Pulse dialing is too fast

[Timusius]

I have created a home PBX system intended for the kids to play around with. It works just fine with newer DTMF phones.
But the idea started because I have a rotary phone that I would like to play around with. I have some ATAs, (Grandstream HT802 and a HT818v2) that both support pulse dialing... and it works... sort of.

At first I thought the phone was faulty, because all it could dial was numbers like 1111 112221 etc. no matter what number was released on the dial. (I can see the numbers in the log in the PBX telephone system.)
So I bought a couple of extra phones, just like the one I had already.
These however turned out to do the exact same thing.

Because one of the phones was a bit broken, I discovered by chance, that when the rotary dial was running a bit slower... everything worked just fine!

So I did a support case with Grandstream, who were first very interested in the problem. They had me run a lot of tests, and send them debug logs from the devices, and record the pulse patterns and let them analyse them.

However in the end they were not really helpful. It seems like the phones all run too fast. Something like 15 pulses per second.And when they worked they were outputting about 12 pulse per second.

I was hoping that this could be adjusted by some some setting in the ATA, but in the end Grandsteam was not interesting in helping in this matter... Fully understandable since it's just a hobby project, and the phones are probably out of spec. (At least compared to what I can read on the Wikipedia page on pulse dialing.)

I am guessing that these 3 phones, all Danish Kirk 73D, probably worked all their early life just fine with the 15 Pulses Per Second, and perhaps in the Danish telephone system that was within specs.


Is there some ATA that is good at supporting "out of spec" pulse dialing?

Or, ss there a good way to change the phone to slow down the dial, so the phones just work instead of requiring special handling?

I've got a picture of the mechanism here. The black wheel is the one that spins and generates the pulses.

I would love to have these old phones working, so the kids can call each other using a good old rotary dial.

[HarrySmith]

Hi and welcome to the forum. Most rotary dials can be adjusted. I am not familiar with any European phones so I cannot offer advice on how to adjust these but I am sure one of our European members will chime in.

[TelePlay]

I have never worked on this type of dial but every dial I have worked on has had a governor to control the dial speed.

10 PPS is the most common speed, 20 PPS was used on switchboards so operators could work more efficiently (but the step system behind the switchboard was set to recognize 20 PPS.

I am not familiar with historical European phone systems but Grandstream seems to be a 10 PPS device.

Your dial seems to have its governor buried beneath gears and brackets, hard to get to for adjustment. Would require taking the dial apart, adjusting the governor, assembling the dial and testing its "new" speed - with repeating those steps if the dial speed is still not 9 to 11 PPS. Not an easy way to service dial speed.

Maybe someone from the area has experience with these "faster" phones, why they were made to be so and if they can slowed down.

Since the image provided does not clearly show the governor, it's impossible to see how it would be adjusted. It seems to have coiled springs acting on the governor brake shoes (wings).

All rotary dials have a mainspring, gears, a speed governor and leaf switches. How that all goes together depends on the manufacturer's propriety, patented design.

[5415551212]

I would verify the dial speeds with another device such as an oscilloscope or just using a computers sound card as discussed here:
  https://www.classicrotaryphones.com/forum/index.php?topic=17011.0

[dsk]

1 sec from you release a 0 to the dial stops is slightly faster than 10 pps. 9 is slightly slower. Probably OK.  Try one, and use it a guide when it works.

[Timusius]

I would verify the dial speeds with another device such as an oscilloscope or just using a computers sound card

This is what I did in cooperation with Grandstream.

They returned this image. (Which shows two recordings from two different phones. A red one and a green one.)
They probably forgot why I mailed them in the first place, and basically told me that the green phone seemed to produce good consistent pulse signals.
So I asked.. "OK, so why doesn't it work with your device?" they quickly shut down the conversation :-D

I looked at the image afterwards and it does seem that the phones are a bit on the quick side on the pulses.

[5415551212]

If you have the audio recordings there are some threads on here about analyzing the sound files with a audacity and a spreadsheet, here is a recent thread discussing the different methods:
https://www.classicrotaryphones.com/forum/index.php?topic=21463.0
https://docs.google.com/document/d/1j-s3V-ZXQO-VSwhKnBU_NT2u9L6Z_VDga-1PRuxdPwI/edit?tab=t.0#heading=h.dlafkaud40vv
I don't any experience with that method, but it seems the majority of folks on here have had great success with it.

[TelePlay]

I have no idea what the circles and arrows on that image are referencing.

The only important part of the wave form in the image is the speed at which the pulses take place.

First image below is the wave form of a working WE dial from the point of lifting the handset, going off hook, dialing "0" 3 times and then hanging up the phone. This dial is operating at 9.44 PPS.

The second image is from you image trying to determine the dial speed. The image only shows 6 complete pulses that occurred over 0.41 seconds which normalizes out to 14.29 PPS, too fast.

The final image is determination of the WE dial speed in the first image which shows it working at 9.44 PPS with a 62/38 Break/Make ratio.


The circles and arrows on your image are pointing out line noise (that can be created by just holding the handset - it is an audio file after all), static noise from the handset cord or any other similar source (red arrows) and the circles covering spaces of dead audio that seems to be from the spike of going off hook to the moment the dial starts being rotated.

Unless your, a, Grandstream is susceptible to line noise (the unit reacts to any line noise), I see nothing of use on that image except for the pulses themselves. The size of the circles has nothing to do with the dial pulsing speed.

Am I wrong? I've been doing this for some time and don't understand what Grandstream is telling you.

[HarrySmith]

Circles & arrows and a paragraph on the back? 

[dsk]

Take a look at these 2 threads.
https://www.classicrotaryphones.com/forum/index.php?topic=8792.msg94483#msg94483
https://www.classicrotaryphones.com/forum/index.php?topic=21454.25

Try to measure depending of what you have available. 

[Timusius]

Am I wrong? I've been doing this for some time and don't understand what Grandstream is telling you.

Well.. perhaps they sort of got lost along the way. I guess different supporters answered the 5-6 mail long case.

I wrote them because I noticed that the dialing worked fine when it was slower.
They asked me to provide a bunch of debug information from the Analog Telephone Adapter. That gave no new information.
Then they asked me to do a video, I guess to understand what was doing when I said "when I make the dial go slower, it works".
Then they asked me to do the sound card recordings. In the mean time I had gotten hold of 2 more phones of the same type. So I did recordings of them all. A brown, a red and a green one.

The brown one was broken, and based on the recordings they said that the pulses were unregular and that there was a lot of noise in the signal.

The images above are from the red and green phone.

• The red arrows indicates some disruptions in the signal.
• Slower dialing apparently lead to fewer disruptions, indicated by the yellow circle.
• Finally the green circles indicate that the signal was good enough.

But they failed to explain to me that the PPS was too fast. (I did guess myself after reading a bit on these forums.)
So thank you for clearing that up, and providing an exact number 14.29  to confirm my guess.

All 3 phones seems to agree that about 15PPS is normal, so I guess in the weekend I am going to look into if I can slow one of them down, just for the fun of it.

Because the ATA does have all kinds of weird settings for different systems around the world, I was hoping that Grandsteam would simply reply with "Ohh it's a 15pps, so you need to set this setting to XYZ"

But apparently not for pulse dialing.. which is sort of understandable.
The devices are built to support companies that has a lot of real phone calls going on, where no one would be using a dial phone.
They are not spending time building features for the occasional hobbyist that just likes the sound of the old dial and ringer.

[TelePlay]

I got on my desktop so I could create a scale to determine the timing of the pulses.

I selected one series from each of the 4 phones. I created a 1 second scale so I could determine pulse times. This is what I got, for what it's worth.

The red slow seems to be right in spec at 10.17 PPS.

[5415551212]

I was hoping that Grandsteam would simply reply with "Ohh it's a 15pps, so you need to set this setting to XYZ"

I think 10pps and 20pps were standards but not sure if 15pps was?
And I am not sure what the standard tolerance range is, is it +/- 10% ?
It would be a nice feature if a ATA could support a 20pps phone.
I think the folks on here can help you adjust your dial(s) to get them tuned to 10pps.

[TelePlay]

US standard set by WE I believe has always been 8-11 PPS. Break/Make is 60%/40% +/- 5%. When adjusting dial, I try to get between 9.5 to 10.3 PPS.

Also, I seem to remember that central office equipment and PBX systems that accepted pulse dialing were more forgiving, had more dial speed tolerance) than ATA's which have a rather small tolerance around 10 PPS.

20 PPS was only used on switchboards, allowed operators to dial faster, and the equipment used was set to accept that faster speed (IIRC they were WE 6E dials and that dial had a curved finger stop - easy to spot on eBay listings).

I've found AE dials have a Break/Make ratio of 65%/35% and many of their systems used frequency ringing, if that means anything. Same for SC.

I have no idea what any foreign standards are/were.

[countryman]

In Germany, Break/Make was defined as 1.6 : 1, what figures out to 62 ms Break, 38 ms Make. Practically the same, and practically any global standards used were within the tolerances of each other. Even with "exotic" dial layouts where the indicated numbers do not directly correspond with the pulses (Swedish or New Zealand dial).