Author Topic: Setting timing on a dial - A discussion of Analytical Instruments and Audacity  (Read 2150 times)

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
I usually do not fix things who works, and to Dan, and Dennis. You will only need this in a few situations, to fix a physical damaged dial, or to cope with some analog adapters.  It is not about luck, but a sudden feeling of whats right, and the most of telephone equipment is made to tolerate really large differences. Usually it works without any hi-tech adjustments.

I actually finds this interesting, and it looked like my Starlite had a slightly different last pulse, an yes it had.  By visual inspection I could observe the last pulse was disturbed by the little finger securing the pulse contact in closed position when the dial is resting. Only tools need was a pair of pliers.   But as you have observed, this was not necessary to do, it worked before too!

What could I to make the audacity visualization better for this thread?  Adding a 1000Hz tone and feeding back to the microphone input.  Here follows my readings, the spreadsheet with values from audacity.  And then I read out pretty good readings of make and break ratio, and PPS value. 
Enclosed pictures.

dsk
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
As you see, the speed is a little slow 9.65 when I would like to have 10, by my opinion good enough.
The same about brake length I would prefer 60-66, but got 66.58.  No reason for adjustments. 

Should I sell it as adjusted, it would be a different case, It had to be within 60-66% and 10-12 PPS.

By theory, it is not good, in practice life, its more than good!

dsk
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Online TelePlay

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 6763
  • A Ringy-Dingy Thingy
    • . . . been good so far
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

Got to the computer I use to do dial tests, where the analysis files are parked so they are attached as well as the spread sheet I use (in a zipped folder) to determine PPS and the break/make ratio for each digit, and then the average of those ratios (the long hand, time consuming way to do it)
            John . . .

              

Offline Contempra

  • Hero Member
  • *****
  • Posts: 926
As you see, the speed is a little slow 9.65 when I would like to have 10, by my opinion good enough.
The same about brake length I would prefer 60-66, but got 66.58.  No reason for adjustments. 

Should I sell it as adjusted, it would be a different case, It had to be within 60-66% and 10-12 PPS.

dsk

Okay but how to adjust the dial ?  all i see is 2 screws !.. As i said , i've never adjusted a dial but all works ... but i'd like to know how to adjust a dial .. thanks
Denis

unbeldi

  • Guest
You can also automate the measurement of the break ration, although I did not find a suitable filter in Audacity, although the function for the first step of that is available.
The principle is the same that most analog break/make measuring instruments used by Western Electric, for example in the No. 2B test sets for pulse train measurements.

The method involves clipping your sound recordings at the top and the bottom to a square wave with the zero line exactly in the middle. That means removing any DC bias from the signal.  Let's say you clip the data to values +1 and 1, centered on 0.  This can be done with the clipping function in Audacity. In addition the data should be clipped to a some number of complete dial pulses, not more and not less.  I am showing the result below for your No. 2 sample.

The next step is to divide all data into sample slots, which they are in already actually, because the data is sample at a certain frequency. I believe it was 44.1 kHz in your samples.  So, all you have to do now, is compute the sum of all samples, i.e. adding the sequence of 1 and +1 values.  Image that the make/break ratio is exactly 50%, this would mean that there are the same number of negative values as there are positive values and the sum would be zero.  For 66% there are 2x as many samples of one polarity vs the other. There the sum is a direct representation of the break/make ratio.  With the polarity in your data, negative being a break, the sum is positive below 50% and negative above 50%.  All you need to do is count the sample properly which is the 100% value and normalize the value.
This should be very simple to implement in an Audacity plug-in, the hardest part being learning how to do that.  I would actually be surprised this does exist already somewhere as a third party module. I will look.

I will attach the promised image in a moment...

The top graph shows the original data trimmed to 9 pulses, and the lower half shows the clipped data normalized to +1 and 1.

BTW, if one uses this clipped signal, the correlation function cleans up amazingly, and provides a nice bell curve for reading the principle frequency.
« Last Edit: October 30, 2016, 09:14:32 PM by unbeldi »

unbeldi

  • Guest
BTW, if one uses this clipped signal, the correlation function cleans up amazingly, and provides a nice bell curve for reading the principle frequency.

Here is the graph for that.

There is now almost no uncertainty in reading the peak value, which has a delay of 0.1258 seconds.  Taking the inverse of that, gives  7.95 PPS.
You can't expect it to be exactly the same as before, but we have to remember that previously, I just took a guess at the peak.

« Last Edit: October 30, 2016, 09:13:00 PM by unbeldi »

unbeldi

  • Guest
Ok, last post today, I promise.

Attached is a paper introducing a break/make ratio meter of 1933 by Ericsson in Great Britain.
I came across it not too long ago while researching other matters.

Ericsson Bulletin January 1933, No. 2, page 16  A new Impulse Ratio Tester

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
Okay but how to adjust the dial ?  all i see is 2 screws !.. As i said , i've never adjusted a dial but all works ... but i'd like to know how to adjust a dial .. thanks

This time, I  shall try to not mix threads, but take a look at this thread: http://www.classicrotaryphones.com/forum/index.php?topic=15443.0

Every dial has their slightly different way to adjust the tension on the centrifugal break, Typical dials used here looks like AE dials were you have to bend on the springs. but anyway we are talking about the same principles. 

dsk
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
I  would like to thank everybody who has contributed to this thread, it has made me making the process of using audacity better.
dsk
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Online TelePlay

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 6763
  • A Ringy-Dingy Thingy
    • . . . been good so far
As an FYI follow up, I have a very slow ASUS Notebook that I bought 5 years ago for a traveling reason but haven't used it for that in 3 years. Use it occasionally to keep it updated and battery charged.

It's now in my dungeon hooked up by hard wire to my internet box. Put a full sized external HP keyboard that someone at work tossed in the dumpster a year ago on it, added a wireless mouse, a large display I had on a now dead computer. Found a Creative external USB sound card I used many years and computers ago (to add 5.1 stereo to my laptop for theater work) and downloaded Audacity.

The external sound card being a nice buffer between a dial and Audacity. Just checked my first dial. Works great - no need to run upstairs to my desk top to check dials anymore.

So, all the junk I had laying around in this box and that box is now a very nice platform for dial checking, and everything else on the internet. I'm happy.
            John . . .

              

Online Pourme

  • Hero Member
  • *****
  • Posts: 1432
  • Benny
Excellent, John. I had thought about doing the same thing, to avoid taking dials from my workspace to my office. I never kept any of my old computer equipment. I want to keep working on my phones restricted to my bench. Whatever I use to work on dials I will have to learn. After considering these reasons I ordered a 930a this week off of EBAY. I have a spot on my bench ready to place it. I have much to learn and all winter to learn it!
The Internet is a telephone system that's gotten uppity

Online TelePlay

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 6763
  • A Ringy-Dingy Thingy
    • . . . been good so far
Had a 10A dial in a late 60s Trimline that was a bit slow when nearing the end of returning when dialing 0. The first image shows the break/make analysis for 10 digits. The slow down at the end is clearly apparent. Analysis of the data shows an overall PPS rate of a bit more than 9 and the break/make ratio near 60/40.

The bottom of the first image shows the break/make digital wave form for 10 digits after the dial was removed, stripped down to it's basic metal/plastic components and cleaned ultrasonically for 25 minutes, rinsed off with water and then distilled water, excess water blown off with compressed air and then heated to about 100-105 degree with a hair dryer for a few minutes. All of the bearing and friction points were oiled with a watch oil. On re-assembly, 1/3 of an extra turn was put into the dial spring. This bottom wave form shows no problem as the dial ends a 10 digit cycle and each pulse cycle looks cleaner that the as received wave form.

The bottom image is a side by side spread sheet analysis of the pulse data, before on the left and after on the right. The dial picked up 1.3 PPS from 9 to 10.3 PPS, some of that increase due to cleaning and oiling and a bit due to the extra tension put into the spring.

What's interesting is the standard deviation for break and make periods for each of 10 pulses before and after cleaning. While the average ratios before and after are acceptable and very similar (note - graph labels for the average break/make are reversed), the standard deviations improve from 1.441% to 0.619% for the measured "make" times and from 0.947% to 0.733% for the "make" times. That would indicate a more consistent break/make period over all 10 pulse periods.

This was done the hard way, data read off of Audacity wave form and manually entered into an Excel spread sheet.

Presented as further anecdotal evidence that dial ultrasonic cleaning works to clean without fill disassembly and Audacity with Excel is a good combination of software tools. While the "dirty" dial may have worked, the "cleaned" dial would produce better results when dialing.
            John . . .

              

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
...

The bottom image is a side by side spread sheet analysis of the pulse data, before on the left and after on the right. The dial picked up 1.3 PPS from 9 to 10.3 PPS, some of that increase due to cleaning and oiling and a bit due to the extra tension put into the spring.

What's interesting is the standard deviation for break and make periods for each of 10 pulses before and after cleaning. While the average ratios before and after are acceptable and very similar (note - graph labels for the average break/make are reversed), the standard deviations improve from 1.441% to 0.619% for the measured "make" times and from 0.947% to 0.733% for the "make" times. That would indicate a more consistent break/make period over all 10 pulse periods.

This was done the hard way, data read off of Audacity wave form and manually entered into an Excel spread sheet.

Presented as further anecdotal evidence that dial ultrasonic cleaning works to clean without fill disassembly and Audacity with Excel is a good combination of software tools. While the "dirty" dial may have worked, the "cleaned" dial would produce better results when dialing.


You did the right thing, all the way.
Adjusting the spring did solve the slow speed on last digit, and that did probably also make the make/brake correct for the last pulse too, if you look at the make/break on the 5 first digits of before and after, it is pretty constant (=right).

The make/break ratio has different standards here and there but all my exchanges, and dial gizmos has accepted them all.  The break is 2/3 in some countries, and 60% in others so if I should adjust a dial in hence of this, I would put it  between 60 and 67 %.

dsk
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Offline dsk

  • Hero Member
  • *****
  • Posts: 3182
Recently it was several B&K Telephone testers at EBAY. But are they good enough to test phones for use on IP adapters?
The Norwegian standard for dial speed was to tune between 10 and 12 pps, witch probably tells us about the exchange should accept e.g. 8-15 pps?? We have brake/make 60/40 UK has 2/3 to 1/3 Never recognized problems with the ratios on regular phones so I just have to guess about it to be tolerant.

dsk
« Last Edit: February 06, 2017, 03:58:26 PM by dsk »
Please notify me of dead links in my postings.

C*NET +1 999-3011


I have even got a regular New York number :-) 646 570 1796

Offline Jim S.

  • Hero Member
  • *****
  • Posts: 3122
I think the BK 1045's use the same dialing  tolerance as the phone company . If that is the case, then I don't think the tolerance is tight enough for IP.

Just a guess tho,
Jim S.

You live, You learn,
You die, you forget it all.