Salas DCG3 preamp (line & headphone)

diyAudio Chief Moderator
Joined 2002
Paid Member
Thanks for the responses. I think I will go high with the bias, and think about a switchable gain giving me two positions. The sockets are tight and were recommended for Pearl 2 resistor changes. Not very elegant, I will explore some sort of switching...for the gain.

Russellc

Maybe a DIP switch and a couple of resistors on a thumb sized board directly to the R6 pads. Don't ever switch the gain when powered on. Also test that the offset is good on all gains.
 
Member
Joined 2003
Paid Member
*BTW I extrapolated the Nighthawk's 1V RMS sensitivity from an Innerfidelity's mesurements sheet I found online. There it took 0.098V RMS to reach 90dB SPL calibrated. Because 20log 0.098V is minus 20dB to 1V then 90+20=110dB SPL/V

I know they have in mind driving it with their Dragonfly DAC, so I figured they were plenty sensitive. BTW, I have both the original Dragonfly as well as the newer Red Dragonfly. Storing my CDS on the laptop and utilizing a free "MusicBee" software (which allows by passing the microsoft sound stack of mess and use of a WASAPI output), sounds way better than it has any right to, given the price

Russellc
 
the above diagram has only balanced in balanced out, SE in SE out but not BAL In SE out.

All you have to do is connect a balanced source at the input XLR and take SE output at the RCA. This will be one phase (positive or negative). The other phase will simply stay dissconnected.
 

Attachments

  • Rear.jpg
    Rear.jpg
    430.4 KB · Views: 437
Perhaps it should be made clear that it's not a differential configuration. We are not talking about CMRR here. It's only a four rail line stage to amplify a weak balanced signal for the requirements of the next stage. CMRR will hopefuly take place there (the amp) provided it has differential input.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I know they have in mind driving it with their Dragonfly DAC, so I figured they were plenty sensitive.

My 110dB/V estimation was based on independent voltage test data. If choosing the manufacturer's claim of 99dB/mW it translates to 115dB/V for 25Ω. That's a very high level for 1V RMS drive. I had similarly sensitive Sennheiser HD8 DJ around and they were subjectively twice as loud on the volume control than the HD600.

But testers use standard spot frequency sound pressure cal or a standard span's average while manufacturers may choose to quote the most sensitive audio spectrum's band or frequency on each model.
 
Member
Joined 2003
Paid Member
Personally, I would choose 3x especially if you get another set of headphones. 3x is just right with f5 turbo (20db gain in amp) a little to much gain for grados. 3x is perfect for akg 7xx. Headphones vary a lot. Those audioquest headphones are meant to be driven from a phone...so not much gain needed.

Mosfets theoretically sound best with more current so I’d max it out if you have the heatsink and your power supply will accommodate it. And with 25ohm headphones you want as much current as you can get.

The power supply will be the dual version from Teabag, the two transformers are the recommended Antek 18 + 18 @ 50VA. Is this sufficient?

Russellc
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The power supply will be the dual version from Teabag, the two transformers are the recommended Antek 18 + 18 @ 50VA. Is this sufficient?

Russellc

Yes. More than enough.

Each transformer is rated .5A. You want the load to be less than half the transformers rating ideally.

So you are in good shape. I run mine with a single 50va transformer at 130ma with no problems.
That could be their rated primary current? They are rated stronger for secondary current I believe.
 
Iam listening to mine now at work.

Sounds damn good. Seriously.

One thing to consider is if you are building for headphones and don’t listen loud (at work I can’t listen loud) you want less gain so the volume pot is not always set so low. Many volume pots have inconsistent tracking between channels at the beginning of the pot.

I have a stepped attenuator in mine so it’s not a problem. But at work I only use 1/4 of the dial with much less efficient headphones than the audioquest. At home I use 3/4 of the dial with akg 7xx.
 

Attachments

  • 76FF6B76-676A-4CCE-BF92-C346AB1CEB30.jpg
    76FF6B76-676A-4CCE-BF92-C346AB1CEB30.jpg
    564.7 KB · Views: 370
diyAudio Chief Moderator
Joined 2002
Paid Member
Just do the initial electrical checks carefully and it will be easy to finish it on first shot with success. Almost all builders managed so.

Its correct you used the original current mirror transistors so to first know the standard sound after 48 hours of play giving a chance to electrolytics for shaping up. It takes you explore well enough the new preamp's impact to your system before attempting to compare next of kin BJT types differences that could prove too subtle or not in this set.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Iam listening to mine now at work.

Sounds damn good. Seriously.

One thing to consider is if you are building for headphones and don’t listen loud (at work I can’t listen loud) you want less gain so the volume pot is not always set so low. Many volume pots have inconsistent tracking between channels at the beginning of the pot.

I have a stepped attenuator in mine so it’s not a problem. But at work I only use 1/4 of the dial with much less efficient headphones than the audioquest. At home I use 3/4 of the dial with akg 7xx.

True, most people will listen at a 90dB SPL median with 100dB SPL peaks and will describe that as loud enough. The Nighthawks will be at 90dB SPL with just 0.1V and at 100dB SPL with 0.32V when using the strictest sensitivity estimation we could find for them. Maybe Russell will have to include a 2X setting also to evaluate with them during his planned gain tests.
 
Hi
Have now power on the pre

1 channel do not work right ,,,to high DC on output

it was a fault 9610
i mounted 2 new..and then it looks alrigth

witout the servo ic i have about zerro to 10 millivolt...not bad

with the servo ic zerro to 2 millivolt

also the volt over R10 is 119mv on both channel

now it is time to mount the in and output

Best Bjarne