Ich brauche Hilfe mit meinem Telefunken M15

[sigitask]

I need help with my Telefunken M15. Strange very low frequency amplitude modulation when recording. Please see the attached video. What can be the cause of this problem?


https://www.youtube.com/watch?v=oochHjkJvaE

[kaimex]

Hi,

I don't see any low frequency amplitude modulation.
I see a 1 kHz test tone with stable amplitude plus some spiky interference with apparently low repetition rate.
I am hearing some background music and wonder what that might have to do with the images on the display.
If the interference is the problem, it would be helpful to show it without the test tone, just triggered on those spikes.

Regards Kai

[Peter Ruhrberg]

Is it possible to have that 1 kHz signal in the video as well?
Have you checked synchronicity of the AM with supply reel rotation?
Does it appear on both channels and with different tapes?

Peter

[andreas42]

Hi sigitask,

from the video, it looks like a modulation frequency of roughly 2-3 Hz - and might correspond to one of the rotating elements in the tape path: Pinch roller, capstan, reels. I'd expect the frequency to change when you change to the other tape speed?

If you happen to have a reference tape, it would be useful to digitze a longer sine wave (e.g. the one for the reference level, or for azimuth adjustments) of about 1 minute, preferably with a frequency of 315 / 1000 / 3150 Hz, both channels in one file, and make it accessible here. We can then have a look at the spectrum and phase over time.

Also without a reference tape, you can record and replay 1-2min of a generated sine wave with 1000 or 3150 Hz, for the same measurement.

Best,
Andreas


edit: of course, Kai and Peter were faster

[andreas42]

Taking another look, I agree with Kai's point: The spectrum shown on the laptop screen with the 1 kHz peak and the waveform below (preiod length of about 1 ms) does not correspond to the audio track of the video; at the same time, the music doesn't seem to be visible in the spectrum. My "2-3 Hz" comment is in reference to the oscillation of the needle on the Sennheiser instrument.

Could you describe the experimental setup seen in the video in a bit more detail, e.g. what signal from where was present a) on the computer b) on tape c) on the sennheiser?

Best,
Andreas

[kaimex]

the oscillation of the needle on the Sennheiser instrument.

that might be just "apparent" due to the slow response of the instrument to the interference spikes

Kai

[sigitask]

Thanks for answers and toughts. I have to explain that this measurement what we see on video was made during recording of 1khz signal and playback it and analising with FFT. This spikes on FFT imidge have no connection with parasitic AM modulation, it was some soft artefact. After reloading soft signal eas clear without any spikes, but with the same AM modulation. Test tapes signal on playback is absolutelly stable without any deviations. This modulation appears only by record-playback. This music what is possible to hear is from radio and have no any connection with signal. I will try to make record with 7,5ips to check period.

[kaimex]

Hi,

are you sure that there is no superposition of recorded and playback signal with a possible beat frequency in case of speed variations ?

Kai

[sigitask]

How can happen this superposition? Im conected my machine Clio pocket and it works perfectly. Theis machine has separate generator and separate analyser.

[sigitask]

Taking another look, I agree with Kai's point: The spectrum shown on the laptop screen with the 1 kHz peak and the waveform below (preiod length of about 1 ms) does not correspond to the audio track of the video; at the same time, the music doesn't seem to be visible in the spectrum. My "2-3 Hz" comment is in reference to the oscillation of the needle on the Sennheiser instrument.

Could you describe the experimental setup seen in the video in a bit more detail, e.g. what signal from where was present a) on the computer b) on tape c) on the sennheiser?

Best,
Andreas

Setup is very simple. Measurement machine is clio pocket, with separate generator and analyser, everything is digital. Results is possible to see on computer screen. Signal 1khz from generator goes to M15 inpt. Output is conected to clio pocket FFt analiser and to milivoltmeter Shennheiser. This slow AM modulation is possible to see not only on arrow of Sennheiser but also on clio pocket osciloscope section, so it is not from sennheiser inertion.

[sigitask]

I tried to make record on LPR 35 tape, and now situation is normal. No harmonic modulation. Probably is some mechanical problem?

https://www.youtube.com/watch?v=0knVmgpd2y4

[sigitask]

Hi sigitask,

from the video, it looks like a modulation frequency of roughly 2-3 Hz - and might correspond to one of the rotating elements in the tape path: Pinch roller, capstan, reels. I'd expect the frequency to change when you change to the other tape speed?

If you happen to have a reference tape, it would be useful to digitze a longer sine wave (e.g. the one for the reference level, or for azimuth adjustments) of about 1 minute, preferably with a frequency of 315 / 1000 / 3150 Hz, both channels in one file, and make it accessible here. We can then have a look at the spectrum and phase over time.

Also without a reference tape, you can record and replay 1-2min of a generated sine wave with 1000 or 3150 Hz, for the same measurement.

Best,
Andreas


edit: of course, Kai and Peter were faster

When I playing MRL test tape signal is absolutelly stable without any modulation. It is only on record-playback. Will make digital record in few days. At hte moment do not have digital recorder.

[kaimex]

In case you accidentally mix a 1 kHz signal with approximately 6 % of a 1002 Hz signal you will end up in something that looks like ~1 kHz signal with 1 dB of amplitude variation/modulation and 2 Hz repetition rate. This could e.g. happen on a computer's audio input mixer or in a mixing console, when your tape recorder has periodic speed variations of this size. Perhaps this already happens in the M15.
You could check this hypothesis by analyzing the 1 kHz playback response more closely with a longer FFT to acchieve better then 1 Hz resolution.

Regards Kai

[Peter Ruhrberg]

I tried to make record on LPR 35 tape, and now situation is normal. Probably is some mechanical problem?

I don’t know which type of tape you used for the first recording.
However, it is quite possible that its sensitivity changes rhythmically (maybe due to irregular casting or some finishing process) which would constitute a production error.
Sometimes a similar effect appears if the machine’s bias level is not adjusted to the tape in use.

Peter

[sigitask]

I tried to make record on LPR 35 tape, and now situation is normal. Probably is some mechanical problem?

I don’t know which type of tape you used for the first recording.
However, it is quite possible that its sensitivity changes rhythmically (maybe due to irregular casting or some finishing process) which would constitute a production error.
Sometimes a similar effect appears if the machine’s bias level is not adjusted to the tape in use.

Peter

Firs record was made on tape SM468. Bias was adjusted to this tape, overbiassing -3db drop from maximum on 10kHz. May explain what you mean "irregular casting or some finishing process" ?

[sigitask]

In case you accidentally mix a 1 kHz signal with approximately 6 % of a 1002 Hz signal you will end up in something that looks like ~1 kHz signal with 1 dB of amplitude variation/modulation and 2 Hz repetition rate. This could e.g. happen on a computer's audio input mixer or in a mixing console, when your tape recorder has periodic speed variations of this size. Perhaps this already happens in the M15.
You could check this hypothesis by analyzing the 1 kHz playback response more closely with a longer FFT to acchieve better then 1 Hz resolution.

Regards Kai

Thank you! Interesting toughts. But how I can mix this 1khz and 1,002khz signals. There is no mixer. Signal from playback amplifier output goes directly to FFT analyser and to milivoltmeter. And this is not computer hardware. This is Clio Pocket machine. Very simple. But I can not make longer FFT sequence because it is maximum what this analyser can.

[kaimex]

Hi,
when I wrote that, I was not aware that you are using Clio pocket (did not yet know that hardware), assumed you were using a PC with some sound-Interface.
"mixing" is meant in the sense of crosstalk, i.e. adding unwillingly.

But I can not make longer FFT sequence because it is maximum what this analyser can.

You could make a recording of several seconds (T) according to the desired resolution (df~1/T) with 48 kSps and analyze it with other software, that can make longer FFTs and zoom into the result at any frequency. Audio analyzers with a capability for "Wow and flutter measurement" do that.

Regards Kai

[kaimex]

You might post a 10 s recording of a 1 kHz testtone at 48 kSps as 16 Bit PCM (wav-file). This will result in slightly less than 2 MB (stereo) or 1 MB (mono). Up to 5 MB are allowed in a single attachment file. A *.wav file is not accepted, must be put in a zip-archive.

Regards Kai

[Peter Ruhrberg]

May explain what you mean "irregular casting or some finishing process" ?

During the production process, a suspension of magnetic particles in lacquer is cast onto the film. This critical step has to be done with utmost precision and evenness for achieving a homogeneous coating with smooth surface.

Variations in casting pressure or film speed for instance would result in sensitivity variations at lower frequencies, while defects in the calender system would influence higher frequencies which are recorded on the surface.

Sensitivity variations of a random sample of SM 468 made by Pyral in 2015:

   

You might check your SM 468 tape by recording a similar series of frequencies.

Best, Peter

[sigitask]

Dear Kai, yes I will make next week digital record with Nagra LB, and put here. Thank you!

[sigitask]

Dear Peter,
Thank you very much! I never was thinking that partickles non regularity can have some regular close to harmonic distributon. But in this particulary case, on Studer A812 the same test goes absolutelly free af any AM modulation.

[Peter Ruhrberg]

in this particulary case, on Studer A812 the same test goes absolutelly free af any AM modulation.

Those things happen sometimes ....
You may do another experiment with your M15 by changing the bias and see if the fluctuations change.

Best, Peter

[sigitask]

Hello Peter,
It is very interesting software. Maybe possible to measure W&F to? May ask to share link to download software.

[Peter Ruhrberg]

It is very interesting software.

Actually, it is standard DAW software (albeit somewhat expensive) called "Samplitude" or "Sequoia", respectively.
Audacity should be able to display envelope curves as well, but I don't know.

There is also a number of W&F software on the internet.
This zipped freeware for instance does the job nicely:

  Wow&Flutter GUI.zip (Größe: 633.54 KB / Downloads: 6)

Peter

[kaimex]

Audacity can display waveform/envelope(double-sided) (linear) and in dBs and spectrogram. In the newest version even both at the same time.
In an extra window the spectrum of a limited time interval (not specified) can be displayed via FFT with maximum size 65536, but it is not possible to zoom into details of this spectrum. Not even the limits of the frequency axis can be set. You can only switch between linear and log scaling. The data can, however, be exported into a file and then processed as you like by other programs like e.g. Matlab, Octave, SciLab, FreeMat, python....

MfG Kai

[sigitask]

It is very interesting software.

Actually, it is standard DAW software (albeit somewhat expensive) called "Samplitude" or "Sequoia", respectively.
Audacity should be able to display envelope curves as well, but I don't know.

There is also a number of W&F software on the internet.
This zipped freeware for instance does the job nicely:
[attachment=34538]

Peter

Thank you Peter!
W&F meter works well!

[sigitask]

Today bring Nagra LB digital recorder to record signal with AM modulation. Put the same SM468 tape, the same settings. And AM modulation disappear compleatelly. Signal absolutely stable. What of ghost effects !

[kaimex]

Elderly people and machines tend to have occasional tremor.
Some get already much better when the doctor comes closer with a syringe...

regards Kai