MIDIBox Mixer - A PGAx311 Audio Mixer

[lylehaze]

It seems we have a few different projects building on the PGA2311 and PGA4311 Audio Mixer Chips.

http://focus.ti.com/docs/prod/folders/print/pga2311.html

Already we have a desire for a small self-contained mixer, a controlless mixer backend, and a foundation for a more capable effects unit.

Variety is a strength, but what I'd like to try is to combine what we can between these projects.

Perhaps we can all agree on the same pins to control the clock and data, so we can all use the same low-level drivers. Are others going to want control over the ZCEN and MUTE inputs, or will these be hardwired

into a single state?

Another issue we will all face is where to get the -5Volts for the chips. And do we want to share the same digital +5 with the analog +5, or should we use separate regulators to keep the costs down?

Mixers? Easy enough to lay out a board (probably separate from the PGA board to be configurable). What board design would serve the most of us with the least excess?

It looks like we all want different things, but let's try to figure out what the common parts are, so we can join together!

Please Comment..

Lyle Haze

[stryd_one]

Some other conceptual modules that have been mentioned: optical IO, balanced IO, matrix switching, summing, PSU.

A few thoughts off the top of my head to get the ball rolling:

It seems we have a few different projects building on the PGA2311 and PGA4311 Audio Mixer Chips.

http://focus.ti.com/docs/prod/folders/print/pga2311.html

I think the 4311 is not really a good idea as it is only available in SMD packaging, for the DIY crowd it's not worth it for the price of extra space on a PCB :) I should note that there is a compatible IC - the Cirrus CS3310-KP, but the 3310 is no longer produced in PDIP, so the -KS variant will not work, it is SOIC.

Already we have a desire for a small self-contained mixer, a controlless mixer backend, and a foundation for a more capable effects unit.

Variety is a strength, but what I'd like to try is to combine what we can between these projects.

:)

Perhaps we can all agree on the same pins to control the clock and data, so we can all use the same low-level drivers.

Pins can be swapped by changing defines, it's very simple, so that shouldn't be a concern. We would need to decide how many pins we need, which is a matter of whether we need core module to control the hardware mute or the ZCEN pin

Are others going to want control over the ZCEN and MUTE inputs, or will these be hardwired

into a single state?

Personally, I'd say pins and jumpers to hardwire them. That way, it's easy to connect a wire to the pin for hardware control if it is required, but if it's not needed you can solder a bridge in there or just jumper it.

Another issue we will all face is where to get the -5Volts for the chips.

I think a dedicated PSU might be a good call here... Noise will be a touchy issue. There are some designs around but they are untested at the moment, although I would imagine that our brethren in Synth DIY would have that taken care of ;)

And do we want to share the same digital +5 with the analog +5, or should we use separate regulators to keep the costs down?

Definitely separate supplies for the sake of noise.

[stryd_one]

Another thought - the SPI buss could also be cascaded to control switching circuitry and digipots (eg AD5204) for other, related modules like routing and control.

[lylehaze]

OK, I'll start with some things I forget to say before..

I am very very new to Midibox and MIOS, so if it looks like I've got

something wrong, please SPEAK UP! I'm here to learn.

Also, Pilo was nice enough to share the work he has done so far on

the PGA4311 with me. Big Thanks Pilo! I'll do more testing when I have

boards built. I'll share with anyone who asks for it, unless Pilo asks me

not to.

Re: Chip Selection.

I agree that the 2311 is a better choice for most people. But for

myself, I'm sticking with the 4311. It's not just board space,

but 4 PGA4311 chips are a bit cheaper than 8 PGA2311s.

I hate to be so cheap, but I really need to save every penny.

Luckily, the software should be the same for everyone. I'll just

let someone else design the 2311 PC board.

I'll hand-wire a separate bipolar power supply board and add a

few op-amps to mix into. That should be pretty straightforward.

Big question on the mixer:

The line mixers I've built before always had a series capacitor at

every input to block DC and pass AC. I would think that this would

be even more important if your sources are coming from different

equipment. The datasheets for the PGA2311 do not include anything

like this. Adding them after the 2311 seems silly, putting them before

would make sense, but I'm not sure how it would affect input impedance.

If I'm not making sense then I'll draw a schematic.

Any mixer experts want to offer an opinion?

Thanks.

LyleHaze

[stryd_one]

I'd be interested to find out about those input caps too, that's a good point...

[pilo]

I use PGA4311, the SOIC package is not very hard to solder, even for "newbie". It may take longer to do, but there's still 1.27 mm (0.05") between pin.

But as it had been said, almost every PGA chip (from TI or not) use the same SPI protocol, so the same driver and same kind of connector.

There was also a relay/resitor board made some years ago by a guy on The labs forum (mcs).If I remember weel dan (D2K) and SmahTV were interested in the design (and made some board maybe already?). They can be drive like the pga chip, there are more expensive, but maybe the very high end of digitally controlled volume.

also some MIC pre from TI (if I remember well, the PGA2500) has a gain control that works with SPI too. And there were line for controlling relay (enabling phantom power, phase inverte, etc) on this chip, I'll find the link.

As thorsten said, making a driver for those chip is very easy (mine was based on the aout module), so it takes no more than one or two hour to make it works with MIOS...

My schematic for this project was very simple (only pga4311 and some power filtering caps) because it was intend to use in my ADAT convertor. S

/edit

http://focus.ti.com/docs/prod/folders/print/pga2500.html

ok, only avalaible in SSOP-28 package.... but still usable!

[lylehaze]

I found an interesting page of mixer circuits.

There are lots of choices for input modules.

mic/line/balanced/unbalanced/Phantom power

See this page

http://users.otenet.gr/~athsam/Audio_mixer_6_ch_ENG.htm

This project just got a little more complicated.

LyleHaze

[stryd_one]

Nicely spotted man! I've been looking for something just like that!

[stryd_one]

FWIW, SMD doesn't bother me much either... Anyone else got 2 cents on that? I know TK prefers MBHP modules to be DIP for ease of use, but it sounds like this package is not too bad...

An ADAT IO would be awesome for me too, I don't suppose you'd be interested in sharing those goodies Pilo? :)

[tos]

Hm, for that price, why not fry few SMD chips? The problem with SMD is that you need etched board where for DIP packages a breadboard works fine.

[pilo]

we can make PGA module for 2310/2311(DIP) and 4311(SMD)  ;)

An ADAT IO would be awesome for me too, I don't suppose you'd be interested in sharing those goodies Pilo

Everything is available here:

http://www.multimania.com/blendinpulse/adat

well, almost everything... I had a lot of problems with this project, I spend a lot of time and force on it, and it still not work properly. But I guess this time I on the right way... (at least I hope lol).

[stryd_one]

Yeh why not both ;)

Thanks Pilo, I will almost certainly give that a shot soon  ;D If I can help with testing or whatever just let me know.

[smashtv]

FWIW, SMD doesn't bother me much either... Anyone else got 2 cents on that? I know TK prefers MBHP modules to be DIP for ease of use, but it sounds like this package is not too bad...

The SOIC is not bad, but the SSOP is from hell.  :-\

I use PGA4311, the SOIC package is not very hard to solder, even for "newbie". It may take longer to do, but there's still 1.27 mm (0.05") between pin.

I agree...with a good tip on a temp controlled iron and a little patience these are not bad at all to solder.

http://focus.ti.com/docs/prod/folders/print/pga2500.html

ok, only avalaible in SSOP-28 package.... but still usable!

SOIC was my stop on the SMD party train....  ;)  Still useable but you will hate life.  Forget it with a contact iron, and plan for some time on an air rework rig to get anything useable.  The pins are -really- close together, a magnifying glass is a must.  My rework setup has a 2mm nozzle and I still have to go over them several times before they will pass testing.

Here is an picture (sorry about the low res) of a PGA2500 sitting atop a 16 pin DIP to give you a scale reference/clue into the world of pain your in for when soldering these.  :)

Anyone know if the toaster oven guys might have a solution for soldering these that's repeatable?

Shifting gears..... ;)

For a look at TI's suggested implementation/layout for the 23xx, check out the doc for the 2310 eval module.  Even though the eval PCB is multi-layer, a SS layout without jumpers would be possible.  It remains to be seen how the lack of optimal ground/power planes will hurt quality/performance.  There is a nice photo there on page one of TI's board layout, and page 6 has the schemo to back it up. 

I found it interesting that they provide jumpers to power the board from the same +5vdc rail on both the analog and digital side, is this a hint that this was much more of a bad idea with the crystal csxxxx devices these are pin compatible with?  :)  (would love some opinions on this, and whoever builds first needs to jumper one up this way to see how bad it would affect quality!)

Gotta run....Best Regards

Smash

[pilo]

I managed without so much pain to solder SSOP package (both for audio clock and dac). The clock module doesn't work anymore (but I can tell where this comes from)... But it's really not that hard, even with a classic iron, with time and a mgnifying glass, it's still possible : iron will follow track and package pin, so a little iron on SSOP will be enough to solder it, then just check there's no short.

Well this is quite of topic, but I think it's doable (I fried some transistor but never a SSOP package).

If I can help with testing or whatever just let me know.

I just need to replace some buffer chip to test it again, and the design should be ok for building :) Again this is a little of topic lol

About the pga stuff, the best is to make the software driver for it (should be easy); and then build module for it : let's see where are the difference between chip and what we want to do, so we can have a module that can work with many different PGA chip.

[Artesia]

Zcen & Mute...

Hardware mute maybe pretty redundant if volume settings can be spun into theoretical silence pretty much instantly..

It may however prove extremely useful as a Hardware 'Panic Button' ..for instantly killing all audio on a stand alone, user interfaceless box.

Zcen is a matter for technical debate as to weather its useful or not.. it should be left as a jumper option for the end user to decide if they want it / want to toggle it on and off..

Power supplies...

For good, noiseless audio.. the power supply rails for the audio section must be sepperate... otherwise a very good circuit.. could just become a mediocre one sonically.

Balanced/unbalanced inputs... an opamp can be configured easily to accept both unbalanced & balanced signals.... and provide an analog 'level trim' ..if required.

[offe]

Ok, I'm waiting for some pga2311 chips to arrive (and hopefully some time to play with them). Well, I was just curious if someone has used mb hardware to control these chips and if there has been any development on a (mb) mixer module?

offe

[lylehaze]

Yes. I'm still working on my PGA4311 line mixer.

It has been stalled for a while as I don't have much time.

So far, the case is 95% complete, and it holds a working

core module and VFD display. I'm building a bipolar power

supply now. The case is a 1U rack mount, so everything

has to be "low profile". Fitting the display was tight.

Basically, 8 stereo pairs are mixed down. I'm hoping for

a few "extra features" too, but I'll not say much until

they are tested.

I have not yet begun the software. I want to complete

the board building first.

As this is a prototype I'm hand wiring the boards, I have

not yet done any Eagle designs.

It all would have been MUCH easier if I just threw it all

together and mounted it in tupperware, but the beautiful

work being shown off in "Midibox of the week" shamed me

into trying to make it look a little better. You people have

built some beautiful projects!

What are you planning for yours?

LyleHaze

[offe]

Sounds like a really interesting project! I've been searching for the PGA4311 but could not find a source here in scandinavia that could provide just a couple of chips. I'm aiming for a (preferably small) mixer with 8 channels but being a novice in assembler, I decided to get a couple of chips just to play with before taking on a full mixer project.

Any ideas where I could get the PGA4311 for a fair price?

[stryd_one]

I wouldn't bother doing something like this in ASM... Just use C mate :)

Yes. I'm still working on my PGA4311 line mixer.

It has been stalled for a while as I don't have much time.

Story of my life  ::) It seems every time I get enough time, one of my workmates goes on leave, and I get more work. Ugh.

I'm hoping for a few "extra features" too, but I'll not say much until they are tested.

Gutless  :P  Spill it!  ;)

It's open source, it's not like anyone will can steal the idea, and it's not like anyone will mind if it doesn't work out. Why do so many people keep mb projects secret? :(

[lylehaze]

offe:

Making it as small as possible? I'll admit that the display on my project, and the

one encoder could be eliminated (assuming you'll control it over MIDI).

But the real space-hog for me was finding roon for 8 pairs of inputs and one

pair of outputs. I'm using a combination of Phone and Phono jacks, and the

back of the case has filled up quickly.

For a parts source I used DigiKey.com. Fair price, good service, great website.

But I don't know how well they do outside the US.

Stryd_One:

Why do I keep it secret? That's easy. It's because I don't know yet how well it will

work, and I don't want to claim that I can do something impossible. Once I have it

working, then it becomes something I can brag about. ;-)

No, I really don't mind telling you what I'll try. If it works, great, if it doesn't, then

I'll just look foolish for trying. (won't be the first time)

I figure the core won't be too busy, since it will only be passing incoming volume

changes to the PGA array. so I'm planning on adding stereo VU meters and also

Peak Program meters (DIN 45406). These will be routed to 4 analog inputs on the

core. Now the user can select which of these can be displayed on the VFD display,

and also which can be (if desired) sent back out by MIDI for remote viewing.

I mean, if we are separating the control deck from the audio processing, we might

need to send back a little feedback on how things are going.

Finally, if I make it that far, the option of doing audio compression comes up. I could

do it first from my Amiga, by watching the "virtual meters" for audio levels above or

below the desired range. If I get the algorithm right, I might even try to re-write it

in the PIC. I could use something like that between my bedroom TV and speakers.

The broadcaster seems to turn the volume up around 3AM. I'd like to fix that.

What do you think?

LyleHaze

[stryd_one]

Rock on man!

You won't look foolish for trying at all!! You know how many people tried to fly before the Wright brothers? You ever hear people make fun of them?

;)

[lylehaze]

Hey wait!

How many died trying?

I'm enthusiastic, but not THAT enthusiastic.

:-)

LyleHaze

[Artesia]

you could try mixing with valves in the signal path...

Then you could indeed play the danger high voltage ! game..

Dont ask me about large frame helium neon power supplies ....you really don't want to know ;)

[pilo]

If you want some example source code for controlling pga with a mbhp core, I have done it some years ago. I have to find where the code is. It's a simple app that display the gain (in db) on the lcd, and you can set it up with an encoder. You can daisy chain PGA, and change the one controlled with some switch. And the gain value are stored in flash memory (or EPROM, can't remember), so even after power down you keep the setting.

So if you want it, I can send it via email, or put it somewhere to be available for download.

[lylehaze]

Pilo,

Thank You. I have your code, and I look forward to getting into it as soon as the hardware is finished.

I worked most of yesterday on the power supply. It is working and mounted in the case. I have also

mounted the core and display. Today I think I'll try making the summing amp for the mixer, and maybe

a few stereo input modules. I doubt if I'll get any PGAs wired up today, but maybe next week.

It is coming along very slowly, but there is some progress.

Thanks for your help,

LyleHaze