Which ADC for Phono Signal into S/P-DIF or Toslink?

[wgh52]

pcm1804 like your idee.. but finish? ...but most of the delicate work is done at the level of your prephono...

@morillion

We'll see how we finish the PCM1804 eventually, after I tested its use (and decided for it...):
A case, to e.g. also house a matching (linear) power supply is available here. The ultimate goal would be to integrate the ADC into the Flex Eight, using its 24.576Mhz oscillator and power supply.

Having used the DEQX PDC ADC input (which employs a AK5393VS ADC chip) with the RIAA Pre line out so far, I cannot see major issues,, will compare the input sensitivities and headroom, though, and take care of these.

Greetings,
Winfried

 

[wgh52]

BTW on Swissonic

- The HPA-1 is an ADC and DAC device. I'd pay full and only use "half" which is not preferrable to me...
- The Converter AD 24/192 has a fixed 192kHz sample rate which would be ASRCed to 96kHz "anayway", but otherwise it may be an alternative (if the PCM1804 "fails" to perform).

Greetings,
Winfried

 

[LTig]

BTW on Swissonic

- The HPA-1 is an ADC and DAC device. I'd pay full and only use "half" which is not preferrable to me...
- The Converter AD 24/192 has a fixed 192kHz sample rate which would be ASRCed to 96kHz "anayway", but otherwise it may be an alternative (if the PCM1804 "fails" to perform).

If you already have an ADC and DAC then just link them together, feed the ADC from your phono preamp and check whether you can can hear a difference listening to the phono preamp directly or through ADC/DAC (level match is essential). If you can't then what you have is good enough for vinyl.

 

[wgh52]

@ LTig

...potential misunderstanding: The DAC is part of a DSP with digital inputs and 4-Way-Stereo outputs to 8 power amps driving the chassis of the spreakers. Therefore I cannot compare with and without ADC scenarios...

Regards,
Winfried

 

[LTig]

A perfectionist perspective:

The key cautions you need to watch for are peak output levels and peak input levels - you want to ensure that the ADC never under any circumstance clips...

The worst case that you need to cater for is the pops/clicks that are unavoidable, and can peak at 20db above the peak level of the recording on the LP...

So you need to have some form of preamp/level adjustment in the chain to be able to adjust the analog signal up/down before the ADC, so that you can properly digitise it without clipping while also keeping it as far above the noise floor as possible.

I'm not so sure about this requirement regarding headroom for pops and clicks. If you give enough headroom through the full chain then eventually you leave the task to handle the overload to the power amp and your speakers, and a 20 dB peak means that the power amp must deliver 100 times more power and the speaker must handle it. In my view this is the worst place in the chain to handle it.

Pops and clicks should be handled at the first possible place in the chain, so either in the phono preamp or in the ADC if this comes first. Safe handling means limiting the amplitude of peaks without lengthening them - this will in fact make them less audible as well so we have a win-win situation. One way to measure this is to feed the input of the chain with a sine (at normal or low amplitude) and 2 embedded short pulses (one going positive, one negative) and connect the output to a DSO to find out how the preamp/ADC behaves.

 

[JP]

I've not found it to be an issue in practice. Above anything I'd want an ADC that is well-behaved with overload. For the event itself, the data is already lost so there's no point in perfectly capturing it. That said, running maybe 12dB of headroom I think I've seen one pop in the last ten years that hit 0dBfs.

 

[dlaloum]

I did say it was a perfectionist posting....

However, when the pop/click clips the ADC, what is lost is not only the recording under the pop/click - but also the quality of the recording around that point in time, as the ADC generates various forms of distortion when clipped, and it takes some ms' for it to settle .

I normally try to provide 20db of headroom, when digitising an LP ....
The next step will be the clean up, where the pop/click is removed - and one that has been "cleanly" recorded, can be cleanly excised - with a minimum of impact to the recording.
I find that if the ADC has clipped, the excision process becomes more complicated - the peak/noise "spreads", and where a peak of this kind excised is normally completely inaudible, trying to clean up a clipped peak, is often audible.

Finally it is played through speakers/amps/headphones - normally post cleanup - and that is the form in which it is saved to my music server

 

[JP]

Hence why I said it needs to be well behaved with overload. Regardless, I've still never taken the measures you want people to take, and have never seen the issues you're claiming. There's perfectionist, and there's overkill.

 

[LTig]

I did say it was a perfectionist posting....

However, when the pop/click clips the ADC, what is lost is not only the recording under the pop/click -

The loss is insignificant because it's swamped by the spike and cannot be recovered. At least to my knowledge the recovering algorithm replaces the data points with the spike by intrapolated data based on the points from before and after the spike.

but also the quality of the recording around that point in time, as the ADC generates various forms of distortion when clipped, and it takes some ms' for it to settle .

Do all DACs show such a long recovery time? Has this been measured?

 

[dlaloum]

The loss is insignificant because it's swamped by the spike and cannot be recovered. At least to my knowledge the recovering algorithm replaces the data points with the spike by intrapolated data based on the points from before and after the spike.

Do all DACs show such a long recovery time? Has this been measured?

ADC Saturation...

ADC Saturation: Why More Gain Isn’t Always Better

Avoiding ADC saturation is hugely important! I show you why more gain can lead to non-linearities which destroy the dynamic range of your receiver.

www.wavewalkerdsp.com

 

[wgh52]

OK folks,

we may be drifting off (my) topic here as my ADC question and application actually did not have digitised recording in mind at all. I'm seeking to AD convert purely for the listening chain and do not plan to "save" the digitised datastream. In the past I just connected the RIAA Preamp to the PDC and it worked fine - Was I just lucky?

Greetings,
Winfried

 

[LTig]

OK folks,

we may be drifting off (my) topic here as my ADC question and application actually did not have digitised recording in mind at all. I'm seeking to AD convert purely for the listening chain and do not plan to "save" the digitised datastream. In the past I just connected the RIAA Preamp to the PDC and it worked fine - Was I just lucky?

No, we're not drifting off, the problems are the same whether you rip or listen. And yes you may have been lucky, or the problem is just a theoretical one.

 

[wgh52]

OK, so let's maybe agree that input level adjustment is strongly recomended as well as monitoring the RIAA Preamp output (maybe an overload indicator is good enough?). My plan is to have a variable RIAA Preamp output: One pot per channel before the preamp output buffers (for constant output impedance) to set levels and compensate for cartridge output imbalances (cartridges' channel imbalance often is 0.5...1dB which shifts "center").

When I do a rough noise calculation: (exceptional) -70dB LP record media noise level and 10dB AD headroom would only require ADC noise level to be lower than -80dB.

Greetings,
Winfried

 

[LTig]

OK, so let's maybe agree that input level adjustment is strongly recomended as well as monitoring the RIAA Preamp output (maybe an overload indicator is good enough?). My plan is to have a variable RIAA Preamp output: One pot per channel before the preamp output buffers (for constant output impedance) to set levels and compensate for cartridge output imbalances (cartridges' channel imbalance often is 0.5...1dB which shifts "center").

Good plan.

When I do a rough noise calculation: (exceptional) -70dB LP record media noise level and 10dB AD headroom would only require ADC noise level to be lower than -80dB.

Yep.

 

[dr0ss]

I have a spare Roland/Edirol UA-1G you can try. I bought my first one to digitize vinyl, my second to get a variable input knob on a microphone, and then somehow ended up with a box of them. It does need a USB connection to a computer to work (I don't know if that can be bypassed on the PCB), but it has decent hardware from AKM and is about as small a device as you're going to find.

 

[wgh52]

Hi to Hawaii!

thank you for your very kind offer! Looking at our very distant locations implementation seems not easy or cost effective to me and somewhat risky due to shipping around the globe...

Thanks anyway!
Winfried

 

[dr0ss]

thank you for your very kind offer! Looking at our very distant locations implementation seems not easy or cost effective to me and somewhat risky due to shipping around the globe...

Perhaps. I buy stuff from Europe pretty often, but the other direction might be worse.

 

[audiopile]

I listen thru Meridian DSP-5000 speakers -which means they only accept 24/96 SPDIF inputs. 90% of my listening is off of LP's -so ADC's are necesary I own and use the Park's audio Puffins and actually -you could set one of those up as a ADC's with some interesting features, I've used the Mytec 24/96 ADC's -own two-but a year or so back ended up buying three very old (20 years or so) MSB PAD-1 . These have a clipping indicator and three analog inputs :balanced ,unbalanced and phono. Sampling rate is switch /button selectable but tops out at 24/96 -fine for my use. So far -the best ADC for my use and pleasure.

 

[wgh52]

I listen thru Meridian DSP-5000 speakers -which means they only accept 24/96 SPDIF inputs. 90% of my listening is off of LP's -so ADC's are necesary I own and use the Park's audio Puffins and actually -you could set one of those up as a ADC's with some interesting features, I've used the Mytec 24/96 ADC's -own two-but a year or so back ended up buying three very old (20 years or so) MSB PAD-1 . These have a clipping indicator and three analog inputs :balanced ,unbalanced and phono. Sampling rate is switch /button selectable but tops out at 24/96 -fine for my use. So far -the best ADC for my use and pleasure.

Thanks for sharing your interesting experience! MSB PAD-1, offered at 200$ used in the USA, could/may be an alternative, I've not found it on offer in Europe yet, though.

Greetings,
Winfried

 

[audiopile]

There is a UK source for a modified/improved power supply which they will make in whatever AC input voltage you want.