Magnetic Tape Harmonics / Modulation

[moamps]

IIRC, in that time, I  didn't do repro check more often, especially if repro head was in very good condition, but recording circuitry alignment was a must when we switched to another tape type. That was a PITA.

I would just add that the heads must be cleaned and demagnetised before the alignment procedure.

 

[abbey road d enfer]

At Barclay studios, we did the whole shebang every monday. We checked repro with a reference tape, but it didn't change much one week to the other. Record and bias alignment was more variable. Although we kept only one type of tape, we sometimes noticed significant variations from a reel to another. I concluded that it was not optimum. For utmost performance you need to do the alignment everytime you change the reel. For reasonable maintenance cost and confidence, you need to do a complete check about every month. That's what I did when I later had my own 16 track facility.

 

[radardoug]

At 7.5 ips, it is not possible to record 10k and up at normal operating level. The deck and tape wont handle it. -10 at 10K works OK, Hence 7.5 ips test tapes are recorded at -10, with a reference tone at 1k at operating level.
The spectrum of normal music rolls off at I think 3dB per octave, so in normal music high levels of hf are not present, and so we can get away with the tape compromises. At 15 and 30 ips you can record 10k and up at full level.

 

[PRR]

> It is instructive that they used 60Hz and 7kHz because even 7kHz was very HF for early tape machines (early everything). Not only did they consider 7kHz HF, they mixed it in 12 dB lower than the 60Hz.  .... SMPTE IMD

Actually 60/7K comes long before tape, is for amplifiers. Specifically Motion Picture (MP) amplifiers. 1939. ("T" came later.)

60cps should be in the passband, and is *cheap*. 7kc is in the passband of good film amplifiers, and its harmonic distortion falls 'above the audio range' (opinions have changed over time). 4:1 is an approximation of real signal spectrum--- nobody will pay to hear 7KHz near as loud as mids and bass.

 

[Bo Deadly]

Updated original post with 0, -10 and -20 dBu two-tone tests. Even at -20dB, distortion is pretty high.

 

[radardoug]

Its not at all clear from your tests what you are doing. Might I suggest doing it the old fashioned way? Buy a distortion meter like an HP 331. Then do straight distortion tests. You will get meaningful figures. It still looks like products are 40 dB down which is 1%.

 

[abbey road d enfer]

Updated original post with 0, -10 and -20 dBu two-tone tests. Even at -20dB, distortion is pretty high.

THD, being strongly related to saturation, increases rapidly above a certain threshold. OTOH, IMD is related to the slope of the B-H curve, thus has a more gentle increase.

 

[abbey road d enfer]

The main reason is wow & flutter. Instant speed variations result in pitch-shifting. Wow and flutter are the same as flutter, but in different frequency ranges, the latter above 4Hz and the former below 4Hz. Today they tend to be aggregated under the generic term of "frequency wobble".
Flutter is more damageable to perceived quality than wow, although the latter induces larger absolute pitch-shift, but the former introduces intermodulation distortion with non-harmonic components in the high-midrange.
A very good explanation is given here Wow and flutter measurement - Wikipedia

 

[abbey road d enfer]

Thanks for answer! I was afraid this thread was too old.
This physics was my first guess too, but measurements suggest it could be more than that.

Any comments/explanations for points below?
1. with same machine (B77) and different tapes the splatter profile is not the same.
- AGFA PE41 - gives me a very narrow peak and very small side bands (or none). Great S/N. Best tape that I found around in my pool.
- BASF LPR35 - gives stronger (larger amplitude) side bands and more +/-Delta_F offset (spaced from 1KHz line)
- TDK LX35B - a mix of the two, with less strong side bands than LPR35 but less strong ones than PE41. Also the spacing Delta_F is at around half between BASF and AGFA.

Different friction coefficients result in different speed variations.

2. with same machine: the sidebands and splatter depend on speed (3,75 or 7,5ips)

W&F is always inversely proportional to nominal speed.

- funny, the PE41 had the beautiful profile at 3.75ips but not at 7.5ips
3. with same machine: staying at 1KHz and varying the bias from min to max, only the amplitude of the 1KHz peak changes, not the shape of the splatter.

That's normal, because the mechanical effects are the same.

4. with same machine, same speed, same tape and Bias aligned: profiles at L, M and H frequencies are not the same, but side-band separation dF also not linearly increasing with frequency (suggesting that is not a simple modulation side band)

Friction and adherence are not linear effects.

I am physicist and not finding (myself or literature for reading) the model for this simple situation is quite disturbing and driving me crazy.

Moving an object under friction is a succession of tension and release.

 

[radardoug]

You refer to scrap jitter or flutter? Suspect is noted, but not all splatter features match this model. One machine has scrap rollers, yet some features are similar and still avoid a reasonable category.
I guess this is what I wanted to say: which parts of splatter do not come from mechanical effects? Can we see some examples of perfect tape slide (anyone knows if air-cushioning was tested and measured, just for fun and reference?).
Considering in real world a good machine (or a better one) and a fresh new tape, what is a reasonable shape and level for splatter? Looking at OP's plots using virginal new tapes - they are not far away from what I've got.
Anyway, true that all my tapes are old (and unknwown, I bought them on Ebay), but obviously I do not put in tests and measurements the squeaky ones. Rather could be reasonable large sized dust, shedding and other micronic clumps. How many such loose particles could be, considering that I have cleaned and inspected the tapes in all possible ways?
This answer suggests that if I would reduce the friction to zero (tape not touching the heads) then I will see a 1kHz clean peak as the source monitor delivers? (actually I come with measurements at the conclusion that all signal path is clean, up to the input of the REC current source buffer/amp stage, bias trap, oscillator and heads + magnetic recording process itself).


I do not know about adherence, but dynamic friction in this case should be a linear effect. Unless you go for flow mechanics, Reynolds conditions etc.
I can try wetting the old tape surfaces with some light oil (disclaimer - I do not care about those tapes, and machine parts are cleaned after each tape measurement - which is not much length).

I have taken today 36 spectra on 6 tapes, but decided will not post them yet.
First, I must buy some brand new tapes and redo the measurements. Probably next weekend, if Christmas comes earlier.
I am very sorry to see I cannot buy virginal PE41 as new.

What you should do is measure the wow and flutter of your deck. The evidence points to it not performing to spec.

 

[abbey road d enfer]

I do not know about adherence, but dynamic friction in this case should be a linear effect.

Macroscopically, it is generally accepted as linear, because it gives reasonably accurate accounts of forces, but at a microscopic level, the travel is subject to fluctuations.
Jut like pressure is macroscopically a steady quantity, when actually it is the result of the Brownian molecular agitation.

 

[radardoug]

Wrong, the W&F is in specs and the Bias is aligned for at least one tape (LPR35). Actually the machine records just fine (to my ears, and for younger ones too).
Analytically... something might be wrong and fixable. I am looking to improve and I am struggling to find out what exactly.
Even my higher expectations and unnecessary curiosity&questioning could be wrong... The curiosity is the worst one of all

So what is your w+F figure?

 

[abbey road d enfer]

Starting with simple arguments: if there was pure mechanical contribution, it would have been fully improved long long ago, before the 50's.

So you think tape recorders have not been improved mechanically between 1950 and 2000?

Indirect support for my thinking: when reading about W&F, we read about spectrum ranges below 100Hz for motors and scrape.

It has been shown that the traditional restriction of W&F below 100Hz is wrong.

That should be enough,

What is enough?

and if I understand correctly, any W&F above 100Hz in a non-servo regulated machine which has scrap rollers - is wrong.

What is wrong? Do you mean effects above 100Hz do not exist?

Can anyone help here with insights or hints?

I'm trying but you won't listen.

Generally speaking, there must be some improvements to get a semi-pro machine near to the pro level. I need to find the short list, and get it understood and prioritized

Just look at the evolution of tape recorders, say at Studer, see teh differences betweeen a J37 and an A827...

 

[Tubetec]

Wow and flutter on the Studer C/J37 is actually worsened by the presence of the mechanical tape counter , I saw one person did a project to create an optical( slotted disc, light source and seven segment leds) based counter for it , they didnt get it to work in the end , some of the bits ended up for sale as non working kits on Ebay

 

[radardoug]

Spec is less than .08%, so your machine is not in spec. Fix that problem first.

 

[radardoug]

Yes, well I have only been working on tape recorders for fifty years, so what do I know?

 

[abbey road d enfer]

Thank you, but I think the topic of this thread is different.
So far: as a side question, but obviously related to the original topic (IMDs), I asked if the spectral splatter causes are known. Because it is known that similar splatter fingerprint can be also typically caused electronically (i.e. not mechanically) by dense IMDs sidebands (so dense that they look continuous) caused by nonlinearities of the active elements and by other non-linear/hysteresis/mechanical (but not exclusively W&F) magnetic interactions.

Well, if you discover what in the electronics or in the process of magnetic recording is producing this additional splatter, you'll cause a significant advance in science.

 

[abbey road d enfer]

No sarcasm here, you claimed that 50% of the effect is due to electronic causes, you should have some data to support it.
Personally, I am not able to bring a single element of data that could support your claim.
You say you have found no evidence of such data on the internet, maybe for the simple reason it doesn't exist.
What I know is that the effects of W&F are known since long and that they have always be analized as frequency wobble, which is what the FFT skirts indicate.

 

[abbey road d enfer]

OK, you always want to have the last word. If you think it's how you'll get answers, keep on.
You win, I'm done.

 

[zamproject]

@Analog Pitstop
The few you'r dealing with in this topic probably cumulate tens and tens of years rolling tape for audio industry.
For what I read you get answer to all your Q... but who I am ...not expert in tape machines.