News:

"The phone is a remarkably complex, simple device,
and very rarely ever needs repairs, once you fix them." - Dan/Panther

Main Menu

Setting timing on a dial - A discussion of Analytical Instruments and Audacity

Started by RDub, October 28, 2016, 04:47:58 PM

Previous topic - Next topic

DoubleTone

Another alternative: a TPM32 Digit Grabber, available on eBay (as in, right now) for about $60+.  I bought one several years ago, and for dial calibration, it's excellent.

.....

Quote from: DoubleTone on October 29, 2016, 07:29:28 PM
Another alternative: a TPM32 Digit Grabber, available on eBay (as in, right now) for about $60+.  I bought one several years ago, and for dial calibration, it's excellent.

Here's one starting at $10.00

http://www.ebay.com/itm/Metro-Tel-TPM32-Touch-Tone-Digit-Grabber-/302118208864?hash=item4657a60560:g:A90AAOSwnbZYD5w9

dsk

Quote from: DoubleTone on October 29, 2016, 07:29:28 PM
Another alternative: a TPM32 Digit Grabber, available on eBay (as in, right now) for about $60+.  I bought one several years ago, and for dial calibration, it's excellent.

This thread has lots of arguments for and against this or that, and we will probably not even agree about the need of accuracy, but does this unit tell how fast or slow the rotary dial are, or the make/brake ratio?

The need of accuracy are not pretty high on my equipment, so I will not tune the 2 dials I posted sound files of, but if I have to tune it I will try to tune it as good as possible brake at 60-66% and speed 10-12 pps will be OK

dsk

DoubleTone

Quote[...] does this unit tell how fast or slow the rotary dial are, or the make/brake ratio?

Yes, and yes.

Stan S

"Yes, and yes."

The answer is no and no.
They made different models all having the same number. That one is Touch Tone only.

It's my auction.
You have to be careful with those.
Stan S.

TelePlay

Quote from: dsk on October 30, 2016, 01:26:55 AM
[ ... ], but does this unit tell how fast or slow the rotary dial are, or the make/brake ratio?

. . . and other than a digital oscilloscope that can capture and display a trace, does or do these "averaging" analyzers tell a dial's full story like this, first the as found state or condition of a dial



and the state or condition of the dial after cleaning and oiling?



Yes, it take a lot longer to hook up the dial after the computer is turned on, the capture the audio file, to determine each event point in time (start-stop and each break/make point) and to put that data into an excel template to analyze the dial, then to do a screen capture of the audio file, past it into Photoshop and annotate the dial analysis for personal use or posting. That would be about 30 minutes work once comfortable doing all that (having done it enough in the past to make it easy).

Certainly not something a high volume refurb person or shop would want to do but in my opinion, very useful to a hobbyist who not only wants to fix/restore a dial but also understand what is wrong with it and see the effects of their work. There's at least 4 unique issues seen in the first image above.

Just saying, a lot of time to get these scans but IMHO, invaluable for the hobbyist trying to understand a dial and fix/restore it by taking a scan after each procedure performed (the difference is the dial between being cleaned and then being oiled and then several days later to see if the oiling has kept the dial the same (I just discovered the oil I used does not hold up over time (months) and as such am in the process of switching to a different, more expensive oil).

Just saying, but I do appreciate all of the information I can capture at different points in time and retain it digitally on any dial in my possession.

unbeldi

Quote from: TelePlay on October 30, 2016, 08:44:36 AM
. . . and other than a digital oscilloscope that can capture and display a trace, does or do these "averaging" analyzers tell a dial's full story like this, first the as found state or condition of a dial



and the state or condition of the dial after cleaning and oiling?



Yes, it take a lot longer to hook up the dial after the computer is turned on, the capture the audio file, to determine each event point in time (start-stop and each break/make point) and to put that data into an excel template to analyze the dial, then to do a screen capture of the audio file, past it into Photoshop and annotate the dial analysis for personal use or posting. That would be about 30 minutes work once comfortable doing all that (having done it enough in the past to make it easy).

Certainly not something a high volume refurb person or shop would want to do but in my opinion, very useful to a hobbyist who not only wants to fix/restore a dial but also understand what is wrong with it and see the effects of their work. There's at least 4 unique issues seen in the first image above.

Just saying, a lot of time to get these scans but IMHO, invaluable for the hobbyist trying to understand a dial and fix/restore it by taking a scan after each procedure performed (the difference is the dial between being cleaned and then being oiled and then several days later to see if the oiling has kept the dial the same (I just discovered the oil I used does not hold up over time (months) and as such am in the process of switching to a different, more expensive oil).

Just saying, but I do appreciate all of the information I can capture at different points in time and retain it digitally on any dial in my possession.


You should be able to shorten the analysis time considerably, by using the built-in tools in Audacity to give an almost immediate read-out of the dial speed.
Audacity can analyze data in time and frequency domains.  By using an autocorrelation function on the trace it should tell you immediate the duration or frequency of the beats in the sample. With some research a plug-in could be written for Audacity that would optimize the settings for dial speed data and provide an immediate read-out of at least dial speed.  But it should also be possible to automatically analyze the break/make ratio.  This would be really no different than what I am doing in my home-built analyzer with an Arduino, only it is analyzing the data in real-time, rather than taking the sample from a file.
My analyzer samples the on-off status of the switch with sufficient resolution and computes the average make and break duration for a full digit. This is exactly the same principle that you are doing manually. The real-time analyzer is just complicated a little by the fact that it has to automatically detect the beginning of a digit, and distinguish that from a on-hook and off-hook condition, and from hook switch flash signals.

If you still have the sound file of this plot, try the following:

a) Plot the frequency spectrum
b) try some of the auto-correlation options in the Algorithm tab:  Try Standard and Enhanced.
It should give you peaks for the frequencies of the sample.

I don't know whether this provides sufficient accuracy, or entails more guess work, as you will have to select the true fundamental frequency from a spectrum, but it should be in the 8 to 12 Hz range.

unbeldi

ok, I just tried the autocorrelation myself on the two files that you provided earlier.

I come up with the impression that the first of files has a lot of jitter in it, and I think it will test mostly fast at around 12 to 12.5 pps, but does have some low frequency components between 7.9 and 8.3 pps.

The second file (AE Dial 02) seems to perform more uniformly, but slow at 7.8 pps.

What does your manual measurement determine?  I have to admit that I am too lazy to do that myself.

It may be that the first dial does not have a constant recoil speed, perhaps it is slowing down during the return rotation.


AE Dial 01 Speed Audio File.wav
AE Dial 02 Speed Audio File.wav

TelePlay

Wow, first, I didn't know Audacity had all of those features. I downloaded it years ago for use making and editing sound effects on my PC when at the theater during rehearsals (I use Adobe Audition on my home main PC to do most of the heavy lifting work of sound design but needed an easy to use audio program when on the road for last minute creations/adjustments). I'll have to look into those features so thanks for the lead/tip.

As for the two files, both are as found condition. One is indeed 12.13 PPS and the other 7.80 PPS. I have the manual analysis image files on my other computer and will add them to this post later today. dsk also got the same numbers so exchanging data files for analysis on other computers to test this method seems to confirm no variation due to systems/software used.

As a side FYI, seems anything can be attached to a forum post by enclosing it in a .zip file. Audio files (wav, mp3, etc) are not allowed and can not be uploaded unless encased in a .zip folder. Good to know for anyone wanting to post an audio file.

unbeldi

Ok, the candidate gets 100 points, LOL.  I call that an unqualified success, as far as dial speed is concerned.
While you posted I cropped the screenshots I took earlier.

After loading your files, I cropped them to the data just before and after the signals.  On that I applied a 100 Hz low pass filter to get rid of all the features that not pertinent to our interest here.  Those are shown in the top portion of these screenshots.

The lower portions show the enhanced autocorrelation functions of each trace.

dsk

I love this kind of threads, and will like to underline a few of my ideas.

Do not try to repair somethings working. Regular cleaning and oiling and service may still be done if you consider it is smart to do it.
Audacity and sound files together with a good spreadsheet is good enough for the hobbyist to tune a dial or two from time to time. The make break ratio will usually not change considerably during regular use, and most equipment will read a zero of a dial using approximately the time it takes to say 1001 from release to stop.  (Siemens dials slightly longer, they have long way to move between 1 and stop)  A rough but good enough for me, spreadsheet is here: https://docs.google.com/spreadsheets/d/1iUKj3jlKp7EoqXAOceQozimP6Pb4OKyP9l5B_KFUztg/edit?usp=sharing

dsk

unbeldi

The manual method that you used to analyze your dial 'sound' trace is the simplest form of time domain analysis. You measured the time between transitions between on- and off-states at the zero crossings, and reduced the data for 10 pulses by averaging to yield just one number.

Autocorrelation is a mathematical technique used in digital signal processing (DSP), that is more sophisticated, and relates to frequency domain analysis by the Fourier method, although it is still a time-domain technique.

I am certain this is also the technique used in the Sage Instruments 930A. It uses DSP algorithms for many of its tasks when measuring frequency.  When you get some experience with that instrument you'll find that it rarely provides a definitive constant number for the dial speed. There is always some fluctuation. This stems from the nature of the signal, it contains noise and old dials do no rotate with a completely constant speed for every digit.   From my autocorrelation graphs you see that the peaks have a certain width, can even have fine structure internally with multiple smaller peaks.  Now the question arises which one do we pick as the dial speed ?  I don't know what the Sage does exactly, but I suspects it computes some statistic on those peaks, perhaps fits a curve to it, and takes some kind of average after satisfying a set of signal thresholds. The results can vary from time to time, and they do.  I have had some dials that even seemed to completely confuse the box, and it was frustrating for this user too. Perhaps an example might be your dial that is mostly fast and partly slow.  In these cases, your method of recording these traces and using a spectrum analyzer should explain the situation better, and actually it did!

One more aspect: The trouble with simple autocorrelation has always been that the result is not immune to the harmonics of the true or expected main frequency.  Peaks show up not only at their "nominal" frequency, but at twice or other multiples of it.  This is avoided by many enhanced autocorrelation techniques that mathematicians have developed, some of which are also implemented in Audacity, hence my recommendation to try those.  I do have some additional screen shots that I took earlier to select those that I presented.  Later...

[PS: for early readers, I misstated the relationship of time and frequency domain in my first draft of this.]

Contempra

Quote from: dsk on October 30, 2016, 11:59:53 AM
I love this kind of threads, and will like to underline a few of my ideas.

Do not try to repair somethings working. Regular cleaning and oiling and service may still be done if you consider it is smart to do it.
Audacity and sound files together with a good spreadsheet is good enough for the hobbyist to tune a dial or two from time to time. The make break ratio will usually not change considerably during regular use, and most equipment will read a zero of a dial using approximately the time it takes to say 1001 from release to stop.  (Siemens dials slightly longer, they have long way to move between 1 and stop)  A rough but good enough for me, spreadsheet is here: https://docs.google.com/spreadsheets/d/1iUKj3jlKp7EoqXAOceQozimP6Pb4OKyP9l5B_KFUztg/edit?usp=sharing

dsk

Hi dsk ,... First , I've never adjusted a dial on rotary phone , secondly , I don't understand the spread sheet . so what is the difference ? i can call every where without problems !, can you explain me ?.. oh I rebuilted some dials but no adjustment on them .

unbeldi

Here are the  "Standard Autocorrelation" plot and the "Enhanced Autocorrelation" (per Audacity) of the first of John's dials.

By immediate visual inspection we see that the enhanced version is much cleaner.  On second view we also see that the low speed (= high lag) component is almost not visible in the simple version, it can easily be just overlooked.  It is obscured by the first "harmonic" with twice the time delay (0.160 s) than the first.

Our dominant time delay is at 0.080 s, which corresponds to 12.5 PPS, which may be identified in both plots, but it is very impressive how the second method enhances that data.

Dan/Panther

Personally I guess that the accuracy of the dial can be set to perfection if one chooses. However like I mentioned, of the scores of phones that I have cleaned and restored to functionality, every single one worked as it should. Maybe I just got lucky ?
D/P

The More People I meet, The More I Love, and MISS My Dog.  Dan Robinson