MixBuzz500 build/support tHread

[Igor]

Hi to all, I had crazy busy week than nice but short vacation.
Back to our Mixbuzz500 stuff.....

but i definitely can see some smoke residue on the sides of the 1246 on each boards
(Rectification of the last post, it's the 1246 not the 1646!)....

OK....I think +48V somehow passed to input. Bad to hear about this.
Again!!!!, tabs for 51X rack, and never power on the rack before the module sits perfect inside the slot!

Well i think one of the DBX is dead  
Cry, without the chips but DBX in i'm reading 27v on the V+ pin,
i really don't get what happened, everything is fine in the box....
and the unit was working properly in the lab...

Something wrong happens inside your 51X rack, or maybe the module inserted wrong way.

+48V is far from input/output connectors. Strange.....

PS: I think there is a way to check the DBX, like its resistance?

Yes. Take it out from PCB (desolder). There are I_IN and I_OUT pins of VCA.
Put in-between "fake unity gain resistor" with value=R3-R6. I.e. if R3=47K, and R6=jumper,
use 47K (DBX202C version); if 202X used, and r3=5k, r6=3k74, use 1k25 resistor.

Put audio board inside the rack, check if all sits fine, connect control board,
only after checking everything is right, power on.

Audio board's gain should be 0db (audio and control boards should be connected!).

Than, check the voltage at VCA's control signal pin, without makeup/compression.
It should be 0V. Power off, take "fake unity gain resistor" out, put VCA. It should work.
If not works, VCA probably (99%) is dead.

hmmm, i Absolutely don't get it, i just connected the boards alone in the rack,
one board with chips out, all had good voltages on right pins,
and tried with the other one that had the components all in, and everything is fine,
no crazy voltages on the pins or whatsoever,
DBX getting correct voltages all Socket pins getting correct voltages and no Smoke  
Huh Huh Huh Huh Huh Huh Huh  

Good news. Use crocodile for -(common) probe of multimeter (in case of whatever);
maybe previous measures were wrong?

at least on one of the boards and Controller board not connected,
didn't check sound passing still but very Weird issue,
maybe a pin got shorted when inserting the card  
Huh, Now i need to check all the configuration in the box and not out of it,
just to have a better look of that.

No, 48V is far from input/output pins. Very strange. Maybe.....
In case rack was powered, module got first (somehow????) +48V than the rest of voltages....
Kill me here, impossible, if all voltages reading OK at the card edge.
If module was reversed (pin 1 goes to pin 15)...but same, impossible!!!!!
Tony, I thought about this issue, one more hour and I will smoke-up like your THAT1246!!

BTW. The audio board's relays are in BYPASS in default state.
This means, to check the audio board, you have to connect control
(front panel) board. For ROUGH check, enough one audio and control board.
For normal test, BOTH audio and control boards should be connected!

/in case of whatever:
CTL1, SC1 goes to CTL and SC left, CTL2, SC2 goes to CTL and SC right.
Or vice verse, but don't mess say CTL1 with CTL2. This way compressor will work wrong.
See first posts of this thread.

One of the ribbon cables had got damaged and the sidechain was going nuts.

Guys! What happens! Somebody PUT A SPELL on those freaky ribbons!!!!
2 ribbons FUBARu'ed for one week!!!!!!
C'mon, easy with'em

any suggestions on the silver mica caps?

Will try to answer in Zen weird way
OK....The story is next. (a bit OT!) Me and my friend from Bardo Mastering
tested Sontec kit stuffed with different kinds of caps some years ago.
Relay based A/B box (my mastering console) and ATC monitors.
Monster/Kimber/etc super-duper BS wiring in case of whatever,
but I don't trust any hi-end cables except military silver-plated teflon isolated,
which never were rated as hi-end audio cables.
We found immediately that EQ stuffed with SilverMica "anti-oscillation",
or "feedback" caps, sounded better than other with NPO's, 63V.
I can hear it as more definite highs. More details. Less fog.
Again, I hate sound talks like punchy-crispy.
So, I ALWAYS use micas at important places.
Second reason, I have about 2 kg of low value SilverMica's from surplus
and (or thus) may be wrong
Silver micas mostly have 7...10mm spacing for 47...82pf (C1); they fit fine on PCB,
maybe, bend leads a bit.

some protection against 48V and RFI is suggested in the date sheet.
I would strongly recommand this modification for reliability:
look at page 9 of:  http://www.thatcorp.com/datashts/THAT_1606-1646_Datasheet.pdf

line recievers/drivers:
THAT1246  IC1: some additional  RFI protection look at page 4 and page 5 fig 4 of :
http://thatcorp.com/datashts/THAT_1240-Series_Datasheet.pdf  .
As suggested do this directly on the connector.
The IC is protected against +48V on the input pins, so no worry there.

Up to you. I personally hate "useless" or "additional" components.
Audiopath should be short and simple as possible. Diodes can be installed on pcb,
there's a plenty of space. THAT1646 doesn't needs capacitors at it's out.
THAT1246's out bypassed with capacitor and it has -6db gain, so, no need for input capacitors as well
even in case of little DC offset at out of device connected to Mixbuzz500 input.
"Fool protection": install diodes. I prefer to check everything before putting module
in the rack and sleep good with mind of lack of additional components in my pure soundpath Mixbuzz

I also suggest to lower R7 to 1K for better S/N ratio.

Nope. IC8 (THAT1646) see at it's input: R7 in parallel to impedance of C3/C5, plus R5=68R,
plus output impedance of IC11B. It is less than 100R in 20-20Khz.
Lowering R7 to 1K: will give audible phase shift at low freq's and little (some db) drop at 20Hz.
Those who good with math can calculate/simulate. No noise floor improvement.

The noise floor actually caused by VCA's impedance...
I.e. using 4 paralleled 2181's and R3=5k IMHO not too bad solution

Somebody tried to leave out (bypass) the coupling caps in front of the linedriver?

Need some clarification, but, if you mean C3/C5, there will always be some tens of millivolts
of dc offset at IC11b out,even in case VCA perfectly calibrated. VCA gain changes, offset changes a bit as well. Little DC offset appears especially when VCA gives positive gain.
IMHO, MixBuzz500 using necessary minimum of parts,
mods and tweaks are at your own rescue and taste, but....feel free to tell if there some ideas
In any case, this is build support thread;
for "mod" people:  first build and make it working 100%, than mod at your rescue

 

[zayance]

Tony, I thought about this issue, one more hour and I will smoke-up like your THAT1246!!

Sorry about this, i found it super weird too, since everything was working good on the bench, and the rack inserted i believe well? Still pretty tight to see tough, but full way at the back and the front panel sitting nicely, i've measured the entraxe of the cards before fitting in....
I'm waiting for some 1246 to exchange, and retry on the bench with and without DBX's and your tip, and carefully will try boards out of metal chassis, in the rack afterwards and we will see.
Thanks a lot Igor.

 

[reanimatorstudio]

Hi Igor, thanx for popping in, I was waiting for some decent explanation.

Quote
some protection against 48V and RFI is suggested in the date sheet.
I would strongly recommand this modification for reliability:
look at page 9 of:  http://www.thatcorp.com/datashts/THAT_1606-1646_Datasheet.pdf

line recievers/drivers:
THAT1246  IC1: some additional  RFI protection look at page 4 and page 5 fig 4 of :
http://thatcorp.com/datashts/THAT_1240-Series_Datasheet.pdf  .
As suggested do this directly on the connector.
The IC is protected against +48V on the input pins, so no worry there.


Up to you. I personally hate "useless" or "additional" components.
Audiopath should be short and simple as possible. Diodes can be installed on pcb,
there's plenty of space. THAT1646 doesn't need capacitors at its out.
THAT1246's out bypassed with capacitor and it has -6db gain, so, no need for input capacitors as well
even in case of little DC offset at out of device connected to Mixbuzz500 input.
"Fool protection": install diodes. I prefer to check everything before putting module
in the rack and sleep good with mind of lack of additional components in my pure soundpath Mixbuzz Smiley

I agree with you but in real life, it is imaginable to patch preamps inputs with phantom power on, on outs and ins of other outboard. I have repaired more than one outboard piece ruined in this way. So like you said; "Fool protection": install diodes

Quote
I also suggest to lower R7 to 1K for better S/N ratio.

Nope. IC8 (THAT1646) see at it's input: R7 in parallel to impedance of C3/C5, plus R5=68R,
plus output impedance of IC11B. It is less than 100R in 20-20Khz.
Lowering R7 to 1K: will give audible phase shift at low freq's and little (some db) drop at 20Hz.
Those who good with math can calculate/simulate. No noise floor improvement. The noise floor actually caused by VCA's impedance...
I.e. using 4 paralleled 2181's and R3=5k IMHO not too bad solution Smiley

You are right again. I'm just sceptic if I see resistors with high values.

Quote
Somebody tried to leave out (bypass) the coupling caps in front of the linedriver?
Need some clarification, but, if you mean C3/C5, there will always be some tens of millivolts
of dc offset at IC11b out,even in case VCA perfectly calibrated. VCA gain changes, offset changes a bit as well. Little DC offset appears especially when VCA gives positive gain.
IMHO, MixBuzz500 using necessary minimum of parts,
mods and tweaks are at your own rescue and taste, but....feel free to tell if there some ideas Smiley
In any case, this is build support thread;
for "mod" people:  first build and make it working 100%, than mod at your rescue Smiley

You see, again it is good that I asked it before. Your explanation makes sense so I don't have to lose time and try to mod it.
Damn it's a good design. There is no room for improvement!!!

 

[Igor]

Ops. Never pretended
Always there is room for improvement. But please, build unit first than modify

High value resistors (in continue to R7 discussion):
it used to pull line driver's input (that1646) to ground after decoupling capacitor.
I fully agree, sometimes high value resistors causing us think about noise
In any case, "worse" impedance-wise point in this compressor is VCA.
The rest of places in Mixbuzz500 kept on low as possible impedances.
IMHO, the total noise of MixBuzz500 is way lower as GSSL, but
this is different story.

BTW. How many people have their units up and running?
Tell us!

 

[reanimatorstudio]

But please, build unit first than modify Smiley

I agree completely. But we want to learn something. And while building, some questions come to mind. This forum, in this case again you Igor, is a nice help to succeed in this. That's the beauty of DIY & a good forum. Something that was impossible in the past!!
Thanx technology, thanx Igor.

 

[dandeurloo]

I just had a friend give me a sifam meter.  It looks like it is the correct one for this build but I'm not sure.  It is the presentor shape, in the bottom left lower corner of the meter it says 19.  In the right bottom corner it says 2.5.  The scale is 0-100 currently.  On the top of the housing of the motor it says 12V / 20mA.

Anyone know if this sounds like the correct meter.  If so I just need a different bezel.

Thanks

 

[Igor]

Hmmm... Not enough info
Dan, please send photo by mail, will try to help.

Tony: crossed the fingers, waiting for "Yes, it works!" from you.

 

[dandeurloo]

email sent.

Thanks Igor

 

[dbeeson]

I spent my long weekend building this work of art and I'm nearly completed.  I have a couple of noob questions that I hope to have cleared up.  Thanks again Igor for such an awesome project.  The PCBs are great to work with!

Diode:
Am I ok to fit an IN4003 for D1?  I only have IN4001 and IN4003.

JP1:
I'm assuming this is a simple jumper?  I stuck a resistor leg there.

Hairball meter:
This meter has led's already inside the meter, and has 2 tabs on the back for power.  I know I need to take a feed from led 2,3,4 network, but where? 

THAT 2180c for sidechain:
These don't fit in the sockets too well, do you guys clip a bit off the legs to fit them in?  I'm worried they'll just fall out if the unit gets knocked. 

What do you guys use to cut the shafts off the lorlin switches?  A rotary tool?

Those are all my noob questions.  I want to be sure of everything before I do the final assembly/calibration.  I can't wait to start smashing drums with this thing!

 

[AudioGimmer]

Hope I can help with some of your questions [only the simple ones! ]

1. Yup, JP = jumper: resistor leg is good
2. THAT 2180C and sockets:
Well, I used the 2181 and they are tight to get into the socket. They do pretty much go all the way in though. If it feels like it could fall out, its not in properly. What did you use for sockets?
3. To cut the Lorlins I used a hacksaw, just a small cheap one.

Re. diodes and meter, I'm not sure, but hopefully someone will answer those for you.

The Gimmer.

ps. Igor, I used my Mixbuzz over the weekend on a session and it ROCKED!!

 

[stereokillah]

dbeeson, 1n4003 is ok  in the bom the posibilty begin with 4002 so you could put all that is more than it 4003 4004 but not 4001.

for the led vu, i propose you  to forget the inside led at time and put big led on the board, like taht will be easyer to change if one blow

 

[dbeeson]

Thanks guys! I managed to get the 2180s in the sockets...just wasn't pushing hard enough I guess.  Nice and tight now. 
On the audio boards, am I supposed to ignore CNB and G pads?

 

[reanimatorstudio]

look at the datasheet all these diodes are the same except voltage rating

page1:
http://www.elfak.ni.ac.rs/phptest/new/html/Studije/predavanja-literatura/osnove/pdffiles/1N4001-D.pdf

They all work. I take 1N4004 or higher. I never take lower numbers in stock, don't see the point of them still existing

 

[Igor]

On the audio boards, am I supposed to ignore CNB and G pads?

C&B pads provided for connection to external C&B.
If you don't use C&B (it is not issued yet, so probably... answer Yep
the bottom layer SMD pads should be shorted...that's it.

Sorry, was crazy busy these days, it is almost morning here and I found some time for forum;
big thanx to all you, guys, for support!

C&B with M/S: coming....soon! Around November, if all goes OK, first proto' will work.

The Gimmer.

ps. Igor, I used my Mixbuzz over the weekend on a session and it ROCKED!!

This did me the day! Really, like a good pipe filled with Latakia smoked at my balcony!!!!
Or, even more.
/a view from here:

 

[dbeeson]

Thanks Igor.  Not sure which SMD pads you're referring to but I'll have a look today.  I finished the build last night, fired it up. No errors, looks great!  I will take pictures today, I just need to fit some knobs.  Then I'll start putting it through it's paces. 

A quick tip for calibrating with the 500 series lunchbox.  Use long molex cables as stated by Igor.  Remove IC11 and cut a small lead (4/5cm), strip and insert into socket 7. Install IC. You can now attach your probe to the tip of the lead. This is much easier than trying to probe Pin 7 if you want to calibrate it in the rack.  I calibrated both boards in 15 min with this trick. 

Once again Igor, Awesome project.  It was even simple enough for a complete noob like me to put it together correctly.  You are gifted man!

 

[reanimatorstudio]

Not sure which SMD pads you're referring to but I'll have a look today.

Crush and Blend option
Quote
SJ points. Mixbuzz done with thought it should be modular, and SJ points reserved for
C&B option. If there's no C&B, these points should be connected, like picture shows.
http://i251.photobucket.com/albums/gg291/diy33609/mixbuzz500/P1040223.jpg
http://i251.photobucket.com/albums/gg291/diy33609/mixbuzz500/P1040224.jpg

 

[dandeurloo]

Hey guys, noob questions here.  I am still slowly sourcing parts for a US build.  Seems a lot cheaper to use parts from mouser to do this.

1. Any reason not to use these epco caps in the sidechain?  I just want to make sure they are quality enough for this build.  This would be the cap for c27 for instance.
http://www.mouser.com/Search/ProductDetail.aspx?R=B32529C223K189virtualkey59250000virtualkey871-B32529C223K189

2. Also, I have a number of higher value uf caps and voltage that will probably fit in the PS bypass section.  I can use those correct just not lower value or voltages is my understanding.

Thanks guys

 

[zayance]

Any reason not to use these epco caps in the sidechain?

No reason

 

[dandeurloo]

Thanks, I can't believe how much cheaper those caps are then the other brands.   

Also, while at it is an AD713 a good chip for IC3 or no reason to use anything special here?

How are these sounding?  I can't wait to try it on some drums and subgroups. 

 

[reanimatorstudio]

Some confusion when installing the VCA boards and looking at Igor's prototype:
http://www.groupdiy.com/index.php?topic=39459.msg487509#msg487509
just to make shure expencive VCA's are not blown away.
In the new kits/PCB's the VCA boards should not be mounted up side down like in Igor's picture but they should be mounted like in my picture, correct?
At least that is what I measure