Attenuators with transformer preamps, a good idea ??

[pcrecord]

Hi gang,

You know I own a few Focusrite ISA preamps. I'm considering buying a few attenuators to allow me to drive the transformer a bit harder and try to pull out a bit more mojo.
I currently record tracks with a peak range between -18db to -10db (on the preamp and on the DAW).

So my question is ; is it worth it ? Do pushing the transformer a bit more allow a more colored sound ?

About this item :
https://www.amazon.ca/dp/B0060GDZTG/?tag=r06fa-20

 

[Boswell]

Yes, attenuators on the output of pre-amps do allow you to push them a bit harder. I made a set of 12dB balanced attenuators for my API3124+ pre-amps, as the output transformers don't start to wail until you get to about +32dBu. This is too much for most recorders, ADCs or mixers that I might want to send them to.

Balanced attenuators are easy enough to to make yourself if you are handy with a soldering iron. I can give you the formulae for resistor values. The advantage of making them yourself is that you can choose exactly the type of resistors that are going into them, and therefore know that they will not wilt or change their value appreciably at these high levels.

 

[Kurt Foster]

why not just attenuate the DAW input?

 

[Boswell]

why not just attenuate the DAW input?

Maybe we could be doing that - but we're talking more generally about attenuating the output of a transformer-coupled pre-amp before it feeds into an ADC.

The ADC could be part of an audio interface in front of a DAW, or it could be (as it often is in my live work) the the balanced insert returns of an Audient ASP880 that feeds via ADAT into an Alesis HD24XR hard disk recorder.

 

[pcrecord]

why not just attenuate the DAW input?

I will try it but if I push my ISA preamps, it might overdrive the ADC

In my case, I only have one setting for all the interface inputs (-10db +4db) and I'm using other preamps which would not be a good Idea to push that far...
I'm thinking of attenuating only 2 or 4 channels.

I made a set of 12dB balanced attenuators for my API3124+ pre-amps, as the output transformers don't start to wail until you get to about +32dBu. This is too much for most recorders, ADCs or mixers that I might want to send them to.

Wow, that would be so nice of you. It would be nice to do them myself and save a few bucks.
Another question would be how much attenuation the ISA's would need ? 10 -20 db ?

 

[Boswell]

Another question would be how much attenuation the ISA's would need ? 10 -20 db ?

I would look up the full-scale dBu value of all the inputs that you might want to feed from the ISA outputs and note the lowest. The only figure I can find for the ISA maximum output is +25dBu, which doesn't strike me as exceptionally high. I would be surprised if the lowest figure on your list of inputs was less than +20dBu, so something like 6dB attenuators could well cover all eventualities.

The 6dB figure is easy, as it's a 1:2:1 ratio of resistors in a U-shape on its side. So, as an example using 1% range resistors, 1K Ohm in series in the signal lines and 2K Ohm shunt across the driven input would give the correct ratios and a reasonable level of load impedance when plugged into line level inputs. If the line inputs are lower than about 10K Ohm input impedance, go up to a 2.2K Ohm shunt. The values are not critical, as long as all the attenuators use the same resistor values in the equivalent places.

I haven't recently been able to find for sale the unwired XLR barrel M-F adaptors designed for this type of thing. I bought some Canford ones many years ago, but I don't see them in their latest web catalogue. If you don't fancy stuffing resistors inside XLR connectors (the two series ones in the XLR-F and the shunt in the XLR-M), then you could get some short XLR patch cords, cut the signal conductors in the middle and insert the three resistors there, leaving the screen wire continuous to take the tensile forces.

 

[pcrecord]

Even if a 6db might be enough, I think I'd like to aim at 10 or 12 db. I like my tracks to be recorded at -18db specially on projects with 30 and more tracks.
Will it be too much ? I'm such a newb in electronic circuit. Do you have a schematic or photos of the circuit I'd need to build.

There is a store in MTL that has a lot of components, I'll check if I can fin the XLR Barrel.

 

[pcrecord]

If i don't find XLR barrel, I will do it in boxe(s) with 2 or 4 channels in the same box..
Chassis mount connectors aren't expensive !

Or if I find Xlr to TRS adapters big enough to be modified, it could also work since I'm going to TRS inputs anyway. I still have many TRS cables I used with the patch bay I don't use anymore...

 

[Boswell]

I would have thought that, seeing these attenuators would be used on the output of the ISA pre-amps, the tracks will be incorporated in the mix at a level appropriate to the mix. The level at an input is normally set by the channel's 0dBFS at the top end and the channel's noise level at the lower end, allowing sufficient headroom (e.g. 20dB). Over-attenuating at an input has the effect of raising the noise floor on that track since you have to use more gain to bring it up to mix level. I've been taken to task in the past about this, as some people argue that it's noise-free digital gain on a 24-bit converted signal, completely missing the point that the converted signal includes a representation of the analogue noise floor at the channel's input.

I'm assuming that the -18dB you mention is an r.m.s. level and not a peak. If you are using your FF800 as an ADC, the line inputs of that unit have an 0dBFS of +19dBu (on the +4dBu setting), so -18dBu input is 37dB of headroom! I'm not being critical, Marco, it's just that I'm not a fan of relating the analogue level going into your DAW to the number of input tracks in a mix.

XLR or TRS connectors are fine for this high signal-level work. Building the attenuator in a box with 4 XLR(F) inputs and 4 TRS jack outputs would be good, and it would do your XLR-TRS conversion for you.

I'll put together a table of suitable resistor values for various different attenuations in the range 3dB - 20dB.

 

[pcrecord]

I'm not being critical, Marco, it's just that I'm not a fan of relating the analogue level going into your DAW to the number of input tracks in a mix.

I get you !

You are very kind of helping me with this..

 

[Boswell]

Here you are:

 

[pcrecord]

Great !! It may take time before I get a chance to go to montreal and buy the parts, but I can't wait to try it and compare the sound !
Boswell, If I do a video about this, would it be ok if I show the schematics you gave us ?

 

[Boswell]

No problem about the schematics. If I get the chance, I'll look up some type numbers for the resistors. All resistors are not equal.

 

[Boswell]

While rummaging in my electronic archives for something else, I came across a photo of the innards of a small 12dB two-channel attenuator box I built in a hurry some years ago. This was for reducing the +32dBu output of my API3124+ pre-amps to +20dBu when I needed to feed the balanced ADC insert returns on my Audient ASP880 for recording purposes. The box also has switchable 0dB/12dB attenuation outputs for routing to a mixer for live use.

The inputs are on the left, the switchable outputs top and bottom, and the fixed -12dB outputs on the right. The switches are adjacent to their respective jacks. The series attenuation resistors are just visible mounted on the switch contacts, and the shunt resistors are across the signal terminals of the output jacks.

I see I chose standard metal film 1/4W resistors for the attenuator components, but they were probably what I had to hand at the time. In operation, the box deals well with +32dBu inputs.

 

[pcrecord]

WO ! this is some tight space work !!

 

[Boswell]

WO ! this is some tight space work !!

I thought that getting the two input jacks and the two -12dB output jacks into that little box would be tight. Then I came up with the need for mixer output jacks as well, plus the level switches. I remember it took a bit of working out how to get it all in there to leave the unswitched jack connection tags accessible.

 

[pcrecord]

Hey Boswell, I received all my parts and I thought I'd add a bypass switch to the design.

I got those switches that seemed to be a good choice, 6 pins ON/ON 2 Position DPDT :
https://www.amazon.ca/gp/product/B01D1D0UAC/?tag=r06fa-20

Would this make sens ?

 

[Boswell]

The switch looks fine, but I would wire it so that the low-value shunt resistor cannot be connected across the source:

 

[pcrecord]

Your THE man !! Thanks a lot, I was wondering about that resistor...
I can't thank you enough.. I'll be starting the build this weekend

Thanks

 

[pcrecord]

I took time to analyse your new design.
So instead of being at the output, the resistor will still be across the 2 and 3 pin but at the input and behind the 2 other resistors.
I get that the sum of all 3 resistors will be higher, but will it still affect the signal coming in ?
I'm planning the 10 db attenuation so 2.2k x 2 and 2k will be the values.