DIR9001 vs CS8416 and poss CS8412

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Sure - I've not used one yet mind, but I am planning a design with one. The reason being the flexibility of the clocking scheme - it incorporates a frequency synthesizer which can be programmed in software. I'd not recommend it for stand-alone duty, only in conjunction with a CPU - in such cases its features really outshine the other two you mentioned in that it can still generate a nice precise clock in the absence of any incoming signal.
 
Reading through it's looks superior but is much more complex to control. At this point, I think I need to concentrate my efforts elsewhere so the DIR might well do. In any case the plan is to make the whole DAC modular. I can revisit the input side later.

So the question changes to dir9001 or WM8805 (in h/w mode)??
 
Perhaps those naive punters think that because there's an XTAL there's a varactor inside the WM8805 and that makes for a lower jitter (from narrower bandwidth) PLL? From reading the datasheet though it looks as though the WM8805's key feature is that the PLL is digital rather than analog.
 
ThorstenL's actually worked with programming the part so I defer to his greater experience.

As regards CPUs, I'm a devotee of ARM so I'm planning to use LPC1313 or LPC1113 in my design. That's because these have the grunt to do a bit of DSP (oversampling filter for example) and are barely more expensive than a PIC (sub $2). They have hardware I2C so at least part of the interfacing is done already.
 
Realistically a PIC10 would be easier to implement and costs less than 1 USD. But... is it worth dealing with the added circuits and RF noise? Do you have the programmer for it? You won't gain performance - 192kHz sample rate does not perform better in modern DAC's - look at their data sheets. Cannot "travel" via most of optical converters either.
 
Last edited:
Realistically a PIC10 would be easier to implement and costs less than 1 USD.

It doesn't appear to have the requisite serial ports for I2S. So as a way of programming the 8805 its fine, but doesn't add much interesting by way of DAC features. Like volume control to offer just one possibility.

But... is it worth dealing with the added circuits and RF noise?

All reward is dependent on risk. Why would there be more 'added circuits' than with your PIC10 idea? Or are you comparing CPU with no CPU at all?

Do you have the programmer for it?

For sure, that's one reason I chose this rather than a family I don't have the kit for. It programs itself though, just needs an interface.

You won't gain performance - 192kHz sample rate does not perform better in modern DAC's - look at their data sheets.

You're jumping to delusions - I wasn't planning 192kHz nor was I planning to use a 'modern DAC' in my design. But I agree with your point - taking the PCM1704 as an example, the THD goes up as the sample rate goes up.
 
I do have a pic programmer and limited experience of programming with it!! (Never used it in ager as I intended!)

I'd not really be interested in the advanced features of the receiver other than the input selector might be useful? I cant help but think external mechanical control would possibly be better?

All this considered, it still sounds like the advanced clocking of the WM8805 even in H/W mode may be prefereable to the DIR9001?
 
WM8805 does NOT have any "advanced clocking", I was sure that you will missunderstand that - see my discution above.
It locks on to the SPDIF signal with the PLL loop, exactly like any other receiver. With 50pS jitter - like DIR9001.

PS: My bad, PIC10 was just an ideea of cheap, simple (6 pin) way to program the ports. Sorry, for i2c lines, there is PIC12F1822 at 1.15$ and 8 pin.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.