Voiceless Pulse Dialer

[MIPS]

Pardon me if I'm a little iffy my knowledge here. I'm generally an old computer guy and normally my only real knowledge of a telephone and line is how you plug a modem into it. 

Anyways, I'm working on a project that has me building a semi-automatic modem interface for an old model 33 Teletype machine. It's semi-automatic in that control of the Hayes Smartmodem is done by pushbuttons but dialing numbers is done with a traditional rotary dial instead of keying in AT D 1-800-BBS-NMBR.

For all my knowledge so far I've been relying on Radio Shack's old book on Telephone Electronics. Good read, but it seems to talk more about DTMF and the exchange more than the inside of a regular phone but what I can extrapolate is that if you take a traditional rotary dial telephone and remove all the voice electronics, you get something like this:



I've been able to build this using an M1A bell, a regular Western Electric rotary dial and two 5V relays paired into what can be used as a hook switch.




Basically the relays are operated by the OH lamp on the modem, allowing you to dial your number when the modem itself is off the hook and waiting for a carrier signal. The bell simply notifies you when someone is calling the teletype so you can turn it on and answer. My problem is that I cannot find the values of that one resistor and capacitor (aside from I assume it's bipolar AC) which from what I read are required both for "anti-tinkle" and spark quenching and every telephone I have peeked inside so far has yielded me nothing I can take note from. I've been asking around for months. Even over at videokarma and badcaps.net. Nobody seems to have even the slightest idea. You guys are my last real hope on this as this is the last part of the entire machine that isn't finished. Any tips?

[unbeldi]

As you stated, the RC network across the dial pulse switch serves to quench arcing and suppress radio frequency generation from the pulses. In your circuit it indeed would probably prevent bell tinkle as the ringer stays connected.

The choice of values for the resistor and capacitor can be calculated from the requirements of the circuit.  Basically this is intended to be a high-pass filter, preferentially shorting RF across the switch.  This leaves low frequencies, including the dial pulses, mostly unaffected, but strongly attenuates high frequencies above the filter cut-off.  The useful range for telephony is up to 3500 Hz, but I would dimension the filter for twice that.

The Western Electric 500 telephone has a similar filter in the 425B network and uses a 180 Ω resistor and a 0.1 µF capacitor.  IIRC, I calculated that this produces a ca. 9 kHz cutoff.

I think it's safe to chose similar values.   Automatic Electric used a 100 ohm resistor in the No. 40 Monophone, but I don't recall if the capacitor was specified in the diagrams or catalogs.


PS:  After rereading, I am a little confused by the hook switch connections in your diagram.  Is the resistor disconnected when the set is taken off-hook?  I believe it should not. It should be disconnected when on-hook.

In you diagram, the capacitor fulfills also the function to block DC from the ringer.  For this, the capacitor should probably be a little larger than I suggested, ~0.4 or 0.5 µF, if the ringing frequency is low (20 Hz).  So the resistor should be adjusted accordingly.

[Phonesrfun]

The ringer capacitor is standard at .5 (.47) mF 250 v. These days there is really no need for RF noise filtering, so there is no need for the resistor, is there?

[unbeldi]

The ringer capacitor is standard at .5 (.47) mF 250 v. These days there is really no need for RF noise filtering, so there is no need for the resistor, is there?

Well, this all depends. Certainly not many people experience radio interference, as very few use an old fashioned radio, but it is still a  good design precaution, to prevent RF getting into other old circuits connected to the teletype, which could be effected.   It also prolongs the life of the dial pulse contacts, as it prevents arcing.

[MIPS]

PS:  After rereading, I am a little confused by the hook switch connections in your diagram.  Is the resistor disconnected when the set is taken off-hook?  I believe it should not. It should be disconnected when on-hook.

Looking through my only book on the subject, that is a no, mind you the diagram was a little different.

[unbeldi]

PS:  After rereading, I am a little confused by the hook switch connections in your diagram.  Is the resistor disconnected when the set is taken off-hook?  I believe it should not. It should be disconnected when on-hook.

Looking through my only book on the subject, that is a no, mind you the diagram was a little different.

This diagram shows it correctly.  It is shown in the ON-HOOK (handset not lifted) state when the contacts S1, S2 and S4 are open. This is when the set can ring.  When the handset is taken OFF-HOOK, the contacts close, connecting the speech and dialing circuit to the line. It also connects the resistor via S4 to the capacitor, but it won't be used at all until the dial is rotated from its normal position. When off-normal, switch S5 completes the filter across S3 (dial pulse switch) and S6 shorts the entire speech circuit so that the dial pulses have maximum amplitude and are not distorted; this also prevents the dial pulses from being heard in the speech circuit.  On the return rotation of the dial, S3 interrupts the circuit according to the number selected.  The ringer is also shunted during dialing by the resistor. Since the resistor has very low resistance, the ringer never receives enough energy to move during pulses.

PS: Your diagram appears to be from Bigelow et al., Understanding Telephone Electronics.  The copy that I have, fourth edition, lists the value of the resistor as 340 Ω in the diagram.  Using a standard value, 360 Ω, should be fine if your ringing capacitor is a standard value of 470 nF.

[MIPS]

Huh. I wonder why mine lacks that resistor value.

Okay, so 360 ohm and 470nf. I have neither on hand so I'll make not to place an order next time I travel to the coast and get back to you all.

[unbeldi]

Huh. I wonder why mine lacks that resistor value.

Okay, so 360 ohm and 470nf. I have neither on hand so I'll make not to place an order next time I travel to the coast and get back to you all.

330 Ω is another standard value.

PS:  there is of course quite a bit of ambiguity in that circuit, so that I am not sure whether it is even wise to show a specific value for that resistor.  The frequency response of the circuit depends greatly on the type of ringer installed.  Both the DC resistance and the AC inductance of the ringer coil play into that equation.  This circuit, sharing a capacitor between the ringer and filter has been fairly obsolete since the 500-type telephone became the de-facto standard in the US. I think it may be more common elsewhere.

[unbeldi]

Here is a circuit label of a Japanese DO-8 decorative 'antique' telephone, which was sold in large quantities in the US and around the world.

It uses a 0.9 µF ringing capacitor that is shared and a 100 Ω resistor.

[MIPS]

Well the components came in and the dialer was finished.
Bad news is that it doesn't function properly.

With the modem and the dialer connected I take both off hook and try to dial. A dial tone is heard and when a number is dialed you get the proper pulse dial clicking sound but the dial tone never stops. Just in case it was some weird issue where the modem was not liking a second phone off hook at the same time I grabbed a second phone, unplugged the modem and listened through the handset as I dialed. Same thing.
For an additional sanity check because some areas are dropping pulse dialing completely I disconnected everything but my spare rotary phone and called my cell. It went through.

I'm still no expert in this but my only suspicion is that the pulsing isn't strong enough for the exchange to detect it, correct?

Second, the hook switch SORT of works. The bell rings when the line does but when the relays take it off hook the exchange does not register that it has done so and continues to try and ring.

[poplar1]

Not sure that this applies here, but here goes: You can't ever dial out with a rotary dial if another phone is off-hook on the same line.

[MIPS]

What?? Oh god that would be a massive problem. I swear I thought about and checked for that exact same possible issue a long time ago when I first started designing the circuit.

Hmm, so I double checked DTMF dialing and yes, that will still work so I guess in theory we could kludge in one of those Pulse to DTMF adapters, thing is, what are you people paying for those? I'm quickly looking around and the prices are all over the place from $20 - $100. I also can't modify this circuit to replace the rotary dial with a proper keypad. The CCU coverplate on my particular model only fits four buttons and one rotary dial.



Anyways, to confirm we could dial standalone (nothing but the dialer off hook) and verify that the dialer itself was otherwise working I tried calling my phone and no, the call did not go through so even though the components should all be there we still have neither a functioning dialer or hook switch.

Edited: for reference, here's the dialer in its current form.

[MIPS]

Okay, so several months later......

As you've stated, you cannot pulse dial when other phone devices are off hook. It simply will not work. DTMF on the other hand uses tones to dial and is immune to other devices off hook, so I redesigned the circuit on a breadboard and eliminated everything but the hook switch, a Dialor 1.2 (mentioned here) and the rotary dial. I'll add back in the bell once I get this working but for now I wanted a proof of concept so there's nothing else attached for simplicity.



Anyways, it doesn't work either. Standalone I can blindly dial my cell and it will dial out but when the modem or any other telephone in the house goes off hook at the same time it emits no tones and therefore cannot dial. I'm not entirely sure how the adapter works internally but given how you seem to have to install it, wouldn't it emit tones regardless of anything else on the line? Input polarity does not seem to affect the adapter.

[luns]

I think that with the modem off hook, the line voltage is too low for the dialor to work.

I would go back to trying to get the rotary dialing working.

First step is to get the actual dialing going. This is really simple to start; just put the DP (dial pulse - S3 in your anti-tinkle figure) contacts in SERIES with the Hayes modem. When the modem is off hook, the dial should be able to dial.

This may already be enough for you to call the job finished, but there's two things which you may or not care about. The more important of them is that the dial pulse contacts have no arc suppression on them this way, so every time they open, there can be a little spark across the contacts. This will cause RF emissions (which you may be able to hear on an AM radio, but will interfere with other stuff too) which you might not care about. The other thing though is the spark can also erode the contacts. It can also split CO2 in the air and leave soot covering them. Between these two, you may eventually end up with contacts that don't make electrical contact.

To deal with the arcing, you just need an RC snubber across the points. The later phone networks (rotary versions of 425 onwards) have a snubber between the F and RR terminals, so you could just connect to those points if you have one available. Or you could just make your own snubber. ITT's circuit diagram for their variant of the 425 gives values of 180Ohm for the resistor in series with a .10-.14 mfd capacitor, and this pair goes in parallel with the DP contacts.

Another issue would be that with only the DP switch in place, you will hear the dial pulses at the modem. As I remember these modems, you would hear the pulses when the modem dials for itself anyway, so I don't think this is really an issue. You may or may not care. If you prefer quieter dialing, you can shunt the modem with the ON (off normal) contacts from the dial.

One more issue might be the tinkling of the bell. The snubber ought to take care of it, so I'll hold off making suggestions for that unless it proves to be a problem. The ringer and its capacitor should be considered their own unit independent of everything else, and just hang off the incoming line. I'll bet they'll do just fine like that.

Notice that I made no mention of the hook switch. The modem takes care of going on/off hook for itself, so the relays you were trying to include for this are redundant.

[MIPS]

I think that with the modem off hook, the line voltage is too low for the dialor to work.

Voltage to both IC's sits at 2.5v regardless of the modem or any other device being on or off hook. That seems low but the datasheets indicate that I'm fine so long as I don't go under 2.2v.

I would go back to trying to get the rotary dialing working.

How? Just a few posts above I was told that you cannot pulse dial with other devices off hook.

Another issue would be that with only the DP switch in place, you will hear the dial pulses at the modem. As I remember these modems, you would hear the pulses when the modem dials for itself anyway, so I don't think this is really an issue. You may or may not care. If you prefer quieter dialing, you can shunt the modem with the ON (off normal) contacts from the dial.

No issue at all if it can be heard. If anything that's how I was verifying that the exchange was detecting my dial pulses (which it wasn't)