tubee said:You use a trannie for i/v? Tell me about inductors..
.....
Trannie does not ring....
Bernhard said:Tubee you could either use 500 ohm I/V or stack PCM56 2x.
I can give you new filter values.
Transformer without termination rings much worse than any filter.
Filter ringing is no problem.
Too bad I don't have a digital camera.
The square wave on my scope does NOT show any ringing.
(how using a trannie without termination escapes me)
When i stack pcm56 i must have some extra. Have been thinking on stacking too btw.
Just got the parts for the filter, have to check them with CL meter.
All those measurements/scope vieuws. My ears tell if it sounds reasonable. I discovered the most important thing wich counts to me now is a lineair response. And then with all gear considered, and a preamp without tone control. Every part has his influence, cable, diode caps, and so on. Considered a baxandall tone control also some times on the preamp.
OT: curious: have a CD880 to mod and all i do is fiddling with a PCM56...
Just got the parts for the filter, have to check them with CL meter.
All those measurements/scope vieuws. My ears tell if it sounds reasonable. I discovered the most important thing wich counts to me now is a lineair response. And then with all gear considered, and a preamp without tone control. Every part has his influence, cable, diode caps, and so on. Considered a baxandall tone control also some times on the preamp.
OT: curious: have a CD880 to mod and all i do is fiddling with a PCM56...
tubee said:
Just got the parts for the filter, have to check them with CL meter.
Then you are back at 100 ohm I/V.
As you have the parts already, maybe built it and report to us what you hear....
QSerraTico_Tico said:stacking smears the highs
![]()
Stacking/paralleling is the only way to get low output impedance and sufficiant output voltage when using passive I/V.
I hear no smear.
Maybe linear interpolation smears
Bernhard said:
Stacking/paralleling is the only way to get low output impedance and sufficiant output voltage when using passive I/V.
I hear no smear.
Maybe linear interpolation smears![]()
Yes of course. You never heard of buffers, post-amplification and so forth....
Bernhard said:
Then you are back at 100 ohm I/V.
As you have the parts already, maybe built it and report to us what you hear....
Yes i will, costed only 6 eu on parts. Not handwinded coils though
I can see if i can crank up the tube amp a little
QSerraTico_Tico said:
Yes of course. You never heard of buffers, post-amplification and so forth....
![]()
I have eliminated all of them
There isn't any part in the direct signal path between the Iout pins and the input of the power amp except for straight wire and wound wire.
How does that sound ?
It plays now with your suggested last posted filter. I have unwinded previous coils to the stated values, and even trimmed the capacitors simply by grinding them off! Even with a 20M resistance measurement i could not measure a resistance.
Great LC meter also i bought, very easy to tweak the filter components with.
Well: Sounds good, but need some further sound testing on my main speakers/amp downstairs.
First impressions: Very wide soundstage, still not deep. All instruments are good articulated and detailed, have a natural appearance. Treble is more obvious now, good detail but it is only 16 bit, can have a very faint sharp edge over it: cymbals have a small sort of noise around it, but on the contrary by a sort of lifted curtain-effect the cymbals are more obvious now. Dig OS has a more "covered" sound. Electric Guitars from Santana screams really through the room as it should be, but mids have a sort of distorsion (a lot of 2nd harmonic?) but some lesser then without filter. Overall i like this dac. This evening i listen to it further.
An externally hosted image should be here but it was not working when we last tested it.
Well: Sounds good, but need some further sound testing on my main speakers/amp downstairs.
First impressions: Very wide soundstage, still not deep. All instruments are good articulated and detailed, have a natural appearance. Treble is more obvious now, good detail but it is only 16 bit, can have a very faint sharp edge over it: cymbals have a small sort of noise around it, but on the contrary by a sort of lifted curtain-effect the cymbals are more obvious now. Dig OS has a more "covered" sound. Electric Guitars from Santana screams really through the room as it should be, but mids have a sort of distorsion (a lot of 2nd harmonic?) but some lesser then without filter. Overall i like this dac. This evening i listen to it further.
You can reduce R1 to 30 ohm, but it will make no real difference.
R1 is there because the average Ri 1165 ohm of the DAC is parallel to Ri/v.
I found that R1 in this less steep filter is not critical yet.
So no adjustment.
But you could check the frequency response with a scope and a test CD: 1k, 2k, 3k, 4k, 5k, 6k, 8k, 10k, 12k, 15k, 18k, 20k. Or a slow sweep.
I don't claim that this filter is the holy grail but it will fix the treble rolloff issue.
IMHO for correct reconstruction of the waveform, nonos has to be filtered with very steep 60dB / oct., no matter how bad phase response may get.
This filter has only 18dB / oct.
Do you use PCM56 with receiver chip or in a Philips player ?
I would make sure there are no problems from implementation with the SAA chip, I had strange and subtle problems even with one inverter for glue logic....
It is very unlikely that a noticeable distortion in the midrange comes from the filter.
I am missing the words how to describe the sound of not completely reconstructed nonos waveform.
Perhaps the treble is a bit dry, not liquid.
Also if you use a tube gain/buffer, that could alter the sound and should be avoided when judging the sound of the DAC.
What coils do you use ? They should have no iron core.
R1 is there because the average Ri 1165 ohm of the DAC is parallel to Ri/v.
I found that R1 in this less steep filter is not critical yet.
So no adjustment.
But you could check the frequency response with a scope and a test CD: 1k, 2k, 3k, 4k, 5k, 6k, 8k, 10k, 12k, 15k, 18k, 20k. Or a slow sweep.
I don't claim that this filter is the holy grail but it will fix the treble rolloff issue.
IMHO for correct reconstruction of the waveform, nonos has to be filtered with very steep 60dB / oct., no matter how bad phase response may get.
This filter has only 18dB / oct.
Do you use PCM56 with receiver chip or in a Philips player ?
I would make sure there are no problems from implementation with the SAA chip, I had strange and subtle problems even with one inverter for glue logic....
It is very unlikely that a noticeable distortion in the midrange comes from the filter.
I am missing the words how to describe the sound of not completely reconstructed nonos waveform.
Perhaps the treble is a bit dry, not liquid.
Also if you use a tube gain/buffer, that could alter the sound and should be avoided when judging the sound of the DAC.
What coils do you use ? They should have no iron core.
Wow 60dB, allmost a brickwall. Have seen that before here in diya don't know who.
What i like on this I/V and filter is the integration of it together.
Have listened to it some longer, it warmed up a bit. I thought a 1541 can give nice stereo effects, but this pcm56 is very nice! Heard detailes before i never did.
I hooked it up in a cdp on saa7210 I2S with glue logic. If i have some time can make some photographs of different sine wave frequencies, and maybe some other waves also.
What i like on this I/V and filter is the integration of it together.
Have listened to it some longer, it warmed up a bit. I thought a 1541 can give nice stereo effects, but this pcm56 is very nice! Heard detailes before i never did.
I hooked it up in a cdp on saa7210 I2S with glue logic. If i have some time can make some photographs of different sine wave frequencies, and maybe some other waves also.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Source & Line
- Digital Line Level
- SAA7210 to PCM56, a Dutch-American connection