My "mastering" D-AOC finally is done ! (recipe for antek power trafo inside)

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Silvas

Well-known member
Joined
Aug 15, 2006
Messages
727
Location
Chia, Colombia
Hello

What a nice compressor. Since the beginning, i wanted to turn it into a "mastering" compressor.

It has all kubi mods, but the output cap is not as big. I used a 22u solen fast.

Input and output transformers are Edcor XSM. 600:10K for the input and 15k:600 for the output. I ran a sweep while watching Waves PAZ, and it´s quite flat up to 20K (+-1dB). Need to do more adequate freq. response tests.

The power trafo, well, it is an Antek 1T250. It is working quite well, with some drawbacks. I stole AC from the heaters secondary to feed PSU3. No problems with that yet. PSU3 is 5V. 7805 is a bit hot, 14 volts at the input.

This trafo has 2x 6.3VAC + 2x 250VAC secondaries. the 250VAC secondaries are wired in paralell. I ended up with a bit high voltage, but after adjusting the voltage resistors on the TL783, i have 247VDC after the HV resistors. Close enough. With the 6.3VAC secondaries in paralell, the rectified voltage is way too low for the heaters reg...And with the 6.3VAC secondaries in series, the voltage is way too high...so, i used a LM338K, mounted on a quite big heatsink. Voltage before regulator is around 14 volts DC. Regulating that down to 6.3V generates LOTS of heat. Even with the big heatsink, the regulator was running too close to its thermal limit (had the thing on for a whole week with no thermal shutdown). I had two ideas : 1) put a power resistor before the regulator and 2) put a fan on the heatsink. I tried (1) but the power resistor was just before the reg input and after the filter caps (bad place i think) so heaters DC voltage was unstable (but the reg was quite cool and the resistor was quite hot). I wanted to try the power resistor right after the diode bridge but didn´t want to cut anything so...i ended up using a computer fan, fastened to the heatsink and the regulator is now happy. The fan is running on the +5V PSU3 and it´s quiet.

Anyway, the beast is running smoothly.

The meters are Nissei. Changed the LED´s for blue LED´s and love it. It´s a nice contrast from all that warm VU´s on the studio. The relay boards? The original relay boards (4 boards, for D-AOC and PM670) were stolen from my car (along with LOTS of parts). So i did a quick layout in eagle and etched them. They work as expected. If anyone wants the layout, please PM or email me.

All controls are 24pos rotary switches, the uraltone ones. Knobs are Davies 1919.

Front panel is laser engraved as on my D-LA2A. Case is also the same.

The sound ? Love it. Invisible compression unless you start to compress more than 12-15dB´s. Super smooth. Just what i needed for my mastering duties. Even with the edcor input iron. I´m going Lundahl input tx in the future. But for now, i´m happy. Love it on lead vocals and the 2-buss. Nice even when the input is driven hard. Very quiet, btw. Even with input and output full open, the noise is reading -78dBFS (0VU = -14dBFS at my interfaces).

I have a few questions for all you D-AOC gurus out there :

I have a small level mismatch between channels, 0,7dB. And the rotary switches are tracking levels perfectly between channels.  I want to fine tweak the output level. Which will be the best place to do so? i don´t think i can live with that level mismatch.

And...the stereo link. Stereo link is working...but matching between channels is not so good when linked (perfect when unlinked). Is there any way of doing "stereo adjust" like in the LA2A ? for now, i´m going unlinked.

Thanks Rowan and Volker for an incredible project ! Had a great time building it. And kubi for the nice mod recipe.

Now, the pics : (crappy camera, sorry)


PANELFRONTAL.jpg


DUALOPTOMASTERINGCOMPRESSOR017.jpg


DUALOPTOMASTERINGCOMPRESSOR013.jpg


DUALOPTOMASTERINGCOMPRESSOR010.jpg


DUALOPTOMASTERINGCOMPRESSOR001.jpg


DUALOPTOMASTERINGCOMPRESSOR005.jpg


DUALOPTOMASTERINGCOMPRESSOR019.jpg


DUALOPTOMASTERINGCOMPRESSOR025.jpg








 
Looks great, congrats!

What is the hand-drawn-like font you used on your faceplate?  I like it for the knobs.  :)  Could you share your faceplate graphic, or parts of it (knobs circles)?

Inspiring!
 
tommypiper said:
Looks great, congrats!

What is the hand-drawn-like font you used on your faceplate?  I like it for the knobs.  :)  Could you share your faceplate graphic, or parts of it (knobs circles)?

Inspiring!

Sure ! Front panel is done in Corel. The font is the same font i used on my D-LA2A. It´s the font from a TV show i like called "Heroes". Mail me and i´ll send what you need.

 
Nice build!

So you need 6,3V for the heaters right? Have you tried both 6,3V secondaries in parallel?
Full wave bridge + capacitive input filter gives you 1,4 X 6,3V = 8,82V.
Is 8,82V volt enough for the M338K to regulate down to 6,3V?
 
radiance said:
Nice build!

So you need 6,3V for the heaters right? Have you tried both 6,3V secondaries in parallel?
Full wave bridge + capacitive input filter gives you 1,4 X 6,3V = 8,82V.
Is 8,82V volt enough for the M338K to regulate down to 6,3V?

Nope. Definitely not enough. I tried. In theory it should but i was getting about 7VDC (before regulator).  :(
 
Looks cool!

I'd think you would love the sound even more with Lundhals in the input.
I recently did that with mine and it sounds so much better now!
I also have stepped switches instead of pots but for input and output i have 12 steps lorlins but 24 steps (elma) on the threshold though that needs smome finetuning since the response is a bit to rude between the steps; a lot of compression or not enough compression so if i need something in between i have to adjust the input as well.

Concerning the level mismatch; are you sure it's not a compression mismatch? mine had quite a difference between L and R which was down to the vactrols. Now with new ones (4 in total, kubi mod) it's totally flat. (ok.. maybe 0.2 db difference)
So you might want to check that first, made a big difference in my case and i supsect it has something to do with the batch numbers.
The old ones (with error) where from batch706, the new ones are different but i don't know what.
 
Noone has yet worked out a procedure how to find a matched pair of optos? I just recently started to think about this project and haven't yet read all about it. It should be easy to build a little test setup - or even use the finished pcbs as such - in order to measure and compare different optos?

Michael
 
[silent:arts] said:
wow :eek: :eek: :eek:
two such nice builds in such a short time, what are you doing ...  ;D
I'm impressed.

DAOC-Silvas.gif

Love my banner ! thanks for the compliments ! For me, it was a bit more difficult than the D-LA2A, but quite fun. I broke my poor wrists doing the VU holes  (i have CTS on both hands....forget about playing drums)

dagoose said:
Looks cool!

I'd think you would love the sound even more with Lundhals in the input.
I recently did that with mine and it sounds so much better now!
I also have stepped switches instead of pots but for input and output i have 12 steps lorlins but 24 steps (elma) on the threshold though that needs smome finetuning since the response is a bit to rude between the steps; a lot of compression or not enough compression so if i need something in between i have to adjust the input as well.

Concerning the level mismatch; are you sure it's not a compression mismatch? mine had quite a difference between L and R which was down to the vactrols. Now with new ones (4 in total, kubi mod) it's totally flat. (ok.. maybe 0.2 db difference)
So you might want to check that first, made a big difference in my case and i supsect it has something to do with the batch numbers.
The old ones (with error) where from batch706, the new ones are different but i don't know what.

Absolutely sure. Not a compression mismatch. In fact, maybe i'm too lucky but the tracking between channels is perfect ! and the tracking between measured GR and the meter GR is almost perfect, like 2dB off at  ca 15dB compression. Definitely a level mismatch in one of the gain stages.

I was looking at the schemo, and i was thinking, maybe i can tweak the 470k feedback resistors in the output (or input) stage to fine tune the gain ? or there's a better way? (the 49,9k grid resistors?)

I forgot to mention, this thing is quite sensitive. I mean, i can get full compression with the threshold in "5". Probably due to the 600:10k input tx and my hot levels.

 
On my build I just have the Threshold control. I put in trimmers  for input and output and use them for "level calibration". There isn't much between the signal input and the input attenuator. I would measure the level after the input attenuator on both channels to make sure there isn't a problem there. After that I would measure at the output with no compression. If the level is off maybe you could put  trimmers before the switches. Then check with various amounts of compression to see if the vactrols are reasonable well matched.
 
Looks great! My only tip would be to try Wima MPK instead of the Solen Fast caps, I found them to somewhat muffle the sound.

I also need to change the pots for stepped attenuators on my unit (Kubi mod as well).
 
I think that the unit would sound much better if you changed the switches on the front-panel to black color (or simply painted them).
 
living sounds said:
Looks great! My only tip would be to try Wima MPK instead of the Solen Fast caps, I found them to somewhat muffle the sound.

I also need to change the pots for stepped attenuators on my unit (Kubi mod as well).

Sure ! i think that experimentation phase is about to begin....

tv said:
I think that the unit would sound much better if you changed the switches on the front-panel to black color (or simply painted them).

How about some switch caps?  ;D ;D ;D


Gold said:
On my build I just have the Threshold control. I put in trimmers  for input and output and use them for "level calibration". There isn't much between the signal input and the input attenuator. I would measure the level after the input attenuator on both channels to make sure there isn't a problem there. After that I would measure at the output with no compression. If the level is off maybe you could put  trimmers before the switches. Then check with various amounts of compression to see if the vactrols are reasonable well matched.

Do you mean a trimmer as a voltage divider or as variable resistance? can you explain?

Mismatch is present even with no compression. 0.7dB, i think it´s just down to component tolerances. I used 1% MF resistors but didn´t match them to tighter tolerances...

 
Sorry if this is a stupid question but where are you guys getting these boards.... D-AOC, D-LA2A & PM660

I've looked everywhere... even silent:arts website...

Was this a group buy and has that ship already sailed?

Just curious...

J
 
Back
Top