Wolfson WM8816 volume control

Brett

Member
2002-01-07 6:02 pm
Has anyone had any experience with this chip? Comments on any implementation difficulties, and of course, it's sonic performance? I note sonnya said he had some samples in an earlier thread. Did you ever get the chance to try them?

Seems to me it has all the advantages of the CS3310, plus it will swing decent voltages and allows the selection of your own opamp or perhaps even a discrete stage. Plus all other things being equal, I usually prefer the sound of an opamp in inverting mode.

I read the long thread about the PGA2310 and got a lot of ideas for the implementation of the controller and display. For the encoder, Bourns (and others too I'm sure) make contacting switches with a binary output of 4 bits, so wouldn't it be easier to use one of the parallel ports on say an AVR/PIC as an input? No need to poll at a very high rate then, and you could run the micro at a low clock speed (min RFI) and perhaps drive the display though latches if there are not enough pins free on the micro.

Whaddya reckon?

WM8816 data at http://www.wolfsonmicro.com/product.asp?pid=15
Price is only $US8.20 each. [Much cheaper than the S&B transformer I'd like to try too.]
 
A very interesting chip. Its appeal (to me) is that it allows one to choose the inverting opamp for completion of the circut as opposed to the internal ones in the CS and PGA chips. This might well allow tailoring of the subjective as well as the objective characteristics of the completed circut.

Thanks!
 
Hi brett,
where can you by them. I managed to get my hands on 5 samples after 2 month .. A bit long i think.

The price you mentioned is also good.

Is it a 100 pcs price or ?? Would like to know more.

For the unit itself. It do have some nice features like the lack of build in opamp. At this point with the right choice of opamp (discrete?) it can maybe outperfom pga2310/2311/cs3310.

I have not got the time to try them yet... I would like to make a doublesided PCB, but if i can't get my hands on more i will not use them. They wanted me to buy like 250 pcs and i only need like 10 - 20 for a balanced preamplifier for me and some of my friends.

mmmhhh.. Maybe i will make a board afterall...

By the way : The performance of the WM8816 shown in the datasheet is with the OP275 from AD wich have got a baad test by Dself.
With a device like AD8610,OPA627,OP2134 it will maybe show even better performance.

One more thing : It inverts the signal.

Sonny
 
Brett said:
I read the long thread about the PGA2310 and got a lot of ideas for the implementation of the controller and display. For the encoder, Bourns (and others too I'm sure) make contacting switches with a binary output of 4 bits, so wouldn't it be easier to use one of the parallel ports on say an AVR/PIC as an input? No need to poll at a very high rate then, and you could run the micro at a low clock speed (min RFI) and perhaps drive the display though latches if there are not enough pins free on the micro.

You could run the control signal trough hc-type buffer before entering the WM8816. The Hc type then need to be run from a analog +5V. It is a good way of blocking noise and big ground loops from a micro if it is done right.

Also keep the micro runing at a low frequency as you mentioned. Keep crystal traces as short as possible to reduce radiation. If you can it would be a good idear to shut the micro into sleep where it turns the oscillator off.

If not maybe build a faraday shield (shielded to digital ground) around the micro and the crystal. Shields can you get remade from Farnell.

We make transponderreaders where i work and when they have to work near micro we do the same thing, because of interferrence.

For the part :
"Bourns (and others too I'm sure) make contacting switches with a binary output of 4 bits, so wouldn't it be easier to use one of the parallel ports on say an AVR/PIC as an input? No need to poll at a very high rate then"

You NEED to pool so that you can average/filter the value to get stable result. Maybe check debounce time on the switches ... If lets say 10msec. I would go for 20 msec. stable time .. 20msec and 10 times oversampling to get 10 equal values before accept => scan frequency of 2msec.

Another way is to use an optical encoder and 2 interrupt inputs to select direction... Not cheap at all but shure they run like in "oil" And they will not give you problems for many many years.

I hope it is usefull.

Sonny