Ldanielrosa

The zener approach sounds simpler and less expensive. Have you considered using an LED for the voltage limiter? One of the older die-type (not the newer film-type, they have higher voltage cost more and are less forgiving of static) in green will limit to 1.6-2.0V, depending on the current and the manufacturer. It won't take much more space and is likely to be much cheaper than a proper zener diode.

That's one way. I haven't looked at a lot of backlights, but those I've seen are series-parallel arrays two long. The 2*20 modules I have use 42 LEDs.

I neglected to label the anode and cathode, but this should help.

For SPST analog switches, there are 4016 and 4066. There are also analog line selectors like 4051, 4052, 4053. Following one of these with a unity gain amplifier should fit the bill.

I was involved in some discussion on H-bridges a little while ago. Before I start repeating myself, read a little from http://forum.microchip.com/tm.asp?m=64492 and http://forum.microchip.com/tm.asp?m=56838 . As for this application, Do you need isolation or do the motor supply and the low voltage supply share ground?

If you send a square wave through an integrator, you should get a triangle out. Sending that through a second integrator should give you something a sine wave with pointy peaks. I think the Exar chip ends the quest here. One more stage may improve the wave noticably. You'll have to correct for gain vs. frequency, but the concept is there.

The voltage you're reading on the neutral is proportional to the current, so there's your current probe. The only problem is that everything that's downstream of that particular breaker will contribute to your reading. LIke the image you posted- it looks like there's electronic equipment charging a capacitor (the lumps), and a motor or lamp dimmer making the spikes (maybe fluorescent lamp). If you want a better idea, read across a low value resistor.

I recommend against having one reference dependant upon another. This can cause a lot of frustration when trying to adjust them. The power supply isn't likely to notice the current required by the references. And using an opamp as a unity gain follower should do a fair job of buffering the references.

GreekPIC, Go with your namesake (PIC). There is an appnote on this very subject at http://ww1.microchip.com/downloads/en/AppNotes/00843a.pdf . You may even be able to do it with one of the 14 bit cores. The PWM will almost certainly be done entirely in software, considering the change in duty cycle. I'd recommend not using a stamp, not from my own experience but from what I've heard. With the real time interpreter going they can't keep up with the task at hand. If you go to http://www.tinaja.com/ and put "magic sinewaves" in the search box, you'll be able to find some articles on PWM sinewaves.

I've been considering an HV probe for my DMM, but it would only be good to 20kV or so. My meter is not 10M input impedance, it varies. One time I tested it indirectly and found it to range from 9M5 to 11M. With this in mind I'd probably put it in parallel with a load resistor of 1M or less to limit the error.

I'm doing something similar with a 16c73. It'll have limited range, but will have a serial interface as well as direct and LCD.

I've mentioned the fellow before, but here's a link to the circuit you describe. http://www.romanblack.com/a04.htm You may have to change Rs to get the current you want.

There are a few variants of a switchmode regulator at http://www.romanblack.com

As for the disposal, first neutralize what you're planning to throw away. One good way to use up it's oxidizing capacity is to drop in scrap aluminum, but do it outside away from anything you value. Once it's inactive flush it down the toilet, never the sink. Flush a few more times. For software I've used MSpaint to make the image, but something else (I forget what) to print because MSP gives no control over print pitch. Yes, it is a good technique for beginners. There are no toxic chemicals until you're ready to etch. If you find that the transfer is incomplete it's a cheap mistake to fix, just clean it off and burn another copy.

Audioguru, I think Hotwaterwizard means for it to be pronounced 'Sheik', though it is unlikely to find oil barons.

More than half the people I've asked said they made an LED blink. So did I.

I don't know if I can attach two images to one post, so here's the other idea I have. This one does not invert, and is a bit more forgiving of supply voltage. The first stage (Q1 and Q2) steers the current away from the second stage bases. The quiescent current, IRc, will flow when the output is high-Z (this buffer's chief liability in my opinion). This one also requires that the input come within a half volt of the supply rails to guarantee reliable operation. Let the opinions roll in!

I'm sorry about the delay, I had a couple other things to attend to. This one is inverting. The output has resistors in the emitter legs, which are optional. I thought that for multiplexed LEDs, some current control in the buffer would be good. If this is not desirable, the resistors (Re) and the diodes (unlabeled) should be reoved. The zeners will probably be LEDs (common, cheap) but leave the device with a narrow operating voltage range. Ideally the 'zeners' should have a voltage drop more than half and less than all the supply voltage. I haven't yet built one. Once I do I'll let you know what I find out.

I'm interested in a parts minimal tri-state output buffer. No enable pin, just high-OC-low output from the input signal. I'm inspired by the capability of uCs to multiplex outputs, but find the use of support chips wanting. For applications like clocks and gauges the uC will spend a lot of time doing nothing, so the space weight and power used doesn't seem justified. I have a couple ideas I'll submit once they've been drawn. I'd like see what others have been using.

I went looking for such a device at my local hardware store. I sure had trouble explaining the function and purpose. I finally found a vacuum cleaner with one built in, but didn't feel like buying that day. Just the same, a slave outlet would be handy. MP and Ante, I would and will use a method similar to what you suggest- I have on hand a few current transfomrers (1000:1 ratio). When I get mine done I'll let you know how well (or poorly) it went.

There are some legacy character generator chips out there, I saw a spec sheet for one a few years ago. You're going to have an easier time programming your own set on an EPROM of some sort. You're going to have to come up with some addressing hardware though. I am curious. What issues forbid a microcontroller for this project?

Let me add to that. 4. Automotive batteries do not take well to being fully discharged. Ideally you should use a "deep cycle" battery. The internal construction is slightly different. If you never use it longer than necessary to save your work and shut down, then this won't be a problem.

Moltra, I'm going to have to rain on your parade. Reconditioning NiCads is temporary, and some say of debatable value. Two failure methods come to mind (I'm no expert)- 1: the electrolyte was cooked dry, 2: the electrodes grew dendrites. The second one can be fixed, but it won't last. The price per launch is probably much higher that the battery. If you can't open the explosion-proof casing without damaging it, you can still do a failure analysis- 30 samples will give a (barely) valid statistical profile. If you have weight to spare (unlikely) and run out of power before all the data collecting ability is used up (entirely possible), you might consider sending up a reconditioned battery as well (with a diode in series, this will protect the main battery from secondary failure but will only start using secondary power when main power is down 0.7V).

A detector will tell you how noisy an appliance is, but will not let you measure the power being consumed. You may be better off trying to find a non-invasive way of sensing the current through the power cord, though I suspect that will be difficult and you will have an easier time plugging the appliance into (through) a current monitoring device.

I'm sorry 12412, I don't particularly care for it. It makes me directly part of the circuit. I was hoping for (have seen once, and hope to find again) one that uses capacitive coupling.