Using telephone testers.

[Greg G.]

Does anybody here use them, such as the Radio Shack 43-114 or the BK Precision 1045?  What is a reasonable amount to pay for either one?  Up until now I've been "testing" my phones through my PBX by simply plugging it in and dialing an extension, which works ok except that now I want to graduated to checking the calibration on dials.  The only way I've "tested" my dials is to simply give it a spin and see if it works.  If it seems to work ok, fine, if not, I've been shipping them off to Steve H.

Learning to lube them doesn't seem hard, but I'm not sure how the calibration is adjusted if needed.  This boils down to two questions.  Are dials difficult to adjust, and are telephone testers worth bothering with?  If a dial needs to be dug into for repair, that is NOT something I want to get into, I would still use Steve H's services for that.

[DavePEI]

Hi Briny:

They will only give a modified go/no go indication of how the dial is working. IN other words, as you dial, the digit dialed will show - it it doesn't it means the dial is out of the set calibration settings of the instrument. Hard to describe. It the dial or TT pad is outside of specs, the digit dialed will not display.

The specs are fairly broad, so it is not a direct indication of how the dial is set, only whether it is working and close to spec. See the chart below taken from the 1045 manual... If the BK can't decode the number, you can be reasonably certain the dial will not work as is on the PSTN.

Used prices - I have owned both, but I wouldn't think one should pay above $50 for either if you can help it. They can, however, be very useful.

Dave

[poplar1]

The 1045 is useful for checking ringers (both low voltage and normal), transmission, and reception, and it will show what numbers are dialed. However, the 8 to 20 pulses per second is quite a spread. While traditional central offices are tolerant of a wide range of speeds, VOIP is not. WE dials were tested as follows:

10 pulses per second dials:  8 pps- 11 (5H for example)  or 9.5-10.5 (6A for example)

20 pps:  16-20

C34.375
C34.376
028-300-501 i4

                       

[poplar1]

If a WE 5H dial tested between 8.0 and 11 pulses per second, then it passed. Whenever it fell outside of that range, then it was readjusted to between 9.5 and 10.5.

If a 6A dial tested between 9.5 and 10.5, it passed. Othewise, it was readjusted and tested again. It now had to be between 9.7 and 10.3 pps.

C34.178-i1
http://www.telephonecollectors.info/index.php/document-repository/doc_details/3341-c34-178-i1-stroboscope-dial-tester-tl

[Scotophor]

For dial testing of pulse rate and break percent, I used a method similar to the one(s) discussed in the topics linked below:

Adjusting WE500 dial speed...
Re: "Calibrating" Dials?

The advantage to me over buying a piece of test equipment is that it was free! I already had everything needed.

What I did was to build a simple 2-transistor audio oscillator (1 NPN transistor, 1 PNP transistor, 1 resistor, 1 capacitor, 1 small 8 ohm speaker, 9-volt battery; see image "Simple Audio Oscillator.png" below). Values I used for the circuit were: R marked 22K ohms {red-red-orange-gold} and C marked 0.05 uF  {503}, which gave a fundamental frequency of about 1750 Hz. (47K ohms {yellow-violet-orange-gold} and 0.022 uf {223} would give a similar tone) The transistors are not critical; any small-signal NPN and PNP will do. I used a 2N3904 for Q1 and 2N3906 for Q2. Then, to make the dial pulses stop and start the oscillator, I opened the circuit in series with the resistor and added a pair of alligator clips there for test leads (Dial Tester.png). Thus, clipping the leads together starts the oscillator and separating them stops it. Clipping the leads to the pulse terminals on the dial of a rotary phone (after removing any wires attached there) allows the oscillator's tone to be interrupted cleanly by the dial. You can also check the pulse rate roughly without disconnecting anything from the dial, or even by clipping to the phone's line cord tip and ring leads (with the phone "off hook"), but because of the inductance and capacitance built into the phone either of these latter two methods may not give "clean" enough pulses to properly check or adjust the break/make.

After building the oscillator and setting it to pulse with the phone dial, I simply placed the speaker near my computer's microphone and, after adjusting the input level, I recorded the output when dialing a "0". Then, I viewed the waveform using Adobe Audition -- but that's just one software solution which I happened to have, out of many that can do so. Another popular choice is Audacity; it's free and open-source. The dial "normal" and "make" periods contain the tone, while "break" periods are nearly silent (almost flat green line near the horizontal center of the attached images).

By adjusting how the software is zoomed into the waveform, deleting extraneous material and making an appropriate selection, we can check the pulse rate: from the start of the first break to the start of the last break (or end of the first break to the end of the last break) should be 0.9 seconds ±0.045 seconds (DialTest1.gif). Remember that a complete pulse cycle consists of both a "make" and a "break", but because we can't differentiate the "makes" at the start and end of the pulse train from when the dial is at rest, it's not really possible to tell exactly where the first cycle begins or the last cycle ends, so we measure 9 complete pulse cycles out of the 10 dialed. I managed to get my dial adjusted to 9 pulses in 0.902 seconds after about a half dozen tries (dividing the pulses by the time in seconds gives 9.98 p.p.s). See the upper right figure in the "Selection/View Controls" toolbar in the image. That "Length" figure is the duration in seconds of the selected area, shaded in gray.

We can also check the break and make percentages: by measuring from the start of one break to the start of the next break, or the end of one break to the end of the next break, we get the total length of one pulse cycle. Then by measuring just the length of the break and dividing the break length by the total cycle length (and multiplying by 100), we get the break percentage (DialTest2.gif and DialTest3.gif). The make percentage is then just 100% minus the break percentage. The pulse shown in these last two images is 62% break/38% make, though there is some variation between the different pulses in this recording.

[DavePEI]

So, basically we are all saying the same thing. The 1045 will only give a go/no-go indication of the dial.  However, it does a number of other tests of other phone functions.

For real accuracy checking dials, you need to go to the method so well described by Scotophor. Very good description. However, I have used both the BK and the RS units and never had a dial passed by them malfunction on the PSTN (likely as the PSTN here is quite forgiving). But use on VOIP isn't as forgiving, so if you intend VOIP use, it might be better to check out timing on a scope.

Still, they are very useful devices.

Dave

[Scotophor]

Are dials difficult to adjust, and are telephone testers worth bothering with?

Oh, BTW, I just re-read this topic and I noticed none of us directly answered those questions yet.

Regarding difficulty of adjustment, it depends on the dial. Some, like the 5H I used in my post above, are fairly easy. The governor has an adjuster held by a set screw. Turning the set screw is tricky if you don't have the special #260 tool to keep the governor from rotating. Also, this governor uses a spring-based one-way clutch, so when you pull the dial quickly, the clutch slips and the governor rotates backward slowly. Then when you release, the clutch grabs and the governor maintains the dial speed. This means that when you go to adjust the governor, you can push it around in one direction to get access to the screw, but if you push it the other way it will spring back. I managed to loosen the screw by using a small flat-blade screwdriver on the set screw, while holding a somewhat larger flat-blade screwdriver beside the armature and against the edge of the cutout in the governor housing to keep the armature from rotating. When tightening the screw I had to swap the screwdrivers to the other hands. I found that moving the adjuster little by little at first had no effect on the tests, then suddenly overshot what I needed, so I had to go back the other way, overshot again, then when going in small steps in the original direction again, I started dialing several zeroes after each adjustment attempt, in order to help the governor "settle" into its new configuration. On the second following test the dial speed was right where I wanted it.

Other governors require bending a spring to change the tension on the brake pads, which may be a bit trickier, but at least you won't have to juggle with two screwdrivers.

I didn't need to adjust the break/make, but it's fairly simple: if you don't know which contacts are the pulse contacts, observe the contact stack while dialing a "0". You should see that right as you start pulling the wheel, one set of contacts will open and another may close. Those are the rest/off-normal contacts, which prevent the dial pulses from making loud clicks in your ear as you dial. When you release the wheel, a completely different set of contacts that had been closed until now, should pulse open and closed rapidly. Now move the dial slowly until the pulse contacts are open, observing which contact spring blade moves. You want to bend the OTHER (fixed) blade of that pair (or perhaps a shorter, thicker "stop blade" in the contact stack, which sets where that fixed blade sits when the contacts are open) so that the opening gets wider to increase the break time, or narrower to decrease the break time. A contact spring tool will make this a little easier but you can get by without one.

So, I consider dial adjustment to be slightly to moderately difficult in most cases but nothing to be afraid of.

IMO, the telephone testers you mentioned are not worth bothering with for the purpose of rotary dial adjusting. They are barely a small step better than "compare it with a working dial by pulling both to zero and releasing at the same time". The problem with that method is, you don't know whether the "working" dial is perfect or barely adequate. Also, if the dials are very close to the same speed, it's difficult to tell which is faster. If you're comparing dials of different design, rotation times may not be equal for equal pulse rates (there may be different durations of "rest" before and after the pulses, while the dials spin). Comparing by eye to a "working" dial is even less reliable for adjusting break/make times, and the mentioned testers again are barely better if they only give a "go/no go" reading without telling you what specifically is wrong, in which direction and by how much.

If there's a cheap dial pulse tester out there that gives actual readings of rate and ratio, that's what I'd look for, but until then IMO the computer audio method is more than adequate.

[DavePEI]

Can I ask a dumb question?

Why do people use an AC audio signal and sound card scopes rather than using a standard triggered sweep CRT scope, feeding the DC Calibration voltage to the dial, and feeding it back to the scope to view the waveform.

Your waveform would then show as simpler square wave DC pulses on the scope and could be counted and timed same as with the sound card/computer scope.

No extra audio oscillator/sound card, computer/software required. Requires only 2 clip leads and a BNC/binding post converter for the scope input end. Couldn't get much simpler than that!

Many of us are Hams and would have access to a true scope. It doesn't need to be dual trace as my Hitachi is, and chances are if you don't have one, a friend will. A scope with one Mhz. bandwidth or less would perform this test.

Now, my idea of a Neo-Luddite is a person who wants to use the simplest technology which will do the job     If you have a scope, why not use it?

Dave

[Greg G.]

...

IMO, the telephone testers you mentioned are not worth bothering with for the purpose of rotary dial adjusting. They are barely a small step better than "compare it with a working dial by pulling both to zero and releasing at the same time". The problem with that method is, you don't know whether the "working" dial is perfect or barely adequate.
...

For my purposes, I don't want to get a dial "perfect", as long as it works, so from what you say, they're not really all that useful for testing dials.  But, there are other things they do that I thought might make trouble-shooting phones a little easier, such as the handset and cord testing functions.

[Scotophor]

Can I ask a dumb question? [...] If you have a scope, why not use it?

To everyone who has a working oscilloscope, yes, by all means use it for dial testing!

For me, while I technically do "have" a "scope" (of sorts...), it's considerably older and simpler than yours, and isn't in working condition (yet?). Even if it were working, it doesn't have any storage function, so would not be useful beyond seeing whether a phone is or is not producing pulses. So for me, not having a usable scope, budget to buy one, or space to keep one, is what led to me using the audio method.

[DavePEI]

To everyone who has a working oscilloscope, yes, by all means use it for dial testing!

It is funny - there are several threads on doing dial timing on the forum, and they all discuss doing it with the tone generator/sound card/computer method, yet none mentioned using a normal oscilloscope to do it. So, I just wanted to make sure people understood they could do it with a normal triggered sweep scope. Over the years, I have used my scopes for many things (I have two identical scopes, one here, and one over in the museum workshop), and just wanted to make sure people don't overlook the obvious simple method using a real scope. Sometimes we all tend to dream up complex solutions to relatively simple problems.

To me, three things can trigger the need to do dial timing tests; 1) If a dial fails on the BK tester and the cause isn't obvious, 2) if it fails afterwards in the intended application, or 3) if after lubrication and cleaning, it produces flaky results on PSTN or VOIP.

I think if a dial was very flaky, I would probably send it away, but have fixed up quite a few and have had good results.

The scopes work for their place in my test arsenal, as does my BK tester. I also had an RS tester, but gave it to another collector as a gift to his new workshop. I find (In order of the amount I use them), that my Fluke DVM is the most commonly used test instrument, followed closely by my BK,  and the scope is used somewhat less commonly. However, it can be invaluable.

One of the weirdest problems I solved with the scope was when I was building one of my switching displays. The 48 volt switching supply on one when in use provided sufficient amperage to do the switching, yet in operation, its voltage would drop to the point where operation became flaky. Upon examining the supply output with the scope, I found the inductive hash put on the regulator feeding back from the switches was enough to make it lose regulation. The solution, placement of a huge electrolytic capacitor on the output of the supply, now resulting in a rock solid output. Perhaps a large choke in line would have also solved this problem, but I had the cap and it solved it, so I went with it.

Another unconventional tool I use frequently on dials is an Ultrasonic cleaner to rid the dial of dried oil and dust. I have a small page on my server showing the process. You can run the dial through several cycles with the cleaner if necessary. A very filthy dial will need dis-assembly, but I find this works the majority of the time:

http://www.islandregister.com/phones/ultrasonic.html

Dave

[DoubleTone]

Another unconventional tool I use frequently on dials is an Ultrasonic cleaner to rid the dial of dried oil and dust. I have a small page on my server showing the process. You can run the dial through several cycles with the cleaner if necessary.

I'll second this recommendation but with one advisory.  Ultrasonic cleaners offer grunge cleaning you can't match with any other tactic, aside from complete disassembly, but they will remove paint from dials, not often on the first pass but eventually, yes.  That capability can jeopardize WE date and model markings, to cite one example.

[DavePEI]

I'll second this recommendation but with one advisory.  Ultrasonic cleaners offer grunge cleaning you can't match with any other tactic, aside from complete disassembly, but they will remove paint from dials, not often on the first pass but eventually, yes.  That capability can jeopardize WE date and model markings, to cite one example.

I have never had paint removed in the cleaner. However, that would largely be a function of the cleaning solution being used, and not of the machine - it the cleaning solution is too strong, it could remove paint. The "L&R #111 Ultrasonic Ammoniated Watch Cleaning Solution" seems to be fairly safe. I have used it on some 10 or 15 dials to this point. But as I say in the document always check in an unnoticeable place first.

Dave

[Greg G.]

I'm changing the basic question to just the usefulness of phone testers in general.  I've only seen the two types I mentioned in my original post, the B&K and Radio Shack models.  Currently, the B&K 1045 model seems to be easier to find but go for around $200 on ebay, give or take.  The RS model is not so common but are generally cheaper, but from what I've heard so far, also come with a warning that they can be a bit beat.  I also stumbled across this older thread when googling around, so I'll link it here so that anybody with the same question can view both threads:  http://www.classicrotaryphones.com/forum/index.php?topic=4386.0

The B&K model seems to be a better investment than the RS model, but I'm a little hesitant to shell out $150+ if I'm only going to do slightly better than my current method, which is to simply plug a phone into an extension on my 308 PBX and dial it.  As most of y'all know, I have no technical training in electrical trouble-shooting and the details involved make my eyes glaze over. 

[DavePEI]

If you watch, you will find the BK unit for around $50 at times. One sold within the past week for $50. Both the RS and BK testers are roughly as utilitarian as the other, performing the same tests, but the BK is a much smaller unit and take up less room. I had an RS tester for a number of years, and only gave it away when I found an affordable BK 1045.

Dave