# Antique AC clock runs too fast

#### michael8

Joined Jan 11, 2015
107
There are analog ways to make 50 Hz from 60 Hz, however I'd guess a small
micro could generate 50 Hz from it's local crystal clock and watch the
60 Hz power line and make adjustments as needed...

#### MrChips

Joined Oct 2, 2009
22,505
Obsolete MM5369 chip generates 60Hz from 3.579545MHz color burst crystal.
You can create the equivalent function with just about any MCU chip. Generating 120VAC or 240VAC from that is a bit more tricky.

Edit: Sorry, brain freeze. You want 50Hz oscillator.

#### Wolframore

Joined Jan 21, 2019
2,183
It's tricky and a little risky... tricsky. I wonder if it needs the full voltage to operate. It would be nice if it was tolerant and ran at under 50V.

#### s14rs4

Joined Sep 15, 2016
53
I would have thought the current draw would be in the milliamp range. A small step down transformer say 240v to 6v, connected in reverse (input to the secondary output from the primary) would be enough to drive it. A crystal oscillator with a divider and a transistor output buffer to drive it would be enough. Try measuring the current draw at 60Hz.

#### MrChips

Joined Oct 2, 2009
22,505
If the motor already runs directly off 240VAC then it would have very high impedance. Reducing the voltage will not work.

#### s14rs4

Joined Sep 15, 2016
53
If the motor already runs directly off 240VAC then it would have very high impedance. Reducing the voltage will not work.
I described using the transformer in reverse, as a step up. 6 volts to 240v.

#### MrChips

Joined Oct 2, 2009
22,505
I described using the transformer in reverse, as a step up. 6 volts to 240v.
Ok, I now catch your idea. It's worth trying.

#### MrChips

Joined Oct 2, 2009
22,505
Ok, here is my solution.

Take the 60Hz AC LINE signal through an opto-isolator to a microcontroller (MCU).
Program the MCU to generate 50Hz. Quartz crystal stability is not required.
At every 6 cycles of 60Hz detected, re-sync the 50Hz oscillator.
Feed the 50Hz output to a step up transformer or similar.

I can do this with a small 8-pin MCU chip.

#### Bottleguy

Joined Nov 21, 2020
14
Ok, I now catch your idea. It's worth trying.
Ok, here is my solution.

Take the 60Hz AC LINE signal through an opto-isolator to a microcontroller (MCU).
Program the MCU to generate 50Hz. Quartz crystal stability is not required.
At every 6 cycles of 60Hz detected, re-sync the 50Hz oscillator.
Feed the 50Hz output to a step up transformer or similar.

I can do this with a small 8-pin MCU chip.
Think I can get the parts at the local Microcenter store, along with some cranial support? Remember, my experience in this field stops just after "gee, you mean a wire wrapped around a nail and hooked to a battery will make a magnet!? Cool..."

#### Bottleguy

Joined Nov 21, 2020
14
Think I can get the parts at the local Microcenter store, along with some cranial support? Remember, my experience in this field stops just after "gee, you mean a wire wrapped around a nail and hooked to a battery will make a magnet!? Cool..."
Oh, by the way, the wire of the coil appears about as thick as a human hair....

#### s14rs4

Joined Sep 15, 2016
53
Ok, here is my solution.

Take the 60Hz AC LINE signal through an opto-isolator to a microcontroller (MCU).
Program the MCU to generate 50Hz. Quartz crystal stability is not required.
At every 6 cycles of 60Hz detected, re-sync the 50Hz oscillator.
Feed the 50Hz output to a step up transformer or similar.

I can do this with a small 8-pin MCU chip.
Ok, here is my solution.

Take the 60Hz AC LINE signal through an opto-isolator to a microcontroller (MCU).
Program the MCU to generate 50Hz. Quartz crystal stability is not required.
At every 6 cycles of 60Hz detected, re-sync the 50Hz oscillator.
Feed the 50Hz output to a step up transformer or similar.

I can do this with a small 8-pin MCU chip.
I like the idea, but I still think using a crystal reference for the MCU would be better. The whole circuit could run from a DC wall wort or even batteries. I would still like to know what current it draws from the 60Hz supply.

#### MrChips

Joined Oct 2, 2009
22,505
You can use a crystal on the MCU instead of AC LINE frequency if you wish but the time-keeping ability will not be as good. I can put together a circuit for you if you wish but you will have to do some assembly.

We suspect that the coil takes very little current. More importantly we would like to know what is the lowest voltage at which the clock will still run.

#### djsfantasi

Joined Apr 11, 2010
7,211
I like the idea, but I still think using a crystal reference for the MCU would be better. The whole circuit could run from a DC wall wort or even batteries. I would still like to know what current it draws from the 60Hz supply.
Not necessarily. Mains 60 Hz is very stable as required by law. Power companies have complex circuits to maintain line frequency at an average of 60Hz. You’d need similar circuitry to adjust a Crystal reference over time.

By using the stable 60Hz frequency as a reference and adjusting your 50Hz output every 0.1s, you’ll get a very very stable output. You’ll have corrected the frequency 600 times a minute. The adjustment is done in software and can account for several variables. Such as MCU clock, instruction times, temperature variations, load variations, et.al.

Check out Romanblack.com for some algorithms.

#### s14rs4

Joined Sep 15, 2016
53
Not necessarily. Mains 60 Hz is very stable as required by law. Power companies have complex circuits to maintain line frequency at an average of 60Hz. You’d need similar circuitry to adjust a Crystal reference over time.

By using the stable 60Hz frequency as a reference and adjusting your 50Hz output every 0.1s, you’ll get a very very stable output. You’ll have corrected the frequency 600 times a minute. The adjustment is done in software and can account for several variables. Such as MCU clock, instruction times, temperature variations, load variations, et.al.

Check out Romanblack.com for some algorithms.
I agree with you on accuracy but it requires a connection to AC mains, my solution could run from a DC supply. Some crystal references are temperature compensated, I have used a DS3231 RTC with an Arduino and it is good to less than two minutes a year. Your design is probably more accurate but I was thinking of avoiding a connection to mains voltages.

#### Bottleguy

Joined Nov 21, 2020
14
You can use a crystal on the MCU instead of AC LINE frequency if you wish but the time-keeping ability will not be as good. I can put together a circuit for you if you wish but you will have to do some assembly.

We suspect that the coil takes very little current. More importantly we would like to know what is the lowest voltage at which the clock will still run.
Unfortunately, I wouldn't know how to test for the lowest voltage. Also, as the motor requires a physical "kick start", often several attempts, I wouldn't know if it wasn't running because the voltage was too low, or if it needed just one more kick to make it start. A circuit drawing would be great...a friend of mine might be able to help me through it. I'm looking for the KISS principle here...plug and play. I apologize if I'm being annoying...I'm really in the dark here, and you folks have given me the best hope so far of getting this fixed.

#### djsfantasi

Joined Apr 11, 2010
7,211
I agree with you on accuracy but it requires a connection to AC mains, my solution could run from a DC supply. Some crystal references are temperature compensated, I have used a DS3231 RTC with an Arduino and it is good to less than two minutes a year. Your design is probably more accurate but I was thinking of avoiding a connection to mains voltages.
Oh, sure... change the requirements. Considering you want to avoid a mains connection, I like your use of a RTC along with the Arduino.

You can still benefit from a self-correcting algorithm when generating to 50Hz. There are plenty on the Internet, including the web site I linked to in a previous post.

#### MrChips

Joined Oct 2, 2009
22,505
Unfortunately, I wouldn't know how to test for the lowest voltage. Also, as the motor requires a physical "kick start", often several attempts, I wouldn't know if it wasn't running because the voltage was too low, or if it needed just one more kick to make it start. A circuit drawing would be great...a friend of mine might be able to help me through it. I'm looking for the KISS principle here...plug and play. I apologize if I'm being annoying...I'm really in the dark here, and you folks have given me the best hope so far of getting this fixed.
Do you have a way of programming any MCU if I send you the file?
Otherwise I can send you a chip or the chip already mounted on a circuit board.

#### s14rs4

Joined Sep 15, 2016
53
This is a quote from wikipedia, maybe this is what you have, or it could be because you are running it from 110v when it was designed for 240v. (see Wikipedia: ELECTRIC CLOCK )

Spin-start clocks
The earliest synchronous clocks from the 1930s were not self-starting, and had to be started by spinning a starter knob on the back.[9] A flaw in the design of these spin-start clocks was that the motor could be started in either direction, so if the starter knob was spun the wrong way the clock would run backwards, the hands turning counterclockwise. Later manual-start clocks had ratchets or other linkages which prevented backwards starting. The invention of the shaded-pole motor allowed self-starting clocks to be made, but since the clock would restart after a power interruption, the loss of time would not be indicated.

#### Bottleguy

Joined Nov 21, 2020
14
Do you have a way of programming any MCU if I send you the file?
Otherwise I can send you a chip or the chip already mounted on a circuit board.
I have no way of programming the MCU. How much would you want for the circuit board / chip? Also, if it's a simple case of "solder the hot wires directly from the outlet here, solder the motor wires there" I can handle it.