16 channel DAW summing amp

[MerlynMetal]

Hi,

I'm considering a 16 ch summing amp for a DAW for warming up mixes and get some iron coloration.
Something simple with minimal parts count (if possible).
After spending couple of hours of searching and reading I compiled the schematic below .
I'll be using Carnhill transformers but I'm not sure which one of these  will be the right choice:

http://www.audiomaintenance.com/acatalog/Input_Transformers.html

For OP1 I was considering OP275. Also I was wondering if this balanced/unbalanced idea would work.
I have experience building various devices but studio equipment is not exactly my thing so I would appreciate your advice and comments on the idea and schematic.

 

[RuudNL]

The second stage doesn't look like a very good idea to me.
With the wiper to ground, you have (close to) infinite gain...
The first stage has a gain of about 2.3 dB. Don't know if that is your intention?

 

[MerlynMetal]

The first stage has a gain of about 2.3 dB. Don't know if that is your intention?

Not really. Maybe 47k5 resistor will be more suitable.

With the wiper to ground, you have (close to) infinite gain...

A resistor in series with the inverting output should correct that I guess.

Or maybe using the first opamp as a buffer/follower and using the 990 for make up gain is better idea?

 

[ruffrecords]

Have you got 16 inputs total of 16 stereo inputs (16 per side)? You schematic implies 16 per side.

With 16 per side I would recommend changing the 10K series reistors to 4K7. This will give you close a 10K input impedance and reduce bus noise by . assuming all you inputs are arranged so they short hot to cold if nothing is plugged in then the bus impedance will be just over 600 ohms. This is probably a little on the high side for many input transformers so you might need to slug the bus at the transformer primary to bring it down to 300 ohms or less.

You only need two gain stages if you want a level control placed so that it maximises headroom. If you want more transformer tone you could consider placing the level control on the output transformer secondary. If you do this then you can get away with one stage of gain for which the 2520 will be which rhe 2520 will be more than adequate. It is basically a fixed gain mic pre and there are plenty of schematics of these using the 2520.

Cheers

Ian

 

[MerlynMetal]

Have you got 16 inputs total of 16 stereo inputs (16 per side)? You schematic implies 16 per side.

I have 16 stereo inputs.

This is probably a little on the high side for many input transformers so you might need to slug the bus at the transformer primary to bring it down to 300 ohms or less.

Does slugging the bus mean adding a resistor in parallel with the input transformer primary (330 Ohms for example)?

If you want more transformer tone you could consider placing the level control on the output transformer secondary.

I saw somewhere in the threads a schematic for 1k attenuator. Maybe that should do?

 

[ruffrecords]

I have 16 stereo inputs.
Does slugging the bus mean adding a resistor in parallel with the input transformer primary (330 Ohms for example)?

Yes. The idea is to make the parallel combination of the bus resistors and the slugging resistor close to the ideal source impedance for the transformer

I saw somewhere in the threads a schematic for 1k attenuator. Maybe that should do?

Yes, that should work. If you make it using a stepped switch then the stereo tracking will be very good. If you use a regular dual gang LOG potentiometer then it will not be so good. LIN pots track much better but only give 6dB attenuation at the mid point so it is common to slug them with a resistor between the wiper and cold. At Neve we used 2K LIN pots slugged with 1K to give 10dB attenuation at the centre point.

Cheers

Ian

 

[MerlynMetal]

ruffrecords, thanks for your help so far!

Would you advise on which Carnhill input transformer is ok for this application. I guess a 600:600 will do the job for line level inputs?
For some subjective reason I don't feel comfortable having an attenuator at the end of the circuit but considering my poor experience with such equipment maybe it's a wrong feeling. Also the two gain stage configuration will require only one 2 deck stepped attenuator and I can go with something already available instead of making the 1k one myself.

 

[RuudNL]

The gain of the first stage in the last design is now 0.3 dB. I think you need more gain.
If you want more gain, you have to increase the value of the 1.54K resistor.
Also I would reduce the value of the 220 ohm resistor at the output.
In general, distortion is reduced when a transformer is driven from a lower impedance.
(But if it is the idea to introduce transformer distortion, of course keep it!)

 

[ruffrecords]

ruffrecords, thanks for your help so far!

Would you advise on which Carnhill input transformer is ok for this application. I guess a 600:600 will do the job for line level inputs?

Although the inputs are line level they get attenuated quite a lot before reaching the transformer. If the inputs are a pair of 4K7 resistors and the bus us slugged to look like 300 ohms then the attenuation is:

(4700 + 4700 +300)/300 = 32.22 times = 30.2dB

So, for the best noise figure I would recommend using the standard Neve style 1:2 step up - Carnhill part VTB9045 wired 1K2:4K8

For some subjective reason I don't feel comfortable having an attenuator at the end of the circuit but considering my poor experience with such equipment maybe it's a wrong feeling. Also the two gain stage configuration will require only one 2 deck stepped attenuator and I can go with something already available instead of making the 1k one myself.

Of course it is your decision. You could always try on at the output as well as one half way - that way you could experiment with altering the drive level in the output transformer which after all is where most of the transformer sound will come from.

The balanced output attenuator also only requires one 2 deck switch. You wire it just like a pot - hot and cold to either end and the output between wiper and cold.

Cheers

Ian

 

[MerlynMetal]

The gain of the first stage in the last design is now 0.3 dB. I think you need more gain.
If you want more gain, you have to increase the value of the 1.54K resistor.

Thanks for that note. Since it's not a clone project there will be some experimenting going on for sure.

Also I would reduce the value of the 220 ohm resistor at the output.

I noticed a 39ohm/3.6uH combo recommended on 990's input and output.

So, for the best noise figure I would recommend using the standard Neve style 1:2 step up - Carnhill part VTB9045 wired 1K2:4K8

Thanks for that suggestion!

So basically the most simple solution would be something like this below.