trafoless microphone preamp wish list

[Svart]

"probably at least +/- 35V at this point."

:shock: :shock: :shock: :shock:

better possibility for a light show! :green:

 

[Gus]

I think a third project could be the green redone a little. I have done some simple things to the two I built.

 

[analag]

I have a FETCODED design on the forum, maybe if we stop f...ing around we can do something with it, because I don't think we will be going much further than the babbling stage with this here thread. Correct me if I'm wrong. Gordon this Gordon that, sounds more like gossip than what we could be doing.

Analag

 

[pstamler]

I've pretty much finished the everything-including-the-kitchen-sink power supply board. It has about every bell and whistle anyone might want to put on: snubbers, room for 3A diodes, 1-ohm resistors before the first set of filter caps, multiple paralleled filter caps, a second stage of filtering with more multiples, and 317/337 regulators with optional fine-adjust trimpots. For really high-end design, as someone suggested, those could be preregulators. The +/- supplies can be anything up to 35 volts, I suppose; the third supply is intended for +48V phantom, but it could as well be 5V for logic for non-mic-pre applications. I suspect a lot of folks would use only some of the bits, but it's designed to allow that. For high current there are terminals to attach TO-3 packages. It will also take the LT adjustable regulators, which are pin-compatible with 317/337s.

This group design we're talking about might be an application where a supply board like this could be used. Its main drawback, which might not apply here, is that it's biggish -- 3x9". But it's designed to be sawed in half, so you can put the first part (3x4") in a remote power supply box with the power transformers, and the other part (3x5") in the main chassis. Or, if you use it in a single box, you can put the two halves side by side.

I originally planned to use a TL783 for phantom regulation, but got foxed when TI dumped the through-hole version of the chip. For the nonce I'm floating a 317 on a 47V diode for the phantom supply. Sounds dicey, but I have a preamp in the other room where I did that, and it hasn't failed since I built it back in 1983.

Right now I'm trying to sort out a couple of issues with the Gerber files, but the basic design's done. Next I gotta write it up.

Peace,
Paul

 

[analag]

Now that's what I'm talking bout, this is a man of my own kidney.
Action.
Analag the instigator
the trouble maker
the DIY junkie, and the list goes on and on and on

 

[Rossi]

[quote author="analag"]I don't think we will be going much further than the babbling stage with this here thread. Correct me if I'm wrong. Gordon this Gordon that, sounds more like gossip than what we could be doing.[/quote]

The Force is strong with bcarso, I'm sure he'll come up with something very interesting. Still, a simpler design would be cool as well. Building something like the Gordon is beyond my abilities, I'm afraid. But I definetly like the multiple fixed gain stages approach. PRR's active ribbon booster + my ancient Siemens V272 was lower noise than any single preamp I tried with my ribbon.
I personally don't mind a little distortion as long as it sounds good. What I want is excellent noise performance, a lot of gain, and a pleasing sound. A little color is fine. Something that's not too difficult to build but is better than your average preamp.

 

[bcarso]

I've heard of audio products described as "revealing." This exchange is revealing as well :wink:

It may be time for the thread to spawn some new ones, devoted to specific design directions.

I see at least two camps:

Numero Uno: A relatively simple, low noise, maybe ~euphonic distortion one;

Numero Dos: A cost-little-object, long-march neutral one. As complex as it needs to be but no more than it has to be. Maybe somewhat modular so bells and whistles can be added eventually to a central core design.

I like number two myself because it is a greater design challenge, so I think that's the direction I'll stick to.

There 's every reason to have the two or more directions cross-couple and benefit from each other.

This all may take a while. Obviously, please don't feel you need to wait for me :grin:

 

[analag]

That's all we needed to hear man, kick your feet up, light your cigar and ponder your design considerations. And remember....the more you know the more taxing it becomes.

Analag

 

[Bauman]

Ok, I can sound wrong here, but I'd like to hear a trasformeless micpre that sound so good like a transformer couple one.
Yea, it needs to be cheap enough, less that the cost of a transformer coupled one!

That's my big wish for a transformeless one!

cheers!
Fabio

 

[chrissugar]

In the first week of August I will have to record some classical music for film. I would be very happy if I will not have to use my transformer based preamps for that project. So hurry up Brad. :grin: :wink:

It looks like I will have some Bruell & Kjaer mics for that session.

chrissugar

 

[bcarso]

Chris, what kind of output levels do you expect from your B&K's?

 

[FredForssell]

Chris, what kind of output levels do you expect from your B&K's?

You best figure a max output level right around 0 dBu... really.

 

[PRR]

> what kind of output levels do you expect from your B&K's?

I figure AKG 414s can make 0.5V in front of an orchestra, fourth-row, and I don't think that is too far wrong.

Some mikes are hotter, and some situations call for putting the mikes closer. I think 3V is a possible level without even considering close-miked percussion and hot guitar amps.

At some point you do have to cry "enough!" and use a pad. And if the input noise is lower than 150Ω source, with big condensers a 10dB pad is no-problem (their self-noise runs 1uV).

Dynamics (and some SD condensers) in similar situations will top out around 100mV and have noise like a 100-200Ω resistor.

I figger full orchestra in the house as about 126dB SPL. That number was derived in the 1930s, confirmed many times, and my observations agree. Actually I can go a whole year without exceeding 118dB SPL, but you never know when the composer and conductor will pull every stop and belt a big chord with every lung and elbow working maximum.

A concert with noise near 14dB SPL and a peak at 126dB SPL, 112dB dynamic range, can't be played back verbatim in any normal listening system. Either we let the ambient and mike noise lay below media and playback-room noise, or we discretely trim a few notes. But the mike-amp should capture that 112dB and leave the trimming to post-production.

In less-loud situations: on women's choir, 2nd row, some Nak condensers slightly hotter than typical dynamics, 54dB gain into a 2V recorder just blipped the 0dB LED, so peak input was around 4mV.

Harpsichord in the big room, 4th-row (really too far back), AKG414, with 40dB-42dB gain, barely touched the -20dBfs lights at 2Vfs, so that works out to 2mV. Knowing AKG414 sensitivity, 80dB SPL at the mike. However the meters bobbled even when the harpsichord wasn't playing: sub-sonic rumble. So the actual musical level may be closer to 74-76dB SPL.

Note of course that if room rumble is near musical level, we ain't gonna get a high S/N no matter how we polish the amplifier. This room has problems; in a room suited to harpsichord we might get 74dB-14dB= 60dB S/N.

In the same situation, dynamics would have peaked maybe 0.3mV, and some ribbons (especially 37Ω types) closer to 0.1mV. Now we will have to polish transistors to preserve the acoustic S/N.

 

[bcarso]

Thanks to all. I guess I'm favoring ~1V rms max and anything above gets a pad. That seems to be just achievable with a fair amount of power dissipation quiescent (assuming class A throughout).

 

[pstamler]

[quote author="bcarso"]Thanks to all. I guess I'm favoring ~1V rms max and anything above gets a pad. That seems to be just achievable with a fair amount of power dissipation quiescent (assuming class A throughout).[/quote]

I think that's reasonable. In Russell Hamm's famous "Tubes vs. Transistors: Is there A Difference?" article he looked at output levels from various microphones; worst case was a U-87 and a loud yell, which came out to 0dBu. For years I took that as a maximum working level, but these days there are mics a good deal hotter than U87s, such as the TLM-103. For those I use pads.

Peace,
Paul

 

[chrissugar]

1Vrms and pad for anything over that level would be good.

chrissugar

 

[Samuel Groner]

I'd like to air one of my prejudices and say that I would like the gain of the electronic circuit to not be variable, at least not via the usual routes. The reason is that most of the time the gain is varied by varying the amount of feedback around the circuit, and in effect you have a completely different amplifier at low gain and high gain. Different distortion for sure, different output impedance, probably different bandwidth and/or noise performance.

Though I do not completely agree with some of the given arguments, it sounds like an interesting approach.

I suggest the following topology:
Connect three +20 dB gain amplifier in series. Add a variable PAD (0 to -20 dB) in front of the last stage. For 41-60 dB gain, route the signal through all three amps, for 21-40 dB trough the last two and for 0 dB-20 dB to the last stage only. This should give excellent noise and THD performance at all gain settings.

The gain stages could be optimized for the given task - i.e. beefier output stage for the last amplifier.

Someone may argue that changing the number of gain stages will lead to a similar variation in performance as changing the gain in a single stage. I do not disagree here, however I would add that to me there seems no way of building a variable gain amplifier with identical performance @ all gain settings.

Example: Two fixed-gain stages and a variable PAD in between. The performance of the two gain stages itself does never change (assuming a proper implemenatition of the PAD), however, we run the first stage at completely different levels with different PAD settings and similar output levels. So there is at least a change in distortion in this approach as well. Noise at low gain settings seems to be an issue as well. I suspect that my approach might even show less performance variation over the whole gain range than the latter topology.

Samuel

 

[Svart]

"I would add that to me there seems no way of building a variable gain amplifier with identical performance @ all gain settings."

true, But I think "very close" is a realistic goal. however I am also interested in how the device being driven acts to the preamp.. that might dictate more change than the output of the preamp design might have...

I don't see a solidstate preamp having an "output impedence" rather I see it in terms of current sourcing. How much current/voltage can it source without starting to get peaky or skewed.. same for a sandstate input that it's driving.. how much current/voltage does it take to drive the inputs without clipping or becoming strained.

all in all I think a multiple stage device with the identical stages paralleled would be the way to go. I proposed that the stages be "always connected" on their front ends and switched in/out of the circuit via JFETs on their outputs. this would of course add to their combined output drive abilities in an expected way..

 

[chrissugar]

If the Gordon is done with only two stages I can only think that probably that was the best compromise and that is the way to go. I supose two stages is the best balance between noise, distortion and change in sound.

chrissugar

 

[tk@halmi]

Chris,

The way I understand this Gordon amp is that there are many gain stages, but only two of them are switched into the circuit at any time. Is this correct?
Or do they change current sources within stages to obtain different gains? They were talking about having ten constant current sources per channel or something like that.

Also, it would be nice to know if every stage is balanced or they unbalance on the front and rebalance at the end.

Thanks,
Tamas