Microphone preamp idea

[NewYorkDave]

TK: Why use one tube?

Lower replacement cost.

Lower noise from fewer stages.

Reliability through simplicity (50% less chance of a tube failure!).

Less power required.

An output transformer is a one-time expense. And if your load is high-Z like the input of most DAWs, you don't even really need one. My One-Bottle does 50+ dB into a high-Z load, 40+ dB into low-Z via a stepdown transformer. Are you guys all recording hummingbirds through ribbon mics or somethin'? Everyone seems to have this mania for more gain.

Really, +40dB is fine for a general-purpose mic preamp, anything more is gravy. The inputs of most recorders have enough gain to compensate for most situations.

If one tube will do the job, then I say do it! If not, use two :wink:

 

[tk@halmi]

It is debatable how important one criteria is more important to one as opposed to another. We could probably debate them forever. Noise is almost universally critical, but that can be helped with picking low noise input tubes. The output tube can be picked for gain/girth. You may still come out ahead of using some of the more obscure tubes.
I understand that we have these wonderful one-bottle things and I am not against them. However, I still prefer most two tube topologies, especially when going into low Z inputs like some compressors and limiters we adore.
Also, note that jsn wanted to use his 6SN7 tube.

Peace,
Tamas

 

[PRR]

Just to be clear: "+4dBu" is a VU meter reading. Peaks on speech/music will be over 10dB but not 20dB higher; we often take 16dB headroom above nominal zero-VU. If zero VU is +4dBu or 1.23Vrms, then the amp has to cleanly handle peaks to 11V peak (7.75Vrms sine test-tone). If zero VU is -10dBV, we have to handle peaks to 2.8V peak (2Vrms sine).

If we have to drive long lines, we need to be prepared to drive a few hundred ohms. For lines less than 25 feet, we only need to drive a few K of line load. Since gain got cheaper than transformers, we rarely see the old 600 ohm (step-up) inputs, most gear is 10K or higher.

This will be a MAJOR design decision for a tube preamp. "+4dBu" in long lines or 600 ohm loads will need 18mA peaks, which needs a very fat tube or step-down from few-mA tubes. 2.8V peak in short lines and 5K-10K loads needs only 0.5mA peak, which can comfortably be sucked from few-mA tubes.

> They had separate line amplifiers

Indeed. The standard mike-amp was asked to drive 600 ohms and moderate lines, but was NOT asked to deliver +4dBm/+8dBm at nominal zero VU. More like -10dBm or even -20dBm. Then a gain-knob (or mix-network) and to a Line Amp.

> How much gain do I need in a mic preamp?... a mic preamp needs to be able to handle signals anywhere from around 2mV to around 1V, and have enough gain to output something like +4dB into 600 ohms. 2mV to +4dB is 56dB of gain. 1V to +4dB is only 2dB of gain!

Don't want to repeat myself.

> less gain since you would get less noise

No, not really, not when you finish mixing the song.

 

[jsn]

Thanks again for all the great info. Thank you PRR for the link to your recording level experiences. I am re-reading that post now and have a feeling it will take me a while to digest it all since I am so new to all this.

I think I might have a neat idea for a low noise, low gain, low distortion mic pre based on a circuit people have started using for RIAA peramps. I will post it as soon as I have a rough draft because I will definitely have questions.

As for the flycutter, aluminum not only grabs the cutter, but heats up quickly and squeals like a banshee. Very not-fun. Clamps and a sacrificial board under your panel are definitely required. The hole does turn out really well though.

I am considering cutting large holes in aluminum rack panels by using a circle cutting router jig to make a wooden template, and then using a roto-zip bit or something designed for metal to cut the aluminum.

Thanks,
jsn

 

[CJ]

The WE 417 circuit runs fine without an output transformer.
I think the input transformer is most important, since it has to step up and be quiet.