Hello, everyone.
I would like to share my experience with dam1021 dac's
currently comparing dam1021 fully balanced version mixed filter with PCM1704 x 4chips in nos mode, sowter transformers. Everything feeds on A123 batteries. Without voltage regulators as possible. Usb interfaces: waveIO 3,3V battery modded / m2tech OEM / singxer F1 battery modded ( F1 is the best, then m2tech, then waveio).
It little bit winning in dynamic over PCM1704, but completely loosing in speed/ resolution. I was hoping for better performance from dam dac.
I would like to share my experience with dam1021 dac's
currently comparing dam1021 fully balanced version mixed filter with PCM1704 x 4chips in nos mode, sowter transformers. Everything feeds on A123 batteries. Without voltage regulators as possible. Usb interfaces: waveIO 3,3V battery modded / m2tech OEM / singxer F1 battery modded ( F1 is the best, then m2tech, then waveio).
It little bit winning in dynamic over PCM1704, but completely loosing in speed/ resolution. I was hoping for better performance from dam dac.
It is not necessary to run the batteries without regulator/chargers for best sound.
How long are the voltages staying constant while playing music? Are you using the onboard regs of the SOEKRIS using the batteries to power the onboard regulators? If this is the case your are missing most of the advantages of battery power.
Is the DAC you are comparing to using the same basic powering scheme as what you have used for the SOEKRIS?
Most people condemn the SOEKRIS for having less than stellar dynamics. I have not been bothered by this but it does stand out in your assessment.
How do you get -4V from a LT3042 ??
The reason to the distortion at -3.2V only is the jfets, their gate-drain diodes start to conduct without the higher off voltage to the gates. Or you could just remove the jfets, but then you get the turnoff thumps.... Or connect -6.4V to J2 pin 9 instead of -3.2V.
The onboard vref regulators are better then an external LT3042, but then, 6.4V will kill an opa365.... If you insists then ensure your regulators are very close to the dam1021 and connected with thick wires.... You might want to verify than vref ripple is less than 1mV pp.
You might want to use the "soft" filter settings, sounds like it will more to your liking....
Since your're using the unbuffered outputs you need to keep the wires reasonable short, in balanced mode the Zout is 1280 ohm.
And btw, your postings really belong in the main dam1021 thread....
Last edited:
batteries should be without center tap
but the scheme become little bit complex, it's easier with 3,2V instead of 4V
Last edited:
Yes, floating supplies will do, just not shown like that on your earlier drawing....
I have been asked for the TXT of the filter because people need to ecualize and use DSD. I try to explain what I've done by putting together filters made by Soekris and Paul. Hoping they not to be bothered by the use of this. Otherwise, the administrator could deletes them please.
F4 is the first that provides Soekris by default. Soekris Linear Phase, 44.1KHz. 0-19.7Khz -1dB 22.3Khz -150dB. The sound seems similar to the integrated commercial DAC. It has a large slope of attenuation and produces an impulsive response of pre- and post ringing and poor phase response at high frequencies.
F5 is "EQHQ_lpbr_b4" Linear Fase published by Paul that has an even greater slope of attenuation. But its the sweetest and most pleasant sound of all filters. It produces a much better impulse response than the F4 with less ringing and a much better phase response at high frequencies. Unlike this filter published by Paul at frequencies above 88.2 Khz, 96 Khz, which dropped from 20 Khz. approximately, I performed the calculations in Excel by decimating and calculating attenuation in the impulse response to 88.2, 96, 192 ... Khz. Etc. So the attenuation at these frequencies is close to half of the sampling frequency. I estimate that: 88.2 Khz 0-40.9 Khz + -0.00000005 dB, 45 Khz -174 dB 96.0 Khz, 0-44.5 Khz + -0.00000005 dB, 48.9 Khz -174 dB The audio pass over 40 Khz. My bat pet are happy.
F7 is a NOS filter. Although some people have commented that DAM DAC an oversampling occurs in the signal. Is not true. This DAC repeat N times the signal without any skip or switching defect in each repetition, it is easily observable on the oscilloscope that the output is a perfect NOS filter with signal identical to that of any other NOS. The NOS have qualities and defects. The impulse response is perfect there is no phase variation of the high frequencies, the percussions and attacks instruments is very natural. But also the progressive attenuation of the signal about 1 Khz. produces a dark sound.
F6 is a filter published by Paul "C128dp Linear Phase NYQUIST" is the most similar to a NOS filter, only uses 8 samples at the front and 8 samples at the back, and the weight of the samples decays very quickly. The impulsive response is perfect. It is the closest to NOS but with the quality of having a flat response to almost 19 Khz. It also has a perfect phase up to high frequencies. And it also keeps zeros every N values in the impulse response which means that it reproduces exactly the same samples as the original signal, which filters F4 and F5 do not. They are several things that I wanted and this filter make. That I like most of how the percussions sound and instruments attacks. The loudspeakers in my house are temporal aligned and have a perfect impulsive response so this filter makes a difference to F4 and F5. Very similar to F7 but with trebles. I have also done with Excel 88.2, 96 Khz and above, decimating and calculating the attenuation from the original to 44 and 48 Khz. Similar in F2 filters.
F4 is the first that provides Soekris by default. Soekris Linear Phase, 44.1KHz. 0-19.7Khz -1dB 22.3Khz -150dB. The sound seems similar to the integrated commercial DAC. It has a large slope of attenuation and produces an impulsive response of pre- and post ringing and poor phase response at high frequencies.
F5 is "EQHQ_lpbr_b4" Linear Fase published by Paul that has an even greater slope of attenuation. But its the sweetest and most pleasant sound of all filters. It produces a much better impulse response than the F4 with less ringing and a much better phase response at high frequencies. Unlike this filter published by Paul at frequencies above 88.2 Khz, 96 Khz, which dropped from 20 Khz. approximately, I performed the calculations in Excel by decimating and calculating attenuation in the impulse response to 88.2, 96, 192 ... Khz. Etc. So the attenuation at these frequencies is close to half of the sampling frequency. I estimate that: 88.2 Khz 0-40.9 Khz + -0.00000005 dB, 45 Khz -174 dB 96.0 Khz, 0-44.5 Khz + -0.00000005 dB, 48.9 Khz -174 dB The audio pass over 40 Khz. My bat pet are happy.
F7 is a NOS filter. Although some people have commented that DAM DAC an oversampling occurs in the signal. Is not true. This DAC repeat N times the signal without any skip or switching defect in each repetition, it is easily observable on the oscilloscope that the output is a perfect NOS filter with signal identical to that of any other NOS. The NOS have qualities and defects. The impulse response is perfect there is no phase variation of the high frequencies, the percussions and attacks instruments is very natural. But also the progressive attenuation of the signal about 1 Khz. produces a dark sound.
F6 is a filter published by Paul "C128dp Linear Phase NYQUIST" is the most similar to a NOS filter, only uses 8 samples at the front and 8 samples at the back, and the weight of the samples decays very quickly. The impulsive response is perfect. It is the closest to NOS but with the quality of having a flat response to almost 19 Khz. It also has a perfect phase up to high frequencies. And it also keeps zeros every N values in the impulse response which means that it reproduces exactly the same samples as the original signal, which filters F4 and F5 do not. They are several things that I wanted and this filter make. That I like most of how the percussions sound and instruments attacks. The loudspeakers in my house are temporal aligned and have a perfect impulsive response so this filter makes a difference to F4 and F5. Very similar to F7 but with trebles. I have also done with Excel 88.2, 96 Khz and above, decimating and calculating the attenuation from the original to 44 and 48 Khz. Similar in F2 filters.
Last edited:
Have you considered a "F5" with for example the filter for Fs=96khz:
a) Keeping same steepness but start filtering from 30kHz for higher attenuation at Fs.
b) use less steep filter for same attenuation at Fs.
a) is probably the most interesting option of the two.
Bat pet still quite happy. Master even more happy?
//
a) Keeping same steepness but start filtering from 30kHz for higher attenuation at Fs.
b) use less steep filter for same attenuation at Fs.
a) is probably the most interesting option of the two.
Bat pet still quite happy. Master even more happy?
//
Hi
Just got my board, its working well, however there is one thing I haven't been able to figure out despite reading hundreds of posts
.
I want to use this in an active crossover, so obviously need to filter for woofer and tweeter. I can see how to do that, however how do you specify different filters for the left and right channels?
I am probably being dim but help would be appreciated
Thanks
Just got my board, its working well, however there is one thing I haven't been able to figure out despite reading hundreds of posts
.I want to use this in an active crossover, so obviously need to filter for woofer and tweeter. I can see how to do that, however how do you specify different filters for the left and right channels?
I am probably being dim but help would be appreciated
Thanks
Last edited:
don't know how I missed it but onethumbs newest filter pak has working 192khz f6 c128 quasinos filter. I should have loaded it a long time ago. thanks onethumb and paul and soren for awesome filter choices. using a usb to aes converter now so dsd less important to me.. quasinos filter rocks
not sure anyone ever posted a video on how to change filters.. this video is of a commercial soekris dac, the vinshine audio dac-r2r-ref which already has a serial port mounted.. hope it can help someone who is scared of terminals and command line
https://www.youtube.com/watch?v=A03jSNR31LA
https://www.youtube.com/watch?v=A03jSNR31LA
