"Perfect" OU tests?

[Enigma516]

I would like to put together a suite of tests and a description on the tests / how to. So that we as a group can be using similar tests (hopefully in stages). These tests should be some combination of scientific tests, practical tests, and incremental steps leading up to a result.

Example (but not likely good tests)

Test 1: measure voltage out and voltage in
Test 2: measure amps out and amps in

etc... I would really like to nail this down so that I can test some devices and designs. This would give us all a set of metrics to compare.

[Madebymonkeys]

I would like to put together a suite of tests and a description on the tests / how to. So that we as a group can be using similar tests (hopefully in stages). These tests should be some combination of scientific tests, practical tests, and incremental steps leading up to a result.

Example (but not likely good tests)

Test 1: measure voltage out and voltage in
Test 2: measure amps out and amps in

etc... I would really like to nail this down so that I can test some devices and designs. This would give us all a set of metrics to compare.

I am happy to help with this.
Power in vs power out is key although measuring current and voltage is only the start of it....for an AC system the phase needs to be considered (search 'power factor' on Wiki). In fact, if you aren't sure wether an output (or input) is pure DC then power factor needs to be considered.

That's it basically - a good PF meter (around a couple of thousand £ or so) should do the trick. Don't be fooled by off the shelf kill-a-watt type devices, they assume some level of compliance with the law and line voltages etc - you need a decent PF meter as used in compliance test houses (the type of test labs who test for (for example) CE compliance or VDE etc).

Many OU designs involve heavily inductive or capacitive elements which all effect PF, there can also be switching elements which inteoduce harmonics - only a high quality power quality meter can give a true reading.

Let me know if you want me to suggest some or point you in the direction of a lab with facilities (in the UK) to test.

MBM

[Enigma516]

Are there any incremental tests we all could be doing at home without that level of expensive equipment? Obviously the end goal would be to send it to be tested at an independent lab with that level of test equipment

[TinselKoala]

I beg to disagree with Enigma's early posts above, a little bit, because what is important is not power in/out..... but rather _energy_ in and out.

It's easy to fool oneself by looking only at power measurements, and there are many examples of this happening right now.

I have SSTCs that have inputs of less than a kW and produce "outputs" of many kW if measured naively. My "tinselkoil 2.0" delivers 7-10 amps at 30 kV or higher, yet is powered by ordinary American mains power, 120V at 7 amps. How can this be? It is because peak power is not the same thing as average power, reactive power is not the same as real power, and proper measurement -- integration of the instantaneous output power over a suitable time period, to yield energy flow data -- is elusive and apparently misunderstood by many experimenters.

If power is the only measurement available, be sure to measure real power by using it in Joule heating measurements, like JLN's coffeepot tests of the induction hob "ou" claims. This kind of measurement, properly done, is a rough quick and dirty physical integration which does eventually yield energy flow data.

[TinselKoala]

Are there any incremental tests we all could be doing at home without that level of expensive equipment? Obviously the end goal would be to send it to be tested at an independent lab with that level of test equipment

There are, certainly, but the specific details will depend on the characteristics of the device under test. For example, JLN's coffeepot calorimetry is a very good test of the _real_ power (as opposed to reactive power) coming out of the bifilar coil coupled to the induction hob cooktop.

Having a high-end DSO do realtime multiplication of voltage and current, and then integrating the result over an appropriate time period will almost always work properly _IF_ the initial measurements are accurately done and the time interval is properly chosen. Examples of doing this improperly can be found in the same JLN work as the coffeepot calo measurements.
This kind of multiplication and integration can also be done manually using analog scopes, as I have shown elsewhere, but the accuracy of the result still depends on good input data, and it seems that this is where many claimants fall off the "straight and narrow" thorny path of truth. Particularly with high frequency circuits, the input measurements aren't actually measuring what they are supposed to be measuring, but rather are contaminated heavily by artefacts, and so... GIGO applies. As I have also illustrated elsewhere.
A simple illustration of what can go wrong, even when using modern equipment on simple circuits:
http://www.youtube.com/watch?v=KWDfrzBIxoQ