I would imagine silence if properly designed, that or it switches to 44.1kHz if that's a valid option.
So, you really ARE the spokesperson for the peanut gallery? (I think Jan and Mooly misunderstood ... 'for' is used here in the meaning of 'on behalf of').
Last edited:
Actually I had this happen. The console was set to 44.1K for that input but for some reason it was expecting 48K like all the other inputs and outputs.
The result was bursts of loss of signal (silence) then back to music. That is because the data was too slow and the buffer would regularly run out of data.
An interpolating rate converter fixed the problem. (Extra credit question should the interpolation be linear?)
Now what happens if the SPDIF CD send signal is running say 44099 samples per second?
Last edited:
With two trims (two degrees of freedom) either Vos AND drift or Vos AND 2nd's can be trimmed without matching.
Overhead, chip design, protocols etc. On paper there seems to be plenty of bandwidth however in practice it doesn't work. The bandwidth is shared and between other reserved use and the way the system communicates 24/96 is the highest you can get on a 12 Mbps link. Also if you have even a mouse sharing the link you will have problems. Most PC's have only 1 or maybe 2 USB controllers so all the traffic passes through the same controller.
With Windows you need 3rd party drivers to get past the 24/96 limit (even on Win 8.1).
Dan-
if you measure the following (assuming you have the resources available) it can be enlightening: First the AC (and DC) voltage between the chassis when the isolator is in the link. Then the ac current passing across the grounds.
if you measure the following (assuming you have the resources available) it can be enlightening: First the AC (and DC) voltage between the chassis when the isolator is in the link. Then the ac current passing across the grounds.
Measured with ITT-Metrix RMS 20kHz+ BW meter across host and device earths on the isolator pcb.
Charger connected - 63uA AC, 0.1uA DC, 36.10 VAC, 0.020 VDC.
Charger disconnected - 0.6uA AC, 0.1uA DC, 0.118 VAC, 0.000 VDC.
With charger disconnected, AC voltage value changes when PC to isolator pcb USB cable is touched/handled, also when touching netbook vga connector metal shroud ....didn't check with charger connected and powered.
System sounds subtly different when connected to powered charger...eg blues fuzz guitar loses definition and slight veiling.
Dan.
Charger connected - 63uA AC, 0.1uA DC, 36.10 VAC, 0.020 VDC.
Charger disconnected - 0.6uA AC, 0.1uA DC, 0.118 VAC, 0.000 VDC.
With charger disconnected, AC voltage value changes when PC to isolator pcb USB cable is touched/handled, also when touching netbook vga connector metal shroud ....didn't check with charger connected and powered.
System sounds subtly different when connected to powered charger...eg blues fuzz guitar loses definition and slight veiling.
Dan.
Last edited:
On second thoughts, the AC current with charger measurement is the leakage current available, not the in practice leakage current.
To measure this I would need to break host earth or device earth connection and insert the current meter.
Dan.
To measure this I would need to break host earth or device earth connection and insert the current meter.
Dan.
Last edited:
SW --- That is very useful, of course. Can you show a sketch of how/where the R's are that would accomplish that?
You can use your own topology or the one JC put up.
How is DC balance/offset affected when trimming for min 2H in the circuit of choice?
THx-RNMarsh
Last edited:
You'll have to be patient it was two or three years ago, I did post some pictures which I think showed simultaneous trim of Vos and Aol on an op-amp. At the time we were arguing about the wide band open loop gain nonsense (gee, what else is new). I was showing that trimming DC Aol to essentially infinite (and even a large negative number) made no difference in the AC response at higher frequencies.
This thread just circles and circles.
I would think it depends on how good the PLL pull in range was on the receiving end. Isn't the clock derived from the incoming data stream ? The receiver doesn't even need be crystal controlled. So 44099 incoming just means the source players internal clock is running slow, and so the replayed disc.
(Ever used two different SPDIF (optical) feeds from two source components in parallel ? I do 😉)
Trick question! If the converter uses a PLL to sync up then that determines the amount of jitter during conversion. If it uses a buffer and reclocks it then the buffer and start up delay must be long enough to cover a fairly long period. A smart algorithm would of course recenter the buffer at a zero level passage.
If it uses a clock and no buffer then there will be bits of silence. So a discrete converter can vary from simple and inexpensive to quite complex.
Now for interpolation, linear is a good fit midrange levels but not near maximum or minimum levels, so a very smart algorithm would actually look at that also.
Now as most SPDIF outputs are within .01% of clock frequency it is not a big issue, just a bit of finesse.
Of course this is absolute accuracy jitter will bounce around that.
I was just playing with the Adjust+ record player measurement system and to my dismay the RPM of my player was shown as way too low, out of tolerance. Yet it sounded fine with music. Hmmm. Then I discovered that Adjust+ was assuming the soundcard sampled at 44.1k whereas the soundcard though it had to do 48k... 😎
Jan
- Status
- Not open for further replies.