Correct electricity supply for slave clock - avoiding overheating

Discussion in 'Electric Horology' started by dw2007uk, Feb 28, 2017.

  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  1. dw2007uk

    dw2007uk Registered User

    Dec 2, 2007
    34
    0
    0
    Hi

    I have a few slave clocks which are fairly standard movements found in GPO clocks etc. They run on 12 volts and I've got a Raspberry Pi set up to provide the pulses at the required intervals.

    However I noticed the coil in one of the clocks gets very hot if the pulse stays on for too long - I discovered this when testing. If something went wrong in operation and the pulse remained on for too long, I'm worried the coil would overheat and get damaged.

    Is the overheating just something that happens on slave clocks if the pulse is left on for too long, or can the overheating be prevented by using a resistor? Or maybe a thermal fuse would be best to cut out the current in such a case?

    Thanks
     
  2. Robert Gift

    Robert Gift Registered User

    Nov 12, 2012
    184
    0
    16
    Male
    Tree service "Branch Manager". Medical courier
    Denver, Colorado
    Country Flag:
    Region Flag:
    Limit pulse duration?

    "GPO" ? "Raspberry Pi"? Can the pulses be limited so that they never stay on too long?
     
  3. flynwill

    flynwill Registered User
    NAWCC Member

    Feb 1, 2007
    1,010
    6
    38
    Rolling Hills Estates, CA
    Country Flag:
    Region Flag:
    Re: Limit pulse duration?

    Indeed I would recommend adding a "one shot" to your interface between the PI and the clock to avoid the possibility of the coil being left on too long as a result of a crash or software fault.
     
  4. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    Re: Limit pulse duration?

     
  5. dw2007uk

    dw2007uk Registered User

    Dec 2, 2007
    34
    0
    0
    Re: Limit pulse duration?

    Thanks for your replies.

    GPO is General Post Office which used many slave clocks and they have a fairly standard movement. Raspberry Pi is a small computer which runs Linux operating system.

    The program I have running limits the pulses to be very brief but, as flynwill recognises, there's a chance there could be a software fault which causes the pulse to stay on too long.


    Do you think a resistor would be the best choice? Would the resistor get too hot though if the pulse went on too long?


    What sort of value resistor would you recommend?

    In the original setup with a master clock, would there have been something to prevent a failure in the master clock which could have caused a pulse being applied for too long?

    Thanks again! :)
     
  6. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    #6 rogerj, Feb 28, 2017
    Last edited: Feb 28, 2017
    Re: Limit pulse duration?

    In the original PO No 36 clock the pulse was generated mechanically by a count wheel turned by the pendulum. See here :
    http://www.hvtesla.com/masters/po36_countwheel.html It's almost impossible to visulalise a situation where the pulse could be prolonged.
    A series resistance is the universal method of current control for slave dials and in your case, with a 12v supply, a good starting point would be 47 ohms.
    If the switch is solid state it should be protected by snubber diode.
     
  7. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    Re: Limit pulse duration?

    Coils are current devices. Some have internal resistance to limit their
    current while others depend on an external current limit.
    I don't know the specifics of your clock but many slave clocks are intended
    to run on a series current loop. Each clock dropping some voltage with the
    master having a limiting resistor to adjust the loop current.
    If the output were continuous, for the suggested limiting resistor,
    with 12V across 47 ohms, it would dissipate
    3 watts. I would recommend using a 5 or 10 watt ceramic resistor ( commonly
    available ).
    That should be enough to protect the coil, should the software leave the pulse on.
    I would also mount the resistor in a location that has air circulation rather than in
    a small closed box, if you don't want to cook everything in the box.
    Tinker Dwight
     
  8. dw2007uk

    dw2007uk Registered User

    Dec 2, 2007
    34
    0
    0
    Re: Limit pulse duration?

    Thank you!

    Apologies for the questions (I'm new to this), what value resistor (as in number of Ohms) would you recommend for each clock?
     
  9. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    Re: Limit pulse duration?

    I assuming you realise that this type of slave clock is always connected in a series loop - and in series with the final switch. This is probably a transistor triggered by the Pi.
    As I said previously, 47 ohms would be a good starting point with one dial..If more than one dial allow 1.5 volts for each dial and subtract the answer from 12. Call the answer "x" Then....by ohms law, 12 volts minus "x", divided by .2 (amps) is the likely value. You must be sure that the pulse from the Pi is just long enough to advance the dial reliably..
    A 50-50 square wave is no good and will likely be the cause of overheating..you must use a "one shot" monostable or trim the software to produce the output pulse of about 200 to 250 milliseconds..
     
  10. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    Re: Limit pulse duration?

    What is the supply voltage and what is the resistance of your coil on the slave?
    Tinker Dwight
     
  11. dw2007uk

    dw2007uk Registered User

    Dec 2, 2007
    34
    0
    0
    Re: Limit pulse duration?

    Thanks for explaining - I'll see what I can come up with!

    The supply voltage I was using is 12 volts. I've just measured the resistance of the four slaves I've had running - one has a resistance of 9 ohms, one is 3 ohms (this one has a resistor soldered in across the terminals) and two are 1,000 ohms (these two also have a resistor across the terminals).
     
  12. ElectricTime

    ElectricTime Registered User
    NAWCC Member

    Sep 28, 2002
    190
    0
    16
    Re: Limit pulse duration?

    Power is the Voltage Squared divided by the resistance.

    Assuming continuous current, and V= 12VDC

    for 1000 ohms P=144/1000=.144 watts
    for 9 ohms P=144/9= 16 watts (this is high)
    for 3 ohms P=144/3= 48 Watts (high and would probably overload your power supply).
     
  13. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    Re: Limit pulse duration?

    The 1K one is likely to use the full voltage but the low ohms ones
    are surely current loops.
    Tinker Dwight
     
  14. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    #14 rogerj, Mar 3, 2017
    Last edited: Mar 3, 2017
    I should have read the original post more carefully..You appear to have a collection of different slave clocks - as evidenced by the variety of measured coil resistances..maybe they are not all GPO clocks..Some pictures would help a lot..
    The GPO slave clocks I've been referring to are all like the ones I have and as pictured in this photo. Made by the Synchronome company, they measure about 9 ohms including the parallel connected resistor. Incidentally, this resistor is included to increase the life of the contacts that switch the dials. It may not be sufficient protection for solid state switches against the spikes produced by back EMF. The value of the resistance is usually between 5 and 10 times the actual coil resistance. I'm not sure exactly what resistance Synchronome would have used.
    All dials in a series loop should be the same. Dial mechs with high resistance (like 1000 ohms) must have been intended for parallel connection and clearly won't be suitable for a series loop. Parallel ccts often worked at 24 volts.
    The current limiting resistor for the loop SHOULD only dissipate power for 1/5 of a second every 30 seconds. In these circumstances a 1 watt resistor should be adequate..

    [​IMG]
     

    Attached Files:

  15. ElectricTime

    ElectricTime Registered User
    NAWCC Member

    Sep 28, 2002
    190
    0
    16
    I have to assume the clocks are not being run directly by a raspberry PI - the GPIO outputs are only 3.3 or 5V and would toast pretty quickly.

    A hardened SSR should work fine for this application - relay contacts are a different story.

    Is the use of the resistor across the coil to prevent relay contact sparking ?
     
  16. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    A small correction..I said in#9 that a pulse of about 200 milliseconds is all that's required to impulse a dial of the Synchronome type. In fact, because I'm currently working on a dial that requires impulses of alternate polarity which DO require a pulse of about 200 Ms I made that mistake. Checking in Hope-Jones book the pulse used to reset the gravity arm in a Synchronome - and hence drive identical dials as those we are considering - is only 60 Milliseconds long. (page 108 in my edition). It won't make much difference to the operation of the dial but it will increase the battery life a little. With a Pi driving the clock it should be easy to experiment to find the length of a reliable pulse.
     
  17. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    #17 rogerj, Mar 3, 2017
    Last edited: Mar 3, 2017
    Yes..It is simple and effective in the original context. It dissipates the back emf. It's worth Googling "contact arc suppression" for the pros and cons of many other methods.
     
  18. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    An electrical lesson:

    Coils create a current in a changing magnetic field.
    When the contact start to open, the field collapses.
    This tries to induce a current in the coil.
    Since the points have just started to open, the current
    has no place to go but the field is collapsing. The voltage
    start to raise as the energy of the coil has to go someplace.
    If the point could open really really fast, the stray capacitance
    would eventually be enough to absorb the current and
    limit the voltage.
    This creates a ringing like a flywheel and spring. Once the field
    collapsed the capacitor would build it back up in the opposite
    direction. This action would continue until the resistance of the coil
    would dissipate the energy in heat.
    The rise in voltage happens in nanoseconds the point just don't
    open fast enough. The voltage quickly reaches the arcing voltage
    of the points. The energy is now turned into light and heat that is
    dissipated quite quickly. Even hardened point will be eroded.
    The resistor gives the coil someplace to dump the energy stored
    in the magnetic field.
    Better results can be had with a small capacitor and a resistor
    but that requires selecting the right combination. You don't want
    a capacitor that will weld the points on closing and you don't want
    it to ring because the resistor is not of a low enough value.
    As Roger has said, you want to use a lower value of this resistor for
    circuits using solid state switch. The resistor has been optimized
    for a contact, like a relay.
    Tinker Dwight
     
  19. ElectricTime

    ElectricTime Registered User
    NAWCC Member

    Sep 28, 2002
    190
    0
    16
    I've been a big believer in using off the shelf Solid State Relays with built in Snubber networks. Had basically 0% failure, though some have leakage current which can drive you crazy at first.
     
  20. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    Good tip ! found some on *bay for <2$.I have ordered one to experiment. They all seem to be rated for mains voltage but I assume they will switch low voltage DC ok. Not sure about the truth of the claimed manufacturer but worth a punt.
     
  21. flynwill

    flynwill Registered User
    NAWCC Member

    Feb 1, 2007
    1,010
    6
    38
    Rolling Hills Estates, CA
    Country Flag:
    Region Flag:
    Many AC Solid State Relays are SCR/Triac based and will not switch DC properly (they rely on the AC zero crossing to shut off the switching device). There are also some that require a relatively large load voltage to trigger. You'll want to check the ratings closely when you get the device.
     
  22. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    If it is rated for mains voltage, it is most likely a SCR type and won't
    work with DC.
    You might get by with switching the AC to the power supply.
    Tinker Dwight
     
  23. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    I did wonder about that..It will take a couple of weeks to come from China. If I don't report back assume you are both right and it didn't work on DC !
     
  24. ElectricTime

    ElectricTime Registered User
    NAWCC Member

    Sep 28, 2002
    190
    0
    16
    SSR relays come in lots of flavors, as long as the OUTPUT is DC it should work fine. An ac output relay typically won't turn off with DC.
     
  25. fdew

    fdew Registered User

    Jul 12, 2007
    210
    0
    0
    One way to limit current is with a light bulb in series. They have low resistance when cold but if the pulse lasts the bulb lights the resistance goes up.
     
  26. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    Light bulbs have a really large resistance swing from cold to hot.
    I'm not sure that is desired for this purpose.
    It might be that the light heats too fast and not allow the coil to pull in.
    Something like a DC motor might work better. On can control the time
    till it limits the current by the mass of a flywheel.
    Tinker Dwight
     
  27. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    I can't really see any advantage in messing with bulbs when a resistance is simple enough and there is one provided in the clocks..In the early days - pre trickle charging and wet cells - Gents sold a bell that could be put in series. As they patented the idea, Synchronome used instead a lamp in a box with a red glass. The idea was to warn of low battery voltage. The way it worked was if the current became insufficient to reset the gravity arm the pendulum assisted it on the next swing. Because of the greatly lengthened pulse the bell would ring....or the lamp would flash to alert the keeper of the clock..With a normal short pulse it did nothing. R
     
  28. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    The thought is not so much to replace the resistor but to give protection
    in case the voltage should stick on, from a failure of the master clock source.
    A dynamic current source that can limit the current would protect the
    slave coils.
    The lamp idea responds too fast and is not likely to actuate the slave.
    Tinker Dwight
     
  29. lmester

    lmester Registered User
    NAWCC Member

    Dec 30, 2009
    442
    2
    0
    I manage HVAC systems for a school district.
    West Virginia
    Country Flag:
    Region Flag:
    Here are two things you could do to protect the clock coils:

    First is to use a constant current source circuit. This would allow you to add or remove clocks from a series circuit without any re-adjustment of resistors. You can build a current source from a few inexpensive components. One of the simplest circuits uses an LM317 voltage regulator and a single resistor. About $2 worth of parts. Do an internet search for "LM317 constant current source" for more info. Since you'll be switching an inductive circuit, a few more components may be necessary to protect the regulator.

    These clock coils are not designed for continuous operation. They may overheat even with the proper current if left energized constantly. You may also want to use a time delay relay to limit the maximum pulse time. Set it to break the circuit when the delay expires. Set a delay of a few seconds. If your clock control computer locks up with the output on, the time delay relay will stop current flow when the delay expires. I have a lot of these relays on the shelf at work. They're available in various operating voltages and with fixed or variable delays.

    The only problem with this solution is that these relays are fairly expensive. Last time I ordered some they were about $100 each. Since you don't need an industrial rated relay there will be cheaper options.

    Just thought of a really cheap way to fix this problem. Use a time delay fuse to protect your clock circuit. Pick a value that will blow with the normal clock operating current. With the low duty cycle pulses that are normally applied, the fuse won't blow. If the output gets stuck on, the fuse will then blow in a few seconds.
     
  30. rogerj

    rogerj Registered User

    Dec 21, 2014
    115
    0
    16
    retired broadcast engineer
    Cornwall
    Country Flag:
    #30 rogerj, Mar 19, 2017
    Last edited: Mar 19, 2017
    To that I could add a third even simpler suggestion. Replace the battery with a capacitor charged via a resistor (from the battery). I haven't tried this but it loosely parallels the way I powered the impulse magnet in my cHipp toggle clock..This https://planetcalc.com/1980/ online calculator shows that a 3300uf capacitor charged via a 1500 resistor will be 99.2% charged in 24.75 seconds when it would be ready to generate the short pulse for the clock and dials. The short circuit current would be in the order of <20 Ma with those values. Use the calculator to work out other values..
    3300uf may not be the right value but as we are discussing a diy electronic clock experimentation is in order.
    A fourth approach might employ a 555 timer wired as a one shot with a timing of , say 0.5 second. Either gated with the signal from the Pi or operating a secondary switch . Homemade clocks are very satisfying projects !
     
  31. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,680
    28
    0
    Calif. USA
    I like the capacitor idea. It is a fully fail safe method.
    It might take a larger capacitor but the method is sound.
    Tinker Dwight
     
Loading...
Similar Threads - Correct electricity supply Forum Date
Gents C7 Pulsynetic Master Clock Power Supply Electric Horology May 15, 2017
Correct power for IBM self winding master clock movement Electric Horology May 23, 2014
Correct BPH for a Schatz Lectronic? Electric Horology Jul 20, 2013
Still can't get beat / impulse correct on Barr Poole clock Electric Horology Oct 17, 2012
correct battery for Schatz wall clock Electric Horology May 1, 2011

Share This Page