Let's start with one DAC first.
What are your requirements? I.e. enumerate the lines that you need along with their specs (rate, voltage).
Now for 2 DACs, you'll have to split the signals to feed each of them. Indeed, the expansion ports will have only 1 LRCK line and 1 bit clock line. So if your DACs come on 2 separate boards, then you need to feed the signals above to each of the boards.
Also, in order to have the main board operating, you'll need first to provide:
- a power supply (5V and +/- 12V)
- a back-lit LCD display - any color any size provided it's the proper type (2x16)
- 8 push-buttons
- a IR sensor (5V 38kHz) if you intend to use a remote controller, which I recommend.
so there is only multiple Sample Data pins present ie. Sdata 1, Sdata2, Sdata3 etc., but BCK, LRCK only 1 each? I will be using 2 or more dac physical boards too, guess i'll have to figure out some sort of buffer. possibly 2 dacs that are close will be ok. do you think you could just put 2 of each ouput in parallel? pretty easy with the u.fl headers
That's correct.
I don't think it's going to be a problem for LRCK and BCK to drive 2 gates in parallel, but I'd suggest to make a clean wiring.
so I take it that you are running the core at one interrupt per sample ? You are not using block dma mode ?
which is what u.fl is all about.
I would suggest you please put 2 x u.fl for each BCK and LRCK like below. patterns are drawn to scale as well as a 1:10 scale and possible suggestions for compact arrangement of 2 in parallel opposed to each other. the small pad is signal, the outer 2 bars are ground. you should just be able to import them, if you need a DXF or something let me know
And if you put 2 ufls on single output, put two series resistors, and keep the lengths of ufl cables equal (load 'em equally).
This is "Y" transmission line configuration. If you won't care to load everything equally, you'll need to go "bus" termination with a resistor on parallel to transmitter or thevelyn biased termination. These aren't nice on the driver's use.
This is "Y" transmission line configuration. If you won't care to load everything equally, you'll need to go "bus" termination with a resistor on parallel to transmitter or thevelyn biased termination. These aren't nice on the driver's use.
Thanks. I keep in mind and I'll look at it carefully - but I don't want to promise anything because it's already crowded around these I2S expansion headers.
How about you having your own little adapter board stacked onto the header? Would that be an issue?
I dont like the idea of a header adapter, almost may as well just use the ribbon cable if you do that and IMO ribbons are pretty compromised for impedance controlled connections and I would never use them by choice. that adapter was created to get around a pre-existing design.
the makers of that design, twisted pear, are now looking at making it available natively due to public demand. all the same Ian has measured an improvement using the adapter, though not as good as having it there in the design
using the u.fl for i2s and clock wipes out any concerns of neat interconnection in one go in a simple and pretty reasonably priced, compact and widely available industry standard package.
TBH I wish ribbon connectors would be banished for eternity. they are convenient and cheap, thats all theyve got going for them IMO.
the makers of that design, twisted pear, are now looking at making it available natively due to public demand. all the same Ian has measured an improvement using the adapter, though not as good as having it there in the design
using the u.fl for i2s and clock wipes out any concerns of neat interconnection in one go in a simple and pretty reasonably priced, compact and widely available industry standard package.
TBH I wish ribbon connectors would be banished for eternity. they are convenient and cheap, thats all theyve got going for them IMO.
Last edited:
I agree, but then it should be as many connectors as many channels
A parallel of two connectors can be useful to drive two external DAC only.
What about using three or four external DAC in a three or four way system?
Yeah that's a bit what I was afraid of ...
So say there's provision for 8 I2S data lines, and you want individual LRCK and BCLK for each, you're ending with 24 connectors... And we haven't talked about MCLK yet.
Also these sockets are relatively expensive, and 24 of them is going to add substantial marginal cost for a feature that maybe not many adopters are going to use.
Last edited:
then leave it to them to populate or not. for me, its only really important for BCK and MCK. I dont think you quite understand how economically they can be placed. you are aiming for high performance here right? talk to marce about what pinheaders do to digital and RF bandwidth high speed signals, its a mess, they were never designed for that. I think it would be a mistake to leave it off at this juncture when its starting to gain acceptance. to have to use an adapter to use it with high end dacs seems wrong to me.
given people who have gone to the trouble to build a separate dac for each of 4 or 8 channels are on the extreme side, comparatively the money for enough cables and connectors is a miniscule amount, they probably spent more on their IEC inlet for one dac =) (joking, but its not unheard of) besides we are a community here, we would group buy them.
where did you look at pricing? it varies wildly. 25 pieces for $18 at digikey and thats just one person buying 25. are you telling me someone who decides to build 8 discrete balanced dacs so he can have one per driver would throw up their hands at that?
it would be very easy to have it right there using the very same pads as a 1mm or perhaps 2.54mm pitch SMD ribbon. illustrative, I didnt bother to do the below exactly, the signal pads are not perfectly aligned
given people who have gone to the trouble to build a separate dac for each of 4 or 8 channels are on the extreme side, comparatively the money for enough cables and connectors is a miniscule amount, they probably spent more on their IEC inlet for one dac =) (joking, but its not unheard of) besides we are a community here, we would group buy them.
where did you look at pricing? it varies wildly. 25 pieces for $18 at digikey and thats just one person buying 25. are you telling me someone who decides to build 8 discrete balanced dacs so he can have one per driver would throw up their hands at that?
it would be very easy to have it right there using the very same pads as a 1mm or perhaps 2.54mm pitch SMD ribbon. illustrative, I didnt bother to do the below exactly, the signal pads are not perfectly aligned
Last edited:
I'm using two BII DACs fed by a single I2S source and each I2S line is routing two channels, and as I understand this is quite a standard
so the connectors needed in a four way system (8channels) should be 12
4 x I2s + 4 x bck + 4 x lrck
if there isn't space for all those mini RF connectors, a second choice could be to have a pin header where to solder a daughterboard with the miniRF connectors and resistors on it
so the connectors needed in a four way system (8channels) should be 12
4 x I2s + 4 x bck + 4 x lrck
if there isn't space for all those mini RF connectors, a second choice could be to have a pin header where to solder a daughterboard with the miniRF connectors and resistors on it
I was going to mention that, but thought it was best to illustrate his example directly. why resistors? the termination should already be in place
what do you mean by i2s line, you mean sdata? buffalo has BCK, D1 (LRCK), D2 (SDATA). but if using this I would be trying to us synchronous clock so I would want MCK too, so 16 total IF I wanted so many discrete dac channels (4 x BII/III), which I dont
Last edited: