Author Topic: Converting a Kellogg Subset to a WE 634A Equivalent  (Read 5296 times)

Offline oyang

  • ***
  • Posts: 102
Converting a Kellogg Subset to a WE 634A Equivalent
« on: November 07, 2015, 06:58:22 PM »
In case someone finds this useful....

I didn't want to dismantle another phone to use as a subset, and early WE subsets for rotary candlesticks and D1 handsets etc seem scarce and expensive. So, I got a couple of Kellogg subsets on eBay cheaply and converted them to 634A equivalents.

I'm attaching 3 pictures: the first is the wiring schematic of the 634A, the second is the original configuration of the terminals and soldered connections in the Kellogg subsets, and the third is a description of the modifications I made to make it equivalent to the 634A (anti-sidetone). It was a matter of using the existing connected induction coil, renaming some terminals so there were enough connectors, and adding another capacitor.

It works! I could dial out and receive calls using a 151AL hooked up to this setup.  I hope someone finds this useful too.

Otto
"In theory, theory and practice are the same. In practice, they aren't."

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #1 on: November 08, 2015, 08:55:37 AM »
What is the number on your induction coil?  Usually Kellogg stamped huge yellow numbers onto the wrappers.

Is it a 103A triad coil?
I have used a 103A similarly, and will check yours against my circuit.   But to use a triad coil correctly in a WECo circuit, one actually has to cut a wire coming out the coil and reconnect the windings in a different sequence.  Yet, indeed it can be made to work for laymen's purposes.  But you cannot call it a 634A subset anymore.
« Last Edit: November 08, 2015, 09:02:14 AM by unbeldi »

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #2 on: November 08, 2015, 09:23:14 AM »
I would think a good Kellogg subset is much rarer and harder to find than a WECo 634A, which can be found every week.
I could use a couple, for sure.

The only reason for my "conversion" was because the 103A coil was mounted inside a switchboard and I wanted to use a WECo 211 hangup handset mounting on the side of the PBX, and thus repurposed the induction coil.

Offline oyang

  • ***
  • Posts: 102
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #3 on: November 08, 2015, 05:46:22 PM »
The coil has 4 connections to the same terminals as the one in the WE subset, and is marked 99-A.

I've recently gotten 3 Kellogg subsets at about $15 each.... So I figured they are common and available.  Are they actually rarer than the 634A?
"In theory, theory and practice are the same. In practice, they aren't."

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #4 on: November 08, 2015, 05:58:26 PM »
The coil has 4 connections to the same terminals as the one in the WE subset, and is marked 99-A.

I've recently gotten 3 Kellogg subsets at about $15 each.... So I figured they are common and available.  Are they actually rarer than the 634A?


Aha!   The No. 99-A induction coil is actually the completely wrong induction coil to make a 634A-class subset.
It is only a local battery induction coil, i.e. not even a common battery coil, and not an anti-sidetone type like that in the 634A subset.
It has only two windings instead of the three necessary for anti-sidetone operation.

I still haven't had a chance to go over your hook-up, but will do so.

Offline oyang

  • ***
  • Posts: 102
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #5 on: November 08, 2015, 10:37:09 PM »
Thanks for the review.

Maybe this is why the box doesn't ring for an incoming call?  The striker does vibrate, but not hit the bells.  I thought this was just a matter of adjustment to be made, but could it be because the coil isn't right?

Showing my inexperience, let me ask: does the induction coil play any role in sound transmission?  I thought it was purely a converter of DC to AC, to allow the bells to ring because the AC current then alternates creating magnetic impulses on each side of the striker.  Speaking of which, can one buy an appropriate induction coil from a modern source like Radio Shack? 

With the setup I showed, the 151AL dialed out and received calls fine, with both transmission and reception of sound that seemed good.

Thanks,

Otto
"In theory, theory and practice are the same. In practice, they aren't."

Offline G-Man

  • Hero Member
  • *****
  • Posts: 1957
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #6 on: November 09, 2015, 12:51:17 AM »
Induction coils are used in the speech transmission circuit and they do not have anything to do with the ringing circuit. The TCI Library is an excellent source of instructional material regarding basic telephony.
 
You are not going to find a modern off-the-shelf source for a telephone induction coil, though some may attempt to hack an audio transformer. An inexpensive subset may be created out of an old 302 or 500-set base. In fact, Ray Kotke manufactures a cover that fits over a 302 base, making it look like a vintage subset.
 
Also, Automatic Electric mini-networks are prized by collectors since they often can be fitted inside of a telephone requiring a subset, though using them externally seems to be a waste of resources.

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #7 on: November 09, 2015, 10:12:23 AM »
Thanks for the review.

Maybe this is why the box doesn't ring for an incoming call?  The striker does vibrate, but not hit the bells.  I thought this was just a matter of adjustment to be made, but could it be because the coil isn't right?

Showing my inexperience, let me ask: does the induction coil play any role in sound transmission?  I thought it was purely a converter of DC to AC, to allow the bells to ring because the AC current then alternates creating magnetic impulses on each side of the striker.  Speaking of which, can one buy an appropriate induction coil from a modern source like Radio Shack? 

With the setup I showed, the 151AL dialed out and received calls fine, with both transmission and reception of sound that seemed good.

Thanks,

Otto

We'll find out eventually what the problem with ringing is.   The good news you are reporting is that the striker does vibrate.  Since it doesn't strike the gongs, I suspect that it is located in the center between the gongs, and simply does not have enough amplitude to make a sound.  This leads me to suspect that it in fact is what is called a frequency-selective ringer. These have the striker mounted on a stiff reed spring in the center between the gongs.  The reed spring in combination with the weight of the clapper ball determines the principle frequency of vibration,  the resonance frequency.  The frequency of the AC current for ringing must match this mechanical frequency or the striker will not operate at full efficiency. Today's ringing uses 20 Hz frequency, which was historically called straight-line ringing.  The frequency ringers used a variety of frequencies between 16 Hz and 67 Hz.  Look on the case of the subset or even on the ringer itself for any numbers in that range, most likely on the lower side.  Perhaps 16 2/3, 25, 30, 33 1/3, 42.

As you know, a telephone line only has two wires, that serve for electronic speech transmission and reception at the same time, two-way traffic.  A part of the art/science of telephony is to efficiently connect the microphone (transmitter) and the ear-piece (receiver) to the telephone line.  Together, the receiver and transmitter have four wires, two each, one pair for reception and one for transmission. The induction coil performs the electronic function of converting the four-wires to the two wires of the line, by combining the reception and transmission paths, or when viewed in the other direction, by splitting the audio paths for transmission and reception.  It does this in an efficient manner so that the least amount of energy that the transmitter generates is wasted in the receiver, and that the least amount of energy that is received from the distant telephone partner is wasted in the local transmitter.  At the same time the induction coil also amplifies the signals produced by the transmitter through its action as a transformer.

I will let you know when I get chance to look at your wiring in detail.
« Last Edit: November 09, 2015, 10:16:32 AM by unbeldi »

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #8 on: November 09, 2015, 04:00:21 PM »
In your pictures of your subsets you are showing us two different types, yet, you labeled them as "original" and "modified final".   I am assuming you're simply using the first one to show the original state of the induction coil terminals, as you mentioned that you had multiple sets.

Picture "Slide2.jpg" shows a "normal" straight-line ringer, while the ringer in "Slide3.jpg" is the type of frequency ringer that I described earlier. This one must be the one that you described as vibrating only without striking the gongs. As can be seen, the clapper sits right in the middle between the gongs. In the other type, the clapper is pulled to one side by the bias spring.

You should have made the modifications to the other set, the bells might just have rung properly.

I don't seem to have the diagram for the 21068, I think this is the drawing number of the circuit.  Are there printed labels with the circuit inside the cover or anywhere?

« Last Edit: November 09, 2015, 04:21:35 PM by unbeldi »

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #9 on: November 09, 2015, 04:47:44 PM »
Considering your admission of little experience, you made an excellent choice for connecting your desk stand to coil and ringer.  Good work, essentially.

What you couldn't know probably is that the black wire from the desk stand isn't needed when connecting the instrument to a sidetone subset.  Its only use is for bypassing a set of switches in the instrument to provide the anti-sidetone compensation.   As your subset cannot provide it, you simply don't need to connect the black wire at all.  So, insulate the spade and store it.  After doing so, you will have your BK terminal point (the second side of the 2 F capacitor) not connected to anything.  Connect it to your L2Y terminal, where the yellow mounting cord wire is already connected. Of course, you can just install a wire bridge from BK to L2Y.

What I can't tell for sure is whether you have used the two windings in the induction coil "properly".  You can test and exchange L1--R  with C--GN, perhaps one or the other configuration gives better audio performance.  Perhaps you already discovered this in your experiments.

Your ringer is properly connected.   I already stated the reason why it doesn't make a sound.  Transfer your wiring to the other subset with the straight-line ringer and it might just work.
« Last Edit: November 10, 2015, 09:49:37 AM by unbeldi »

Offline oyang

  • ***
  • Posts: 102
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #10 on: November 09, 2015, 11:56:41 PM »
Thanks for these last 2 detailed posts!

You are correct about the two different ringers.  The "slide 3" is the final tested rewired subset; "slide 2" is a different subset that was identically laid out and so I used that picture for convenience.  I had not noticed that the ringer types were different, because everything else was the same.  I was planning on rewiring the "slide 2" subset the same way, so sounds like it should work as intended.

The sound quality was quite good, so I haven't tried switching the terminals of the induction coil.

Question: Once I have added the second capacitor and there are now a 1F and 2F wired the same as the 634A subset, doesn't this now become an anti-sidetone configuration that can utilize the black wire from the desk stand? Or is the problem that the induction coil doesn't work for anti-sidetone?

You guys were right that this is addictive.... It's very zen to sit quietly studying a wiring diagram, stripping wires, crimping, testing my connections with an ohm meter, ending up with a physical object that functions.  This contrasts with my job sitting in an office, staring at a computer screen, writing papers or grant applications, getting interrupted constantly by my students and lab staff to figure out why their experiments failed, going to meetings that accomplish almost nothing: the academic rat race!

I think the take-home message from this thread so far is "a little knowledge is a dangerous thing."  I've learned just enough about telephones to mess around and think I've accomplished things I haven't (i.e. making a 634A-equivalent subset).
"In theory, theory and practice are the same. In practice, they aren't."

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #11 on: November 10, 2015, 09:32:03 AM »
Thanks for these last 2 detailed posts!
You're welcome!

Quote
You are correct about the two different ringers.  The "slide 3" is the final tested rewired subset; "slide 2" is a different subset that was identically laid out and so I used that picture for convenience.  I had not noticed that the ringer types were different, because everything else was the same.  I was planning on rewiring the "slide 2" subset the same way, so sounds like it should work as intended.

The sound quality was quite good, so I haven't tried switching the terminals of the induction coil.
I found that the old catalogs provide conflicting information about the 99A induction coil. In 1928 it was reported as a common battery coil while 10 years or so later it was described as a local battery coil.  Since I don't have one, I don't think, I cannot confirm one or the other. Since the quality you obtained is good, I suspect it is in fact a common battery coil.  Only a resistance measurement of the windings can confirm that for sure.

Quote
Question: Once I have added the second capacitor and there are now a 1F and 2F wired the same as the 634A subset, doesn't this now become an anti-sidetone configuration that can utilize the black wire from the desk stand? Or is the problem that the induction coil doesn't work for anti-sidetone?
The anti-sidetone compensation requires a third winding in the induction coil. In the most common implementation, this third winding is wound or connected in a way so that the induced signal at its exit wires has the opposite phase relationship to the signal generated by the transmitter, so that when it is connected straight across the receiver it cancels the transmitter signal to high degree. Thus, no or very little energy is wasted in the receiver coming from the transmitter and more can be transmitted to the other party over the telephone line. This is the principle of anti-sidetone compensation.  The capacitor has to be inserted into the feedback loop to prevent DC from entering the receiver. The value 2 F is just large enough to not cause excessive signal damping. In theory, an infinite capacitor would be ideal, but until the 1930s capacitor technology was poor and a 2 F capacitor would be larger than a pack of cigarettes.  Automatic Electric used 4 and 5 F capacitors for the same purpose in the Monophones in the 1930s, but by 1934 the sizes could be reduced with new materials.

In your adaptation, technically you don't strictly need the second capacitor, but since you already have it, I'll let you keep it there, because electrically it is a better choice than the 1 F capacitor that would normally be shared between the ringer and the audio circuit.

Quote
You guys were right that this is addictive.... It's very zen to sit quietly studying a wiring diagram, stripping wires, crimping, testing my connections with an ohm meter, ending up with a physical object that functions.  This contrasts with my job sitting in an office, staring at a computer screen, writing papers or grant applications, getting interrupted constantly by my students and lab staff to figure out why their experiments failed, going to meetings that accomplish almost nothing: the academic rat race!
I remember!

Quote
I think the take-home message from this thread so far is "a little knowledge is a dangerous thing."  I've learned just enough about telephones to mess around and think I've accomplished things I haven't (i.e. making a 634A-equivalent subset).
You've done quite well, and this exercise is a good starting point to learn about the circuit theory of sidetone vs. anti-sidetone circuits, and telephone circuits in general.  Dig into the TCI library and other websites and study the diagrams.  There are also many excellent handbooks on the web from the 1900s to the 1930s that discuss early telephony better, certainly in more detail, than the scarce modern texts.

In Western Electric product terminology, the differences between the sidetone and anti-sidetone circuits is exhibited in the differences between the 102 and the 202 type telephone sets, discussed here on the forum many times.  Here is a recent topic where I highlighted this circuit evolution, and it is directly applicable to the conversion you performed:  102 versus 202



« Last Edit: November 10, 2015, 09:54:15 AM by unbeldi »

Offline oyang

  • ***
  • Posts: 102
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #12 on: November 11, 2015, 11:44:44 PM »
Here are the wiring diagrams inside two of the Kellogg subsets, for documentation.
"In theory, theory and practice are the same. In practice, they aren't."

unbeldi

  • Guest
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #13 on: November 12, 2015, 08:00:55 AM »
Ah, thanks for those!
From the little schematic in the upper left corner you see that the induction coil has only two windings and diagram also tells us that it is a common battery circuit, meaning the current to run the transmitter is supplied by the exchange.
Kellogg actually used the same terminal designations and wire colors here as did Western Electric, so this diagram is directly applicable to connecting WECo desk stand or desk top sets with a mounting cord having red, green, and yellow insulation on the conductors.

Compare with this diagram from the topic "102 versus 202":

All you need to do is flip one of them horizontally.  The switches are different, but that is a minor aspect.

So, in hind-sight this topic should be renamed to ".... to WE 534A Equivalent".   In fact, it doesn't need to be converted at all.  For all practical purposes, it is the Kellogg equivalent.  Simply connect the yellow, red, and green mounting cord wires to the same colors as indicated on the wiring diagram in the center.
« Last Edit: November 12, 2015, 08:08:33 AM by unbeldi »

Offline oyang

  • ***
  • Posts: 102
Re: Converting a Kellogg Subset to a WE 634A Equivalent
« Reply #14 on: November 13, 2015, 12:28:04 PM »
Interesting.... so is the 2F capacitor/condenser I added actually doing anything? Or is it just a nice decoration?
"In theory, theory and practice are the same. In practice, they aren't."