Question about multistage preamps

Hi folks,

just wondering, it seems general practice in preamp design (those with more than one amplifier stage) to have either the first stage doing the low gain work (with the second or third kicking in at higher gains ala Neve) or a 50/50 distribution of amplification (ala JLM99V??).

Is it bad practice to have the first stage running at say unity gain and have the second do the grunt work with the first kicking in by a fixed amount when higher gains are needed?

It seems backwards but the only prob I can think of would be some sort of noise issue. Though I suppose it could be negligible when running at low gains presuming a hot source (sauce? :grin is feeding the pre....

Any thoughts or am I over thinking again?

Cheers Tom

I thought common practice was to have one, or sometimes two, pre stages doing ALL of the voltage gain and then follow it with a single output stage contributing current gain only. Seems like Neve had their own way of doing things... but I wasn't around back then so what do I know?

Oh, and actually the output stage on the Neve input modules did have a fixed (yes, fixed, NOT varied via the sensitivity switch) amount of voltage gain along with its current driving capabilities, but it was usually at least 10dB less than the pre stages. It was only there as "fader gain" as far as I know.

A few things to consider when distributing gain between two stages:
* noise
* distortion
* frequency response

Personally I like preamps that do not change this parameters much as overall gain changes; so far I found a 2:1 distribution the most satisfactory solution, e.g. the first stage has 20 dB and the second 10 dB gain.

A 1:1 distribution optimises frequency response, but it performs worse with respect to distortion and noise. Remember that the first stage runs at mucher lower signal level, thus it does need less feedback for good linearity.

A 1:0 distribution (all gain in the first stage) has best noise performance and is good for driving heavy loads, as the output stage has very high feedback all the time. However, there is significant variation in frequency response and distortion in the first stage. It might be a good option for CFB amps.

Running the first stage at lower gain than the second seems to have nothing to recommend, at least with conventional amplifiers.

Samuel

Thanks for the replies guys, I'm actually looking at the Calrec1061, and my new version of it.

[quote author="Family Hoof"]I thought common practice was to have one, or sometimes two, pre stages doing ALL of the voltage gain and then follow it with a single output stage contributing current gain only. Seems like Neve had their own way of doing things... but I wasn't around back then so what do I know?SNIP[/quote]
Yeah that makes the most sense to me. In the original module, the gain is spread over two amps, one doing all the work upto 40dB and the 2nd kicking in at 45dB, increasing both together to create 5dB steps upto max gain of about 70dB. In the end the 1st stage does about 30dB and the 2nd does about 35dB.

Oh, and actually the output stage on the Neve input modules did have a fixed (yes, fixed, NOT varied via the sensitivity switch) amount of voltage gain along with its current driving capabilities, but it was usually at least 10dB less than the pre stages. It was only there as "fader gain" as far as I know.

Click to expand...

Yeah, actually what I'm going to do is add a third stage as an output amp @ 10dB with a full fader in front, with the first two stages doing the preamp work. However the three amps contribute to an overall gain of 75dB....there is about 90-100dB available though.

I've calculated a couple of arrangements, the first using a 2pole 12position switch which dealt with the gain setting ala the original - a split distribution with the two stages supplying varying amounts, both adjusted simultaneously via the switch.

However I had another idea as I wanted to use a single pole switch for ease of availability and reduced cost. In the next arrangement the first amp is unity gain (normal gain mode) and the 2nd has programmable gain from 5-30dB in 2.5dB steps. The 1st stage can be switched to +25dB (Hi-gain mode) to increase the total amount (a range shift). Its sort of cyclical I suppose and offers an increased resolution of 2.5dB steps with a continuously variable fader. There is also a pad between the 2 stages (attenuated mode) which should allow for really hot signals to be accomodated without a pad infront of the input transformer.

The nice thing about this besides the cheaper switch and resolution is that fact that it should allow ways to vary the drive in each stage to colour the sound. It was also the only way I could get the gains to play well and accomodate a 0dB padless mic input and accept +10dB signals on the line in.

However I'm wondering if its better practice to have the first amp variable between 5 and 30dB and have the 2nd stage switchable between two static values, 5 and 25 dB.

I suppose that fits Samuels thoughts better as is closer to the originals distribution. The amplifiers in question are only capable of about 30dB gain. The 2nd and 3rd stages are beefier amps and capable of 35dB.

I think I'll recalculate the stages to have the first do the stepped work as I think that seems better practice.

Any further thoughts?

Cheers guys.

All amps as a general rule, distortion goes up as gain is increased, the trade off is if your adding a 2nd stage which adds a few db of noise. You could build a 2 stage amp with any type of Op Amp and set the gain so that the first stage has say 50db of gain max and the 2nd have about 10 db of gain and work great. One very good advantage is the 2nd stage could have a slide fader so you could ride the level on a vocal for example.
I built a mic pre like that for that purpose and everyone loved it, didnt have to go thru the console for the fader section.

Thanks Mike,

I just recalculated my values and am going for stepped gain on the first amp with the second switchable between the two values....should work out fine and in practice not be too different from the original.

Tom