Investigating if ultrasound is harming LPs

Much ado over...

+ An Additional Conclusion.

Sorry could not hold myself on your absurd method especially your weighing methods. First you let it soak and wonder why it's heavier. Vinyl record isn't completely water-proof and able to soak water especially if it's completely submerged into the powerful bath. The numbers of decimal points on your scale are EXTREMELY FAR to reflect any differences in weight even after cleaning vinyl manually with dedicated liquid!!!

If you would compare other cleaning methods vs. Ultrasonic, you would be able to present a lot more understandable proof than what you’ve just accomplished.

It’s a lot more notable to accomplish what matters rather than accomplishing what’s not all in all.

start over with a very important class called ; Design of Experiment, one very important step is expert review of the DOE..before proceeding… to in your case, generate junk science….

You want to hold the vinyl rigidly during the cleaning, to mimic the situation with an actual LP. Otherwise the US energy is dissipated kinetically as the 5g piece of vinyl is tossed about. Some energy is lost that might otherwise hypothetically be effective in grinding the vinyl.

@lewm @tomic601 @czarivey ​​​​@fuzztone are all correct.

I’m the guy that usually signs off with "happy listening" because sounds that make us happy are what the audio hobby is all about. This thread touches on my professional world of metrology and begs more rigorous attention, however.

Examining the mass of a vinyl record before and after ultrasonic cleaning might produce some interesting data. It’s far from certain the observed data would be meaningful.

It must start with defining the capabilities of the mass measuring device: Tolerance, resolution and gauge repeatability & reproducibility. Those must be known to generate validation of the recorded statistics.

Very few consumer devices are capable of tolerances better than 0.1 gram. This presumes the device is tared (validated) using a known standard traceable to an accredited, recognized standard (NIST in the US or NPL in the UK are examples). Generally, they can resolve to 0.1 gram (my Ortofon DS-1 and by Ohaus triple-beam lab scale are both examples that do so). Their tolerance is affected by a number of other issues and is emphatically not stated by their manufacturer (marketer) for these consumer grade devices.

Tolerance is the statistical measure of confidence (aka uncertainty, which has it’s own associated statistical measure of confidence). To that must be added the aggregate of standard deviation of repeatability (the ability of the same operator using the same device to achieve the same result) and standard deviation of reproducibility (the ability of multiple operators using the same device to achieve the same result). Only then can an approximation of measuring precision be determined.

Returning to the example posted, it would be very surprising indeed if the gross tolerance of the device employed is anything better than 0.5 grams, excluding repeatability and reproducibility variables.

Sample preparation has already been observed as insufficiently precise. Using a non-destructively obtained sample (i.e. an entire vinyl record) would be more representative of any potential real-world implications.

Repeating the experiment with sufficient device, sample, technique and documentation exactitude would be needed to examine if the results produce meaningful data.

ah….    @effischer Yes, an excellent start….  Certain experts in metrology and the other sciences certainly rescused ( or not ) my bacon when someone decided to use a tool where cal / cert had expired….

Enjoy the music ;-)

Wow!

effischer. No problems with what you wrote, except this: "Very few consumer devices are capable of tolerances better than 0.1 gram. This presumes the device is tared (validated) using a known standard traceable to an accredited, recognized standard (NIST in the US or NPL in the UK are examples)." While what you say about taring is probably correct, in this case, since we only want to know whether there is a change in weight, all we care about is the repeatability and the sensitivity of the scale and other elements of the weigh procedure. One other point, my Ortofon DS3 reads out to 3 decimal places (as do many other digital VTF gauges); i therefore would think that, best case scenario, it is accurate to 2 decimal places, at least.

In general, if US cleaning clips the HF response of the signal encoded on the LP (which is what some audiophiles fear), I would guess the weight of vinyl removed to achieve that damage would be infinitesimal. Better way to do this is to start with an LP that has test tones on it, say 1000Hz, 10kHz, 20kHz. Document its frequency content electronically, say with a storage oscilloscope, US clean it, then play it again comparing the data recorded before vs after. The relative levels of the various frequencies, before vs after, would constitute a nice additional internal control. (For example, compare the 1kHz signal amplitude to 10kHz and 20kHz signal amplitudes, before vs after. If US is harming the HF response, you would expect the 10kHz and/or the 20kHz signal amplitudes would be decreased after cleaning by more than any effect on the 1khz amplitude.)

Methodology is a problem. Measurements need to be more precise. More sample points to do statistics. And then, have to decide what is gone - dirt or vinyl, and why?

Too many variables to draw a conclusion.

I've done a better experiment. I cooked one half of an LP for an hour without rotating it. Then I cleaned the whole thing and played it. Any problem would have appeared and disappeared at 1 Hz frequency. Found NO DIFFERENCE through resolving electronics and ESL's.

oh Werner H.

The thread is too much like work…

Wow.  I am wildly profoundly perplexed.

Yea son, get her!

Way out of the limits of your scale. FYWIW scales are robots that measure the voltage or current to hold them in position and translate to mass. That would be a very expensive scale at that weight and resolution. 

Yes, I addressed some of the issues myself in the text above:

About I noticed that I as operator got better and better in the weighting procedure that the first ones is not that repetition accurate.

Yes, we are not interested in if the piece is weights on that scale 5.02 gram when it actually on ANOTHER weight scale that is tared and calibrated and show that the same piece actually weight 5.15 gram. That is not interesting when nobody cares about what that specific scrap part is weighting.

The interesting part is the repetition accuracy between measurements and as we can see of the 40 measurements the VARIATION between all of them is only 0.02 gram from the heaviest and the lightest measurement result. And those 40 weight data points is spread over 5 sessions and a operator that getting better and better at it, so the first session that resulted in the lowest reading (5.00) maybe would not appear when it were in the first session that I were not as good as in the later sessions. Then we would be down to a variation of 0.01g. (So we could have a operator and a weighting scale that can reproduce weighting with a repetition deviation of ~ ±0.005 gram. if we consider that there were no reduction in mass from the US cleaning sessions..)

So I have seen weight scale with 3 decimals of a gram but I guess that it really doesn't matter if we see deviation of weight in that higher resolution for only a 5g piece, then we are down to reading more of the operator repetition accuracy between measurements. And repetition accuracy between sessions, room temperature and below 0.1% in total weight difference that maybe theoretical 1h of US cleaning might have caused(!) 

And no record in any shelf on the planet has undergone that heavy US cleaning in my believe. 

In other words higher resolution will more showcase other external factors (some of them accounted for above) than what we are really trying to get a answer to.

Methodology is a problem. Measurements need to be more precise. More sample points to do statistics. And then, have to decide what is gone - dirt or vinyl, and why?

 

Too many variables to draw a conclusion.

Regarding if it is dirt or not that were removed then read the text again I stated that I used a initial US cleaning before I started the testing, so that ALL dirt were removed when I am only interested what US did with the plastic and if it eroded the material. So the method and variable of regarding dirt were accounted for and removed.

Yes I did note that the piece were hanging lose that were nothing that I had thought about when I started but noted that it might be a issue. I think that maybe the ultrasound is acting on all sides simultaneously (the scrubbing bubbles). And I have previously when testing the efficiency of the US unit that I have tested with a aluminum foil dipped in the bath and the US shredded holes in the metal in less than a minute. As just a example of something else that were not rigidly fastened the US had no problems to work on the aluminum foil.

But of course I believe it is better to maybe use steel wire to hold the piece steady and mimicking how a record is held!

So there is potential methods improvements to be done. But I have not seen someone else trying to do the same thing at all when we have all of the tools and the ability to make this rather easily.

A YouTube friend said that Michel Fremer had US cleaned a red vinyl and tried to see if he got some red residue in the bath afterwards as a method.

I have read somewhere someone that had a "precision" scale and if that it were PVC pipe and not LPs that they treated with US session.

Anyway when seeing what US can do with metal and the power of it then that is raising the question for me what US is doing with LP records.

to in your case, generate junk science….

I hope that this can be considered as a starting point and show case that it is possible to do. Plus that it might inspire someone and giving some ideas. If someone would like to try and test. Let's us be positive and creatively contribute. But thanks for regarding my experiment 🥼 as science. 

But the good thing with this first shoot don't reviled that it is a disaster to use US cleaning at this point maybe if we get some process/methodology improvements might change that view, but hopefully not.

Using a non-destructively obtained sample (i.e. an entire vinyl record) would be more representative of any potential real-world implications.

I see that there is a issue with the method that I needed to take. When I don't have any wight scale that has enough precision and repeatability accuracy to weight a whole LP (~120-200 g). 

So the scale i used can measure max ~5g. And I still believe if I get the 5g piece to be more rigidly suspended in the bath. Then it is maybe good enough and a step better than other previously mentioned methods above. The positive thing is that 30 minutes in the bath is more than 1 hour if real-world implications is used. And when the theoretical erosion is so small then and hard to detect so we need to run US on the material for a extended duration to be able to detect that small erosion if there's any erosion that is occurring at all to begin with.

First you let it soak and wonder why it's heavier. Vinyl record isn't completely water-proof and able to soak water especially if it's completely submerged into the powerful bath.

Yes, i did let it dry after the "soak". But are you implying that the "vinyl" material is porous? (Remember that the stylus read ridges smaller than one micron then the porosity need to be smaller than that and not detectable by the stylus)

And I need to dry out liquid that is "induced" into the material and not only on the surface of the vinyl. Do you have issues with surface dry records that were soaked cleaned and afterwards put in a plastic bag (Inner sleeve). With some condensation and moisture in the inside the bag should be seen? I have not experienced that anyway? Should be seen if liquid is trapped in a bag!

I don't believe that LP material is soaking up liquid so that is in my eyes only spreading FUD in a sense. And lastly it got lighter (0.1%) after the second US bath so the theory that it holds water fails in that aspect also. But you might have other experience?

On the other hand the drying step is important and I did not weight wet material of course.

This was the another reason I did the experiment to remove any FUD about usage of US cleaning method. Regarding if or if not it erodes material from the grove.

Somebody has way to much time on his hands…

Pitter patter. 

Hey, everybody needs a hobby!  

My guess, like someone above mentioned, is that the record is slightly hygroscopic and the very long soak for your test is different from your initial cleaning.  Try also soaking the pre test sample in water for the same period and weight the sample before subjecting it to the ultra sound cleaning.  That will at least reduce that potential variable.  

@larryi thanks for the input:

My guess, like someone above mentioned, is that the record is slightly hygroscopic

I thought I provided several different reasons why it can't be hygroscopic in my previous post.

But we will go the extra mile here and try to debunking that myth, I took the out the part and the scale again. And I were thinking that after 24 hours drying time it should have got dryer. I got 5.02 grams again. So not there is nothing that has evaporated the last 24h. 👍

I think you are pushing the accuracy of the scale and the amount of change beyond a reasonable level. The ambient temperature, relative humidity, and atmospheric pressure will all be an affect at the precision you are attempting. They should all be a part of the reading at each step of the experiment. The drying time (as you said) will also affect the reading, So to continue your experiment... Try your 8 measurements at 1 hour, 4 hours, 12 hours, and 24 hours to try to narrow done the ambient variations that occur and stabilization of the vinyl after US cleaning.

I really think the better test would be the frequency test of before and after as lewm said.

Scales are LEAST accurate at the extreme ends of their range. In this case 5g and 0grams. You are better off using a piece of vinyl coming in about 2,5grams. Also unless you are using a Mettler scientific laboratory scale in a temperature controlled environment, your accuracy at the .00X level is not going to be good. Interesting experiment though.

added weight could be remaining moisture? Or a coating of surfactant?

Interesting approach.  But if audio, specifically hi-freq loss is the issue then it only matters if that is the outcome.  So measure what is the issue; sound (hi-freq).  As far as material loss, what if loss occurs on all grooves all shapes of waveform surfaces equally?  Like a sharp knife getting sharpened again, you have loss of metal but it stays equally sharp.  Not a perfect analogy but potentually similar.    Secondly, and I'm no materials expert, but vinyl is obviously flexible so it will be less likely to be removed than a rigid, brittle material.   Lastly, perhaps people are visualizing a force that is along/across the groove surfaces.  But as it's done in a liquid, the forces of the US wave is pressure, not shear.  Fluids, gas, cannot support shear.  Back to the variable in question; measure before and after freq content.  

I ultrasound my records and play them just once in order to digitize them to my disk storage for playback from there.  This is for real…..no joke.  Records are delicate little creatures….just one of their faults. Would not want to subject them even to a stylus tearing through them any more than once if possible. It’s a very savage way to create music! What were they thinking 100 years ago when they came up with the idea of records? Too precious! No telling how much body mass a record looses when played. Someone ought to look into that!

You've discovered the methodology for removing the warmth from vinyl.

Like I said above, but which gleaned no attention at all, if you want to find out whether US cleaning can damage the encoded music signal on an LP, the way to go is to look directly at the signal on an LP before vs after US cleaning, not the weight of the LP before vs after cleaning. With respect for your diligence, it is misguided, because your method is faulty to begin with. Which is to say that if you conclusively find no difference in weight, it does not mean necessarily that no damage was done to the music signal. And you cannot even hope to detect such damage with a listening test, as someone else seems to claim. You need the proper electronic tools and the capacity to collect the data for comparison (e.g., an oscilloscope with storage function, a test LP with pure tones recorded on it in stereo, and probably other devices I have not thought of).

It’s true that dirt and grime on a record results in less treble and detail and adds noise resulting in a “warmer” sound ie relatively more midrange. Take away all that mess and vinyl starts sounding more like digital and vice versa. If you want warmth, better off pursuing it with the appropriate amplification known for that.  You can have more warmth with everything that way including with impeccable  digital sources. 

My records sound better after a session in my DeGritter than they did before. Surface noise much reduced. That's all I care about... 

I ultrasonic each vinyl once, or twice for old used ones as needed, then place it in an inner plastic sleeve and the album cover is then put into an outer plastic sleeve with the sleeve opening in the BACK of the album. Hence zero further dust accumulation. Point being, with one or two ultrasonic wash, no way I have to concern myself over ultrasonic cleaning damage.

@larsman yes I agree.  Can’t handle record noise anymore now that we are well into the digital age.  I just got an ultrasonic cleaner and finally I can stand to listen to records more once again. 

@lewm  " And you cannot even hope to detect such damage with a listening test, as someone else seems to claim. "

If it can't be detected by a listening test, then just how bad can the damage be?

How bad can the damage be? Good question. But I thought the question was how to detect damage, not whether it’s bad (or noticeable) or not. Whether it's noticeable in your test of course would be a function of the listener’s acuity as well as the degree of any damage. Besides, I did not say that I know the damage could not be detected by a listening test; you said that you could not detect damage based on apparently one experiment with one LP using one machine under conditions that you typically use for cleaning (I am guessing). I think you would admit that your study was not exhaustive.  In my opinion, to address the issue you'd want to do it scientifically to get an objective result, not a subjective one.

At my age it was taking too long to read the entire thing but I guess it is ok to keep using my ultrasonic cleaner. Most record only need one cleaning. The US machine was the best investment I have made for enjoying vinyl.

I've been using the Klaudio ultrasonic cleaner for years without any issue at all. The difference a truly clean LP can make can be astounding.

How about a before and after scanning electron microscopy image?  View the “damage”?

@lewm Agree, not exhaustive. There are many formulations of vinyl, for one thing. But then, cooking one side for an hour, reaching high temperatures, is a torture test by any standards - compare a with normal procedure of 6 rotations per hour, which leaves each small portion submerged for at most 4 minutes. Compare 60 minutes at temperatures exceeding 60 C. Frequency was 40 KHz, which is common and the most damaging to records.

Not publishable, perhaps - the only unscientific thing about the test was my generalizing to all vinyl records and all US frequencies. The human ear is a good transducer and it correlates well with other human ears, which measurements may not.

@terry9 & @lewm,

This is such a can of worms.  There are so many variables involved such as the record itself, of which we do not know the actual composition, but you may segregate to virgin and recycled, clear, black and colored.  But this would not cover the older composition that may have anti-static ingredients that may be exhausted from extended UT exposure.

Then you need to establish control for the ultrasonic tank.  If it's an inexpensive tank, can you believe the 'published' frequency and power; and other aspects such as power/volume and sweep frequency. Then you have rotating speed and of course duration and temperature.

Then you have the process chemistry - what exactly has been added - actual chemical and concentration.  To make it easier - just start with lab grade high purity to reduce the variability.  And the 800-lb Gorilla is how to dry.  The 'dry' surface will be influenced by the background environment.   

But, after all is said and done, what damage do you expect?  Assuming you eliminate water/solution residue as a variable, it should come down to damage of the thin high frequency side wall ridges or general surface erosion.  Damage to the side-wall ridges should result in loss of high frequency data.  But the damage may not be acute, but more chronic.  So, the damage does not become evident until after multiple plays.  Surface erosion may have no damage to frequency output, but it could increase the surface roughness which would then raise the background noise level.  But measurable weight loss is unlikely.  You could try sampling post cleaning the water for particle count, but you need to be able to differentiate record plastic from tank metal, possible but quite scientific.  So, @lewn is right that one of the best ways is to compare digital files before and after since hearing is so subjective.  

However, the 800-lb elephant is that unless the 'control' record is very clean, UT may alter the output but beneficially.  So, you need to first establish initial cleanliness w/o using UT?.  Too many pressing plants are very dirty with atmospheric contaminants including condensing organics from the lubricants used by the pressing equipment - all those moving parts are generally lubricated.

However, can UT damage a record, yes.  I am working with someone right now who was cleaning 25-min with a very powerful unit, very slow rotation and allowing the temperature to reach about 55C.  Chemical reaction rates generally double for each 10C temp increase, and cavitation intensity for lower kHz can also increase.  One recent late 1960's record finished with a mottled dull finish surface.  He said it still sounded OK.  He is now installing a cooler and increasing the rotation speed.

Just some casual thoughts.  

No objection to any of your finer points.  But you posit all good reasons why the weigh method is not the right approach, on which we agree.

@antinn Points taken. Elmasonic commercial unit (and now that reminds me, 37KHz, not 40, 37 being even more energetic), VersaClean lab detergent from Fisher Scientific, air dry in a clean environment. Control was the other half of the record.

Thanks for weighing in @antinn . Always good to hear from you and always good to hear from an international expert. 

"It is up to you guys to decide what you believe the data means. "

Considerably less than I expected.

Almost nothing.

At last an attempt at science on AG.  Congratulations.

But most AGers would prefer to clean the LP, listen to it and see if they can tell any difference.  You can read them above: 'if it's not detected by a listening test then there's no damage'. etc etc etc

clearthinker

At last an attempt at science on AG.  Congratulations. But most AGers would prefer to clean the LP, listen to it and see if they can tell any difference.

The notion that listening is itself unscientific and is an obstacle to science is misguided. If science is the goal, listening is necessary to correlate the results with the measurements. This is such a fundamental premise that I feel like Captain Obvious for mentioning it.

Cleeds, if you want to know whether your US RCM can damage your LPs, then you have to design an experiment to answer THAT question on the physical level. If you want to know whether you can hear a problem, then just listen. Aural memory is so poor that you’re likely to hear what you want to hear. OK with me but just be aware of the bias issue.

lewm

Cleeds, if you want to know whether your US RCM can damage your LPs, then you have to design an experiment to answer THAT question on the physical level.

I shared my listening experience and have conducted experiments to my satisfaction. Please feel free to conduct your own tests. Also please kindly note what I wrote in the post directly above yours:

If science is the goal, listening is necessary to correlate the results with the measurements.

Note very clear thinking, @clearthinker . You manage to mischaracterize my point and misquote me, all in only three lines.

The damage to the record can alter its physical dimensions or its physical properties.  Damage to the physical dimensions - any dimension change to the groove which can include change to the side wall modulations, or the material surface roughness 'may' be audible depending on the reproduction playback equipment (starting with the stylus shape), the listening environment and finally the listener.

Damage to the physical properties is much more insidious.  This can alter the durability of the record.  This damage becomes cumulative and if the degradation is graceful the listener may not notice unless they periodically compare the record to a control.  

UT is a high energy process and while there is no evidence that the record is heating on any bulk scale, that does not exclude what may be happening at the surface boundary as its bombarded with high energy implosions.  Heat is the main enemy of records.  If they overheat (beginning about 50-55C) PVC releases hydrogen chloride gas which if not scavenged by the internal stabilizer (part of the composition) the surface embrittles.  Once pressed, the record only has very limited stabilizer left.  Once its exhausted, that's it, any further excessive heating and its record damage.

So, because of uncertainty, being 'reasonably' conservative by keeping temps <40C, not spinning too slow (<~0.3 rpm), not using aggressive chemistry, not exposing for extended periods (>~20min); and at <60 kHz no crazy high power, there should be reason to suspect that UT is damaging the record.   But get on the wrong side of being conservative and UT can damage that record.  But trying to determine absolute thresholds of when and what damage will occur would be a serious scientific endeavor.  Although Olympus has some serious digital microscopes (magnification >15,000X) - check out the LEXT™ OLS5100 3D laser microscope that precisely measures shape and surface roughness at the 
submicron level.  But you can only measure a very, very small area, but probably better than scanning electron microscopes.

Just some casual thoughts.

@cleeds     I think more Captain Subjective.  The 'science' of listening will always be flawed.  Listening is always subjective and the results never scientific unless conducted under scientific conditions.  OP is to be praised for trying to build an objective scientific examination.  Brave, very rarely seen here.  If you can't follow that, then just dream on.......................................................................................................................and on.......................and.....................zzzzzzz

I believe both Rega and Ortofon do not believe in cleaning records with any kind of solvent or liquid. Dry cleaning or surface dust removal is ok, same for stylus cleaning. I personally use the record DR as well as manually cleaning with vinyl revival product. Generally though, I just dry clean them with a record brush. If you think about it, adding a liquid could just be causing mud/sludge, which may or may not be fully removed. Maybe best to let alone and dry brush and enjoy the music. I gauge it by examining the stylus upon completion of play. If it is consistently gunked up with crud from a certain record, then I proceed to use liquid(s). A lot of the time, the stylus appears rather clean, even after playing a thrift store find. It just depends on the record. My ears are old, and I honestly cannot hear most of the surface noise present most of the time, maybe it’s s good thing!😁

Honestly, each time you play a record, there is some minute amount of material being removed...how can there not be? A hard diamond running through the grooves (softer material) creating friction and of course heat. It is inevitable degradation.....with thousands of records, I doubt I’ll wear any of them out during the rest of my time. I own so many now that I honestly Forget what I have! Not a bad problem really....wifey thinks otherwise...🤨. Now to sneak the new Accuphase into the house...🙄

I commend your effort to try the experiment.  Given the tiny scale of weight we are interested in, I think this experiment is impossible to do reliably without very expensive lab equipment and carefully conducted in a controlled lab environment...

Hypothetical: If there was evidence a UT Cleaning was a cause of a loss of recorded data, as a result of effecting the Modulation in the Groove.

Would not a Test LP be a better to use as a LP Sample, as a Few Samples could be acquired and used in a manner that could at some point enable for differences developing between the LP's to be distinguished. 

Say a particular Khz, Rotational Speed and Temperature were discovered to have detrimental effect, as a result of Comparisons of the Test LP's, would this then be one of the cleaning environments to be avoided.

Is there not a Computer Software today such as 'Analogmagik', that would help with substantiation of  differences being detected between the Test LP's, if something is also being suggested there is a audible difference occurring.

For the Record, I abandoned the UT Cleaning and use the PAVCR Manual Cleaning Method, with solutions available to be produced in the UK. This has proved most satisfactory and Purified LP's are certainly a end product.