Driving a 60Hz clock

[PJL]

I have just picked up a USA clock radio and want to see if I can get the clock to work on 50Hz. Does anyone have a circuit using a multiplier/divider before I start to play.

One way would be to divide MSF by 1000, you would need a pll to get the carrier first as it drops out for the data bits. It could use a 60kHz crystal, then you would have a fairly accurate backup during mantainence periods.

[PJL]

It's a telechron (GE) and I just tested it and it does work though needs a clean. It already has a replacement rotor unit which is bodged in a bit but does the job.

There is a 50Hz rotor unit around but seemingly very rare ($120). I wonder what european clocks have it in?

[PJL]

This is what I am thinking.
Multiply by 6 using a VCO/PLL chip locked to 100Hz
Divide by 10 using something
Use a 240-120 transformer for both clock & radio
Bridge rectify for the clock
Drive the clock from a square wave via a capacitor tuned for series resonance at 60Hz

Hopefully not too many parts and minimal heat

[Darren-UK]

Originally Posted by PJL

It's a telechron (GE) and I just tested it and it does work though needs a clean. It already has a replacement rotor unit which is bodged in a bit but does the job.

How long did you test it for? Just a thought, but I don't suppose the bodged replacement rotor is a 50Hz one? After all, it's been changed for a reason.

[PJL]

I did for a few glorious minutes think it might be a 50Hz unit but checked as it has a secondhand and this is confirmed by a mark of 60C. It looks like Sony were using Telechron movements in the USA around the 60/70's. Maybe they were used in Europe too.

[neon indicator]

Cheap Standard crystals with Integer division to 60Hz
1.843200 MHz / 307200 Awkward as it's 3 x 10 x 10 x 1024
2.457600 MHz / 409600 2^12 = 4096 (or 2^10 = 1024, 2 and 2) , 10 and 10, good choice
4.332000 MHz / 722000 (Not a PAL crystal!)
4.915200 MHz / 819200 2^13 = 8192 (or 2^10 = 4096 and 2) 10 and 10, good choice
6.0 MHz of course
6.144000 MHz / 1024000 2^10 = 1024, 10, 10 and 10

I think some of these are Baud Rate Generator crystals.


I have my crystal stock in a spreadsheet with divisions and harmonics listed in table!

/ 10 is a standard chip 74LS90 or dual decade is 4518
the CMOS 4040 chip is / 4096

So a 2.457600MHz is a good choice
Or 4.915200
/2, /10, /10 and /4096 = 3 ICs
4th IC as crystal osc or buy a 4.915200MHz osc module. You can get TCXO modules quite cheap and they are very good.

[valveaudio]

You can generate 60Hz by using a 7.864320 MHz crystal (available from www.surplectronics.com) and dividing it by 131072 (17 bits) using 2 x CD4040
The CD4040 is a 12 stage counter which can be driven directly by a crystal.

You will need to filter this with a simple low pass or band pass filter to obtain an approximate sine wave and then amplify it to feed a step up transformer.

A TDA2030 or TDA2050 should do the job as an amplifier, it really depends on the transformer you use and how much power you need to run the clock.

Trevor

[neon indicator]

Will a CD4040 work as high as 7.8MHz?

With the 4.915200 Osc or crystal I'd use a 1/2 a 74HC74 first giving 2.4576MHz, below the normal 2.5MHz limit (I think) for CMOS. Then 4518 dual decade and finally 4040 12 bit counter.

Low pass RC filter..

A reversed mains transformer driven by TDA2030 as suggested...

[Ed_Dinning]

Hi PJL, you may find that a divider type of circuit will output a square wave that will cause the clock parts to "rattle" badly. A divider from a clock crystal, then fed through an simple
L/C filter and a small audio amp should work fine.
I presume that it is 110v, if fed a square wave this should only be about 100v amplitude or you may overheat the coil. All related to form factor (sine 1.11; square1.0)

Ed

[valveaudio]

CD4040 is speced to 8.0 MHz so should be OK but take care with wiring layout and supply filtering.

I have used this configuration to drive a synchronous turntable motor using a 4MHz crystal.

Trevor

[Andy Doz]

Use a PIC with appropriate timing instead of all that tedious dividing down?
One crystal, one device ... then on to the amp and shaping to suit..

Good starting place: http://www.sm0vpo.com/power/pic_dc-ac_01.htm

[Kat Manton]

Personally, I'd go with something locked to 50 Hz mains as it'd keep time better (long term) than a crystal oscillator derived source.

But, if you're happy enough with 60 Hz derived from a crystal oscillator, here's the circuit for a turntable PSU for inspiration. It does pretty much exactly what you require, using a crystal oscillator, divider, filter and amplifier to drive a small motor.

You'd still need some of it (filter and amplifier) if you derived a 60 Hz square-wave from a VCO/PLL/divider locked to 50 Hz.

Kat

[Kat Manton]

Also, this article may be of interest: http://sound.westhost.com/clocks/freq-changer.html

That looks like quite a good way to get 60 Hz from 50 Hz (or 50 Hz from 60 Hz) without using a VCO and PLL. I might try that method myself (for something other than a clock.)

Kat

[PJL]

Thanks Kat. Interesting but the component count is quite high and I suspect the VCO approach might be more temperature stable.

[Kat Manton]

There shouldn't be any issue with temperature stability with Rod Elliott's frequency changer as the 60 Hz output is generated from the 50 Hz input, not an oscillator. The UK mains frequency doesn't, AFAIK, drift much with temperature...

A 60 Hz synchronous electric clock run from this will keep time as accurately as a 50 Hz clock run from UK mains.

A VCO/PLL circuit would need two chains of dividers (rather than one) to produce 60 Hz (for the clock) and 50 Hz (to compare with the mains) from the VCO, making it more complicated.

Generating harmonics of 100 Hz, filtering the 6th harmonic out, then dividing that by 10 is actually a lot simpler (and quite elegant IMO.)

However the 60 Hz signal is produced, it's likely to be a square wave at 60 Hz at a low voltage. It'll need filtering and amplifying in order to drive the clock motor smoothly without noise/vibration, so a filter and amplifier are going to be required regardless.

FWIW I think a hybrid of Rod Elliott's frequency changer and the amplifier from the turntable PSU is as simple as it gets. It won't even need a mains transformer if it's powered from rectified mains as per the turntable PSU.

Kat

[jimmc101]

If you can handle PICs then there are some interesting ideas here
http://www.romanblack.com/one_sec.htm
'50Hz to 60Hz sinewave converting inverter' dated 9th Dec 2009
(just over halfway down the page).

Jim

[Andy Doz]

This will give you some idea of the frequency stability of the UK mains:

http://www.bmreports.com/bsp/bsp.php#chartelement