mic pre concepts

[dale116dot7]

[quote author="tk@halmi"]
I have been entertaining this idea for some time. Using a double pole rotary switch for gain control one circuit could adjust Ccomp and the other Av. This way at higher gains the amp would be gradually decompensated as needed.[/quote]
I was thinking about this too, but I never got the board design done yet. I was planning on a transformer coupled preamp, AD797, and one deck of the switch would shift in different R values in the feedback circuit, the other would trigger a couple of relays (RF or something like that) to switch the decompensation cap. The cap is small so I wanted to keep trace lengths short.


Kev: I like transparent, so I'd like to see a decent square wave response, and one that doesn't change significantly with gain. And reasonable noise, but that's not a deal breaker for me, anyways.

Perhaps programmable reactive loading (either real or simulated) on the mic would be nice? Dial the real and reactive impedance in... ??? Hmmm... 500 ohms, 10 henries, and 4200 pF... ok just dial them in. Could be interesting on dynamic mics.

 

[tmbg]

I don't fully grasp how the front-end of a preamp should load the mic... I always assumed the ideal would be high-Z inputs for as little loading as possible, but I guess that's more applicable to guitar-type signals, and not microphones...

So, what exactly are we trying to achieve in the front end?

 

[PRR]

> Our aim may be to be higher spec'd than most microphones

But mikes are not designed in a vacuum. Most "good" mikes deliver a S/N that won't be degraded by a "good" preamp.

> I always assumed the ideal would be high-Z inputs for as little loading as possible

Some mikes, it hardly matters. Others are normally speced for "best" results with load over 1K or 2K. "High" compared with 150Ω, though not fabulously high as amplifier inputs go.

> So, what exactly are we trying to achieve in the front end?

Who the hell knows? Everybody rides different horses. Some people listen to square waves. Some make semi-square waves out of nice sounds. I doubt any single design is right for everybody, and the wide variety of boxes on the market agrees.

 

[tmbg]

So is there actually any BENEFIT to having a low Zin? or is it just, design the preamp so it does what you want, and so long as the Zin is above 1k or so you should be fine? or what...

I could see a couple of ways to approach it, and I'm very very green, which is why I'm asking these questions

one way would be to use a unity buffer up front to present as high an input impedance as possible, then have our first gain stage be right after that... I can envision in my mind why that might not be a good idea, but I don't know if it is or not.

Another way would be to do a 10k Rin, Xk feedback inverting amp as the first stage, so our Zin would effectively be ~10k. What's wrong with this? Perhaps that we'd have to use somewhat large values of feedback resistance to get an appreciable amount of gain? 1M nets us 40dB of voltage gain, no? Do these large resistances cause input bias current issues? If so, can they be minimalized with ye olde 9.9k compensating resistance on the non-inverting input?

I may be completely talking out my arse here... I'm frantically trying to apply theoretical book learnin' that I've picked up to these oh-so-real-world situations...

I guess what I'm really asking is not what works, but what DOESN'T work

 

[PRR]

> use a unity buffer up front to present as high an input impedance as possible, then have our first gain stage be right after that...

Then you have the noise of two amplifiers. You almost never want unity-gain input in low-noise application. Anyway it isn't impossible to get a satisfactorily high input Z with a gain stage. With tube or FET it is trivial. With BJT, the Base impedance tends to be plenty high; the hard-nut is the base bias resistors and current offset errors. And many good designs exist.

> a 10k Rin, Xk feedback inverting amp as the first stage, so our Zin would effectively be ~10k

You don't have to go that high.

But this way adds the noise of a 10K resistor. Noise will be much higher than mike self-noise.

I have seen exactly this done, but with 600Ω input. It loaded the mikes, but not real bad. It had more noise than a "perfect" mike amp, but not so anybody complained. It gives very simple wide-range gain control.

 

[tmbg]

What about using a non-inverting gain-stage where the feedback resistors aren't contributing to the input impedance? In this case, I could use a much lower feedback loop resistance, say 100ohm - to ground, and 10k output to -, which gives us a gain of 101...

Are either of those feedback resistors going to contribute to noise in that way?

Thanks for bearing with me, I'm learning a ton

 

[Kev]

[quote author="PRR"]But mikes are not designed in a vacuum. Most "good" mikes deliver a S/N that won't be degraded by a "good" preamp.[/quote]

that may well be ...
and I hope to now begin to show what I am driving at.
Many of the posts here have already assumed the pre-amp exists. The moment you do that you run the risk of copying or re-running an old idea. ( and that's ok in the long run - nothing inherently wrong with re-inventing the wheel - that just proves a good idea.)

In that list above I was only trying to formulate a list of requirements we might need to describe this pre-amp ... yet to be designed.
Very Hypothetical so far.

[quote author="PRR"]Then you have the noise of two amplifiers. You almost never want unity-gain input in low-noise application. Anyway it isn't impossible to get a satisfactorily high input Z with a gain stage.[/quote]

This is good but I think we are forward of the game ... as I said before some of the comments here are already assuming an issue exists before we have identified a test.

Let me explain.
If we have an existing group of mics that need a pre-amp. By identifying their requirements we can set a spec for the mic-pre to meet. Some of those spec's might be dead easy to meet with a particular topology ... maybe not. Lets find out.

... So, what exactly are we trying to achieve in the front end?
and replied too
Who the hell knows?

good question
and
good reply

Front end need to interface with the chosen microphones
provide gain and drive the next section
after that it can get tricky as people have different ideals

which is why I want to keep away from them for now.

Perhaps it is easier to chose a few microphones as examples of real word units and use them as yard sticks. Microphones that are typical in someway but also bring a specific requirement.

So far I haven't brought anything specific to the thread. I wanted to get things back to first base. Not working so far so I'll try second base.
SM57/58 Dynamic Mic with transformer output. No Phantom required
AKG414 Condenser Transformer-less and requires Phantom
Tube Mic Has it's own power supply and it probably Transformer'd.
Back Electret ... any special requirements
mmm ?

Ribbon ... someone tell me about ribbon mics

The mic-pre might ..
have 50 to 60 ... 60 to 70 dB of gain.
Provide Phantom
Load the mic - how should it load the mic ?

I didn't want to get down the design line this far yet ... but ...
should it be able to handle unity gain or should we pad the input so as to be able to run some gain even with very loud signals.

remember that ... 0.2 to 3,000 milliVolts.

I'll stop there.
So many timnes in these threads a level of knowledge is assumed and we have some readers that are taken down a line without knowing why the direction was taken in the first place.

OH ... :shock:
I just thought
The mic-pre should be balanced input and have a degree of common mode rejection.
Even that will have some readers wondering what Kev just said.
Why don't we have unbalanced Mic inputs ??
My MiniDV Camera has unbalanced inputs .... :roll:

... should our mic-pre have unbal inpouts too ?

 

[CJ]

What about changing from fet to junction via a switch? I think the junctions would be lower in input impedance?

 

[tony dB]

is it difficult to have variable impedance on the inputs?

I've seen a tool with this funtion and the guy who showed me this said it really was GREAT gear as you could match (or purposely mismatch) your mike to a pre to color and change dynamic response? he wasnt selling it btw

maybe i talk garbage bigstyle here, my knowledge of impedance is not deep enough to answer, for that i ask :?

Cheers,

Tony

 

[Svart]

I want clean sound and frequency linearity (don't we all...) through the gain range. Simplicity is another key factor for me. how come everything seems overly complex unless it has some transformers then it's not complex, just expensive as HELL. then it's also "colored" with that "iron" sound..

FET input? FET output? (why not?) other than personal preference, why WOULD anyone choose any certain parts over another?

 

[pucho812]

I think the one thing over looked is that one preamp is not going to be an end all to the rest of preamp designs. Granted there are certain design ideas currently I really like but to come up with one that would be an end all to end all? I don't think I would want that. IF I have to pick from already out there pre's my vote is on the Vipre. But would have to have a stereo one.
But if we looking for something new?
I dunno if I can think of new micpre design but why not have an A/D-D/A converter with an all tube analog path on the in's and outs? hmmmmm

 

[Kev]

[quote author="pucho812"]I think the one thing over looked is that one preamp is not going to be an end all to the rest of preamp designs. [/quote]
... 'twas never my intention
just to direct the thread to look at some of the boundries and requirements that a Mic-pre might have.
:roll:
in a hope to show how some people might have got to their chosen topologies.

 

[Svart]

thanks Kev, figuring out WHY someone has chosen some type of topology over another is almost impossible IMHO. Someone might choose something like tubes because they are used to that kind of design, and naturally believe that tubes are superior for some reason or another. same with BJT and/or FET based stuff.. Hell even IC opamps might be chosen (ssl9k?) because someone either knows enough(or not enough in some cases... :green: .) to make a certain part work well, maybe even better. but alas, we always seem to get back into the arguement of what is "better" which is almost always based on opinion.

so i feel that a perfect Preamp would be one with NO color, NO artifacts, NO noise(at least a reasonable amount..) and extreme headroom since this seems to be the hardest type to get right.. Everything else that deviates from this we can just call "colored" or a specific "sound".. because that's what we do anyways if you really look at it.

I've always thought that we were attempting to get our sound from the instruments, not the gear, right? adjust and tweak at the source and all that business?

ah hell i'm sick and it's too early for this.. /tirade :green:

 

[PRR]

> extreme headroom since this seems to be the hardest type to get right..

Assuming there is a Gain control, this is entirely the operator's responsibility.

 

[Svart]

true, true.

:guinness:

 

[Gus]

I don't know if you can make a microphone preamp that can do all you might want unless it is built as a cost no object device. I think thta is what the Gordon preamp is about if I understood CJs posts about it.

Put a loud singer infront of some microphone and you can get close to line level without a microphone pre.

Then you want a pre that does ribbon to line input with good noise?

In the AES microphone book there are schematics of higher voltage solid state preamps.

Someone posted at the old place (I forget who) about building a transformer in to a EF IIRC it was a step up transformer. This was for higher level microphone signals

IT ,might be better to have maybe 3 design concepts low, mid and high level inputs. Then you could maximize the circuit for noise and headroom something that would be expensive to do other wise.

One could make plug ins like 7th circle API etc. One tube voltages and other solid state voltages or even hybrid.

 

[PRR]

> One tube voltages and other solid state voltages or even hybrid.

Power voltages are generally about output voltages. And there is never a need for more than 1V nominal 10V peak levels on interconnects inside a recording studio building. So there is no need for more than around 40V, +/-20V on direct-coupled or 1:1 outputs.

Feeding a line driver +/-100V just ensures that your board line inputs will smoke before the preamp clips.

> better to have maybe 3 design concepts low, mid and high level inputs

Yes. If you look at the German Broadcast preamps made for the high-output condensers, they are balanced quite differently than contemporary US preamps aimed at ribbon and dynamic levels. The way many folks use hot condensers on loud sources, the "preamp" should really be a slight-gain line input with Phantom, not a high-gain lowest-noise design.

 

[Svart]

so are you really saying that our mics aren't outputting hot enough and our preamps don't need to be as "great" as we think they need to be?

i'll :guinness: to that!

 

[rafafredd]

How about a box with just a 150:600 (1:2) transformer and phantom, for those hot condenser tracks?

 

[tk@halmi]

[quote author="Svart"]so are you really saying that our mics aren't outputting hot enough and our preamps don't need to be as "great" as we think they need to be?
i'll :guinness: to that![/quote]

This actually makes sense to me. It does seem right to add amplification to the signal where it starts, the microphone. That is where you could omptimize noise and apply the appropriate amount of gain for each type of mic element.

Since we DIY our stuff I would go as far as eliminating phantom power and move to a separate power line. This would help make expensive input transformers and esoteric input capacitors optional.

Tamas