WIREGAIN-DC: PGA-2500 based preamplifier project for DIY

[Igor]

Hi people!

Let me introduce new funny and great sounding project-

WIREGAIN-DC

It is digitally controlled PGA-2500 based pre-amplifier,
very clean sounding, with excellent transient response, ultimately low noise, etc...
WIREGAIN-DC can be compared by specs and sound-wise to top range
clean commercial stuff in k$ range.

Gain: 0...75 dB, in 1 dB steps, controlled by rotary encoder and displayed on
2-digits 7-segment display or....IN-12 Nixie tubes in other test version.
-20 dB pad automatically inserted at 0...19 dB, and switched off at 20...75 db.
No clicks, no pops.

As pure benefit of digital control, PHANTOM switch is smart as well.
When pressing 48V button, pre's gain set to minimum, then phantom relay goes on,
then gain softly set back to original value with speed of 5dB/sec.

120 Hz High Pass Filter placed at the input of preamplifier.
PGA2500 has constant input impedance, and great feature - common mode servo.
It allows to use simple passive filter.  With a little trick, no clicks or pops as well.
BTW, most pre's using filter after gain stage...and I can't see any reason why to
amplify all this rumble-mummble then cut with active filter.

Next feature is kinda limiter, we called it OVRP.
When OVR button is pressed, controller watching OVR signal coming from PGA2500.
If  maximum level less 3dB reached, signal OVR changes from 0 to 1.
It used to take gain down. When OVR disappears, gain returned back.

Gain structure: input circuit-phantom-pad-HPF-PGA2500-balanced output stage (+10dB).

Pre-amp runs on +/-22v if NE5532 used at the output or +/-24V in case of OPA2604.
Power supply has +5v used for relays and +48V for phantom.
Maximum output level of WIREGAIN-DC about +26dBM.
It is unbalanced-compatible, i.e. when pin 3 of XLR shorted to ground, by changed feedback
"minus" don't care, and "plus" signal jumped up by 6dB remaining gain same
as in common line drivers like SSM2142, DRV134, THAT1606.

Schematics:

http://i251.photobucket.com/albums/gg291/diy33609/pga2500/contr_schem.gif
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/pre_schem.gif

(please note, some errors can appear beacause all you see on pictures
is preliminary version of full system's pcbs adapted for single pre-single controll purpose)

More images and video:





http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010370.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010372.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010374.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010376.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010368.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010379.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010381.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010382.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010383.jpg
http://i251.photobucket.com/albums/gg291/diy33609/pga2500/P1010394.jpg
http://i251.photobucket.com/albums/gg291/diy33609/in12.jpg
http://i251.photobucket.com/albums/gg291/diy33609/plasma2.jpg

Only 4 chips used in this pre-amp: PGA2500, OPA2604 (or NE5532), 
microcontroller and BCD-to-7-segment decoder.
Project is really simple at all. One only hard thing- is soldering the PGA2500 on pcb...
with some patience, and Legendary Pink Dots playing from LP on background, possible.

I checking the option to make this project DIY, and in case at least 50 pcb's
will be ordered, you will find it in my sticky list of projects at this year.

Price of PCB's and controller chip will be about EUR40-45 include shiping, with qty discount.

Are there people interested?

 

[promixe]

WOW!!! =)  Are there 4 boards in one panel/kit (preamp, psu, control panel, pic)?

What would be the character of the pre, comparing to some high-end stuff? "Very clean sounding" is like True Systems stuff? Prism Preamps?

As always - awesome stuff, Igor! You are the man! =)


PS: Are you getting my PMs? I've sent you two but since this is a new forum I don't know what's up...

PPS: What was the reason to not use OPA1632 in place of 2604?

 

[Igor]

Hi, yes, I got your PM, looong one, have to think on 33609 issue.
PSU will be separate PCB, because these pre's are great to
make cheap eights or fours....
Cannot compare it now to Prism one by one, but as I remember reminds
same surgical clean sound.
Compared to Millennia HV.
Liked them both as clean pre's....
On vocals, PGA2500 was really good on highs with U67.
I did not expected these results.
It was short test, for serious results, statistics, etc.
have to give the proto to good equipped studios to compare with,
and receive feedbacks from people.

 

[Igor]

PPS: What was the reason to not use OPA1632 in place of 2604?

Main reason, pin3-to-gnd compatibility.
Second reason, I really don't know how will bahave 1632 with long cable.
It isn't built as line driver.
Third, ability to use high power supply....
1632 can't handle more than +/-15V, at +/-16V it runs very hot.

 

[promixe]

Ok Cool!  What is a current draw per channel (preamp with 7-segment LED and pic boards together)?

Can we stuff 8 channels into a nice 1RU chassis with medium sized toroid trafo?

Is the PIC source code a closed source? Or can we have a "pic" at it? =)

 

[Igor]

What is a current draw per channel (preamp with 7-segment LED and pic boards together)?

About 40mA on each rail of +/-24V; 200-240mA max with very high brightness 7seg,
and all relays ON-current draw from 5V.

If IN-12 (Nixies) used instead of 7seg's, additional current draw for power converter 5->170V is ~120-180mA.

Can we stuff 8 channels into a nice 1RU chassis with medium sized toroid trafo?

Yes, it built for this purpose.
In this case, power supply have to be externall.

Is the PIC source code a closed source? Or can we have a "pic" at it? =)

Plans are each pcb set for pre' will include "burned" PIC.
Actually, the software is simplified version of 16-channel system we designing at this time....
Source code is closed, but maybe we will offer pre-programmed microcontrollers
separately from pcb's.

 

[promixe]

Thanks for your answers, Igor!  When can we start ordering, and more importantly, when can you be ready to ship the kit?

I could build a 2-channel unit and give it to some high-end folks around town, as well as test it myself, hell, even use it on a commercial project =) It's always nice to have feedback from engineers spoiled with expensive gear though =)


One more theoretical question: in case of a multiple channel design, wouldn't it make sense to use a single PIC to control all of the pres? How significant is cost reduction if you have 7 PICs less and a single PIC board instead of 8 individual ones?

 

[Igor]

Hehe...let's see what happens with this world's crisis caused by hadron collider
First, I need at least 50 pre-orders to cover minimal PCB batch costs and shiping etc.
If it happens within a week or two, just have to re-draw some wires, order pcb's,
receive first test unit, if all OK-in december 50 pcb's and PIC's are ready to ship.

I could build a 2-channel unit and give it to some high-end folks around town, as well as test it myself, hell, even use it on a commercial project =) It's always nice to have feedback from engineers spoiled with expensive gear though =)

This sounds really great. We have to talk about this.

One more theoretical question: in case of a multiple channel design, wouldn't it make sense to use a single PIC to control all of the pres? How significant is cost reduction if you have 7 PICs less and a single PIC board instead of 8 individual ones

?

Nope, cost reduse isn't so significant. Reason why we using "master-controller" system is only
remote TX/RX interface.

 

[promixe]

Yep, the collider is definitely at fault! How no one can see this is beyond me!  ;D  ;D  ;D

PS: I edited the post while you were replying to it, so there is that question about single PIC =)

I'd be the first one, I guess, to sign up for 2 kits! =)  I'll see, maybe later I'll add 8 more... but for now count in two

Thanks

 

[PRR]

> If  maximum level less 3dB reached, signal OVR changes from 0 to 1.
It used to take gain down. When OVR disappears, gain returned back.

What attack/decay speeds?

I want true-stereo: one knob and CPU controlling two analog channels. Would it be sufficient to OR the two OVR outputs, or are there other signals to combine? I don't have to daisy-chain SDO/SDI if I want both PGA2500 to do the same thing?

If I accept unbalanced 2Vrms max output, I can omit the opamps and +/-18V supplies?

 

[natek]

Igor strikes AGAIN! Wow this is very good job impressive! I would love to test one out in the studio! I am interested but will have to see where my $ is in the next week as i am trying to finish alot of half started projects and clear the board!

 

[e.oelberg]

Yes, I'm very interested. These preamps rock!! I could do also a direct comparison to the prismsound orpheus

nicholas

 

[Igor]

Yes, I'm very interested. These preamps rock!! I could do also a direct comparison to the prismsound orpheus

nicholas

You mean...

http://www.prismsound.com/music_recording/products_subs/orpheus/orpheus_spec.php

Just...by specs, it 99.99% using PGA2500 chip. See:

MICROPHONE PREAMPLIFIERS
Configuration Electronically balanced, with fully balanced analogue signal path
Gain 10dB to 65dB in 1dB steps (0dBFS = -56dBu to -1dBu)
Gain accuracy ±0.05dB
Input impedance 5.5kR
THD +10dB gain: -116dB at -0.1dBFS (0.00016%)
+40dB gain: -110dB at -0.1dBFS (0.00032%)
THD+N +10dB gain: -108dB at -0.1dBFS (0.00040%)

I can argue with gain accuracy...or my old good Boonton has some error while switching
beetween ranges, or PGA2500's accuracy is within 0.3 dB.
It was really smart to measure distortion on 10 db of gain....yes, wery usefull in real live,
in my  head just came perversive idea to record snare drum of heavy-oriented drummer
with TLM103.... Anyway, I can measure distortion as low as my analyzer's residue
(0.0012%).

> If  maximum level less 3dB reached, signal OVR changes from 0 to 1.
It used to take gain down. When OVR disappears, gain returned back.

What attack/decay speeds?

Prob is OVR signal holds for about 2-3mSec...
Then, attack rate is about -1db/msec, and release is 5db/sec.

I want true-stereo: one knob and CPU controlling two analog channels. Would it be sufficient to OR the two OVR outputs, or are there other signals to combine? I don't have to daisy-chain SDO/SDI if I want both PGA2500 to do the same thing?

No daisy chain, CS used to select crystal. I have to think on linking issue... IMHO possible,
OR is easy implemented with 2 diodes then 4k7 pulldown.

If I accept unbalanced 2Vrms max output, I can omit the opamps and +/-18V supplies?

Yes, all you need to do is to install some jumpers on PCB and run it on +/-5V.

Igor strikes AGAIN!

Please see my signature....

STRIKES

is absolutely inadequate definition

 

[matta]

Igor....

You are INSANE and I mean that if the nice possible way! Man too many projects, so little time... enough pres for now but can't wait till the 2254C compressor next month! Still after one of your lower cost monitor controllers if you ever get back to them... apparently there is race going one right now here: Class A transparent Control Room Unit

Again kudos, it looks amazing!

Matt

 

[flaheu]

F*****g how !!!!!

Igor, STOP !!! It's too much !!!

I'm more than    interested ?

I hope we can buy some anytime soon.

I'll for sure follow this thread very close.
 

 

[natek]

Yes!!! The video showing the gain pot with the display is sweet!

 

[Igor]

Matta, please show one who is NOT INSANE on this forum
posted some stuff on CRM tread as well..too much things to do.
BTW, any interest to build this PGA2500 thing with freaky IN-12 Russian Nixies?

(ME ADDICTED TO THIS NEON GLOW)

 

[natek]

BTW, any interest to build this PGA2500 thing with freaky IN-12 Russian Nixies?

Yes! I have been wanting to do a Nixie Clock but this would be much much better! ;D

 

[j.frad]

Hi!
would be interested in a pair... oh and nixies rock...

 

[Rochey]

Igor,

I'm very very impressed. Congratulations.

To all on here... The PGA2500 is an awsome chip - very very low noise, it'll most likely be one of the cleanest pre's you have.

Do me a favor though - buy them
http://octopart.com/info/Texas+Instruments/PGA2500IDB