You can follow the link to the ADI chip I listed. It has SPDIF out which would be the logical way to use it. The ADI chips are 28/56 bits internally so that should not be an issue. If Scott can cram the EQ in (again trivial) the project gets pretty easy. Click-fix is a different challenge since its got to be completely automatic. If we add a surface noise suppression we could make records almost as clean as CD's. That's getting very Steampunk.
Scott
If the price is good, it will be a blessing (and it will become a hit).
As the front end will be an instrumentational amplifier (INA), there is no point to fit it at the headshell (I am experimenting with an AD620, I’ll report soon).
If you are seriously considering it, think of it as a complete unit: R, C loading in front, antialliasing filter/buffer, INA, ADC/DSP.
Just by adding a few more pcb pads (*) this unit may interest many more people, as they may use it as an analog head pre-amp and as a differential probe for audio lab work.
PSU needs I think mate well with what Jan’s Silent Switcher provides.
George
(*) for accessing INA’s analog output and gain select pins
If the price is good, it will be a blessing (and it will become a hit).
As the front end will be an instrumentational amplifier (INA), there is no point to fit it at the headshell (I am experimenting with an AD620, I’ll report soon).
If you are seriously considering it, think of it as a complete unit: R, C loading in front, antialliasing filter/buffer, INA, ADC/DSP.
Just by adding a few more pcb pads (*) this unit may interest many more people, as they may use it as an analog head pre-amp and as a differential probe for audio lab work.
PSU needs I think mate well with what Jan’s Silent Switcher provides.
George
(*) for accessing INA’s analog output and gain select pins
I think there might be noise advantage (hum rejection) in the headshell, because input connection can be balanced without rewiring the tonearm (not enough wires inc power otherwise). Also avoids unscreened tonearm wiring, and exposed loops.
I love the idea of a pcb clipping on to cartridge pins, btw..............
Yes.
LD
Last edited:
I had the idea some years ago to put SMT input devices on a small PCB that would plug directly onto the back of the cartridge, and rewire the arm (I think only 6 wires required). The advantages I thought it would have are less noise, because the signal is amplified as close as possible to the source, and the arm wiring being inside the feedback loop...
The noise benefit is down in the dBGnatFart region unlike say, a satellite TV receiver where you have to get the LNA as close to the receiver as possible, but getting as hot as you can as early as possible is nice. For those who want to play early there is alway Russ White's balanced phono amps Retro - Balanced RIAA Stage
Not pennies, but you can order with SM parts soldered down and with a few tweaks to drop gain you should be able to rip the RIAA out (note have not done the sums). For me I am so far behind on projects I can wait to see what Scott comes up with down the line.
Of course this will mean Jan will need to do a second kickstarter for the silentswitcher
I cannot see your problem.
Chopper - Test circuit for my experimental ultra low 1/f noise preamp.
Designed, laid out, home-etched and populated last weekend.
GerhardAttachments
I figure anything over $250 or so would start to eliminate a lot of folks. Too much stuff in those balanced pre-amps. I'll start with something folks can use even if the whole thing does not get done. I have an idea using a couple of FET's and the THAT chip to do a MM/MC pre-amp simply with switched gain. I won't have the BW to do much more right now.
Gerhard No QFN's there, child's play
Back in the day, I did a Headamp In A Headshell design for a Japanese cartridge manufacturer, which ended up being picked up by a US dealer. The quietest FETs we could find then were CM860/2N6550, we drilled holes in the headshell to insert the metal-canned FETs. The US company that picked it up just encapsulated the circuit so it could be attached to a variety of headshells. Phantom powered, it worked quite well given the technology of the time, but the commercial viability was limited.
A modern instrumentation amp chip would be a much better solution.
A modern instrumentation amp chip would be a much better solution.
Yes, no feature creep! A complete thing that 95% of people can use; no arm re-wiring, that turns off a lot of people who would otherwise be interested.
Accept 12 inch of tonearm wiring. A small box near the arm post. Run it off the 5V USB from the connection cable.
Use 1206 SMD that everyone can easily solder, still will be a small unit.
For inspiration look at Bill Waslo's Omnimic. That has been so successful because he made the right choices (in addition to a great circuit of course).
BTW I lay out a mean PCB if that would help.
Jan
Last edited:
LP noise suppression
As an engineer at a major media plant for almost 30 years starting with Sonic Solutions, I've used Cedar, Pro Tools and a bunch of other plug-ins and software approaches available for noise reduction, and I have to say I disliked some aspect of all of them. By this I mean the side effects on the rest of the mix were as objectionable or worse than the noise being reduced. I and others in the mastering community have discussed this issue and we often conclude that it is easier for the brain to ignore natural noises (which hiss and crackles can approximate) than it is to ignore the unnatural spectral or dynamic compromises the NR can cause. Of course restoring a cracked LP or shellac this was not the case, everything has to be taken in proportion...
Maybe this is just a quirk of mine, but I felt my responsibility to the artist was to try and deliver as much of the original performance and sound of the instrument as possible without a blanket thrown over it to quiet it down.
The approach which I felt resulted in the best compromise of unwanted impulse noise reduction to artifact production is the ultrasonic sideband detection and selective muting approach used by both Packburn and Dick Burwen. The Packburn 323 when critically adjusted could really diminish the impulse without significant coloring of the source, as could some Burwen units like the TNE7000. The TNE unfortunately was not dead quiet itself, but its use often was an improvement over the original despite this limitation. Both require adjustment for each piece of media being processed, which some may find onerous.
After hundreds, perhaps thousands of hours preparing masters for replication, my conclusion is that the main problem with the software approach is due to it inability to distinguish HF program content from HF impulse noise. I have yet to hear a software NR program which does not affect the crack of snares, the rapid rise associated with a pick on a string of a mandolin, cymbal spectra or other essential aspects of a dynamic recording. Since the detection of the impulse to reduce is done by analyzing baseband material this is understandable. Using supersonic spectral content this limitation is greatly reduced. Thinking about the proposed digital NR approach, if we perform 4x or greater oversampling without a band-stop LP filter at 22KHz, applying this technique would seem possible. Not being a software person, I leave the intricacies of coding this up to someone else.
Just my $0.02 worth,
Cheers!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
As an engineer at a major media plant for almost 30 years starting with Sonic Solutions, I've used Cedar, Pro Tools and a bunch of other plug-ins and software approaches available for noise reduction, and I have to say I disliked some aspect of all of them. By this I mean the side effects on the rest of the mix were as objectionable or worse than the noise being reduced. I and others in the mastering community have discussed this issue and we often conclude that it is easier for the brain to ignore natural noises (which hiss and crackles can approximate) than it is to ignore the unnatural spectral or dynamic compromises the NR can cause. Of course restoring a cracked LP or shellac this was not the case, everything has to be taken in proportion...
Maybe this is just a quirk of mine, but I felt my responsibility to the artist was to try and deliver as much of the original performance and sound of the instrument as possible without a blanket thrown over it to quiet it down.
The approach which I felt resulted in the best compromise of unwanted impulse noise reduction to artifact production is the ultrasonic sideband detection and selective muting approach used by both Packburn and Dick Burwen. The Packburn 323 when critically adjusted could really diminish the impulse without significant coloring of the source, as could some Burwen units like the TNE7000. The TNE unfortunately was not dead quiet itself, but its use often was an improvement over the original despite this limitation. Both require adjustment for each piece of media being processed, which some may find onerous.
After hundreds, perhaps thousands of hours preparing masters for replication, my conclusion is that the main problem with the software approach is due to it inability to distinguish HF program content from HF impulse noise. I have yet to hear a software NR program which does not affect the crack of snares, the rapid rise associated with a pick on a string of a mandolin, cymbal spectra or other essential aspects of a dynamic recording. Since the detection of the impulse to reduce is done by analyzing baseband material this is understandable. Using supersonic spectral content this limitation is greatly reduced. Thinking about the proposed digital NR approach, if we perform 4x or greater oversampling without a band-stop LP filter at 22KHz, applying this technique would seem possible. Not being a software person, I leave the intricacies of coding this up to someone else.
Just my $0.02 worth,
Cheers!
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
If we digitize at 192KHz we will have ultrasonic content and a window detection that mirrors the normal hf rolloff of music Energy that is supersonic and strong where instruments (recorded with real microphones) won't be to trigger a click removal. The surface noise issue many not be resolvable, certainly not in real time.
If we keep this simple- differential amp input with less than 2 nV/rtHz and adjustable gain, 120 dB SNR ADC and a DSP like the ADI it does become simple. The ADI DSP self boots which does streamline everything. I would stick with SPDIF out since a USB interface adds a lot of overhead. We could make a board like the ADI programmer board for these DSP's so a user can change the eq and even the functionality on the fly. Bypassing or a different EQ and this becomes a mike preamp or a tape head preamp.
If we use the AK5394A it will be the most expensive part but it works really well.
We can provide for external clocking pretty easily so several of these can be locked together for microphone type applications.
I have always questioned a differential input for a phono cart. It adds 3 dB of noise and the potential for a lot of common mode for the input to deal with. What are the counter arguments? However we need a differential drive for the ADC anyway and it makes mike applications much easier.
If we keep this simple- differential amp input with less than 2 nV/rtHz and adjustable gain, 120 dB SNR ADC and a DSP like the ADI it does become simple. The ADI DSP self boots which does streamline everything. I would stick with SPDIF out since a USB interface adds a lot of overhead. We could make a board like the ADI programmer board for these DSP's so a user can change the eq and even the functionality on the fly. Bypassing or a different EQ and this becomes a mike preamp or a tape head preamp.
If we use the AK5394A it will be the most expensive part but it works really well.
We can provide for external clocking pretty easily so several of these can be locked together for microphone type applications.
I have always questioned a differential input for a phono cart. It adds 3 dB of noise and the potential for a lot of common mode for the input to deal with. What are the counter arguments? However we need a differential drive for the ADC anyway and it makes mike applications much easier.
Balanced doesn't necessarily add 3dB of noise- see, for example MC step up transformers. If the noise of the input is already significantly below the thermal noise of the cartridge (easy to do with MM and MI), the noise difference is buried in the power summation.
As a practical matter, transformers and good INamps handle the real-world common mode beautifully. It's always fun to demo by touching the cartridge pins and watching everyone reflexively jump, followed by relieved laughter when there's no noise.
There's a reason that microphones and other low level transducers (e.g., strain gauges) use balanced.
As a practical matter, transformers and good INamps handle the real-world common mode beautifully. It's always fun to demo by touching the cartridge pins and watching everyone reflexively jump, followed by relieved laughter when there's no noise.
There's a reason that microphones and other low level transducers (e.g., strain gauges) use balanced.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.