[GreenKeys] Please help me diagnose a TT model 33

[Steven First]

Hello guys; first time GreenKeys user here looking for some
help.  I will keep this short and simple
but if you need more details please let me know. I am trying to get this running as a gift to my long time girlfriend who is all things retro computing.


I purchased a Teletype machine that was modified and put in
a Honeywell case.  I purchased it in
hopes to get it running but did not know the condition or even make of it.  When I got it I found that it was a TT 33
with no answer back drum, no tape reader, and no call out equipment (phone) or
selector magnet driver electronics. 
Judging by the wiring it looks like it may have been interfaced to a
computer at some point.


I went the route of creating my own send and receive
electronics that provides the 20MA current loop, reads that data into a Arduino
then puts out the same data out to a high current driver circuit that can drive
the coil at 500mA DC.  


Using the manual and my electronics and mechanical
background I used an oscilloscope to find that bit 8 was not working
correctly.  I found that the leaver was
missing from the code bar on the keyboard. 
I make a jumper that provides bit 8 with an always HIGH (marking) state
as bit 8 is always marking.  This and
cleaning up the brushes and contacts gave me a very good and strong signal with
all 11 bits perfectly shaped.  I have
verified that every ASCII code the system uses can be properly broadcast to my
Arduino (all 11 bits).


Unfortunately I have not had the same luck feeding those
signals back to the systems receive mechanics and getting good results.  To start I have a very good 500mA DC driver
capable of dealing with the voltage and current spikes associated with driving
a coil and have scoped it to ensure acceptable signal quality with clean square
waves. But I have not been able to receive these 11 bits back in a way that will
drive the coil in such a manner as to produce repeatable characters.  I can get some U’s, P’s, and some Z’s etc but
I can’t get the associated ASCII codes (11 bits) to give me the proper printed characters.


The way my electronics work is such that whatever the code
wheel sends out the selector coil receives exactly.  The timing difference is sub 1mS delay and I
get the same results regardless of if I am using the drive motor or freewheeling
the system by hand; my fingers are so worn out from turning the main shaft…. I
have adjusted the selector cam timing from one extreme to the other trying to
tune it in but alas I am still not able to get it working.  The best I have gotten is a sent ASCII U will
give me a typed U on paper…but so will many other ASCII codes.


The only thing I have noticed that was unexpected that I
have seen is that the code cam receives and cams the first 4 bits lifters in
sequential (right to left looking from back of machine) order and then does the
last for bits in a ping pong fashion. 
Seeing this is set by the cam I can’t see that being the issue however.


So that is where I am at. 
I am running out of ideas on what may be the cause of this.  I have made the assumption that all 11 bits
are sent to the selector coil but can’t find any documentation that says one
way or the other.  Any help on this front
would be much appreciated.

 

Steven First 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/greenkeys/attachments/20151203/04b46387/attachment.html>