electricclockcollecting.myfastforum.org Forum Index electricclockcollecting.myfastforum.org
A forum to discuss the collecting, history, construction, maintenance and repair of electric clocks. Feel free to contribute or make suggestions for improvement of the forum. Be polite, friendly non-combatative. Spammers will be banned permanently.
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   Join! (free) Join! (free)
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

for cerebral exercise !

 
Post new topic   Reply to topic    electricclockcollecting.myfastforum.org Forum Index -> Anything else.
View previous topic :: View next topic  
Author Message
Please Register and Login to this forum to stop seeing this advertising.






Posted:     Post subject:

Back to top
hugozair



Joined: 21 Apr 2015
Posts: 134
Location: near Penzance UK

PostPosted: Fri Oct 21, 2016 9:40 am    Post subject: for cerebral exercise !  Reply with quote

Since I became interested in electrical horology the Hipp Toggle has been of particular interest and is at the heart of my DIY clock.
I have often seen the HT described as an example of a negative feedback loop and once as a feedforward servo.
I know nothing about the latter and haven't found anything understandable by Googling. As far as negative feedback is concerned I vaguely remember the definition learned in my 1960's basic training days (relating to a proportion of the output being fed back in antiphase to the input) and can't make any connection with that and the operation of the Hipp Toggle.
I can see that it is an amplitude control system working at very low frequency so possibly a servo of some kind.
So, the question, who knows how to label the operation of the Hipp Toggle system !!
Roger
Back to top
View user's profile Send private message Visit poster's website
hugozair



Joined: 21 Apr 2015
Posts: 134
Location: near Penzance UK

PostPosted: Fri Oct 21, 2016 1:06 pm    Post subject: more Reply with quote

The act of posting the question has got me Googling again...this time with different search terms..One result is this paper: www.hsn161.com/HSN/HansenClockPart1.pdf
here, the writer Mr Hansen calls it a Bang-bang servo - a term which I had  come across before but forgot to include in the original post. It seems Mr Hansen is not impressed with this "crude" control.
However, the system he goes on to describe appears to use a crystal controlled timer. To me this is cheating because if you include a precision timing source there seems no point in using the pendulum in the first place ! In fact, if you had to use a pendulum in a control system with quartz oscillator control, it would be quite possible to to phase lock it to the crystal (synchronize) and not worry about any external errors at all (ie temp and barometric.)
So..any advance or comment on Bangbang servo ?
Edit...Later same day...more googling and found reference to bangbang servo..here on the forum ! by miken68b at the beginning of his excellent pc monitoring scheme... Thread "Monitoring a 30s master clock by computer" I knew I'd heard of that somewhere ! Roger
Back to top
View user's profile Send private message Visit poster's website
miken58b



Joined: 29 Apr 2015
Posts: 46

PostPosted: Mon Oct 31, 2016 8:52 pm    Post subject: Reply with quote

In theory I studied control systems engineering for my degree.  However, I was a poor student and spent much more time teaching myself computing; no regrets!  Bang bang control is a pejorative term for a very simply kind of control system where the controlled variable is allowed to wander between 'too high' and 'too low'.  A simple room temperature thermostat is a good example.  By the time the thermostat has cut out the room has exceeded the desired temperature and continues to become even warmer because the radiators contain hot water and take a while to cool down and stop heating.  Once the thermostat decides the room is too cold no instant heating occurs because the water is cold and the boiler remote so the system takes time to heat up.  Hysteresis is a common attribute of a simple thermostat and is desirable to prevent too rapid changes from 'on' to 'off' and back again.

More sophisticated control systems would aim to have more precise control of the heat input to the room and some 'understanding' of the relationship between the heat input to the desired temperature.  It's not hard to see this becoming almost of arbitrary complexity: measure the outside temperature as it affects heat loss, count the people in the room as they emit heat (together with other heat sources such as light bulbs or electrical equipment), solar gain, window and door opening, air flow and so forth.  In the final analysis what sophistication of control is desirable/necessary is dictated by the accuracy with which it is necessary to control the temperature.  If a £20 bimetallic thermostat can achieve +-3 degrees say then that might be deemed adequate.  For some critical chemical process applications then control to within a fraction of a degree may be required.

I think the term feed forward servo in my room temperature example would be taking a control input from the outside temperature (or met office forecast).  My recollection of negative feedback is that it has more to do with compensating for non linearity in a system.  My immediate thought is it's application to amplifiers where it is used to reduce distortion that would otherwise occur.

The Hipp Toggle in it's simplest incarnation would appear to have some bang bang like attribute.  The desired variable is the amplitude of swing of the pendulum.  If it falls too low the pendulum receives an impulse.  The strength of the impulse is not controlled and is a function of the duration and magnetic force which are generally not dynamically varied.  Following the 'free pendulum' mantra there is some compromise between activating the Hipp Toggle (and driving the nib into a spring) and giving the pendulum such a hefty impulse the arc of swing varies over an undesirably wide range introducing more of the dreaded 'circular error'.

From what I can immediately recall of Roger's clock the pendulum is given a tug every swing that is proportional in some way to the velocity (and hence arc of swing).  This would be a much better way of keeping the arc of swing as near constant as possible.  Such a control system is not bang-bang.  

Industrial controllers are often referred to as PID - proportional, integral, differential (ie. a mix of simple functions of the controlled variable).  Control theory is steeped in solving differential equations which is fine as long as one can model the relationship between input and output.  Once that relationship has been established the perfect control algorithm is the functional inverse.  I have spent years as a school governor and find many parallels (desired outcome is 5 GCSE's grade A-C) but modelling the forward transfer function (that's called a learning environment) is a bit tricky let alone figuring it's inverse.  Frankly what actually happens is more like bang bang (increase maths tuition until English fails, then reverse)...

Not sure if this helps but at least it's warmed up the grey matter in time for dinner...

Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    electricclockcollecting.myfastforum.org Forum Index -> Anything else. All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

Card File  Gallery  Forum Archive
Powered by phpBB © 2001, 2005 phpBB Group
Create your own free forum | Buy a domain to use with your forum