During the last two weeks i spent some time playing with my first digital project - a simple Kusunoki/Daniels testbed in which i was hoping to find my own answers to what makes digital tick 
I'll describe the various experiences in a separate thread but my immediate concern right now is the front end. So far i've tried cap-coupled consumer interface, SN75179 line receiver/driver and DS8923 which is my current favourite. All three sound annoyingly different.
To my surprise i didn't like the 75179 at all. It was obviously very ps sensitive (i only use a 78L05 on it so far) and generally did very bad things to the bass. The cap-coupling was pretty much fine with REL polyprops (never tried other types), resulting a very smooth, laid back and slightly boring sound. The National chip, otoh, although similar elecrically to 75179 performed really well.
The best sounding connection was dc and differentially coupled to the SPDIF source without ground connection to 8414 ground and again dc coupling between the driver differential output and RXp/RXn. I tried cap-coupling between the driver and the 8414 and it sounded obviously inferior.
So, here comes my question: is it ok to dc couple a driver to the 8414 schmitt trigger inputs? This will obviously change the bias condition at the inputs and possibly change the triggering point. Are there other considerations? Kusunoki apparently thinks it's ok as that's what he does with the 75179. I assume the input caps in his circuit are only for safety.
Thanks for any opinions and suggestions. Over the next few days i'll compare Elso's AD8561 and Fred's 74hcu04/74hc86 circuits.
I'll describe the various experiences in a separate thread but my immediate concern right now is the front end. So far i've tried cap-coupled consumer interface, SN75179 line receiver/driver and DS8923 which is my current favourite. All three sound annoyingly different.
To my surprise i didn't like the 75179 at all. It was obviously very ps sensitive (i only use a 78L05 on it so far) and generally did very bad things to the bass. The cap-coupling was pretty much fine with REL polyprops (never tried other types), resulting a very smooth, laid back and slightly boring sound. The National chip, otoh, although similar elecrically to 75179 performed really well.
The best sounding connection was dc and differentially coupled to the SPDIF source without ground connection to 8414 ground and again dc coupling between the driver differential output and RXp/RXn. I tried cap-coupling between the driver and the 8414 and it sounded obviously inferior.
So, here comes my question: is it ok to dc couple a driver to the 8414 schmitt trigger inputs? This will obviously change the bias condition at the inputs and possibly change the triggering point. Are there other considerations? Kusunoki apparently thinks it's ok as that's what he does with the 75179. I assume the input caps in his circuit are only for safety.
Thanks for any opinions and suggestions. Over the next few days i'll compare Elso's AD8561 and Fred's 74hcu04/74hc86 circuits.
I compared SN75179 driven receiver with coupling caps only and with Scientific Conversions transformer instead, and in both cases SN75179 was superior.
However, it is very power supply sensitive and only certain arrangement makes it sound good (at least IME)
So, AN8005 with 15u OsCons before, and after the regulator produced so far best results. Also the supply line to the driver goes through ferrite bead (it sounds better with it).
Main filter caps are 1000u in CLC balanced filtering setup. Now, if this makes any difference, I'm using through hole CS8412.
I'm using ROE/Vishay coupling caps as they are good sounding and small enough in that application.
However, it is very power supply sensitive and only certain arrangement makes it sound good (at least IME)
So, AN8005 with 15u OsCons before, and after the regulator produced so far best results. Also the supply line to the driver goes through ferrite bead (it sounds better with it).
Main filter caps are 1000u in CLC balanced filtering setup. Now, if this makes any difference, I'm using through hole CS8412.
I'm using ROE/Vishay coupling caps as they are good sounding and small enough in that application.
I suspected that much seeing how sensitive it is to decoupling. In my setup, which is very basic, there is a separate transformer/rectifier, 1000uF followed by choke 4u7 - 78l05- 4u7 + ceramic triplet. Under the same arrangement the DS8923 seems to work much better.
Still, i wish someone will answer my question about the coupling caps on the 8412 inputs.
I removed all the ceramic triplets. I find it much better without them.
I don't recall the exact setup, but in some instances I couldn't get any signal when I was using CS8412 without coupling caps.
Since DS8921 comes in a similar package as the driver I'm using presently I will try this one.
I don't recall the exact setup, but in some instances I couldn't get any signal when I was using CS8412 without coupling caps.
Since DS8921 comes in a similar package as the driver I'm using presently I will try this one.
Hi Peter
Did some more listening and building today as i realised that the previous comparison between the DS and the SN was not fair.
Previously i never compared them in the same setup - the SN was as per Kusunoki and the DS - without input caps and with the input spdif signal connected between the two inputs but not to ground.
I still prefer the DS chip but only marginally. Most of the improvement seems to be due to removing the caps and the ground connection at the input.
I also built Fred's circuit
Yesterday I tried to plug DS8921 in place of SN75179, but although they seem similar on paper I couldn't get any signal with DS chip, I have no idea why. I also tried to reverse pins 5 & 6 ( as they seem to be opposite comparing to SN chip).
The caps will make some difference, but the ones I'm using are pretty good in this location (the best I tried so far): Vishay/ROE MKP1837 http://www.mouser.com/index.cfm?&ha...*&terms=mkp1837&Dk=1&D=*mkp1837*&N=0&crc=true
Also, Caddock as 75R shunt resistor fits (sonically) well in here.
The caps will make some difference, but the ones I'm using are pretty good in this location (the best I tried so far): Vishay/ROE MKP1837 http://www.mouser.com/index.cfm?&ha...*&terms=mkp1837&Dk=1&D=*mkp1837*&N=0&crc=true
Also, Caddock as 75R shunt resistor fits (sonically) well in here.
Did you hook the enable pin to 0? Did you remove the input caps?
The datasheet mentions a 'fail safe circuit'
which brings the outputs hi if the inputs are not connected. If you want to retain the input caps you might need to place some large (100k ?)resistors between each input and ground.A very simple experiment for you is to remove the input caps and ground connection between transport and dac ps retaining the 75179.
Reversing inputs to 8412 should not make a difference
analog_sa said:
Did you hook the enable pin to 0? Did you remove the input caps?
The datasheet mentions a 'fail safe circuit'
I didn't notice it was mentioned here: http://www.national.com/ds/DS/DS8921.pdf
But I will try without caps.
analog_sa said:
Hi Peter, I'm curious how is your progress with a DAC?
Yesterday I removed ground connection at the DAC's input and indeed sound improved; seems like it's more spacious with better defind 3-dimentionality and clarity. I think I will leave it this way.
I also removed coupling caps, and it seems like slight improvement as well, with more resolution, although probably less smoothness and a bit harsh. Output coupling caps are more critical now, and BG N seem to be a good match.
I'm playing a lot with Rubycons ZL in the DAC now, and again those caps are very good and more neutral than BG N and they seem like a good match in quite a few positions in a circuit.
Peter
Which ground did you remove? The ground wire of the coax cable from transport to dac, or the connection from the 75R spdif input resistor to ground?
TIA
My favourite coupling cap is 10nF NPO 1206. These sound sharp compared to polypropylene films, which are a bit oily sounding. I'm surprised the cap actually changes the sound, but it seems it does.
Just as a general tweaking note, I find a 1k0 carbon resistor (cheap 0.125W from Farnell) on each CS8412 digital output adds a lot of slam. I'm not sure the 22R-100R resistors I've tried do anything at all, so may as well go all the way
Which ground did you remove? The ground wire of the coax cable from transport to dac, or the connection from the 75R spdif input resistor to ground?
TIA
My favourite coupling cap is 10nF NPO 1206. These sound sharp compared to polypropylene films, which are a bit oily sounding. I'm surprised the cap actually changes the sound, but it seems it does.
Just as a general tweaking note, I find a 1k0 carbon resistor (cheap 0.125W from Farnell) on each CS8412 digital output adds a lot of slam. I'm not sure the 22R-100R resistors I've tried do anything at all, so may as well go all the way
Peter Daniel said:
Well, the thing is not that simple. While this seemed to work well with my ML transport, today, I had a chance to check the DAC with a friend's Shanling.
With coupling caps shorted, we didn't get any signal. With a ground disconnected, there was slight hum, and my friend (a hard core audiofile) actually preferred the sound of input ground hardwired.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.