Mystery solved - I had accidentally switched the scope probe to 1X mode. Thus, the power on spikes are not 10-20Vpp but rather 1-2Vpp which makes a lot more sense. Does not make the power on plops any less nasty though...
A pin18 driven muting relay is working pretty well, just need to make the power-on glitch go away. Here is my ghetto-style prototype
Attachments
I am looking forward to your manual soldering of a few hundred smd resistors.
Actually desoldering is the problem. You can't do it without accidental remaining soldering material, and I'm not sure how this degrades the better resistors.
Otherwise my diyer fellow would do soldering job, I can't do such a precision work. I'm not a diyer, but an audiophile.
Thank you for the resistor selection tips.
Hmm, they may used by the TOTALDAC
Do you have a real experiences with them and how easy & pricey are they and may not easy to adapt to the DAM layout
Hp
I have some problem with self-made filters (with the new version). If I make of the attached simple example (NOS) a filter with MKROM, the volume with that filter is much too low. If I change the multiplicator from 1.0 to e.g. 0.9 I get no sound at all (although when checking with filters the filter is loaded). I also tried it with type 4/8 ... that is not the problem.
Most likely I made a formatting error, but I tried for hours, and do not see it.
Any hint?
Most likely I made a formatting error, but I tried for hours, and do not see it.
Any hint?
Code:
dam1021,44100,8,5,8,1.0
05 Input FIR, 44.1 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
dam1021,48000,8,5,8,1.0
05 Input FIR, 48 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
dam1021,88200,4,5,4,1.0
05 Input FIR, 88.2 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
dam1021,96000,4,5,4,1.0
05 Input FIR, 96 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
dam1021,176400,2,5,2,1.0
05 Input FIR, 176.4 Khz Samplerate, NOS
1.000
1.000
dam1021,192000,2,5,2,1.0
05 Input FIR, 192 Khz Samplerate, NOS
1.000
1.000
dam1021,352800,1,5,1,1.0
05 Input FIR, 352.8 Khz Samplerate, Bypass
1.000
dam1021,384000,1,5,1,1.0
05 Input FIR, 384 Khz Samplerate, Bypass
1.000
dam1021,352800,8,9,8,1.0
09 Final FIR, 352 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
dam1021,384000,8,9,8,1.0
09 Final FIR, 384 Khz Samplerate, NOS
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
dam1021,352800,8,29,5,1
29 DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
1.00
-1.00
0.00
0.99996438167
-0.00
dam1021,384000,8,29,5,1
29 DC Blocking IIR, 384 Ksps, 2 Hz HP 1st order
1.00
-1.00
0.00
0.99996727561
-0.00
dam1021,352800,8,30,5,1
30 Deemphasis IIR, 352.8 Ksps, 50/15 uS
0.331085040573029
-0.577213330670734
0.251294171647457
1.85212451656979
-0.857290398119544
dam1021,384000,8,30,5,1
30 Deemphasis IIR, 384 Ksps, 50/15 uS
0.329617410213504
-0.581107678819127
0.255876017438852
1.86372441280949
-0.868110161642715Hmm, they may used by the TOTALDAC
Do you have a real experiences with them and how easy & pricey are they and may not easy to adapt to the DAM layout
Hp
They offer the model VSMP in Z foil, 0805, .01% tolerance. Around $13.00 each.
That would make for an expensive DAC modification, wouldn't it?
Hmm, they may used by the TOTALDAC
Do you have a real experiences with them and how easy & pricey are they and may not easy to adapt to the DAM layout
Hp
Price is rather high, but you will never reach the Totaldac level (see No.1 comment) without adio grade components imho.
Specification is just as high as the price is. See the link provided by tvicol previously.
Price is about $8 each with quantity discount, slightly depending on the selected subtype.
I had not used the very latest version of MKROM. With this the 0.9 scaled filters now are OK. The filters containing a 1 (as coefficient*multiplier) cause an overflow warning of MKROM, but write a filter where the "1" becomes a "0x8" ... grrr.
@Soren when I said 1.31 fixed floats are enough for the coefficients I also said MKROM should automatically scale by (2^31-1)/2^31
Coefficients 1 come naturally e.g. in NOS, bypass, Nyquest filters and should not cause an overflow/wrong-value with MKROM nor need a manual added multiplier.
(2^31-1)/2^31 is a nasty number to type especially if it can not be entered as hex quotient.
Lately I was toying with EQHQ variations v1, which features FIR2 at bank 01 only. After switching to another bank, the bank 01 FIR2 is still in the pipeline.
(dac)pi@raspberrypi ~/dac/filtry $ dam1021.py -c
INFO:dam1021:
FIR filters:
--Bank 01:
----FIR2: type(08) EQHQ_lpbr_b0, 352.8 Khz, 0-30.00Khz +-0.00000004dB, 264.60Khz -176.11dB
--Bank 02:
----FIR1: type(05) EQHQ_lpbr_b2, 44.1 Khz, 0-19.50Khz +-0.00000005dB, 22.50Khz -174.48dB
IIR filters:
--type(29) DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
--type(30) Deemphasis IIR, 352.8 Ksps, 50/15 uS
This might be or not what you would expect from the DAC.
(dac)pi@raspberrypi ~/dac/filtry $ dam1021.py -c
INFO:dam1021:
FIR filters:
--Bank 01:
----FIR2: type(08) EQHQ_lpbr_b0, 352.8 Khz, 0-30.00Khz +-0.00000004dB, 264.60Khz -176.11dB
--Bank 02:
----FIR1: type(05) EQHQ_lpbr_b2, 44.1 Khz, 0-19.50Khz +-0.00000005dB, 22.50Khz -174.48dB
IIR filters:
--type(29) DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
--type(30) Deemphasis IIR, 352.8 Ksps, 50/15 uS
This might be or not what you would expect from the DAC.
Due to the filter banks it is now possible to pair to each FIR1, per frequency, with a suited FIR2 filter. I.e.
bank1 contains FIR1 only for 44.1 and 48 kHz and their suited FIR2
bank2 contains FIR1 only for 88.2 and 96 kHz and their suited FIR2
and so on.
The only problem is that you have to switch the filters "by hand" if you need an other bank (e.g. frequency range).
@ Soren, could it perhaps be arranged that if one bank does not contain the needed frequency, automatically the first (and in the above example unique) bank, that does contain the needed frequency is selected.
You already have a nice feature in that direction, if a bank does not contain a FIR2 filter, automatically the one of the first bank is selected.
bank1 contains FIR1 only for 44.1 and 48 kHz and their suited FIR2
bank2 contains FIR1 only for 88.2 and 96 kHz and their suited FIR2
and so on.
The only problem is that you have to switch the filters "by hand" if you need an other bank (e.g. frequency range).
@ Soren, could it perhaps be arranged that if one bank does not contain the needed frequency, automatically the first (and in the above example unique) bank, that does contain the needed frequency is selected.
You already have a nice feature in that direction, if a bank does not contain a FIR2 filter, automatically the one of the first bank is selected.
@ Soren:
After loading a new set of filters (via: +++ download etc during playback) the old trick to switch the inputs to trigger the filter to become loaded, does not work any more reliable. I frequently ended up in obscure mixtures of filters. The filters are correctly "somewhere", the latest when I do a power cycle everything is as it is supposed to.
Perhaps a dedicated command to trigger a reload of all filters to the FPGA would be the safest.
After loading a new set of filters (via: +++ download etc during playback) the old trick to switch the inputs to trigger the filter to become loaded, does not work any more reliable. I frequently ended up in obscure mixtures of filters. The filters are correctly "somewhere", the latest when I do a power cycle everything is as it is supposed to.
Perhaps a dedicated command to trigger a reload of all filters to the FPGA would be the safest.
The isolated option I don't tested, but the non-isolated RS-232 option degrades the low level performance by about 20dB at least with noise garbage, but for me are OK because I use DAM stand-alone. A thing to investigate...
I have sucessfully updated the DAM, after one failed tentative (I have forgot to put "update" command and DAM not worked...).
About the sample rate change noise, this also occurs in S-PDIF mode (I only use S-PDIF in my listening equipment, due to my various digital audio sources), but in S-PDIF the change/skip of tracks/pause/stop is completely silent.
The measured performace (and audio result) is the same so the audio processing is AT LEAST equal to before, if not better, at least with new EQHQ filters I have measured.
About the sample rate change noise, this also occurs in S-PDIF mode (I only use S-PDIF in my listening equipment, due to my various digital audio sources), but in S-PDIF the change/skip of tracks/pause/stop is completely silent.
The measured performace (and audio result) is the same so the audio processing is AT LEAST equal to before, if not better, at least with new EQHQ filters I have measured.