Very intresting, great effort. But what's not clear to me is that you are measuring mV, while we want to know mA.., or is the last comment w.r.t 5mV increments a type-O (although you menation current in all other places.). Sorry if I'm nitty picking..(i'm usually the one making all tht type-O's)
However I'm not convienced it's realistic represesntation of what the dac chip is doing to fluctuate 5mV at 10KHz. I thought we'd estblished it was a pretty constant 2.4V.
It would be intresting to hook the scope up to the dac pin 20 and put a 10Khz audio signal on it and see what portion of it shows up on pin 20 to see the impact on the real behavior agianst your current measurements.
There must be a mV fluctution at pin-20 otherwise no SQ gain would be heard swapping out the resistor for a CCS..
Last edited:
On the multimeter pin 20 is constantly 2.4v no matter what the analogue v, ccs current, vcom output bias is etc etc
I've been playing with it
The ccs current value doesn't directly have an effect on sq
But it has an indirect effect on vcom and output bias which does matter
Suppose I should put it on the scope
I've been playing with it
The ccs current value doesn't directly have an effect on sq
But it has an indirect effect on vcom and output bias which does matter
Suppose I should put it on the scope
Excellent work, Ross - thank you for putting together these results.
In my book, the cascoded JFET configuration is typically a top choice (I'm using it in most of my LV mods, together with cascoded depletion-mode MOSFETs for HV applications; they both perform excellent), though I'm weary of the very low voltage here - 2.4V - over the assembly. Your plot though indicates very good behavior.
Radu.
The voltage measurement across a resistor is a current measurement, V=IR. I adjusted each CCS circuit so that the voltage across the 100 ohm resister was 40.6mv or 0.406ma.
If pin 20 operated at a rock steady 2.4VDC, then we would hear no difference replacing the 6K resistor with a CCS - but the change in SQ is significant and probably different with each CCS design. I'm sure that the 0.5V 10 KHz from my signal generator is more severe to these CCS circuits than what pin 20 is putting out, but it allows us to differentiate the potential performance of each CCS device.
I hope someone will put a scope on pin 20 into a resistor and display it on the forum. I downloaded a free program called Sweepgen which I use to adjust my DSP equalizer for my speakers. Music would generate random signals at pin 20 whereas a single frequency from Sweepgen should produce a repeatable signal which could be captured on a Scope. Sweepgen worked fine through my WaveIO card.
If pin 20 operated at a rock steady 2.4VDC, then we would hear no difference replacing the 6K resistor with a CCS - but the change in SQ is significant and probably different with each CCS design. I'm sure that the 0.5V 10 KHz from my signal generator is more severe to these CCS circuits than what pin 20 is putting out, but it allows us to differentiate the potential performance of each CCS device.
I hope someone will put a scope on pin 20 into a resistor and display it on the forum. I downloaded a free program called Sweepgen which I use to adjust my DSP equalizer for my speakers. Music would generate random signals at pin 20 whereas a single frequency from Sweepgen should produce a repeatable signal which could be captured on a Scope. Sweepgen worked fine through my WaveIO card.
the obvious question is how you measured - ac over dc?... depending on your DMM you may have different provisions on how ac riding on dc can be measured
a scope pic would be great
Scope Pin 20
I put a scope on pin 20 feeding a cascoded CCS and also a 6.04K resistor.
The scope pictures didn't show much because the signals smeared together, but this is what I saw.
1. Scope displays showed a variation of what looked like a blend of periodic and random sine waves varying by no more than +/- 5mv. Nothing that resembled square waves, but the sine waves sometimes appeared distorted.
2. The clearest sine waves were displayed at 0.5usec time/div. These could be clock signals or digital noise from the DAC. There were also smeared sine waves at 10usec time/div.
3. Interestingly, the signals for the CCS were about twice as large as the ones for the 6K resistor. I guess this makes sense because a CCS allows the voltage to float without any change in current.
4. The signals were slightly more messy looking with a 1 KHz square wave being fed into the DAC versus no signal being fed.
There is something going on at pin 20 besides a steady 2.4VDC, but the variations are small and at high frequencies. This doesn't tell me what to do but to continue to do listening tests.
Just for fun I scoped the RCA output for the 1 KHz signal. This proves that I was actually putting a signal through the DAC. Maybe someone else can get more definitive scope displays from pin 20.
I put a scope on pin 20 feeding a cascoded CCS and also a 6.04K resistor.
The scope pictures didn't show much because the signals smeared together, but this is what I saw.
1. Scope displays showed a variation of what looked like a blend of periodic and random sine waves varying by no more than +/- 5mv. Nothing that resembled square waves, but the sine waves sometimes appeared distorted.
2. The clearest sine waves were displayed at 0.5usec time/div. These could be clock signals or digital noise from the DAC. There were also smeared sine waves at 10usec time/div.
3. Interestingly, the signals for the CCS were about twice as large as the ones for the 6K resistor. I guess this makes sense because a CCS allows the voltage to float without any change in current.
4. The signals were slightly more messy looking with a 1 KHz square wave being fed into the DAC versus no signal being fed.
There is something going on at pin 20 besides a steady 2.4VDC, but the variations are small and at high frequencies. This doesn't tell me what to do but to continue to do listening tests.
Just for fun I scoped the RCA output for the 1 KHz signal. This proves that I was actually putting a signal through the DAC. Maybe someone else can get more definitive scope displays from pin 20.
Attachments
if logic from attached picture is applicable (I didn't got source of that schematic ) - fiddle with pin 21 an 22 decoupling (Vcom pins) , not pin 20
it's pretty much irrelevant observing what's happening on pin 20 in AC domain; in fact - greater amplitude probably means that servo is working more effectively
it's pretty much irrelevant observing what's happening on pin 20 in AC domain; in fact - greater amplitude probably means that servo is working more effectively
Attachments
That would be brilliant, thanks.
I had a quick ponder over the 2 versions if the mainboard schematics today at lunch and it looks reasonably straightforward, but I couldn't quite make a direct comparison as I don't exactly understand what happens with the bclk out of the new i2s/spdif switching chip, so a diagram would answer all my questions there.
Cheers
I asked this question months ago and did not get an answer from Doede.
I wonder if it has something to do with the SPDIF input? But have no idea.
Maybe with two asking we will be told?
That would be something if THAT is where the real action is!
ZM do you have any idea what could be an improvement over a capacitor?
I know I do not know!
I posted a while back about using smd film caps and working well for me
Which may only work for me because of the lifepo4 ps feeding the 1794 chips
But to be honest it wasn't a huge area of improvement
Maybe ZM has something else in mind?
Last edited:
There's 4 extra chips on the newer blue 'S' mainboard, 1 is the spdif driver, 1 is to handle the switching between i2s and spdif, but then there are 2 extra chips and it's these I'm interested in.
On each dac deck there are 3 series resistors through which the data and clock signals flow.
With the older red main board, the resistor on the 'data' line is 1k. In Doede's documentation he says
Doede said:
