Hi,
I have posed this question over in the analog board, but noone seems to know the answer: I plan to try vinyl playback through my PC (Linux/SoX), with an older RME-Card (Digi96/8 PST)
The Anti-RIAA shall take place by software filters, so the unfiltered input would go into the ADC of the card. Apart from having -20dB from 0dB @1kHz due to the riaa-curve, what are the requieries for the ADC:
What is the best ouput impdance for a preamp (between cartridge and ADC), as low as possible, or what would you consider reasonable?
Can an ADC take residual DC (when using a servo, for instance), or must it be cap-coupled (or are there coupling caps onboard)?
One of my biggest concerns is, how to handle plopps from scratches, any ideas? Could it be filtered by software? How does the ADC behave when the signal input voltage is exceeded?
There are piles of threads for DIY-DACs, is there anything comparable for ADCs?
Any insight is welcome,
Rüdiger
I have posed this question over in the analog board, but noone seems to know the answer: I plan to try vinyl playback through my PC (Linux/SoX), with an older RME-Card (Digi96/8 PST)
The Anti-RIAA shall take place by software filters, so the unfiltered input would go into the ADC of the card. Apart from having -20dB from 0dB @1kHz due to the riaa-curve, what are the requieries for the ADC:
What is the best ouput impdance for a preamp (between cartridge and ADC), as low as possible, or what would you consider reasonable?
Can an ADC take residual DC (when using a servo, for instance), or must it be cap-coupled (or are there coupling caps onboard)?
One of my biggest concerns is, how to handle plopps from scratches, any ideas? Could it be filtered by software? How does the ADC behave when the signal input voltage is exceeded?
There are piles of threads for DIY-DACs, is there anything comparable for ADCs?
Any insight is welcome,
Rüdiger
I've commented on using a single pole in the analog front end to maintain dynamic range in a few of these digital riaa threads:
http://www.diyaudio.com/forums/showthread.php?postid=1165181#post1165181
http://www.diyaudio.com/forums/showthread.php?postid=1500611#post1500611
http://www.diyaudio.com/forums/showthread.php?postid=1165181#post1165181
http://www.diyaudio.com/forums/showthread.php?postid=1500611#post1500611
The impedance requirements depend a lot on what cartridge you're using. If you're using MM, you MUST use a Hi-Z input (input impedance in the MOhm range). If you're using MC, especially LOMC, generally anything will work, although you will likely get improved noise performance with a preamp that has a a lower impedance, down to 10x the cartridge resistance.
The easiest way to handle scratches is just to reduce the gain until they're covered. My headroom is really high - about +25db ref. 5cm/s. This sounds pretty insane until you notice that the noise spectrum of the cartridge+preamp is still way, way above the noise floor of the ADC.
ADCs vary widely with respect to coupling and overload conditions, but generally, you'll probably know quite well when they become issues....
Note, you already have your ADC in the form of the PST (good card btw!). You're looking for a flat preamp.
Have you considered rewiring your tonearm for balanced cabling and XLR?
The easiest way to handle scratches is just to reduce the gain until they're covered. My headroom is really high - about +25db ref. 5cm/s. This sounds pretty insane until you notice that the noise spectrum of the cartridge+preamp is still way, way above the noise floor of the ADC.
ADCs vary widely with respect to coupling and overload conditions, but generally, you'll probably know quite well when they become issues....
Note, you already have your ADC in the form of the PST (good card btw!). You're looking for a flat preamp.
Have you considered rewiring your tonearm for balanced cabling and XLR?
jcx is making the right suggestion IMHO. While it's true that good modern A/D converters probably have a low enough noise floor to do a pure software riaa approach, it really leaves you very little headroom to accomodate pops clicks etc. Consider that the standard riaa curve imposes about a 40dB cut in the high frequencies. Pops and clicks have very significant high frequency content, and so by bypassing analog domain eq these artifacts are coming in 40dB hotter than they otherwise would. I think there is a very real problem with overload margin when doing it this way.
Even aside from pops/clicks, adopting an approach where you are *intentionally* throwing away 40dB of S/N seems less than ideal.
By using a single 50Hz pole, you get most of the benefits of both approaches. It's still a very very simple analog circuit, but you get the benefit of much easier post-processing due to the supression of HF energy and only have to apply 6-8 dB of digital eq.
As for the circuit specifics, what I would look at is a circuit centered around the OPA1632. This appears to be designed with the idea of an A/D front end (look at the datasheet), and due to the ability to tie Vocm to the reference of the A/D, can be done with direct coupling. Implement the 50Hz pole in the feedback loop, and you have a pretty nice looking circuit IMHO. You can probably even hijack the TP Ivy board to do this. Since you have a good pro card with (presumably) balanced inputs, it would be a shame IMHO to use an unbalanced approach.
I guess the downside of this would be that you now need your own custom eq module, or an off-the-shelf RIAA eq that allows bypassing the 50 Hz eq.
Even aside from pops/clicks, adopting an approach where you are *intentionally* throwing away 40dB of S/N seems less than ideal.
By using a single 50Hz pole, you get most of the benefits of both approaches. It's still a very very simple analog circuit, but you get the benefit of much easier post-processing due to the supression of HF energy and only have to apply 6-8 dB of digital eq.
As for the circuit specifics, what I would look at is a circuit centered around the OPA1632. This appears to be designed with the idea of an A/D front end (look at the datasheet), and due to the ability to tie Vocm to the reference of the A/D, can be done with direct coupling. Implement the 50Hz pole in the feedback loop, and you have a pretty nice looking circuit IMHO. You can probably even hijack the TP Ivy board to do this. Since you have a good pro card with (presumably) balanced inputs, it would be a shame IMHO to use an unbalanced approach.
I guess the downside of this would be that you now need your own custom eq module, or an off-the-shelf RIAA eq that allows bypassing the 50 Hz eq.
Thanks to all for your suggestions so far. As I thought, there are several issues that have to be taken into account. This RME card is unbalanced, so I have to stick with an unbalanced circuit this time.
Yes, the whole point of my question is that I will design and build the preamp for myself. Do I get it right, then, that you are talking about a low pass filter with its -3dB-corner at 50Hz?
The goal of it all is to check if a software riaa filter can put up with a good analog-only solution.
Rüdiger
Yes, the whole point of my question is that I will design and build the preamp for myself. Do I get it right, then, that you are talking about a low pass filter with its -3dB-corner at 50Hz?
The goal of it all is to check if a software riaa filter can put up with a good analog-only solution.
Rüdiger
If you can read Dutch language (or try a Google translate), this site has some articles about MD-Preamp (DIY).
It discusses about Mechanical resonance MD cartridge etc. and an good riaa
processor.
Just look at http://www.breem.nl/pgartikelen.htm#Mechanische resonantie bij pick-up elementen
almost at the bottem the article from Hans v. Maanen in PDF (also in Dutch) form.
It discusses about Mechanical resonance MD cartridge etc. and an good riaa
processor.
Just look at http://www.breem.nl/pgartikelen.htm#Mechanische resonantie bij pick-up elementen
almost at the bottem the article from Hans v. Maanen in PDF (also in Dutch) form.
I think the S/N arguments can go both ways. It's true that groove velocities tend to be highest around 10khz. But, human hearing is most sensitive to the 1-10khz range - so in that sense, it makes a lot more sense to make that range more sensitive than the 10-1k range, as far as the analog circuitry is concerned. The ear is simply less sensitive to noise at those lower frequencies. So why not remove the low frequency eq boost?
Pure Vinyl for the mac, I believe, can natively handle the 75us time constant being done in analog. But I don't think it supports the 50hz constant.
Its author submitted an AES convention paper arguing that objections to flat preamps on resolution grounds are generally not important enough to dismiss flat eq: http://www.channld.com/aes123.pdf
Pure Vinyl for the mac, I believe, can natively handle the 75us time constant being done in analog. But I don't think it supports the 50hz constant.
Its author submitted an AES convention paper arguing that objections to flat preamps on resolution grounds are generally not important enough to dismiss flat eq: http://www.channld.com/aes123.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.