Claimed OU circuit of Rosemary Ainslie

[poynt99]

For those that may be curious how my testing has been going...

It's always good practice to verify measurements more than one way if at all possible, especially when the measurement is critical. This usually means comparing one piece of test equipment with another of similar or different type (i.e another scope or meter), or by using some other indirect means (such as current clamp).

In my case so far I have not been able to sufficiently validate my data, and not for lack of trying. For example Supply current measured 4 different ways produces 3 different values:

Scoped (data dump) Vshunt (ave)/0.25 = 60mA
Fluke DC ammeter used directly with internal shunt = 60mA
Fluke DC voltmeter across 1 Ohm shunt in supply "add-on" = 76mA
Scoped mean differential volts across 1 Ohm shunt in supply "add-on" = 120mA

SPICE is very close to 60mA supply current with the same settings.

So that's 3 measurements at 60mA, but I can not explain at the moment why the scope differential measurement is double. Nor can I explain why the DC volts across the 1 Ohm shunt (with two 10000uF caps) reads 76mV (and hence 76mA). I have verified that the meter reading is not being influenced by induced switching noise.

I have a TDS3012B oscilloscope I will try also to see how the differential measurement comes out.

Once I am able to overcome this challenge, I will be confident in my numbers.

Regards,
.99

[poynt99]

Yet one more way of measuring the supply current seems to confirm the 60mA number.

Two voltmeters, one before the add-on, and one after yield 25.06V and 25.00V respectively. A 60mV difference.

It's very odd that measuring directly across the 1 Ohm yields 76mV with these same meters.

Tried the TDS3012B oscilloscope, and the results were worse. It's becoming apparent that these two TEK scopes don't do simple math (A-B) and averaging (on these wave forms at least) very well in real time.

Don't fret, I think I have the solution at hand. Good results should be coming up soon.


.99

[MileHigh]

Dipping my toes in the water!

Glen:

How many samples are you recording per cycle?  I am wondering if you are getting too granular in your sampling per full cycle because your scope captures look like you are recording 80+ cycles.  I think that there is an issue there with undersampling exaggerating the energy in any thin spikes (positive and negative) in your waveform.

If you are a real keener you can look up "Nyquist sampling theorem."  - actually, I'll take the plunge:

The Nyquist frequency, named after the Swedish-American engineer Harry Nyquist or the Nyquistâ€"Shannon sampling theorem, is half the sampling frequency of a discrete signal processing system.[1][2] It is sometimes called the folding frequency, or the cut-off frequency of a sampling system.[3]

The sampling theorem shows that aliasing can be avoided if the Nyquist frequency is greater than the bandwidth, or maximum component frequency, of the signal being sampled.

The Nyquist frequency should not be confused with the Nyquist rate, which is the lower bound of the sampling frequency that satisfies the Nyquist sampling criterion for a given signal or family of signals. This lower bound is twice the bandwidth or maximum component frequency of the signal. Nyquist rate, as commonly used with respect to sampling, is a property of a continuous-time signal, not of a system, whereas Nyquist frequency is a property of a discrete-time system, not of a signal.

This is the biggie:

Nyquist rate, which is the lower bound of the sampling frequency that satisfies the Nyquist sampling criterion for a given signal or family of signals.

Now for a simple example:

To sample a 10 Hz "pure" sine wave you need to sample at 20 Hz.  The reason for this is:

This lower bound is twice the bandwidth or maximum component frequency of the signal

However, this is for the "ideal" case.  In the real world you need to sample at about five times the Nyquist rate, which is 100 Hz.  This will give you a good description of the signal for calculating the energy in it.  (This assumes the "signal" is the highest frequency component of the overall signal which consists of "slow" and "fast" signals mixed together.)

Going back to what you are really doing in your experiment, in theory you need to know the waveform's frequency spectrum to know the maximum bandwidth of your signal.

That's the theoretical and practical background to all of this.  Since you are trying to make precise energy measurements it is actually very important.

Now forget the theory and how do you really do this with your DSO?  You look at your thinnest spikes and you want to have at least say... seven or eight samples minimum to properly record the energy in the thin spike.  You can just look at your Excel data dump to confirm this.

Alternatively, of you set up the DSO time base so that you only see a few complete waveforms on the display and record a data dump.  Then you find a thin spike in your data dump and see how many samples were recorded to see if you are sampling the data fast enough.

There is another way to do this.  You look for the sharpest rising or falling edge in your waveform and measure how much time it takes to go from 20% to 80% of the rise (or fall).  Then you look at your data dump and see what the time step is.  If your time step is such that you make about five samples during the fastest rise time in your waveform, then you should be fine also.

As you can imagine, the more and more complete waveforms you record per data dump, your sampling per each full cycle becomes grainier and grainier.

Well, I didn't expect to do all that, but for anyone with a DSO-type digital waveform recording device, you can apply this Nyquist sampling concept to any measurement you are doing.

MileHigh

[poynt99]

Problem solved.

I have restored faith in the Tek scopes, just not so much in the manual's recommendation for NOT calibrating your passive probe. I did anyway, and it made a significant difference. I now have the expected 60mV across the 1 Ohm add-on resistor as confirmed by 3 other measurements. Apparently, the scope can do the math in real time, provided some settings are set just right.

So onward and forward.

Indeed Glen, MH's post is excellent. I have had no problems with this as I am very aware of the issues. I would recommend you increase your sample rate quite a lot; 10MS/s minimum.

.99 

[fuzzytomcat]

Dipping my toes in the water!

Glen:

How many samples are you recording per cycle?  I am wondering if you are getting too granular in your sampling per full cycle because your scope captures look like you are recording 80+ cycles.  I think that there is an issue there with undersampling exaggerating the energy in any thin spikes (positive and negative) in your waveform.

If you are a real keener you can look up "Nyquist sampling theorem."  - actually, I'll take the plunge:

This is the biggie:

Now for a simple example:

To sample a 10 Hz "pure" sine wave you need to sample at 20 Hz.  The reason for this is:

However, this is for the "ideal" case.  In the real world you need to sample at about five times the Nyquist rate, which is 100 Hz.  This will give you a good description of the signal for calculating the energy in it.  (This assumes the "signal" is the highest frequency component of the overall signal which consists of "slow" and "fast" signals mixed together.)

Going back to what you are really doing in your experiment, in theory you need to know the waveform's frequency spectrum to know the maximum bandwidth of your signal.

That's the theoretical and practical background to all of this.  Since you are trying to make precise energy measurements it is actually very important.

Now forget the theory and how do you really do this with your DSO?  You look at your thinnest spikes and you want to have at least say... seven or eight samples minimum to properly record the energy in the thin spike.  You can just look at your Excel data dump to confirm this.

Alternatively, of you set up the DSO time base so that you only see a few complete waveforms on the display and record a data dump.  Then you find a thin spike in your data dump and see how many samples were recorded to see if you are sampling the data fast enough.

There is another way to do this.  You look for the sharpest rising or falling edge in your waveform and measure how much time it takes to go from 20% to 80% of the rise (or fall).  Then you look at your data dump and see what the time step is.  If your time step is such that you make about five samples during the fastest rise time in your waveform, then you should be fine also.

As you can imagine, the more and more complete waveforms you record per data dump, your sampling per each full cycle becomes grainier and grainier.

Well, I didn't expect to do all that, but for anyone with a DSO-type digital waveform recording device, you can apply this Nyquist sampling concept to any measurement you are doing.

MileHigh

Hi MH,

Thanks for your comments I do really appreciate them ..... first I rehung the load resistor vertically as you originally suggested because of what appears to be stray RF or some type of magnetic  interference that I was able to see on my TV set next to the RA COP>17 circuit when it was horizontal ..... it made my TV have weird radial lines in it about 1 1/2" apart so I turned it back again vertical and it solved the problem     .....  second the wave form data acquisition method sounds interesting and it seems that .99 is in some agreement there, so I will look into it further and see what I can come up with, as you indicated earlier this is a "complex wave form" ..... not your every day sine wave by any means ..... 

Fuzzy