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

RDub

Is there a scientific way to calibrate the pulse timing on a dialer or is it an experienced eye type of thing? And does timing, only, become an issue if you disassemble the gearing while cleaning / lubing during the life of the dialer? RDub

Ktownphoneco

Good question.    I repair and restore telephone dials.   There are a number of scenarios floating around the antique telephone world that would lead one to believe that any number of methods can be used to determine if a dial is functioning correctly or not.     The only way to accurately test a dial to see if it's performing within the require parameters, is with some sort of dial test equipment.
I use a Sage Instruments 930-A telecommunications test instrument.      It gives me precise and accurate measurements of how a dial is performing, whether it's a pulse dial or a touch tone dial.
As I previously mentioned, a pulse dial has 2 parameters it needs to satisfy, speed and the break / make ratio.     The only way to do that in any sort of precise manner, is with an instrument designed for that purpose.   But the bottom line is, that if your telephone is dialing numbers accurately, and that is to say, if every time you dial a number, you reach the person or party you were trying to contact, then there's nothing wrong with the dial.     If that's in fact the case, leave it alone.   It's working correctly.

Jeff
 

TelePlay

There have been some two dozen topic on the forum which discuss the use of Audacity, a free download audio software package, to check dial speed and the break/make ratio.

This is one of my posts showing a really dirty dial's performance and the improvement after ultrasonic cleaning and oiling.

http://www.classicrotaryphones.com/forum/index.php?topic=11721.msg156265#msg156265

It's so easy to use and all you need are one shorter male/male phone plug cable, two alligator clip wire connectors and the dial (in or out of the phone).

Just put one 1/8" plug into the microphone input jack (have to use the more sensitive microphone audio input, the line in will not work) of a computer, connect the tip and one of the rings to the pulse contact terminals on the dial, if removed. If still in the phone with the phone closed, just attach the alligator clips to L1 and L2 (red and green line cord wires). Start the software, select microphone, start recording, turn the dial to zero, with a second or so and release the dial. Stop recording



and record the time of the first break and the 9th make. That would be for 9 pulses. Divide 9 by the total time recorded to get PPS. 9 pulses divided by .9 seconds is 10 PPS.

You can then measure the time for each break and make section of the 9 pulses and divide those my the total seconds of one pulse cycle to get the break/make ratio, which should be around 60%/40%.

Based on what you find out for the dial, you can then take appropriate action, if any is needed.

There are also other testers out these to get the same information. Not having one, my computer with Audacity works just fine.

Search the forum for Audacity and read those related to dial speed testing.

Ktownphoneco

John   ...   If that works for you, and your happy with that arrangement, by all means, keep using it.     I just have some issues with that method.    First of all, and correct me if I'm wrong, what you've described is actually placing a direct short across the microphone input circuit of either a sound card, or an on board sound circuit built into the motherboard (MB).    I'm no computer expert, but I rather doubt that sound cards, add-on or built into the motherboard, were designed to accept a direct short.    Microphones are made in such a fashion as to provide a specific impedance, which in most cases, is acceptable to the input audio circuit of the card or MB.   The pulse springs of a pulse dial make a direct electrical contact with each other.      Albeit not a lot, but there's still voltage and current involved.    I'd prefer to play it safe and not cause damage to my sound card or MB by making it do something that it was never designed to do.    I went to the Audacity web site, and it doesn't explain how the software can be used to test telephone dials.    Most instruments designed to test and calibrate telephone dials, were obviously made for the purpose.    The machines or instruments that do that, are also calibrated at the factory to do that job.    Who calibrates the sound card and Audacity software in a computer to take accurate measurements of the pulses being produced by a telephone dial ?
I sometimes get 8 to 12 dials at a time, from flea market sellers who want dials cleaned, lubricated and checked for correct operation before they sell the telephones.   Some worn dials can take 5 or 6 speed adjustments to get a consistent dial speed.    My dials are fastened into a jig, that allows me to turn the dial towards me to dial numbers, then turned 180 degrees to make the adjustments on the back, all the while still connected to my test circuit, which is made up of 302 components, minus the ringer, mounted on a test board.
It all works very well for me, and I know once a dial is calibrated, it's operating perfectly.

But if anyone prefers to use their computer, Audacity software, and wire a pair of pulse springs to the microphone input jack, that's fine with me.    I just have no desire to place what amounts to a telegraph key across the mic input jack on my computer and start tapping out a message.

I trust no one will mind if I keep using my Sage 930-A to calibrate dials.

Jeff

 

dsk

I do it the same way as john, and it is important to not add any other voltages to the microphone than what the pc delivers on the microphone jack. (exception may be a tone generator, it should not deliver a signal strength large enough to ruin the sound card.)

To be completely sure, a $5 usb soundcard will be an insurance for not ruin the pc sound card.  I use the pc microphone input, and has done this for years. 

This way to do it helped me also to identify my one dial with one longer pulse, witch I did not know ever had been made.

Do you actually need to calibrate?  Well, its wort trying if your dial does not dial out on your line.
Speed is essential, make break ratio will normaly not change during the years, and the "exchange" tolerates pretty much. eg foreign dials with other ratios.

dsk

twocvbloke

Quote from: Ktownphoneco on October 28, 2016, 07:30:57 PMFirst of all, and correct me if I'm wrong, what you've described is actually placing a direct short across the microphone input circuit of either a sound card, or an on board sound circuit built into the motherboard (MB).    I'm no computer expert, but I rather doubt that sound cards, add-on or built into the motherboard, were designed to accept a direct short.

Having been a computer repair tech for a number of years, I've yet to come across a blown microphone input that is the result of a shorted out connection, soundcards for the past 20-odd years have come with protection on inputs to prevent damage to their circuits, and of course, the pulsing is for at the most 1 second, and for the occasional dial test every other month or so, it's not doing any damage to our computers. When I first had the idea to do it, I was worried about causing damage too, but, it didn't, and repeated tests didn't cause damage either, so the conclusion is, it works and isn't going to kill a computer...

In fact, the only sound card damages & faults I ever had to deal with was from people physically damaging the sockets on their sound cards or motherboards (usually tripping over wires and snapping sockets off the board, or ramming a computer so far back into a desk's cubby hole that the plugs were crushed and snapped off in the socket)...

It's great that you have a professional tool to do the job, but as hobbyists, we can't justify spending money on such equipment, so, we devise other contemporary means to analyse dial speeds, and this one works nicely. If you study the screenshots above, you'll note sections where the waveform seems to fall off in a tapered manner, that's the protection doing its job, when a short is detected it shunts the power to a capacitor and then to ground, and in the process audacity records the make/break "clicks" and hey presto, we get a nice reading to look at and determine if a dial is within spec or not, all for the cost of a couple of wires and a free piece of software. Again this is a contemporary solution to repairing dials, it's not up to spec with the official repair houses that are long gone, granted, but, it is a hobbyist-level fix... :)

Though I do see that recommending a professional level tool to do a couple of tests here and there is like an electronics engineer saying you need a $2k oscilloscope to do a continuity check on a simple DIY circuit that has a broken connection, yeah, it works, but, it's an expensive solution to something that has far simpler hobbyist-level means to achieve the same result, such as a couple of wires, an LED and a battery to make a continuity tester, saving a cost of $1999.50 (if that!), it goes back to the old World War II slogan, "Make do and Mend"... :)

Oh, and Audacity is used for a lot of things that they don't mention on their website, they just provide the tool for the job, to record audio signals, the end user provides the functions, operations and inputs... :)

dsk

Jeff:
If I had a Sage 930-A to calibrate dials, I would have used it, but it will cost me pretty much to get one.  But just shipping and customs will be approx $200 to Norway, I feel I could spend that better.
I would end up with one in used condition to 2-300!

dsk

Ktownphoneco

As I understand it, the microphone circuit associated with a sound card, or the onboard sound circuitry, is designed to produce somewhere around 10mV to power the microphone inserted into the jack.    I suppose that isn't a lot of voltage to short out, but I would imagine it's important to remove the male jack as soon as the dial test is completed, rather than leaving it plugged into the mic jack while the graph is analyzed.    However without comparing the results of the same dial connected to an actual dial testing instrument, how do you establish that the results obtained through a PC, using the Audacity software are accurate ?     I can carve out a handle on a 2"x 4" piece of lumber, and use it as a baseball bat, but my accuracy when hitting the ball will be way out of whack.   As I see it, using a software program designed to analyze audio through a properly working microphone, isn't the same as asking it to measure the speed and duration of a series of short circuits.      If the results can't be gauged or compared against the results achieved through a proper piece of equipment designed and calibrated for the specific purpose of measuring the speed and break / make ratio of a pair of pulse springs, the accuracy of those readings can't be confirmed.
But if you feel it works, and your happy with the situation, then use it by all means.     But I'm still not convinced.

However, out of curiosity, I'll do this when I have a few minutes to spare.

I have an old computer in my shop that has either a sound card, or sound circuit built into the motherboard, ( I just can't remember which ), and I really don't care one way or the other, whether or not the test damages the sound circuit.

(1)  I'll download the Audacity software and install it on the old computer.
(2)  Make up a microphone jack with cord and alligator clips on the 2 conductors.
(3)  Select an "un-serviced" Western Electric dial in good shape, and which appears to be operating well.
(4)  Test it on the Sage 930-A, and record the results.
(5)  Then connect the same dial to my old computer, and record the results obtained using the PC and the Audacity software.
(6)  Then post the results of the comparison tests conducted using both methods.

Who knows, maybe I'll only manage to shoot myself in the foot.

By the way, Sage telecommunications test sets get obsolete fairly fast, and they can be purchased off eBay fairly cheap.    I bought mine for $95.00 plus shipping.    The thing one needs to know is if the 930-A set has the add-on card for the tests that the purchaser wants to perform.     Pulse speeds and break / make testing and DTMF frequency tests with the 930-A requires "Option Card - 01".    There are approximately 31 Option Cards for this unit.     The instrument itself, operates on either 48Volts DC, or 115 or 240 Volts AC, but not both.

Results to follow when I have time to do the comparison tests.

Jeff






unbeldi

The microphone input on PC sound cards is highly current limited, so I don't see any reason for concern in shorting it with a dial.  The small voltage that is present on the input is only there to power the field effect transistor that is present in most electret microphones to reduce the very high impedance of the microphone to the few kiloohms of the sound card input.

Audacity has nice application features for measuring timing, which are really native to its intended use as a sampling application.  In addition, sound card sampling rates are very high with modern PCs to generate better than CD quality recordings. I think Audacity has been used for many such measuring tasks, and there is no reason to expect that application to be explicitly outlined in the manual.

I would venture to say that the sampling of a PC sound card is far superior to what a Sage 930 does. Of course, such high accuracy is not even needed.

The manual way of recording the sound with the application and analyzing the trace is quite time consuming, I would say, not something I would want to do for routine work when adjusting dials, which may take a half dozen or more tweaks and measurements.  This may be ok if you do your ONE dial, and may be fun to do once a while as a hobby, but if you want to help others and tune their dials, something more expedient is needed. There is more to life as a collector than tuning dials.  Sage 930 units are dirt cheap these days, I got mine for $20 and again ca. $20 for shipping. It is easy to quickly run through all digits on the dial and display rate and break/make ratios.

I have also used an Ameritec AM7 Central Office Simulator, which writes a timing analysis to the serial port after dialing for both pulse and touch tone, and a complete frequency analysis for TT.
Mine has 20 FXS ports for use a programmable intercom, much like a little PBX, such as the Panasonic types.

I have also built my own dial measuring device, based on an Arduino chip and board for less then $5, but the rest of the cost is in writing the program to do the job.


PS: I see that Jeff posted in the meantime. Only thing I have to say is that the accuracy of a Sage for measuring dial speed is in no way superior in terms of accuracy to recording the sound trace. I don't understand the doubt of accuracy of the sound method, it is by nature of much higher accuracy even on slow, old PCs.

PS: CORRECTION:  I got my 930A for $9.99 plus $14 for shipping.

dsk

Quote from: Ktownphoneco on October 29, 2016, 10:12:44 AM
As I understand it, the microphone circuit associated with a sound card, or the onboard sound circuitry, is designed to produce somewhere around 10mV to power the microphone inserted into the jack.    I suppose that isn't a lot of voltage to short out, but I would imagine it's important to remove the male jack as soon as the dial test is completed, rather than leaving it plugged into the mic jack while the graph is analyzed.    However without comparing the results of the same dial connected to an actual dial testing instrument, how do you establish that the results obtained through a PC, using the Audacity software are accurate ?     I can carve out a handle on a 2"x 4" piece of lumber, and use it as a baseball bat, but my accuracy when hitting the ball will be way out of whack.   As I see it, using a software program designed to analyze audio through a properly working microphone, isn't the same as asking it to measure the speed and duration of a series of short circuits.      If the results can't be gauged or compared against the results achieved through a proper piece of equipment designed and calibrated for the specific purpose of measuring the speed and break / make ratio of a pair of pulse springs, the accuracy of those readings can't be confirmed.
But if you feel it works, and your happy with the situation, then use it by all means.     But I'm still not convinced.

However, out of curiosity, I'll do this when I have a few minutes to spare.

I have an old computer in my shop that has either a sound card, or sound circuit built into the motherboard, ( I just can't remember which ), and I really don't care one way or the other, whether or not the test damages the sound circuit.

(1)  I'll download the Audacity software and install it on the old computer.
(2)  Make up a microphone jack with cord and alligator clips on the 2 conductors.
(3)  Select an "un-serviced" Western Electric dial in good shape, and which appears to be operating well.
(4)  Test it on the Sage 930-A, and record the results.
(5)  Then connect the same dial to my old computer, and record the results obtained using the PC and the Audacity software.
(6)  Then post the results of the comparison tests conducted using both methods.

Who knows, maybe I'll only manage to shoot myself in the foot.

By the way, Sage telecommunications test sets get obsolete fairly fast, and they can be purchased off eBay fairly cheap.    I bought mine for $95.00 plus shipping.    The thing one needs to know is if the 930-A set has the add-on card for the tests that the purchaser wants to perform.     Pulse speeds and break / make testing and DTMF frequency tests with the 930-A requires "Option Card - 01".    There are approximately 31 Option Cards for this unit.     The instrument itself, operates on either 48Volts DC, or 115 or 240 Volts AC, but not both.

Results to follow when I have time to do the comparison tests.

Jeff
I guess you have one point here who can make this inaccurate, and that is the human reading of the data. 
I have often done the measurements just connecting the telephone directly to the mic jack without using time to disconnect the dial, and even that makes a good reading for me. 

When you use an analyzer you will probably get the same reading at least 8 of 10 times testing in series. 
If you please just make recording, e.g. as an MP3 or WMA file and put out a link here or email it or store it on sound-cloud etc we could try to analyze it on different computers and see how much our readings vary.  One recording of a zero with at least 1/2 second of silence in each end will do.  This could interesting! since you have the tester we could compare at the end.

(My interface is a 3.6 mm mono plug connected to an old-fashioned telephone jack)

dsk

dsk

You may try on these two, I have not analyzed them yet.

https://soundcloud.com/d_s_k-2/ericofon

https://soundcloud.com/d_s_k-2/starlite

dsk

Edit: forgot the links.

Dan/Panther

I dial the wheel all the way to '0', then release, and count 1001. It should make the entire travel in that time frame. One second or so. I find that is almost universally the case, with a clean dial. I think they are somewhat forgiving.
D/P

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

dsk

That well be good enough for the most. Only some equipment will need a more accurate tuning.
dsk

TelePlay

Here are two wave sound files (attached) in case anyone wants to analyze them. All I will say is that one tested slow and the other fast.

dsk

Both ways of transferring the files works, Ill send you my results as a PM, and you may make it offical when you feel you have the right feedback.

dsk