Phonograph Stylus Wear Experiment

[JP]

I thought the purpose of REF was to continually validate the test record?

The response from T24-DUT-P2 does not look like a VM95 - the HF peak should be circa 10.5kHz, like the VM95E (and the same thing we see on the ML, SH, and C styli) plot below. I can't say that I've ever seen a VM95 show a peak at 16kHz. T0-REF-P2 shows the same thing.

May be time to halt the experiment, get your electronics figured out, and then restart it start over.

 

[ray_parkhurst]

The REF stylus purpose is twofold: ensure test record consistency; and give a reference vs the DUT. How do the DUT and REF responses compare at this point?

 

[JP]

You've completely incorrect HF response for the cart that is unexplained - I didn't bother looking any further, and really don't see the point in doing so. All the data is suspect at this point.

 

[Icewater_7]

It looks like we need to reset the process a bit. @BMRR had to give up the original Pluto preamp to get access to the Pluto2, and was not able to complete #2 and #4 from the original recording list. The Pluto2 looks to be giving a flatter response at LF, and eliminates the odd HF response, so we'll use it going forward, but we need to get a comparison between the DUT T24 and REF T0 to be sure we have consistent results and a valid reference. So here is the new list, all linked to in @BMRR's post above:

A. T0 on DUT with Pluto 1
B. T24 on DUT with Pluto 2
C. T0 on REF with Pluto 2

I am now most interested in comparing #B and #C to make sure the two styli are performing about the same, which would give us confidence in the measurement as well as the future usefulness of the REF to compare with the DUT as it wears. I don't think anyone expects any difference between T0 and T24 on the DUT, so as long as that holds true with DUT vs REF, we can continue with the experiment.

Can we get a comparison between #B and #C above?

So far, no plots of Pluto 2 data have been published here yet, correct?

 

[BMRR]

So far, no plots of Pluto 2 data have been published here yet, correct?

I believe this plot was made from the recording of the experiment stylus and the Pluto 2:

 

[Icewater_7]

I believe this plot was made from the recording of the experiment stylus and the Pluto 2:

It was not explicitly identified in that post. All plots should clearly identify the DUT.

 

[JP]

It was not explicitly identified in that post. All plots should clearly identify the DUT.

It's not an 'official' plot as an experimental result - the labeling was very intentional.

 

[Icewater_7]

It's not an 'official' plot as an experimental result - the labeling was very intentional.

Ah, my bad again, I didn't recall that experimental plots had "BMMR Test" identifying the DUT and it's signal chain in the plot itself, so not required in the author's textual portion. I'm still not sure if Pluto 2 had the 16K peak or not and if the newest proposal from ASR is implying that the test should start over completely with the Pluto 2 or some other equipment.

 

[JP]

Ah, my bad again, I didn't recall that experimental plots had "BMMR Test" identifying the DUT and it's signal chain in the plot itself, so not required in the author's textual portion. I'm still not sure if Pluto 2 had the 16K peak or not and if the newest proposal from ASR is implying that the test should start over completely with the Pluto 2 or some other equipment.

"Pluto 1" plots show a significant LF roll-off, slight mid range elevation, and as-expected HF with the 10.5kHz peak that is characteristic for the VM95 series. "Pluto 2" shows a much smaller LF roll-off, as-expected mid range, but a +6dB peak at 16kHz where with the loading as guesstimated (250pF) we should see a ~+1dB peak at 10.5kHz.

It appears something is rather significantly off, and that should absolutely be identified. I think the advice is as it always was - verify the electronics and measurement data before beginning an experiment, multiple laybacks to identify and average-out errors like the ~3dB left channel drop in the HF of the first plot, etc.

I'll not be spending anymore time on this as-is.

 

[Icewater_7]

"Pluto 1" plots show a significant LF roll-off, slight mid range elevation, and as-expected HF with the 10.5kHz peak that is characteristic for the VM95 series. "Pluto 2" shows a much smaller LF roll-off, as-expected mid range, but a +6dB peak at 16kHz where with the loading as guesstimated (250pF) we should see a ~+1dB peak at 10.5kHz.

It appears something is rather significantly off, and that should absolutely be identified. I think the advice is as it always was - verify the electronics and measurement data before beginning an experiment, multiple laybacks to identify and average-out errors like the ~3dB left channel drop in the HF of the first plot, etc.

I'll not be spending anymore time on this as-is.

Thanks for the well reasoned summary. I agree that scientific test methodology should aggressively control variables and validate all equipment in the test setup to be certain that it is capable of measuring the information sought without introducing uncertainties of its own into the output data.

 

[phn]

What an interesting topic, just found this, subscribing. Hopefully you'll have the new tests running again soon! Thank you for your efforts, it is amazing something like this has not really been made earlier.

 

[Icewater_7]

What an interesting topic, just found this, subscribing. Hopefully you'll have the new tests running again soon! Thank you for your efforts, it is amazing something like this has not really been made earlier.

Stylus Wear Study 2 - Page 14- Vinyl Engine

www.vinylengine.com

 

[BendBound]

Thank you Icewater_7 for posting this link. I've been traveling overseas since August and only now am able to update this thread.

Our team will not longer be posting updates to our stylus physical wear experiment on this website. Please pick up this topic here, primarily:

Stylus Wear Study 2, On Vinyl Engine.

On the Vinyl Engine thread there are a few posters, one in particular, who demonstrate capabilities to evaluate on-going sonic distortion in our dubs of CBS test records, while evaluating those results is only a secondary goal. Presently, we are at T192 hours of physical wear. Recall, the primary and intial purpose of this experiment is to catelog progressive physical wear of the stylus.

Also, follow this thread on the Steve Hoffman Forum for the same experiment.

Phonograph Stylus Wear Experiment, Steve Hoffman Forum.

Thank you for your interest.

 

[Audiofire]

On15k most cartridge har 10+% distortion….3% is not a realistic number old or new

Something like that. It is why I gave up on trying to take vinyl records seriously. Incidentally, does anyone know if those spectrogram programs on computers show the same amount of distortion (about 10%, or the spectrogram 1920x1080 of a FLAC file is less than 1% distortion)?

 

[dlaloum]

Something like that. It is why I gave up on trying to take vinyl records seriously. Incidentally, does anyone know if those spectrogram programs on computers show the same amount of distortion (about 10%, or the spectrogram 1920x1080 of a FLAC file is less than 1% distortion)?

Most cartridges today have relatively (!) heavy aluminium tube or solid rod exotic cantilevers - their cantilever/tip resonance is often around the 15kHz mark - and that resonance will increase mistracking, resulting in increased distortion. (all theoretical...)

It would be interesting to get something like the Dynavector Karat and measure distortion at 15kHz.... we might find that the distortion is an order of magnitude lower simply because the cantilever/needle resonance is up at circa 50Khz and of no consequence to tracking at 15kHz...

 

[LTig]

Most cartridges today have relatively (!) heavy aluminium tube or solid rod exotic cantilevers - their cantilever/tip resonance is often around the 15kHz mark - and that resonance will increase mistracking, resulting in increased distortion. (all theoretical...).

What about those MC pickups the frequency response is specified to 45 or 50 kHz, usable for CD4? They cannot have such a low resonance.

 

[dlaloum]

What about those MC pickups the frequency response is specified to 45 or 50 kHz, usable for CD4? They cannot have such a low resonance.

Actually those exotic tube cantilevers and the line contact needles were all designed to better pick up the carrier frequencies of CD4.

With regards to effectiveness of CD4 tracking - well how many people today bother trying it? How many actually test that frequency response? - There are very few test records with tracks extending up to or past 50kHz... and I doubt most vinyl presses today would be capable of producing records with frequencies going that high (reliably).

I've measured those resonances on V15V's, AT440, AT152 and a bunch of other cartridges.... courtesy of a spreadsheet which models the electrical parts of the response, deducts that from the measure F/R and provides me with the "raw" F/R - where the resonance is usually obviously exposed (I'll need to dig up some of those measurements I made 10 to 12 years ago and post some examples...)

 

[LTig]

Actually those exotic tube cantilevers and the line contact needles were all designed to better pick up the carrier frequencies of CD4.

Certainly. But they also track better in the audible range, I think.

With regards to effectiveness of CD4 tracking - well how many people today bother trying it? How many actually test that frequency response? - There are very few test records with tracks extending up to or past 50kHz... and I doubt most vinyl presses today would be capable of producing records with frequencies going that high (reliably).

Even in the hidays of CD4 those quadrophonic records probably were good for much less number of plays than stereo recordings.

I've measured those resonances on V15V's, AT440, AT152 and a bunch of other cartridges.... courtesy of a spreadsheet which models the electrical parts of the response, deducts that from the measure F/R and provides me with the "raw" F/R - where the resonance is usually obviously exposed (I'll need to dig up some of those measurements I made 10 to 12 years ago and post some examples...)

Did you measure one of the AT MC pickups like the AT33PTG/II (the one I bought this year) or the AT-OC9XML, by change? Or one of the Van den Hul pickups (those are all specified to 50 kHz and higher)?

 

[dlaloum]

Certainly. But they also track better in the audible range, I think.

Yep

Even in the hidays of CD4 those quadrophonic records probably were good for much less number of plays than stereo recordings.

Yep

Did you measure one of the AT MC pickups like the AT33PTG/II (the one I bought this year) or the AT-OC9XML, by change? Or one of the Van den Hul pickups (those are all specified to 50 kHz and higher)?

No, I don't own an AT MC... I have measured a Sony MC, and I also have a close relative of the VanDenHul - which is an Empire MC1 (late 80's) - it is however one of the more basic models, without an exotic cantilever, and with only an elliptical tip. At the time Emprie and VdH partnered up and did a series of MC's which were effectively re-badged VdH MC's, the current "Frog", MC10, and MC1, all still use the same housing as my own 40 year old Empire MC1, I have no idea whether the innards are also the same or have changed over the decades.

The specification to 50kHz and higher is a little meaningless on many of these (as per my previous comments) - one way of measuring at a higher frequency is to run the test record at 48RPM rather than 33... which extends frequency range by 1/3rd roughly... but it gets complicated as you then need to adjust for RIAA decoding. I did a few wider range tests that way about a decade ago.

The test to 50Khz would be relatively meaningless with my MC1 given the elliptical tip isn't capable of tracing that high a frequency, so it absolutely will be mistracking, affecting the results accordingly (and wearing the test record in the process as well!)

 

[Audiofire]

Incidentally, does anyone know if those spectrogram programs on computers show the same amount of distortion (about 10%, or the spectrogram 1920x1080 of a FLAC file is less than 1% distortion)?

On second thought, that resolution shows about 0.1% distortion (if one compares the same vinyl record from another playback of course).