OK, start with this circuit...

I used my DVM freq counter and verified it is easy to adjust and will hit 60 Hz. The pot adjustment is really slow (a good thing).

How can I use the lights 60Hz to calibrate this? My early idea (which I haven't tried yet) is to use a CDS cell (aka LDR) to switch a transistor and beat it against this oscillator. The LED will slowly blink, the closer to 60 Hz it gets the slower it will blink. This way, a total noob with no test equipment can calibrate this to 60 hz.

Does anyone know how fast a CDS cell responds to AC? I would also like any other ideas that might apply, keeping parts cheap and parts count low.

 

You forgot something...

 

Are you talking about the CDS cell circuit? I haven't drawn it yet.

 

Don't lights flash at 120Hz (in the US), because of the positive and negative peak in the cycle?

 

Yep, fluorescent lamps flash at 120Hz.

CdS cells won't be anywhere near fast enough.

 

A CdS cell may be just about fast enough (I've heard they're good up to 200 Hz or so for small changes) but really, a solution would be to measure the mains frequency directly, not the lights, because what if you have a CFL (these flicker at 1kHz+) or an LED lamp (flashing at whatever frequency they feel like with a SMPS or even at 100-120Hz if they're cheap and only have a resistor)...

One way would be to measure the signal from an el-cheapo wall wart that puts out DC. These wall warts have a lot of ripple, at around 120Hz (in the US) and 100Hz (elsewhere).

I can think of a circuit like this, as attached.

This has the added bonus of also getting its power supply from its frequency source, so the 555 can be powered from this. The 5V regulator will hopefully prevent the filter capacitor on the 5V side interfering with the frequency measurement. There cannot be too many filter caps on the input, because then the circuit will not work, as it is measuring ripple. Also, I am not so sure how stable the regulator will be with an input from as low as 10V to as high as 14-15V.

 

Upon simulation I have decided it would probably be best if the above circuit used an open collector instead of an emitter follower. Also, the circuit must be given about half a second (from applying power) for the capacitor on the input to zero out. Before then, it does not output a useful signal.

 

It may complicate things, but one of these light-to-voltage converters is a lot cheaper and faster than a CDS cell - http://www.taosinc.com/Productfamily.aspx?id=1&SD=ltv

 

Have a look at the attached.

A 6.3v secondary transformer will give roughly 7-8v DC out when lightly loaded. I used a fairly large cap instead of a regulator, just to keep the circuit simple.

D5 outputs a half wave; so 50 or 60Hz. I decided to use a 20k pot instead, to give more finagle room in case the cap tolerances are sloppy (they frequently are).

The two transistors basically create a discrete AND gate.

The 555 timer will need a 0.1uF (100nF) metal poly or ceramic cap directly across it's power pins. The 470uF (or larger) cap should be located within a couple of inches of wire length from the 555.

 

Many people do not use old-fashioned incandescent light bulbs anymore, nor old-fashioned fluorescent tube lights. Instead they use compact-fluorescent-light (CFL) bulbs that oscillate at 40kHz, not 60Hz and not 120Hz.

Bill, are you making a 12VDC to 120VAC inverter? Then use a CD4047 oscillator with a internal divider giving a perfectly symmetrical square-wave and it has two opposing outputs, instead of the 555.

 

A silicon pv cell has fast enough response for mains x2.

 

A silicon pv cell has fast enough response for mains x2.

Tim,
Bill Marsden tries to base his circuits from items that might be purchased by hobbyists from local suppliers, such as Radio Shack stores which are prevalent here in the States.

Radio Shack doesn't have much of a supply anymore, and most of it is quite high-priced, but it's a place for hobbyists to start.

I have some of the Radio Shack CdS cells; they are quite slow, and vary in their response time. I would not recommend them for the kind of test that Bill is trying to accomplish.

The circuit that I posted should be more than adequate to synchronize a 555 timer to the mains frequency. It will not remain accurate, but will be better than nothing.

If accuracy is required, a crystal oscillator is a huge improvement over the 555.

If even more accuracy is required, a TCXO (temperature compensated crystal oscillator) or OCXO (oven controlled crystal oscillator) will prove to be far more suitable; a 1000% or better improvement.

 

I remember filing off the top of a metal 2N2222A. That might be an alternate method of doing it, it becomes a cheap and dirty phototransistor with a really short lifespan. During the week my job and life keep me too busy to do much except post, weekends are where experiments get done.

I'm going to try a CDS and an incandescent bulb just to see, it's a question I haven't asked before. I realized Beenthere's point long after the fact, it is a pulsating 120Hz signal.

It will be for the inverter ultimately, whatever I come up with, if I come up with a workable solution. Maybe I'll use a flip flop, maybe I'll try a long and a short 555 monostable to trim it down to a narrow 60hz pulse based off of light.

Audioguru has a point, an inverting output is a good thing for this. Add a quick RC integrator on each output and a schmitt trigger and I'd have excellent shoot through protection.

 

Hello,

In the old days I used an OC71 as phototransistor.
You scraped the black lacquer from the glass and it works.

Bertus

 

How about an LED for a light sensor? Unfortunately I didn't find too much information when I had a quick search, but it should be possible.

 

How about an LED for a light sensor? Unfortunately I didn't find too much information when I had a quick search, but it should be possible.

Yep, LEDs could be used. You'd need to amplify their output quite a bit.

Bill, I think you'd be a lot better off using a couple of opamps to generate a 50Hz or 60Hz sine wave reference, and use a couple of comparators to test the output voltage vs the reference. You'd basically wind up with a class D type amp.

Have a look at the attached. It's kind of rough yet, because R9 and R15 violate the isolation of primary vs secondary. However, it's not a bad start.

V1 could be simply an inverting supply rather than a battery. I just used a voltage source to keep things simple.
V2 is a single 12v lead-acid battery.
V3 represents 3 lead-acid batteries wired in series. You wind up with V+ being roughly 50v when the batteries have an 85% charge. Using a higher voltage primary reduces switching/transformer current requirements considerably, along with decreasing losses in wiring. You can't run much of a load with just a single battery anyway.

U1a and R1-R3 and C1-C3 are a ring oscillator; output is around 600mV p-p. Frequency can be adjusted via R3 alone without too much distortion penalty.
Instead of an LF353, just about any voltage feedback opamp could be used like a TL081/TL082/TL084 or TL071/TL072/TL074, except I wouldn't recommend a 741.

U1b amplifies the ring oscillator's output to ~7.5v p-p, centered around 0v.
U3a simply inverts the output from U1b to provide an inverted reference signal for the right-hand side.

LM2903's are dual comparators, basically LM393's with higher temperature rating. LM393's could be used, or a pair out of an LM339. Basically, whatever comparator you have handy should work. Don't use opamps unless they have extremely high bandwidth; they'll be too slow.

R10 and R17 provide hysteresis. Increasing them to 1 MEG wouldn't be a bad idea; that would increase the average PWM rate.

Q1 through Q4 are used as MOSFET gate drivers.
R13, R14 allow the comparator output to sink to Vee without drawing excessive base current. The way it is now really needs to be fixed, but I wanted to finish it up quickly. A diode clamp to limit the excursion to -0.7v would work there.

M1 and M2 are N-ch power MOSFETs with a low Rds(on). I really should've used 100v MOSFETs here, as the TVS suppression diodes don't absolutely clamp at 54v like they're rated. There are no gate resistors; they're needed for ringing suppression and a fail-safe in case the transistors croak.

L1 through L4 are a 1:5 broadband transformer with center tapped primaries and secondaries. I used perfect coupling, because being more realistic causes the simulation to take 10x longer.

C4/L6 and C5/L5 are low-pass filters.

 

I actually plan on the Class D approach for my final circuit. Remember though, all we have is +12VDC. I will also point out the resemblance to a UPS to a inverter.

Parts count is also critical, but nothing is written in stone yet.

I'll look into the LED approach. Even if I get a 120Hz signal I think I can beat them successfully together and get the calibration effect I'm looking for.

When I get done with this I'll pick back up on the inverter thread.