Hi ... in being a bit generalistic I'd say that the combination of RF & CF that gives the lowest load on the FF/output device. Some time ago I measured distortion on a DSD converter I made with potatosemi FFs ("differential" out) and when the current got above ~1 mA distortion rose significantly. I reckon something similar could be the case with other FFs.
So, personally, I'd choose RFs as high as feasible and then select the CF value to comply with this.
Of course there may be other perspectives on this ...
Cheers,
Jesper
Help!
I can only get the left DSD channel to work properly on the SE Relay RC filter board that I got from dwjames for my JLSounds USB board. There is no measurable analog signal after the RC filter coming from the RDSD - just the 1.65V DC offset. The RDSD output pin measures no variation and displays a much lower frequency than the LDSD signal. The LDSD outputs a large and lively analog signal after the RC filter.
The relay clicks on and the 5V ma reading more than doubles when playing music. I am using HQPlayer on a core i3 laptop running windows 7 at less than 15% CPU usage with all other programs turned off. The green LED on the relay board never turns on. The DSD clock on JLSounds pin 15 measures the same frequency as the one chosen in the HQPlayer settings. Do I have a bad JLSounds board or have I failed to configure HQPlayer properly? is there a jumper on the JLSounds board that might be in the wrong position? The settings are at the lowest speeds and I am playing a CD. Earlier, 96Khz FLAC files on the computer HD did not play any better.
Any suggestions?
I can only get the left DSD channel to work properly on the SE Relay RC filter board that I got from dwjames for my JLSounds USB board. There is no measurable analog signal after the RC filter coming from the RDSD - just the 1.65V DC offset. The RDSD output pin measures no variation and displays a much lower frequency than the LDSD signal. The LDSD outputs a large and lively analog signal after the RC filter.
The relay clicks on and the 5V ma reading more than doubles when playing music. I am using HQPlayer on a core i3 laptop running windows 7 at less than 15% CPU usage with all other programs turned off. The green LED on the relay board never turns on. The DSD clock on JLSounds pin 15 measures the same frequency as the one chosen in the HQPlayer settings. Do I have a bad JLSounds board or have I failed to configure HQPlayer properly? is there a jumper on the JLSounds board that might be in the wrong position? The settings are at the lowest speeds and I am playing a CD. Earlier, 96Khz FLAC files on the computer HD did not play any better.
Any suggestions?
@carlsor: ... Hmm... when reading your post I'm not entirely sure where you've measured the RDSD signal - on the JLSounds board or after the RC filter? If you measured it after the RC filter I may suggest measuring directly on the JLSounds board RDSD output (with the SE board disconnected to eliminate any potential interference here) e.g. using an identical sine wave (wavegene works well for me) and then comparing the channels. If there's still no RDSD signal (or it's at a lower frequency) what happens should happen either in the JLSounds board or in the computer. Is it possible to swap channels in HQPlayer so that you can see if the same happens to the left channel when the channels are swapped?
About the lower frequency present I once experienced this with my own setup (I was using windows media player 😱) ) and it turned out that it was a reflection down in frequency due to too low sampling frequency which WMP was not considering (or maybe I weren't 🙄).
Otherwise I am thinking that maybe JLSound have experienced this before?
@Theodosis O.:
You are welcome 😉 ... I reckon that whether 3k3/1 nF is suitable also depends on your ancillary equipment: E.g. what is the input impedance of your subsequent amplifier? What type of cable do you use? Overall - how "wideband" is your system in practice?
3k3/1 nF gives about 0.5 mA current draw (appr. +/- 1.65V output relative to the FFs ~ 1.65V DC level when a DSD signal is present) and a cut-off around 48 kHz (as you may know).
Personally I'd move the cut-off frequency up but I prefer "wideband" systems and maybe your preferences are different ... A suggestion could be to experiment a bit and then see where you like the sound best? For trials I might suggest Vishay's RN60/CMF60 resistors - available at an IMO quite reasonable price and with reasonable sound. Again, personally, I would use C0G capacitors for this (> 100 VDC specification; courtesy Samuel Groner) size 0805 or 1206 (TDK or Kemet manufacturers). All is available from e.g. Mouser or Digikey, or when it comes to the resistors a search for " 40x Dale Resistor (RN60C/RN60D) 1% Customerized Values " on ebay allows for lower quantities.
Cheers,
Jesper
About the lower frequency present I once experienced this with my own setup (I was using windows media player 😱) ) and it turned out that it was a reflection down in frequency due to too low sampling frequency which WMP was not considering (or maybe I weren't 🙄).
Otherwise I am thinking that maybe JLSound have experienced this before?
@Theodosis O.:
You are welcome 😉 ... I reckon that whether 3k3/1 nF is suitable also depends on your ancillary equipment: E.g. what is the input impedance of your subsequent amplifier? What type of cable do you use? Overall - how "wideband" is your system in practice?
3k3/1 nF gives about 0.5 mA current draw (appr. +/- 1.65V output relative to the FFs ~ 1.65V DC level when a DSD signal is present) and a cut-off around 48 kHz (as you may know).
Personally I'd move the cut-off frequency up but I prefer "wideband" systems and maybe your preferences are different ... A suggestion could be to experiment a bit and then see where you like the sound best? For trials I might suggest Vishay's RN60/CMF60 resistors - available at an IMO quite reasonable price and with reasonable sound. Again, personally, I would use C0G capacitors for this (> 100 VDC specification; courtesy Samuel Groner) size 0805 or 1206 (TDK or Kemet manufacturers). All is available from e.g. Mouser or Digikey, or when it comes to the resistors a search for " 40x Dale Resistor (RN60C/RN60D) 1% Customerized Values " on ebay allows for lower quantities.
Cheers,
Jesper
If the SE board you have from James is the one I'm thinking it is it worked previously and doesn't need any jumper settings for the relay section but if you post a picture I can confirm.
I've seen this sort of behaviour before with some of my early experiments; I was convinced that I had damaged my JLSounds board but it turned out to be OS/driver problem.
I have found that Linux distributions, with the recent kernel updates for DSD, have been better than Windows, they seem to work more consistently and with higher rate DSD too.
For what its worth I still run HQPlayer on Win10 but use an HQPlayer NAA device running Linux connected to the JLSounds; this gives me a nice easy interface and the advantage of some isolation of the 'grunt' PC from the audio system.
Anyway, if Linux is of interest maybe consider something like this;
AudioLinux - The audiophile realtime plug & play operative system
or back a while I posted a document on how to install/configure a suitable Linux distro yourself, though that was aimed at setting up an NAA.
@nautibuoy:
... very interesting link I would say - have you any experience with using it? Or maybe others here have experience with using it?
Cheers,
Jesper
... very interesting link I would say - have you any experience with using it? Or maybe others here have experience with using it?
Cheers,
Jesper
Yes, I was very impressed with it, easy to use and very good sound quality. I only stopped using it because I planned to go to Roon (which at the time
was not available on Linux) but it is back on my 'radar' to try again now.
Ray
Sob, yes, it is! Sorry for this error, again.
I've built it with "symple" Tyco MF and WIMA, but planning to swap resistors with a pair of RN55 (71-RN55D3900FB14), if I get to properly desolder them.
I have a SET with Zin=35k and use short RCA cables. I have in hand:
3k3 , 1nF mica
120 , 33nF mkp
390 , 10nF mkp
1k , 3n3 mkp. All resistors are melf 0204.
I found these combinations from previews posts.
From where to start ?
Thanks Merlin, I'll try this.
Did you try other combinations and found this is the best ?
Did you try other combinations and found this is the best ?
Last edited:
@ghiglie:
Hi ... Just in case it is interesting to you ... simon7000 (a diyaudio member as well) made quite some distortion measurements on resistors and found that for the same types doubling the resistor's wattage rating reduced distortion by 6 dBs. The RN55 is to my memory a 0.25 W or 0.125 W resistor - I personally would go for the RN70E types which he found to have the lowest distortion readings. But, again, there are many ways in audio ;-)
Cheers,
Jesper
P.S.: For resistors it may actually be an advantage with a big size as this increases inductivity and thus "passive" HF filtering. For capacitors on the other hand they - generally speaking - should be as small as possible (low inductance) because the aim here is to conduct, or short, as much as possible of HF energy to ground. Thus, the HF energy does not continue to the rest of the system.
Hi ... Just in case it is interesting to you ... simon7000 (a diyaudio member as well) made quite some distortion measurements on resistors and found that for the same types doubling the resistor's wattage rating reduced distortion by 6 dBs. The RN55 is to my memory a 0.25 W or 0.125 W resistor - I personally would go for the RN70E types which he found to have the lowest distortion readings. But, again, there are many ways in audio ;-)
Cheers,
Jesper
P.S.: For resistors it may actually be an advantage with a big size as this increases inductivity and thus "passive" HF filtering. For capacitors on the other hand they - generally speaking - should be as small as possible (low inductance) because the aim here is to conduct, or short, as much as possible of HF energy to ground. Thus, the HF energy does not continue to the rest of the system.
Last edited: