My New 1U GSSL Design (AKA SB4000) RELEASED!!! - GOTO PAGE 9 FOR ORDERING INFO!

[DanH]

Count me in for one set!

 

[ruckus328]

OK, so unit is working now.  There were a number of changes I had to make for the Black Can DBX202 to work (surprised how little talk there is on the board about this).

I'm trying to figure out the best way to incorporate provisions for these modifications.  At the moment I'm just driving the CV input directly, and not buffering it through a 5534 as is done with a THAT2180/81.  Reason being is that the 5534 for driving the CV input is underneath the DBX202.  I'll leave provision so you can do it both ways, however if you do drive it off of the 5534 you'll need to stand the can off of the board to clear it with millmax pins or similar.  I was hoping to avoid having to do this kind of stuff but there's really no way around it.  I still have some testing to do but seems that running it directly without buffering is not having any noticable effects, but I have to do some listening.

Next thing I'd like to get some opinions on - the bargraph.  There's 2 versions.  The one I'm using has the top 3 lights red, next 4 yellow, bottom 3 green.  This means however that the "0" level is a yellow light, and the first red doesn't light up until +1db.  This is if you have it set as a classic VU meter where the top light would be +3db.  It would be possible to skew each LED by 1db (I'm pretty sure this will work, my brains fried right now though), so that the red light lights up at 0db, and it maxes out at +2db, and the bottom LED indicates -21db.

So, as opposed to a standard VU it would read like this:

+2 (Red)
+1 (Red)
0 (Red)
-1 (Yellow)
-2 (Yellow)
-4 (Yellow)
-6 (Yellow)
-8 (Green)
-11 (Green)
-21 (Green)

There is an alternative bargraph which has 2 red, 2 yellow, 2 green, 2 yellow, 2 red.  This means that it would remain a standard VU range and read like this:

+3 (Red)
+2 (Red)
+1 (Yellow)
0 (Yellow)
-1 (Green)
-3 (Green)
-5 (Yellow)
-7 (Yellow)
-10 (Red)
-20 (Red)

Thoughts?  Opinions?  I need to make a decision on this as my front panels need to be labeled accordingly.  I'm leaning towards the top solution, as this keeps the 0 indicater a red light, which is typically what you'd see on equipment.

 

[Kingston]

There were a number of changes I had to make for the Black Can DBX202 to work (surprised how little talk there is on the board about this).

What this also allows people to do is to use custom VCA configurations made using the DBX VCA footprint.

For example I never liked the somewhat noisy and smudgy DBX202 and always thought a single buffered THAT2180/81 sounds better.

...until I tried 4X THAT2181A in parallel. I created a quick proto PCB just for this, buffered it with LME49710. There's probably no better VCA available anywhere. Sure 4X THAT2181A sounds expensive, but it's still cheaper than some flaky used DBX202 from the black market. And it's technically superior in every way. Sounds obviously better to me as well, completely devoid of the vintage mud and noise. Not a good upgrade for the people who actually like the standard SSL 4000 crud of course.

related discussion and schematics: http://www.groupdiy.com/index.php?topic=1250.msg16279#msg16279

 

[ruckus328]

What this also allows people to do is to use custom VCA configurations made using the DBX VCA footprint.

Yes that was the plan, problem is, trying to make it so I have the provisions to easily use any type of VCA = nightmare.

THAT2180/2181 - this is the default I designed to because 95% of people will use these.  Everything is good there.  I have a table on the schematic for different supporting component values for the different grades.  The default values on the schematic/silkscreen are for the THAT2181A, under the logic that this is probably the most used VCA.

DBX202C (Gold Can) - Couple value changes (Values needed are instructed on the schamtic) need to be made but other than that it's drop it in and it works.

DBX202 (Black Cans) - Some value changes but 2 resistors need to be added that are under the shadow of the VCA.  I don't have enough room to bring them out of there.  And this is a sensitive area with routing so I don't want traces coming out and back under the cans again either.  How I did it for my prototype was installed these 2 resistors on the backside and made sure my leads weren't long.  Made it so I didn't have to raise the 202's.  Works and in my opinion is acceptable.  Going to add a jumper so you can run the CV input direct into the 202 like I did or if you don't install the jumper you can run it off of the 5534 buffer but then you'll have to raise the can like I mentioned above.  This jumper will also be on the backside with clear instructions what these 3 items are for.

Of course there's the ability to use the parralleling VCA boards too, since they mount into the 202 footprint (Is anyone even selling them still though?)  I wanted to get a few but didn't see anything on the white market.

 

[Ptownkid]

Of course there's the ability to use the parralleling VCA boards too, since they mount into the 202 footprint (Is anyone even selling them still though?)  I wanted to get a few but didn't see anything on the white market.

Easy way to fix that one...on it.

 

[Kingston]

Is anyone even selling them still though?

Easier just to draw your own PCB while you're at it. That THAT schematic is very simple. Trimmers in place for the post-VCA buffers feedback resistors are a good idea for perfect gain matching. If you use THAT2181 you can skip all the VCA trimmers seen on the schemo.

 

[Neil]

Thoughts?  Opinions?  I need to make a decision on this as my front panels need to be labeled accordingly.  I'm leaning towards the top solution, as this keeps the 0 indicater a red light, which is typically what you'd see on equipment.

I like the top one as well, seems more instinctive.

 

[Ringwraith]

^me too

 

[ruckus328]

^me too

Yea, I already to decided to roll with it.  Front panel & schematic are updated.

 

[ruckus328]

OK, so done testing Rev 2.  Analysis was done with RMAA using my GSSL as a comparison.

For ratios I got Ratios:

2 : 1.1

4: 1.9

10 : 0.9

Noise Floor is nearly identical to my CRC GSSL.  Al functions are verified operational.  Bargraph values have been skewed so my front panel now reads

+2
+1
0
-1
-2
-4
-6
-8
-11
-21

This puts the first red LED indicater at 0db.

The only place I score lower is crosstalk.  Overal score is inline with my GSSL but when I look at the graph - L>R crosstalk exceeds my GSSL across the board, however R>L is about 10db higher at 15kHz.  It's at -80db at about 8kHz, -78db at 10kHz, and peaking at -73db at 15kHz, and then quickly drops off, back down to -80db at 20kHz.  This was improved upon from Rev 1 through different pcb routing where it peaked at about -66db, as that number is on the high side imo.  From my understanding, at these frequencies the -73db peak number is still completely acceptable however I'd like some 2nd opinions on this.  Can't really find much info as to what is definiately acceptable.

I'm going to increase gnd plane isolation slightly in the audio section for the production version in hopes to bring the number down.  It annoys me more than anything.

 

[DanH]

Hey just want to commend you... its clear you're really committed to this project. Looking forward to starting a build!

 

[Ptownkid]

Indeed! I can attest to the fact that Mike has put a lot of effort into making this a GREAT revision on many fronts. This will be a fantastic project in the end and will be well worth the wait.

In case anyone has wondered...kits will be available, just trying to finalize some minor decisions on a few things. We're trying to keep the cost down without compromising on the important pieces of the puzzle. That being said, the kits won't be rock bottom cheap, but they will be very reasonable and high quality in all the important areas...

stay tuned.

 

[evilcat]

In case anyone has wondered...kits will be available, just trying to finalize some minor decisions on a few things. We're trying to keep the cost down without compromising on the important pieces of the puzzle. That being said, the kits won't be rock bottom cheap, but they will be very reasonable and high quality in all the important areas...

That's a very good news ! Now I need to be patient, patient...

 

[rrs]

Of course there's the ability to use the parralleling VCA boards too, since they mount into the 202 footprint (Is anyone even selling them still though?)  I wanted to get a few but didn't see anything on the white market.

Easy way to fix that one...on it.

Always wanted this. Keep us us the loop on this please.

 

[Igor]

Cool. 1:1.5 ratio and C&B are IN, IIRC, in your design.
Very nice space saving layout. Like it!!!

 

[Ptownkid]

Wow...just wow. Igor, do you have absolutely zero respect for anyone other than yourself?

 

[ruckus328]

Nope, but doesn't matter at this point.

 

[diyfanatic]

Evil spirit leave Igor's body, I command you....

 

[3nity]

I think human ambition will go to war for money.
I dont blame Igor.

Thanks

 

[ruckus328]

Final High Res Front Panel attached