all boards need to be identical
like tug o war
the team needs to pull together
or it just doesnt work right🙂
Nige2000, Ross,
Thanks for your advice, I have a 5V reg, but only at one deck the 100r. Next weekend I plan to change the other decks.
Paul
Recently i tinkered a bit with the CCS adjustment. Prior the the adjustment, all my CCS were calibrated outside the board for 0.4mA operation with 2.4V. After putting in I am getting exact 2.7v on one channel and 2.73v on the other. So I attempted to fine tune the channel with higher output to exact 2.7v by adjusting the CCS. After adjustment, the channel balance actually shifted and sounded worse off.So after this exercise, more or less I concluded that one should not tinker with the CCS once dialed in 0.4mA outside the board.
Doesn't make sense to me. You should be able to dial in 2.7v with your CCS trimmer which means you will get the requisite 40ma through the CCS. Any other voltage and your current will be slightly adrift.
What I can't get my head around is what happens when I use lower resistors than 134 ohms on one deck? Carlsor says don't touch the CCS adjustment, which implies it is working like a regulator but we know that is not the case. Perhaps someone can explain what's happening in more detail.
David
On page 25 of the datasheet for the PCM1794, you will see that the current output for 0 signal is 6.2mA. Following Ohm's law, V = IR, the voltage you measure across the I/V resistor is 0.0062 X your I/V resistor value. If the resistor is 134 ohms, you would measure 0.8308V for a single output.
However, the current for 0 signal is changed by the reference current (pin 20 current); also, the outputs are paralleled so the current is half per output. If you have a single board with 134 ohm I/V resistors and are measuring 2.7V, then the output for 0 signal is approximately 20.1mA, or 10.05mA per output pin. This is the centre point, and the current output will swing 7.8mA per output pin around this, double in this case because of the parallel outputs to 15.6mA, from 12.3mA to 27.9mA. The CCS current (reference current) affects the centre point of the output, and thus the centre voltage after the I/V. You will always get 0.0156 X <I/V resistor> in voltage swing from minimum value to maximum value (which is 2.0904V with a 134 ohm resistor).
The answer: if you have the CCS set so your outputs are giving 20.1mA total, you will measure 2.7V with 134 ohm resistors. If you have 130 ohm output resistors, you'll measure 2.62V. If you have 100 ohm resistors, you'll measure 2.01V.
Having a slightly different offset (current at 0 signal) between channels should not be noticeable, as only the DC offset changes, not the voltage swing between min and max. What will affect the channel balance is how close the resistors for each channel match each other.
Note: all these values are actually negative, not that it really matters.
However, the current for 0 signal is changed by the reference current (pin 20 current); also, the outputs are paralleled so the current is half per output. If you have a single board with 134 ohm I/V resistors and are measuring 2.7V, then the output for 0 signal is approximately 20.1mA, or 10.05mA per output pin. This is the centre point, and the current output will swing 7.8mA per output pin around this, double in this case because of the parallel outputs to 15.6mA, from 12.3mA to 27.9mA. The CCS current (reference current) affects the centre point of the output, and thus the centre voltage after the I/V. You will always get 0.0156 X <I/V resistor> in voltage swing from minimum value to maximum value (which is 2.0904V with a 134 ohm resistor).
The answer: if you have the CCS set so your outputs are giving 20.1mA total, you will measure 2.7V with 134 ohm resistors. If you have 130 ohm output resistors, you'll measure 2.62V. If you have 100 ohm resistors, you'll measure 2.01V.
Having a slightly different offset (current at 0 signal) between channels should not be noticeable, as only the DC offset changes, not the voltage swing between min and max. What will affect the channel balance is how close the resistors for each channel match each other.
Note: all these values are actually negative, not that it really matters.
Last edited:
An update for the guys that are waiting the boards with the Ross's parts soldered.
The mail from Ross was in my mail box last night.
During the week-end I will prepare your boards and by Monday I should be able to send you all the soldered boards.
Regards,
Enrico
The mail from Ross was in my mail box last night.
During the week-end I will prepare your boards and by Monday I should be able to send you all the soldered boards.
Regards,
Enrico
That was my point. I was able to get 0.4mA externally for each of the 4 pair of CCS and after putting them back into the board (I forgot to mention here that I am using 4 decks), I tried to fine tune to get exact 2.7V. I choose one of the 4 CCS in that channel that is not showing up 2.7 to fine tune to get the exact 2.7V with all the 4 boards hooked up but it seemed to cause channel imbalance after fine tuning (based on listening) even after both left and right are reading 2.7V exact. I am using 33ohm Caddock resistors.
Adjusting it back to the default CCS settings solve this issue.
I cannot explain why I am getting this.
How accurately matched are your i/v resistors? I'd measure the 2 i/v resistors for each channel and see how far off being a perfectly matched pair they are
On paper its 1% tolerance. If they adhere to specs, they should be at most 2% off from each other assuming the worse case which comes up to 0.7ohm. With 80.4mA flowing through (4 deck), so the difference should be about 0.05V volts max.
I wonder if by not having similar amounts of current flowing from each DAC have any implication. Anyway, setting it back to the default trimmer setting worked for me. I think I will stop here. Got other things to pursue like adding a buffer board to my DDDAC. 🙂
DDDAC Buffer PCB
Just getting caught up on this excellent thread, kudos to Carlsor. I keep missing out on great group buys, any possibility to get a buffer PCB?
Just getting caught up on this excellent thread, kudos to Carlsor. I keep missing out on great group buys, any possibility to get a buffer PCB?
measure vcom voltage, pin 20 voltage, analog voltage
and check iv resistor values on affected board
a 40ma ccs is a 40ma ccs
error must be elsewhere
the ccs is a real good mod you shouldnt miss out
I did measure pin 20 and found some having slightly higher voltage from 2.4v thus did the adjustment and ended up sounding worse.
Anyway, CCS is working already and sounding awesome. I am just being nitpicking trying to achieve 2.7 exact from 2.73 with further adjustment.
Last edited:
naimster,
If you can call a difference of 0.05 volts and have a difference of 0.03v, doesn't thaindicate that it is likely resistor variance?
Laters,
Drew.
If you can call a difference of 0.05 volts and have a difference of 0.03v, doesn't thaindicate that it is likely resistor variance?
Laters,
Drew.
You are absolutely right... Silly me. After all the calculations the resistor variance is an apt explanation. But it also seemed to show one should not use CCS to compensate for the variance in IV resistor.
Hey, someone had to try at some point!
Congrats on having a crack and being able to acknowledge it didn't work!
Congrats on having a crack and being able to acknowledge it didn't work!
balance shifted
The same thing happened to me. After checking the CCS current, I found 0.400 mA on one side and 0.401 mA on the other.
The same thing happened to me. After checking the CCS current, I found 0.400 mA on one side and 0.401 mA on the other.