OK, I think I understand what is happening with the last schematic. Basically move one side of the component voltage wise and the other side moves too.

So here is my next approach, I don't know if it will work or not, but it is worth a try.

When the discharge goes low it will discharge the coil, bringing the output high. Without C1 this would be a stable condition, with C1 it is a crude monostable, which turns off after a short amount of time and leaving L1 discharged. L1 starts to charge again until it hits the upper 2/3 Vcc point and repeats.

 

Here is the simple inductance meter
Lx = 1mV = 1uH

Or simpler version for L = 0.1uH..50uH
http://forum.allaboutcircuits.com/attachment.php?attachmentid=22606&stc=1&d=1284461383

 

Vary the freq to change the range? You know I have to adapt that to a 555/556 don't you?

 

Vary the freq to change the range?

Yes, for calibration/ and change the range.

You know I have to adapt that to a 555/556 don't you?

I expected this

 

what about RF filter, ferrite beads? I seem to run into them on lots of junk equipment AC inputs. They would be much better to wind a coil on, no?

 

It is not the winding, it is the measuring after you've wound them. Ideally I would like to come up with something for the AAC experiments page, besides being darn useful in and of itself.

 

I tried breadboarding this,

It was a total flop. I have some serious thinking to do.

I wonder why it didn't work? Both of the relaxation oscillators shown in the attached image work (in simulation, at least; I didn't bother breadboarding them).

 

Really? What did it show for the input waveforms?

My current theory is the coil would move the voltage on both ends the same amount, then slowly equalize. The drive on the 555 may also be too weak.

I haven't given up, but it is on the back burner.

 

OK, another thread the problem came up, and I had a bright idea. This looks much more promising overall.

Inductor-based comparator oscillator?

I really think this one will work, but there are some minor bugs to work out. R2 is the range, and also controls the current through the coil, a major variable. I mentioned this on the other thread, but I need to allow time for the coil to relax through CR1 somehow.

I'm picking this up on this thread since this was also my project. If I get something worked out I will post the results on the other thread.

 

I have drawn up the following experiments. The first is to answer some questions how the 555 handles odd states with its internal flip flop. The second is to see if the 555 can do some of the odd internal states it need to do the LR oscillator simply. If it can't then I will go another way.

...................Experiment 1

......................................Experiment 2

 

Experiments 1 and 2 are complete, and the results are what I was hoping for.

...................Experiment 1

......................................Experiment 2

The next step is to build the circuit. Here is what it will look like. R2 can be adjusted so a simple 1:1 correspondence between the frequencies period and inductance exists, as in 100µs = 100µH. The real measurement will be the time measurement of the positive part of the square wave, but I suspect the approximation of the period will suffice.

 

It looks workable.

Your Vdd will have to be regulated, L / time will be based on Fixed Vdd through R2, but L \ time is based on the fixed diode voltage.

If you put a series resistor with the diode, the \ ramp can be made to approximate the / ramp. You could tune it easy on the scope so / and \ periods match?

 

I'll put pictures up from my new (to me) scope when I have something.

I had a throught I was going to put on your thread.

You can ignore the flyback altogether (though the diode is still needed). Use a signal from the PIC to trigger the MOSFET, then time the duration of the pulse on the output of the (the negative side using a 555). The duration of this pulse is a direct measurement of the true inductance. I expect the flyback duration to be very short, so it may not matter either way.

The Vcc isn't important, as the 1/3 and 2/3 set points of the 555 should be the same. I suspect the frequency formula will be the same. We'll see.

Your kids remember Father's Day?

 

...
You can ignore the flyback altogether (though the diode is still needed). Use a signal from the PIC to trigger the MOSFET, then time the duration of the pulse on the output of the (the negative side using a 555). The duration of this pulse is a direct measurement of the true inductance.
...

Yeah the current up ramp / period should be a good indicator of the inductance if there is sufficient current to get the magnetic properties of the inductor working properly.

That was the problem I had, the PIC was able to do very accurate period measurements but the 5mA from the PIC comparator was not enough for the inductors to do their job properly. Depending on core type the period was hugely different!

...
I expect the flyback duration to be very short, so it may not matter either way.

Actually wouldn't it be longer than the / ramp? Operating like a SMPS buck the duty would be the inverse of Vin:Vout and Vin is high and Vout is 0.6v.

...
The Vcc isn't important, as the 1/3 and 2/3 set points of the 555 should be the same. I suspect the frequency formula will be the same. We'll see.
...

Hmm, what I said was in reference to your comment of "just using the whole period" and in that case the period of / depends on Vin and the period of \ depends on fixed Vout of 0.6v so the total period will change with Vcc.

If you put a resistor in series with that flywheel diode the \ period becomes MORE like the / period but will never be fully linear due to the fixed voltage loss of the diode.

I think your initial judgement is correct you need to measure just the period of the / ramp.

As a solution to the diode issue, you could use 2 FETs, as a push-pull to produce both / and \ current ramps. If the FETs were balanced (and act as resistors) the period would become linear to inductance.

Actually, if you changed the circuit be a buck converter and used 2 FETs (to simulate the switch and the diode) the voltage on the output resistor could trigger the 1/3:2/3 points. The output resistor is included in / and \ periods so oscillation will be 50:50 duty AND have a combined period linear to the inductance...

 

OK, I'm going to use this schematic, since I find 555's easier to work with. For one thing I have the pins thoroughly memorized.

I'm going to use a logic level MOSFET to allow increased parameters for the power supply.

The variables will be

Power Supply Voltage: 6V, 9V, 12, 15V

R2: 10Ω, 100Ω, and 1KΩ

Measurements will be simple, Freq Out.

A couple of scope shots to verify it is working the way I think it ought.

I have a rough selection of coils I'll be trying out, most in the 100µH to 1mH range. I'll try some outside those ranges.

The diode will be a 1N4007. I have literally thousands of them. I'll probably try a Shottky later just to see how much difference it makes.

If you have any suggestions let me know.

You know, if this doesn't work I am going to be highly aggravated.

It wouldn't be the first time.

Oh, BTW, I bought a LC meter a couple of months ago. It will be my transfer standard.

 

It will be interesting to see the results. You should do duty cycle too, even a rough measurement.

I just made some mods to your schematic to show the "synchronous buck" version, it uses 2 FETs instead of a FET and diode. Current path is always through one FET and the load resistor R2, and with the 555 timer thresholds it should oscillate with 50:50 duty with the voltage on R2 going from 1/3 Vcc to 2/3 Vcc.

That should give quite a linear response of inductance to period.

As a synchronous buck the bottom FET acts like the "diode" as it turns on when the top FET is off. You may need to fiddle FET gate resistors a bit to make sure there is no shoot through. (also invert D and S on the top FET as I just cut'n'pasted the symbol!)

 

have never given up on this project. here is my latest brain fart.

 

I am unable to bread board for quit a while. any bored techies out there?

 

Hi Wendy, It is good to hear from you again. I can't offer to breadboard, but I can offer to help on the layout so maybe someone local can plug the wires in.

Here is a recent project done with Eagle (not quite like Fritzing, but it includes schematic capture):


Unfortunately, I only have 2 signal layers in my version, so I assign red to existing traces in a typical solderless breadboard and blue to added wires. Of course, I have to break those rules for readability and obvious exception. X's are connections that one might assume from the Eagle package, but are not. In a later version, I use an X for no connection and a short Bar for a cut trace.

If that will help, let me know, and I will send it off to you.

Best regards, John

 

Under LT SPICE the output of the NE555 stays high and the PNP stays on. I imagine this is not the desired result.