@TK
 
I do not have an accurate "known low voltage source " at home.  Since the problem appears to be with the Atten Scope and at least one of them is still under warranty, I shall get the manufacturer to check it out with their experts.
 
All my three Atten Scopes showed the crossing 0 reference line behavior.  If it were a bug, the manufacturer would be and should be interested.  They need to fix it or provide a solution for their customers.

As I posted and as PW corrected with proper resistor values, a simple voltage divider made of a couple of resistors in series, and connected to a AA battery or to your power supply set to 1.5 volts, will be just fine, as long as you have an accurate DMM for a crosscheck.

And I believe you have some kind of Platinum Visa card to draw upon. Please spend a few dollars and obtain a voltage reference, like the 35 or 61 dollar boards or the 3 dollar chip that I told you about earlier. Surely you can see that a person who claims OU based on small voltage measurements MUST be able to show that the voltage measurements are correct, and this requires measurement of a KNOWN STANDARD. The only known standard that we have seen is your scope's calibrator output... which cannot be trusted, apparently.

The ATTEN scopes are what they are. The manufacturer may be interested in the poor performance of the one channel on the one scope, but there are reasons why the ATTEN scopes are so cheap and the Tektronix scopes are so expensive.... and one reason is accuracy.

You are getting lots of expert comments here already.

Let me review a few points.

1. The AC vs. DC input coupling issue. You didn't understand the significance or use of this setting and many of your initial experiments and data captures were made using an AC coupled channel, which invalidates that data for power measurements. Hopefully you now understand the coupling issue, how it's done in the scope, and the significance and usage of each input coupling mode.

2. The probe compensation issue. Evidently, much of your data was taken with large probe errors caused by improper compensation adjustment and some kind of large voltage discrepancy. This invalidates _all_ of the data taken up to the point when you checked and adjusted your probe compensation, and even afterwards if that probe/channel voltage problem.... which you _still_ haven't explained or even addressed.... isn't cured.

3. The issue of the directionality of current at the input current sensing resistor. It seems that the "negative" current values were being incorrectly interpreted by you to indicate reversed current flow and hence your "negative" COP numbers resulting. Now I think we've gotten that bit straight, and you realize that negative values here mean normal ordinary current, and power flowing out of the battery to dissipate in the circuit, not the other way around. Many of your recent experimental analyses claiming "negative COP" values, up to the point where this issue was understood by you, are invalid.

4. The issue of the need to adjust the input battery voltage value measured at the probe, by subtracting the (negative) voltage drop at the CVR. This required some review of basic algebra (subtracting a negative number is the same as adding its absolute value) and also means that your earlier, uncorrected data reflect an input battery voltage that is lower than the actual input battery voltage. The proper correction, when made, takes your measurements further away from OU because the true input power is greater than what you calculated and reported. Your data and conclusions up to the point of this correction are invalid because you don't have the correct input power figures.

5. The presently discussed issue of the DC offset, or rather the voltage measurement accuracy of the Atten scope/channel/probe combos. Apparently you shipped a board to .99 that DID show OU performance on your scopes and spreadsheet analysis, but when he tested it using a more accurate oscilloscope he found no OU indications, and particularly not the "reversed" current implied by a zero-crossing measurement. Do you have any reports at all from anyone you've sent your boards to, who have tested them on more accurate equipment than the Atten scopes? We are in the middle of this issue at present so one can't yet say just when or how this issue invalidates your experimental data..... but until we get some measurements from known accurate test kit, we have to take the position that your OU readings are likely due to the inaccuracy of your low-end DSOs. Precision and accuracy are not the same thing, but both are clearly needed to support any revolutionary claims.


Other issues regarding the spreadsheet data analysis are also relevant, but I haven't spent time analyzing them.... yet.

The channel BW limit can be set individually for each channel. But as you can see in the scopeshot below, it is the channel WITH bandwidth limit "on" that has the noisier signal, complete with some small positive DC offset.
The channel vertical settings are inside the red box, and the inverse highlighted "B" indicates BWL on.

I've also attached the Atten manual for easy reference.

BW level OFF for both channels.
 
Noisy channel behavior disappeared.

Here are my DPO4054 BW vs. Noise results, with probe shorted and set on 20mV/Div.:

20MHz ~ 600uVrms
250MHz ~ 1000uVrms
500MHz (Full) ~ 1000uVrms

There is a noticeable visible difference (increase) in the noise floor going from a BW of 20MHz to a BW of 250MHz. This should be as expected.

Of the above readings, about 130uV is DC offset.

Lawrence,

I suggest the following:

1.  Perfom another board test using your scopes just as you normally do.  If your input scope continues to show the input current crossing above the zero line, save the capture and immediately do the following:

2.  Without changing ANY scope settings, disconnect the battery B+ connection and perform another capture.  If a trigger is required, select "line" or "free run" in trigger settings but make no other changes, particularly with regard to the vertical channel setings.  The reading you receive at this point should be zero volts DC or very close to it.  Any measured/displayed voltage would be DC offset.

3.  Again, without changing any scope settings, and while the battery B+ remains disconnected as in the previous step, clip a 100 ohm resistor between the battery B+ terminal and the end of the input CSR opposite the battery B- terminal.  If necessary, move the channel 1 probe which is measuring input voltage to the end of the 100 ohm resistor that is connected to B+ so that channel 1 continues to measure the battery voltage.  The voltage measured on channel two, the input CSR, should be equal to -Vbatt divided by 101 (that is, -14.8mv with 1.5V Vbatt).  This test allows a determination of any channel 2 gain error.

Perform these three tests and then post screen captures of your results.


PW

 

 

God may help those who help themselves... but You, Lawrence, have received a LOT of help, real help, from the people on this thread, notably PicoWatt, .99, and me. Who helps us?
I myself have actually spent a few dollars of my own money and quite a bit of time helping you to understand your own tools and measurements. Imagine your position had we here not pushed and prodded and pleaded and explained and reproduced and tested your claims here in this thread. You would be claiming your extraordinary results, based on uncompensated, uncalibrated, error-prone and inaccurate measurements, obtained incorrectly and analyzed incorrectly, and as soon as anyone competent who received your board hooked it up to a proper oscilloscope they would know your errors.
It's too bad that you already have all these boards out there in the wild.... every one with your name on it, every one with your claims attached to it, and every one of them NOT overunity in any way. Some of us urged caution and restraint, you may recall, but you in your enthusiasm wanted to sow your seeds. It is really a shame that your seeds are all rotten, not what they are claimed to be, and in the final end-- are an expensive waste of time.