It was. Must have been Columbia plant contamination stuck on when the vinyl was still hot.
But after finally falling off, there was literally no audible trace that it had ever been there.
If you can avoid clipping under any circumstances that’s better IMV. As soon as you clip, you get loads of harmonics. But with heavy EQ feedback, some phono amps will ‘hang’ to the rails and take longer to recover.
I think these types of clipping events are rare if not impossible with active RIAA, but easy enough to cater for with active RIAA eq.
Here is the Tomlinson Holman graph of peak vinyl velocities. I added the red line to show a typical conservative RIAA gain profile vs frequenc, so about 32-35 dB in the mid band. The green line is the typical 5cm/sec velocity.
Last edited:
For a useful signal - yes. But not for clicks. It would be ideal if the amplifier did not pass strong clicks to the output.
Why play clicks correctly? Let their spectrum expand, but the amplitude will decrease during clipping. It is better.
Yes, I understand that some LPs can be recorded at high peak velocities. Therefore, an overload margin is necessary. It is quite reasonable to take this graph as a basis, but not the possible amplitude of clicks.
The biggest ennemy of a glass cutter ' s life is dust as aluminium oxide lies everywhere around us.Artificial diamonds found in cutters aren't very strong.You should always clean the glass of any dust trace before cutting it.
A 20 db louder signal will clip any signal reproduced by an op-amp supplied below or equal to +-5V. Not sure if you wanna give up 20db of dynamics when the groove noise lets you make use of only 40dB of vinyl's dynamics, but that brings the vinyl into type 1 cassette tape SNR range without Dolby...that's good enough compared with most cd's though...
The real thing with an adc scale is that once digitized you can remove ticks and pops, get it very silent and make it sound like a poorly mastered cd...
It's easy to forget where the noise treshold level lies in a vinyl ,and that's 60...70db below line level.you don't cut 20db out of a 90 db cd SNR range...you bring a vinyl below 53 db type 1 ferric cassette can provide...
Last edited:
I suggested digitizing the vinyl at -20dB just for experimentation to disprove your statement "hard clipping does happen a lot with dust on records".
If the dynamic range of the ADC is well above the dynamic range of the record plus 20 dB, then recording at -20 dB doesn't cost you dynamic range (well, practically not anyway).
There was no suggestion to always digitize at -20 dB. This was only suggested as an experiment to debunk a common myth that vinyl stage often clips in real life.
Back to the O/P's original question, you might have a look at the new phono preamp from SOTA, designed by Wyn Palmer who did a very high end DIY design here at DIYAudio as well as on AudioKarma. The Pyx phono pre is in production and should be available before the end of the year. Final pricing is still unavailable, but it should be in the $300 range; I doubt you could build any of the DIY kits and put it in a nice enclosure for that price. It is very low noise (MM & MC) and the RIAA compliance is within ±0.075dB or better. O/L margin is 27dB.
I collaborated on the PCB layout for this project, but have no commercial interest in or receive no compensation from this project.
I collaborated on the PCB layout for this project, but have no commercial interest in or receive no compensation from this project.
To get back to the poster's topic, here is another recommendation for the Salas phono pre. It made a significant improvement in my system. You can buy parts kits and boards here on diyaudio, the instructions are really good, it can be tuned for your type of cart., and there is great support if you run into problems. Of course DIY experience is a major plus when tackling a project like that.
I would also recommend Salas' Ultra Folded Simplistic Phono. I didn't find it difficult to build (and the cost, with two chassis and nice milling and engraving, was in the $1,000 range) and it replaced a Pearl (MM) / Ono (MC) phono stage I built about 12 years ago. The improvement was noticeable and I am very pleased with the outcome. I also agree with 6L6's recommendation (upgrade the turntable instead); my experience has been that turntable upgrades have been significantly more compelling than changes to my phono stage. To each his own, of course.
If :
-20db is reffered to 0dbfs
0dbfs is 2...2.1V for the newest akm ADC
5mV at 1kHz is found at -52 dbfs
Measurements done by random guys tell a 48db digitally filtered to a maximum of 68 db SNR for vinyl relative to 0dbfs or 2.1V.
That means that the lowest noise signal found on a vinyl groove has 0.000836 V .
It's found by some that at 10khz a vinyl can have 90db SNR.That's fantastic only if we can ignore the low frequency noise found at -48 dbfs before digitization...
ADC's resolution only goes downscale which is irrelevant when sampling signals found on top of -48dbfs noise ...
You choose to go -20dbfs at digitizing ? Cool! Now you're 28db SNR left AFTER all the digital filtering and processing cause you can't add to the filtered signal what you already cut!
It doesn't matter if your ADC can listen the emptiness of the dark outerspace ...your neighbour's car will kill your ADC headroom at the first strike.
We still fight mm cartridge own SNR with 75...90 db SNR of a phono preamp that can't even reproduce those clicks and pops when using today's best op-amps supplied at max +-20V.
Last edited:
I looked up the first 3 pages showed by google with vinyl snr measurements ...I hope you're not going the misticism path with that...Find better numbers if you can!
Last edited:
Suppose you have an ADC with 2 V full-scale input level and 100 dB(A) dynamic range. Its input noise level is then 20 uV A-weighted.
Now imagine you have a cartridge with 5 mV at 1 kHz, 5 cm/s nominal output level. If you connect it to a MM RIAA correction amplifier with 40 dB of gain at 1 kHz, the nominal output level is 500 mV. You then need a factor of 2.5 of attenuation between the amplifier output and ADC input to reduce the nominal level to -20 dBFS.
If the record surface noise is -70 dB RIAA- and A-weighted with respect to 5 cm/s (which would be very low, @xx3stksm measured -60 dB on records in very good condition), it will be about 158 uV A-weighted at the RIAA amplifier output and about 63.2 uV A-weighted at the ADC input. Adding the 20 uV from the ADC increases the total A-weighted noise by about 0.4 dB.
Now where is the problem?
Of course one could argue, based on similar comments from @Nick Sukhov, that even though the A-weighted noise doesn't increase much, the sound of the noise may change because the ADC noise and RIAA-corrected record surface noise have different spectra, but that doesn't seem to be @dreamth 's concern.
Now imagine you have a cartridge with 5 mV at 1 kHz, 5 cm/s nominal output level. If you connect it to a MM RIAA correction amplifier with 40 dB of gain at 1 kHz, the nominal output level is 500 mV. You then need a factor of 2.5 of attenuation between the amplifier output and ADC input to reduce the nominal level to -20 dBFS.
If the record surface noise is -70 dB RIAA- and A-weighted with respect to 5 cm/s (which would be very low, @xx3stksm measured -60 dB on records in very good condition), it will be about 158 uV A-weighted at the RIAA amplifier output and about 63.2 uV A-weighted at the ADC input. Adding the 20 uV from the ADC increases the total A-weighted noise by about 0.4 dB.
Now where is the problem?
Of course one could argue, based on similar comments from @Nick Sukhov, that even though the A-weighted noise doesn't increase much, the sound of the noise may change because the ADC noise and RIAA-corrected record surface noise have different spectra, but that doesn't seem to be @dreamth 's concern.
Last edited:
Absolute maximum modulation of a groove is about ±75um, vinyl surface roughness is about 25nm, maximum theoretical SNR is about 70dB.
If I understand correctly, that's without A-weighting and assuming that the RIAA correction curve is approximately -20 dB/decade over the entire audio band, so the signal after RIAA correction depends on the excursion rather than the velocity. Are 75 um and 25 nm both RMS values?
My concern is the absolute noise found at -48dbfs related to 2.1v full scale... you can't defeat that before digitization...that noise is found in a whole range of frequencies, most of them being below 1khz...so they can't beneffit of the SNR figure of 1khz at 5cm/sec