[GreenKeys] Current limiting

[Randy or Sherry Guttery]

Chris Elmquist wrote:

> but is the high voltage really neccessary if you are
> not running a long distance with the loop?

There isn't a simple "pat" answer to this - some of us have indeed been 
"playing" with "low voltage" loops - and the answer is:
It depends.

> In other words, could a modern constant current source using say, LM317
> be built that provides a nice stable 20 or 60mA but with only 12, 24 or
> 48v DC source?

Again - that depends - and might actually be less successful than a more 
simple (passive vs. active) source.

> Is there something in the older TTYs that requires a higher voltage on
> the loop?

Yes, but again it depends...

> 
> My ASR-33 runs with 20mA sourced from 12v but that is pretty "modern"
> compared to the 5-level stuff you guys are working with I am sure.

Different critter, though there are possibilities here as well...

So - let's talk about what's going on - and why it "depends"...

Preface: I tired to keep this simple (some might think over-simplified) 
and some numbers are generalizations - but the principles are solid and 
can lead to successful implementations in real life.

The selector in the older teletypes are made up of electromagnets, who 
(depending on wiring, etc.) are designed to "follow" either a 20ma. or 
60ma. signal.  This "signal" is NOT a simple series current source, but 
delivered by a group of components which are optimized to deliver as 
faithful of a "signal" through the selector as possible. That signal is 
a series of intervals of current and no current which convey the 
"intelligence" of the signal.  In the case of the older machines - the 
intelligence is 7 - 8 intervals of current and no current - whose 
characteristic is a beginning interval of no current (start interval), 
then 5 intervals of current / no current - the pattern dictated by the 
"character" being conveyed, followed by 1 - 2 (or something in between) 
interval lengths (stop bit(s)).  The primary point of concern here is 
that word "faithful".

Ideally - the keying circuit would merely have to make and brake the 
simple series circuit containing the selector magnet to accomplish the 
above info transfer... but several real life factors work against this 
simple solution.

First is the selector magnet assembly itself. In most teletype machines 
(at least of the vintage we're talking about) the selector magnet is 
comprised of many turns of fine gauge wire wound on a bobbin, which then 
energizes pole piece(s) to alternately attract and release an armature - 
in step with the current in passing through the windings. In the case of 
the original model 28 - the two coils are wired in parallel (for 60ma 
circuits); and consisted of 3600 turns of #33 gauge magnet wire which 
results in (approximately) 132 ohms in each coil (i.e. 66 ohms in 
parallel; or 264 ohms in series). Both coils are "rated" at 115VDC (with 
a HIPOT A-C withstand rating of at least 500V). I don't have the 
inductance figures at hand for the 28 (IIRC the 15's coils are around 
4Hy)- but with 3600 turns on an iron core - it's substantial. Point 
being - in addition to the resistance of the winding(s) you also have 
the inductance of them which tends to distort any square wave applied to 
them.  This distortion manifests itself as a phase displacement between 
the voltage applied to the selector magnet coils - and the resulting 
current through them.  If we assume for a moment - that our selector 
coils have no DC resistance - then when any source is applied to them - 
the voltage across the coils would "lead" the current through them by 90 
degrees. The higher the resistive components of the circuit - the less 
the current "lags" the voltage.  If we them add a significant resistance 
to the loop - say ten times the effective inductance - we minimize the 
phase lag (distortion) in our loop. However - this higher resistance 
requires higher voltage in order to achieve our desired steady-state 
current (say 60ma.). So with  say 1000 ohms added to the circuit - that 
requires an additional 60V just to "drive" that additional resistance - 
PLUS whatever voltage would be needed to drive the coil(s) (which if 
calculated as pure resistance - would be roughly 4V more).

Then there are additional factors: how "accurate" is the keying device? 
  The contacts in a teletype's transmitter - being mechanical - are 
excellent.  Tubes (and transistor) circuits are good, long parallel 
wires start having an effect - and being a series circuit - everything 
contributes something. It's helpful if these do not contribute a serious 
amount of additional distortion, but sometimes that can't be avoided.

In the early days of teletypes - these factors had been discovered and 
explored quite a bit.  Obviously adding resistance to a loop contributes 
quite a bit.  Another factor, of course, is how good in the mechanics of 
the keying system (tape readers, keyboard, etc.). All of these ideally 
don't "push" the envelope - but sometimes they do. To deal with some of 
these variables at the receiving end - virtually all machines have what 
is called a "ranging" adjustment. This adjustment accomplishes it's 
purpose in varying ways - but in simple terms - it allows the operator 
to manually advance or retard the "reading" of the incoming signal in 
relation to time - to physically displace in time the moment the 
selector mechanism samples the magnetic state of the coils (you could 
say "shifts" the *moment* it makes up it's mind - either advancing it - 
or retarding it).


So... where does all of this bring us with regards to your question? Yes 
- some teletypes - depending on several factors coming together - can be 
run on embarrassingly low voltages powering it's current loop.

In my case - I run my model 28 on 24VDC, with the loop "keyed" by a 
Blackbox CL50 RS-232C Serial to Current Loop converter box attached to a 
computer. The CL50's is rated to key a maximum of 30V; hence the use of 
a 24VDC supply. It works, quite well - within limits.

  But several things have to come together to make it work:

1) The better condition the selector mechanism is in - the more tolerant 
of distortion it will be and still copy OK.  This means the armature 
spring and clearance adjustment(s) have to be as close as possible to 
"spot on". Also - my own tests show that the more modern armature styles 
(two button, quick release armature - similar to that supplied in kit 
195634. There is also a replacement armature for the older style 
assemblies P/N 160180).

2) Speed- I can get consistently good results copying at 60 and 75 WPM - 
but things start getting "on the margin" at 100WPM. I don't like running 
at 100WPM anyway - (too frenetic compared to the "comfortable" 60WPM 
pace). With the higher "data rate" - it'd be better to have more 
resistance (with resulting higher voltage).

3) More in the loop: Obviously - with only 24Vdc - there is little room 
left to add more "stuff" - I wouldn't expect my URA-8A (CV-89A) to be 
happy keying such a loop...


As noted above even with teletypes "designed" for low voltage keying - a 
certain "RiceGuy" has been known to replace the low voltage coils in his 
35 and run it on a traditional loop (URA-8, etc.). Obviously - there is 
"wiggle room" here - particularly with the later machines where things 
weren't pushing the limits quite as hard as say the 15.

Hope this enlightens more than confuses.
-- 
randy guttery

A Tender Tale - a page dedicated to those Ships and Crews
so vital to the United States Silent Service:
http://tendertale.com