PCM4222 not used in commercial products, why?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

rkn80

Well-known member
Joined
Sep 20, 2009
Messages
410
Location
Germany
Hi,

just wondering, the last few years I've seen many professional ADC boxes (Avid, Horus, Metric Halo etc.) but I have never seen them using the PCM4222 or similar from TI. They are all built with chips from Cirrus, Arda or AKM for ADC. So why? Did I not see the right ones? Or is it a stupid reason like, 'we always used AKM...' or is something else? Any ideas?

R.
 
I had a conversation the other day about how musicians often prefer to twiddle knobs and switches rather than fooling around in software. He complained about how difficult it was to work typical software to achieve the results he was looking for and lamented the lack of an ADC that had all manual controls. So, the PCM4220/4222 appear to be controllable via mechanical switches (or relays as required) and I've been toying with the idea of building such a beast but the learning curve for me in this is pretty steep. I suppose I'll just have to sit down and dig through all of the information very slowly, but as a short cut I thought I'd ask here if anyone has seen such a beast before? As anyone built, or know of, an ADC that is entirely manually controlled?
 
Most TI high performance ADC's (PCM4202, PCM1804 and PCM422x) have a hardware controlled mode. (meaning, pull up/down on pins to configure).

However, changing sample rate isn't as simple as switching one pin in many cases, as various clock dividers need to be configured.
e.g. your divider settings for the sample rate will be different if your clock source is a 24.576MHz clock vs. a 12.288MHz Clock.

In addition, your wordclock interface that sync's to other parts of your studio is a separate circuit. That will also need configuration.

Then don't forget your digital interface out... such as a S/PDIF transmitter.

What I'm trying to say, is that there are way too many variables in play in the design of a system to simply have one switch that jumps between sample rates. This is typically one of the roles of a microcontroller, to take user inputs (such as you pushing a button) and then step through states and flip multiple output pins to configure the ADC, S/PDIF and PLL circuits.

An Arduino, or an MSP430 can do this easily. Your process psuedocode will look something like this:

1. Detect user button press
2.Debounce button press (count to 1000 and see if the button is still being pressed)
3. Lookup next sample rate settings (which pins need to go high, which go low)
4. Put the ADC,S/PDIF and PLL into RESET (by holding a certain pin high/low)
5. Update your control output pins on your microcontroller.
6. Release RESET allowing each of the devices to start running again
7. Enjoy the lovely tone of these ADCs  ;)

 
Thanks for your replies.

Well, relays could be used for the multiple items that need to be changed for each change in sample rate, etc. Each switch position could actuate multiple relays, for example.

Obviously, the bare chip is useless without input circuits and digital interface chips.

Maybe this isn't the greatest idea on Earth, but I think it's worth pursuing some more.
 
Dirk,

definitely don't give up. Just realize that there is a system level solution at play.

My suggestions is to simplify as much as possible to begin. e.g. make the decision that your circuit will always be master... let other commercial systems slave to your clocks. That will immediately remove your need for the PLL circuit, and for the ADC to switch between master and slave.

Then decide if you need both 44.1 and 48kHz support. If you do, then you'll need 2 different crystals for master clocks, and need to switch between them. Yuck.

Then decide what your output format will be - left justified, right justified or I2S. I like left justified. it just *works* in my head.

Get your S/PDIF reciever working to simply take data in and spit S/PDIF out. No tricks, no gimmicks.

I'd start with an eval kit, then remotely control the Eval kit with a microcontroller.



 
Thanks so much for your help! Much appreciated.

I talked with him and now he says something completely different, or at least I didn't hear it properly the first time. His problem is clipping the ADC on music peaks. I suggested that I build a mixer with built in compression chips, such as those by THAT corp.
 
dirkwright said:
Well, relays could be used for the multiple items that need to be changed for each change in sample rate, etc. Each switch position could actuate multiple relays, for example.

Relays are the wrong thing to use here. You're switching digital logic. Use digital logic gates, or even a small micro (SiLabs C8051F850 series costs less than a buck!). You'll spend dimes instead of dollars.

-a
 
Ross Martin builds 2 channel and 4 channel ADCs based on the PCM4222. 

http://rossmartinaudio.com/index_files/Page318.htm
 
prh said:
Ross Martin builds 2 channel and 4 channel ADCs based on the PCM4222. 

http://rossmartinaudio.com/index_files/Page318.htm

this guy is doing what I want us to do.
https://www.facebook.com/SwetoneAudio
We're just fooling around right now.
 
dirkwright said:
His problem is clipping the ADC on music peaks.

That may well be operator error.

If you see frequent clipping on a modern ADC, the solution is to run your input signal less hot. Like all analog signal chains, you need sufficient headroom to stay away from clipping and/or other forms of gross distortion. An ADC like the 4222 (or any other recent design) will have plenty of dynamic range to not require you to work that close to clipping.

(Ah, but now he says the signals in his DAW don't have the 'right' level. In that case it might be a good plan to reinvestigate his gain structure in both analog and digital -- why do the average levels in the DAW need to be that close to 0dBFS?)

dirkwright said:
I suggested that I build a mixer with built in compression chips, such as those by THAT corp.

That will work to some extent (and like any compressor it will alter the dynamics of the signal being recorded). Consider time constants here: a really fast compressor may well introduce artefacts that sound worse than the clipping, a really slow compressor is no different than manually turning down the gain.

(Depending on what he's recording a compressor may be called for in any case -- either analog/OTB or digital/ITB. IME making the compressor optional is fine, bolting it to your ADC suggests a problem elsewhere. Some would argue it's best to capture the signal as un-altered as possible and then adding compression and other effects as desired)

Ask if he can turn down the gain some. If there's no knob for that (or if he likes the sound of his chain upstream of the ADC with this high gain setting), consider building him an L-pad or a U-pad to see if that fixes the clipping.

JDB
[if there's no gain setting where the clipping goes away without making the ADC noise floor too noticeable then he does need a new converter with more dynamic range]
 
The missing USB connectivity can't be the reason. No other ADC playing in the High-End range does offer it (Cirrus, AKM). They all have usually an I2S interface and you have to develop your interface to the outside yourself.

R.
 
rkn80 said:
The missing USB connectivity can't be the reason. No other ADC playing in the High-End range does offer it (Cirrus, AKM). They all have usually an I2S interface and you have to develop your interface to the outside yourself.

Daffyd can probably confirm this, but my guess is that it might be difficult to meet the noise performance one gets with the PCM4222 if it had the USB connectivity built in.

I think if you really want USB, you should just choose one of the many cheap ARMs which offer USB and I2S ports and be done with it.

-a
 
Andy Peters said:
rkn80 said:
The missing USB connectivity can't be the reason. No other ADC playing in the High-End range does offer it (Cirrus, AKM). They all have usually an I2S interface and you have to develop your interface to the outside yourself.

Daffyd can probably confirm this, but my guess is that it might be difficult to meet the noise performance one gets with the PCM4222 if it had the USB connectivity built in.

I think if you really want USB, you should just choose one of the many cheap ARMs which offer USB and I2S ports and be done with it.

-a

Nailed it.
 
Well many USB/I2S solutions for stereo only can be found on the internet. Multi-channel interfaces are the challenge.

But that still does not explain, where all the PCM4222 TI produces are used.

R.
 
I think you'll find PCM4220 designs in more places.
The only difference between the two is that the PCM4222 also outputs DSD and Modulator out.
Most designers don't have the skills or need for anything but PCM.
 

Latest posts

Back
Top