoogie]

First thought: 'a nice proposal' ... but after some reflection I see that what is excellent in the Altec 458A is not going to work in my preamp. 'Cross-coupled neutralising' done this way would be disturbed by the +-12dB global NFB in the circuit.

Ahhh!  For some reason, I had assumed your pre was fixed gain.  Even though now, I see it isn't.
Sometimes I'm an idiot!

Yes, cross-neutralizing is optimized to compensate for Miller capacitance.  Which, as you say, will change depending on the gain of the device being neutralized.

Good catch 

 

[aazaa]

Ahhh! For some reason, I had assumed your pre was fixed gain. Even though now, I see it isn't.

I could have explained better...

Your assumptions are right, Winston. The preamp as drawn/built has a fixed gain of +48dB on the fixed output.
The variable output has full gain control from 0 to +48dB. This control is connected to the output of the output-transformer. As there is plenty of headroom, this stepless control is largely sufficient for me.
The headroom can be further augmented by raising B+ to 300 à 330V together with the adjustment of a few resistors.

The fixed gain of the preamp can be easely made variable if wanted. The 100k resistor can be replaced by a rotary switch to change the NFB in steps for a change in gain between +36dB to +54dB. The second pole could be used to correct the high roll-off to perfection.

Yes, cross-neutralizing is optimized to compensate for Miller capacitance. Which, as you say, will change depending on the gain of the device being neutralized.

Was not exactly the idea behind my saying: "'Cross-coupled neutralising' done this way would be disturbed by the +-12dB global NFB in the circuit."

It is strange to add positive feedback by cross-coupling the anodes to the opposite grids while there is already a cross-coupled global negative feedback connected to this grids. Both grids are receiving at the same time positive and negative feedback from the opposite side. The stronger negative feedback is supposed to cancel the correction done by the added positive feedback.

But there is more. Even if it would work - beyond my understanding - what about stability???

Surprised by the post of Winston I decided to try things out.
8 pF was the lowest value of caps I could find at home.

Connecting the two cross-coupled 8 pF capacitors was fun... It was creating a very nice, perfect oscillator...


Paul

 

[abbey road d enfer]

Surprised by the post of Winston I decided to try things out.
8 pF was the lowest value of caps I could find at home.

You may want to try a home-made capacitor, which consists in about one inch of magnet wire twisted together. An old trick from wireless days.

 

[Winston OBoogie]

Surprised by the post of Winston I decided to try things out.
8 pF was the lowest value of caps I could find at home.

Connecting the two cross-coupled 8 pF capacitors was fun...  It was creating a very nice, perfect oscillator...

Let's call the oscillation an added feature then 

There is a fine line in value for sure, you see values being used in the 6 - 7 pF range (depends on gain of course and your first stage is a higher mU than I've ever tried it on) , and it would also depend on any stray capacitance from wiring, socket etc.

Abbey's trick is a good one for creating a small value cap.

On whether positive and negative feedback connected at one point is an issue, well it depends on the ratio of the two if course. 

If you have the margin to spare on your negative feedback, then it will be stable if you keep within that margin after adding the positive.
Go over that, and you get a nice new oscillator feature to use for lining up the meters on your recorder 

Edit:  I began trying to estimate how much stability margin you might have in your pre-amp and, after 5 minutes I decided that, in the end, the way you've compensated for the high end droop is easier, so probably the best way to do it.  And how I'd probablyend up doing it too after all is said and done.

I'll admit that I have only ever cross-coupled on a global loop free circuit, using medium gain valves of 40 or so. 

 

[aazaa]

Let's call the oscillation an added feature then 

If you have the margin to spare on your negative feedback, then it will be stable if you keep within that margin after adding the positive.
Go over that, and you get a nice new oscillator feature to use for lining up the meters on your recorder 

I'm afraid this is not how the new superb oscillator is triggered.
The oscillation in the preamp when experimenting with cross-coupled positive feedback, has nothing to do with the global NFB. It is generated in V1 (ECC83), the first double triode.

In the circuit I use large common cathode resistors. This results in a strong interaction between the two triodes. Both triodes are correcting/ helping each other. One of the qualities of the preamp!
When cross-coupled positive feedback is added, this strong interaction between the two halves of the ECC83 is causing the first stage to collapse into an oscillator.

Again I did the test.
The 12dB global negative feedback was completely removed. I shorted the 100k resistor and added a 1µF cap between this point and ground.
The two 8pF cross-coupling caps were added. The result was even better: an oscillator with 12 dB more output.
Very usable as it is stepless variable.

Paul

 

[Winston OBoogie]

Yes I took note of the 'long-tail-pair' resistor which serves to share current between both halves.
We also have degenerative current feedback at that point.

It was when trying to estimate the margin of degrees left for stability in that dual stage (we can also think of it as two common-anode stages feeding two common grid stages) that I gave up and decided the value of capacitance needed for neutralizing could very well already be exceeded by the strays and inherent capacitance and therefore, the way you did it was the easiest and better way.

Cross-neutralizing a current-sourced diff-pair such as yours is certainly possible if that's what you're doubting.  I'm not sure?

Anyway, why mess with what works, and you have it working   

 

[aazaa]

Anyway, why mess with what works, and you have it working

I totally agree, Winston!

There is no intention to make fundamental changes to the circuit posted in the very first post. The only change I plan is to create maximum headroom by adjusting the preamplifier to a B+ of 300 à 330V.

Thank you for your input!

Paul

 

[Winston OBoogie]

There is no intention to make fundamental changes to the circuit posted in the very first post. The only change I plan is to create maximum headroom by adjusting the preamplifier to a B+ of 300 à 330V.

Hey, if your output valve can handle it then yep, why not.
With a coil/inductor/transformer loaded, push-pull stage - a 300V HT supply provides a massive amount of headroom.
This might be seen as excessive to some.  Not me  .

I'm guessing part of your reasoning is that the ratio of signal current to voltage headroom also determines, to an extent, the THD etc. 
With an effective doubling of the H.T. voltage due to the energy storage of the output transformer, AND the fact that you're using push-pull so another 6dB more headroom there...
your thd is pushed down even more.

Do it! 

Edit: quick maths sums in my head seems to conclude that max output might be in the vicinity of + 29 - 30 dBu and possibly then determined by the transformer (which I don't know max output specifications of) before the valve poops out and runs out of juice.
You'll have to check me on my quick sums though as I'm only guess-timating what the actual DC coupled voltage across the output valve is. 
Anyway, cool. 8)

 

[ruffrecords]

Max. Vkf of the 6922 tube is 60V.  For safety it is best to connect the heaters of this tube to at least +60V.
Max. Vkf of ECC82 is 180V.

Paul

Not sure which data sheet you are reading but the E88CC one (which is the same tube) says its 150V (cathode positive wrt filament) and 100v if negative.

Cheers

Ian

 

[Winston OBoogie]

The E88CC can be a beefy little bugger. 
It'll work with a relatively small H.T.voltage across it.  But it'll also work with a 300V or so B+ voltage, dropping half that onto the anode with 8 or 9  mA per triode.

I know Ian has had good luck with one of the newly manufactured and available versions and I should really take another look at what's out there.
My experience is limited to what was out there about 15 years ago and,  N.O.S. was better then.  Of those available at a decent price,  old stock Tesla were good enough for me without paying silly money.