Need to convert 240v 50Hz - 110v 60Hz (newbie)

Thread Starter

malcolmt

Joined Dec 3, 2008
12
CVM
I kind of see what you mean, I am slightly more newb than yourself, I shall now investigate H Bridge.
Malcolm
 

SgtWookie

Joined Jul 17, 2007
22,230
I had to go out for a few hours.

Basically, a UPS puts out an approximation of a square wave. The output transformer is driven in a series of small steps (on;off;on) by an H-bridge. It isnt a perfect sinewave, but it is close enough for running computers and small appliances from.

But, something like that could be done using a microcontroller that was synchronized to the 50Hz signal. It wouldnt be absolutely perfect either, but as long as it output six sinewaves per every five on the input side, it would be as accurate as could be reasonably expected.
 

Thread Starter

malcolmt

Joined Dec 3, 2008
12
Hi Sgt

That sounds do-able and very interesting, I guess now is the time to learn about microcontrollers or at least the basics, Our 50Hz supply is very accurate so i am thinking that i need to multiply by 1.2 to achieve my goal.
Do you have any thoughts as to how i go about doing this, I am not asking you to provide a design service but would appreciate a nudge in the right direction.
Many thanks
Malcolm
 

RiJoRI

Joined Aug 15, 2007
536
Ah! I was thinking of someone's beloved clock-radio!

The idea that first came to mind was to get a 240VAC/50Hz motor, and connect it to a 120VAC/60Hz alternator. Don't know if this would be practical, however.

--Rich
 

peterb33

Joined Jun 10, 2009
10
I am working on the same problem as Malcolm who originated this post, I would also like to get a low power 60Hz 120v supply going to drive a clock.

I have tried various circuits with limited success as suggested in the posts, it really needs to be a sine wave as square waves make it very noisy. I have driven various transformers from amplifiers including an emmiter follower from a good 60Hz sine wave source and I can produce an ac wave of good shape and voltage until I connect the clock to the secondary of the transformer when it then becomes very distorted on my scope. The clock motor is around 1k resistance and rated at 2.5w.

Anyone any suggestions?

Thanks Peter
 

Bernard

Joined Aug 7, 2008
5,784
This project sounds familiar, early '70s "we" had 7 crystal controlled master school clocks, using 50Hz motors. In spite of ovens they either gained or lost a few minites a month, resulting in many lost tech. hours. I multiplied 60 Hz X ? and divided it down to 50 Hz. Worked fine but the details escape me now. Used another frequency dividing system knoun as locked oscillators, where a cristal oscillator synchronized lower frequency oscillators; As Sgt. suggested, might multiply 60 Hz with a locked oscillator then divide it down to 50HZ. Three OP amps can convert square wave to triangle, then to sinewave, followed by power amp.
 

thingmaker3

Joined May 16, 2005
5,083
What IC chip back from the 30's was able to increment time from the 60hz i'm intrigued by this, timing from the hertz cycle of the power supply I always assumed they all had some kind of crystal / oscillation routine...

how's the circuit detect the hert rate? any ideas and would be as simple as modifying the circuit to trick it into going faster.. something new :O
hso what on earth do you have in a 1930's clock?...
Back in ye olden days of yore, electric clocks used a mysterious contrivance called a "motor." This "motor" turned at a rate very precisely proportional to the line frequency. (Lots of AC motors do that. It's a motor thing.) To get the "hands" (clocks had "hands" back then, instead of LEDs or LCDs or plasma displays) to turn at the appropriate ratios, they used a complex mechanical arrangement called "gears."
 

eblc1388

Joined Nov 28, 2008
1,542
This is an interesting project.

In order for the 60Hz-clock to be accurate running on 50Hz AC power, the "generated" 60Hz must be "locked" to the AC 50Hz. The power station controller at night would purposely adjusts the grid frequency up/down to compensate for the net frequency drift during the day so any electric clock would seldom need timing adjustment.

For this project to work, there are three main parts.

1. the frequency multiplier, which multiply the 50Hz into a frequency several times higher. 600Hz is a good target as one can derive 60Hz from it easily.

2. the frequency converter, this convert the output of (1) into 60Hz AC low voltage

3. the amplifier or buffer that would boost the low voltage 60Hz AC into line voltage suitable to drive the clock motor.

For (1), it is straight forward to use 50Hz zero crossing circuit to get pulses of 100Hz. Multiply this by 6 using a PLL+counter is simple and easy. The 100Hz zero crossing pulses is fed to the phase locked loop CD4046 phase detector#2, along with the divided-by-six signal(using CD4018) from the output of the 600Hz VCO. The error voltage from the detector is then filtered and use to control/regulate the VCO to oscillate exactly at 600Hz.

We are then left with a 600Hz square wave "synchronised" to the 50Hz. Using the following ring counter circuit from Don Lancaster's CMOS CookBook, it is easy to turn it into 60Hz. The resulting 60Hz is surprisingly good, after using just a simple RC filter, as can be seen from the FFT waveform.

The ring counters formed using three CD4013s can be replaced by a single chip CD4018, which is also a 5-stage Johnson counter. However, there isn't a 4018 SPICE model available so I choose to build the ring counter using 4013 instead. Therefore part (2) can be done using a single cheap CMOS IC and few resistors.

For part (3) it is simple and easy for me but not for everyone else. I have and so can use an high voltage Opamp PA85 with +/- 200V DC supply and immediately I have the required voltage output, with a load current of 200mA or less. This is sufficient to drive the clock motor. However, the PA85 is not cheap. I have checked APEX(now Cirrus Logic) site and the PA240CX is more affordable choice.

Without using high voltage Opamp, one have to implement voltage buffer using ordinary low voltage Opamp. This too is not difficult to do.

I hope this would be helpful.





 

Attachments

Last edited:

peterb33

Joined Jun 10, 2009
10
Thanks for your reply.

I did consider an approach like yours but I lack the tecnical depth to come up with a workable solution. It looks very neat.

What I have got working is a 60hz xtal timebase followed by an active filter which gives a decent sine wave out. I thought this should be good enough to try but the problem I have is getting up to 120v to drive the clock. Your high voltage ic is ideal but unfortunatly out of reach for me, do you have any further suggestions or is there an altrnative to that ic?
 

Bernard

Joined Aug 7, 2008
5,784
Might try an IC power amp like TDA2003, TDA1518BQ, NTE1288, etc mostly operate from 6-18V DC, 8-24W, & feed into 6.3V @ 1.2A AC filament transformer. Amps. run from US $.50 to $ 2.00, transformer ? I just reach into my junk box and pull one out.
 

eblc1388

Joined Nov 28, 2008
1,542
Your high voltage ic is ideal but unfortunatly out of reach for me, do you have any further suggestions or is there an altrnative to that ic?
There are voltage booster circuits on the web but many of them are too complicated for this application.

I managed to come up with the following circuit which seems to produce an acceptable output, albeit it with a little crossover distortion due to its simple configuration. It would be more than sufficient for driving your clock motor. Your clock motor drive voltage will not affect clock timing while frequency would. One can always get away using lower than required drive voltage.

The high voltage DC supply does not need to be regulated and can be obtained simply by using a small transformer connects in voltage step up mode, with low voltage winding driven by the AC of the main transformer secondary.

If you want to use NPN/PNP as final output transistors, you can. See next image. However, the NPN/PNP models I have in LTSpice only goes up to 150V Vceo so its only good for demonstration of concept. For practical usage, one would need to pick NPN/PNP with Vceo of 300V or higher.





 

Attachments

Last edited:

peterb33

Joined Jun 10, 2009
10
Thanks for your reply and circuits. I thought that there had to be a discrete solution when I read the spec for the high voltage op amp. I will need to gather the some of the parts before I can construct anything so a little delay is not too important to me.

Would the mosfet version have any advantage over using npn/pnp transistors in the output?
 

eblc1388

Joined Nov 28, 2008
1,542
Would the mosfet version have any advantage over using npn/pnp transistors in the output?
Not really, unless very high output current is needed. In fact it is more difficult to make it works than the transistor version because of the MOSFET gate capacitance.

The NPN/PNP version is preferred.
 

someonesdad

Joined Jul 7, 2009
1,583
US companies who make products that they want to sell into European markets have to test their products at 50 Hz (for EMI, safety, and other things). Now, this was 25 years ago, but IIRC one of our engineers decided the best way to allow us to do our testing was to use a 60 Hz motor to drive a 50 Hz generator. Obviously, correct pulley sizes were needed to have the proper speed ratios. But if appropriate parts can be found (try a surplus place), it's pretty easy to build and fairly foolproof (of course, remember that fools can be pretty ingenious ;)).

Of course, if you're willing to shell out some granola, google "50 Hz to 60 Hz conversion" and you'll find people willing to sell you turn-key solutions.

For a do-it-yourselfer, I like SgtWookie's suggestion of multiplying the frequency by 6 and dividing by 5.

BTW, I once made some rough measurements of the AC powerline frequency over time and its time average is very stable over hours to days -- to something on the order of 0.01% IIRC.
 
Top