Kellogg (sort of) works but won't hang up/disconnect

[scottfannin]

I want to get some ideas for a phone that won't hang up.  Seems to work otherwise.  Rings, answers call, speech in/out sounds fine.  Then I hang up and it just sits there unless the other end hangs up.  If I work the switch hook slowly on and off, the bells do one tinkle sound each time and after a few tries at this it stops (suggesting a capacitor?).  The phone was assembled from a completely rusty non-working wreck.  So it's a "Kellogg" but it was literally stripped down to the bare wood, processed, and had all the guts put back in with modern pieces for ruined wires and that sort of thing.  I added a ringer autoswitch, which I've used with success previously, and since it's a pain to put them in and out, I always test before screwing those down and this one worked fine.  Basically what the switch does is let you show visitors the "magneto" so that a regular call rings on L1/L2 but if you crank the magneto it uses a bridge rectifier to power a relay which then puts the magneto power to the bells, thus shielding the cablemodem while letting visitors hear the bells ring locally when they turn the crank like in some old movie.  The handset is a three-wire red yellow green and it was tested and found to work after rewiring to replace the crumbly stuff.  The "condenser" is just a 0.47 and 2.2 capacitor in a decorative wooden box with their leads sticking out and they test fine.  So I'd like to get ideas on what could be causing this behavior.  I've attached a picture, mostly for your amusement since it's all very dense and hard to see visually, and a diagram of what I intended to create and probably have created--not that it's correct but I'm pretty sure I wired it like I intended to. 

[Key2871]

That tink the bell makes is normal, if it has a bias spring  I'd adjust to eliminate that.
As for the phone not hanging up, I'd say there is enough resistance in the circuit that's holding the line open.
I'd check for a capacitor or a connection to a component that will hold the line even after you return the receiver.

Do you have this problem if you disconnect the line from the phone and reconnect the phone again?

If so that's the issue, there is a component that's giving resistance and is holding the line.

[scottfannin]

We're working on it.  It does hang up if I just plain disconnect L1 and L2.  I found an interesting complication that some of the "new old" wires used aren't so reliable.  One can be made to be continuous or not by moving it around gently with alligator clip tester leads on both ends.  The wire was a gift, which I appreciate, but it fails continuity checking once it's been bent a few times.  My plan is this weekend to test using "vintage guitar" wire samples that should arrive tomorrow and see if that provides something that looks authentic but handles being stuffed into a ringer-box-sized phone box and dressed harshly around a big magneto better.  Or it might be horrible tinsel that spits glitter when trimmed, who knows?  Meanwhile this concept that capacitance, resistance, and the overall circuit affect the local exchange (or whatever passes for one now) recognizing the hangup is new but makes a lot of sense to me.  I had kind of thought that you push the switch and it either connects or it does not.  I am definitely still trying to understand the phone circuit and though it's obviously simple I do not yet understand it properly.  I can make one work, but it has an element of voodoo to it still.  Like I imitate the diagram from somewhere and then it works but I couldn't do it from scratch yet.  So what is it that needs to change for the other end to think it's a real disconnection?

[poplar1]

Does the ringer autoswitch, when the generator is not operating, connect C1 to L1, and B1 to L2?

If so, the ringer is connected directly to L1 and L2, without the benefit of a capacitor in series. This will probably keep the line off hook all the time (or, at least, trip ring then keep it off hook, on an incoming call). (Assuming the line is connected to L1 and L2?)

B2 ringer terminal connects to L1. B1 ringer terminal connects through closed contact in autoswitch to L2.

[scottfannin]

You're right, it did show continuity between the NC terminals of the relay, which of course had L1 and L2 connected to them (duh).  So that means L1 and L2 were connected to the bells routed around the ringer capacitor no matter what.  BUT, I tried a different circuit and it didn't seem to work, plus it seems to allow connection of the magneto to the L1 and L2 phone lines, which the switch was meant to prevent, to avoid frying the broadband modem.  Here is an updated drawing, I'm sure there is some kind of arrangement that can work but so far playing around with it hasn't found one.

[scottfannin]

Newest attempt.  Should it work?  So far it does not, but I wonder whether that is not figuring it out yet or (hopefully?) some mistake in my eyesight/fingers.

[poplar1]

Where is the hookswitch?

[scottfannin]

Sadly had to focus on the day job for a bit....  I should perhaps have used "HS" for the switch.  H1 and H2 are the two independent switches in the hook switch.  As is before I took it apart, the middle ones had a jumper between them, making it connect 4 wires at once.  That didn't seem to fit the circuit for modern common battery so I removed the jumper.