Claimed OU circuit of Rosemary Ainslie

[TinselKoala]

Joit:

It's hard to not like you when you make statements like that.  You are a happy shiny person.

When logic and proportion
Have fallen sloppy dead
And the White Knight is talking backwards
And the Red Queen's "off with her head!"
Remember what the dormouse said:
"Feed your head
Feed your head
Feed your head"

MileHigh

So, for example, according to Joit the Cakeman, when I take a 12 volt battery, run a small VDG machine with it, and use the output of the VDG to charge a capacitor to 60,000 volts or so, that's beyond classical understanding. Gee, where do I apply for my Nobel, I sure could use some of that Danish beer.
And my Jacob's ladder, using essentially the same Ainslie circuit with a secondary on the "load inductor", which is making 20,000 volts at least, with pretty good current, is only running on 12 volts too. Darn, I wish I understood this mysterious process better.

[MileHigh]

TK:

I would take relatively new but not brand new battery and discharge it under heavy load with an appropriate load resistor.  When you hit the turning point and the voltage starts to drop, then switch over to a lighter load resistor, and then say stop the discharging at 11.5 volts.  That is your consistent methodology for creating a "discharged battery."  That is your reference datum point.

Supposing the battery holds 1 Megajoule of energy.  The next step is to hook the discharged battery up to a charger and charge it for a fixed amount of time, TIME being the variable that you have complete control over.  You charge your battery with 800 Kilojoules.

To figure out how to put 800 Kilojoules of energy into the battery you have to experiment with timed-charge-discharge cycles where you measure the energy that you can extract from the charged battery vs. the amount of time you have it hooked up to the charger.  That is a whole separate investigation.  To measure the discharged energy without fancy tools, you take voltage measurements across the load resistor every 10 minutes and punch that into your spreadsheet.

Finally you arrive at a point that you are quite certain that your battery has 800 Kilojoules of energy in it if you measure it as per your standard protocol.

So then you do a form of run-down test where you hook up the device under test and run it for a certain amount of time.  Again, we are working with time here.  You make a rough estimate of how long the device under test should run to dissipate 400 Kilojoules, 1/2 of the energy that you "know" is in the battery.  I forgot the error tolerance, say it's +/-3 percent.

Then you disconnect the device under test, and then hook up your battery to your standard discharge protocol and measure how much energy is left in the battery.

It may sound complicated and it is real work.  Starting off with a battery with a known amount of energy in makes the most sense to me.  It also makes sense to stay away from a potentially nonlinear "fully charged" battery state state.

Then you do a "burn" with the battery on the device under test.  Then you burn off the remaining energy in the battery to calculate how much energy was consumed by the device under test.

This whole process stays away from the potentially nonlinear fully charged state and stays away from the definitely nonlinear nearly depleted state of the battery.

The only time you really measure a voltage is at the end of the test when you have to discharge the battery to see the remaining energy in it.  You are in a Catch 22 here, you have to define some sort of method to determine that you have a "discharged battery", your reference datum point.

So those are my ideas about the issue.  I confess lest I be struck down in my tracks by a Tesla lightning bolt that I have no experience doing it.  I am just applying my knowledge instead of going into the woods to look for lime and tin and whatever else so that I can make my own batteries.  You have to be a doer!  lol

Anyway, I am not a fan of batteries as I have stated before.  To make the test "real", one of my favourite words, you would have to do something like 20 runs of the same test to get more accurate data just by averaging your data.

The real way to "navigate" is to use the DSO for the battery output power calculation and the thermal profiling for the thermal calculation.

You can't forget that on top of all of the battery baggage that goes along with the run-down tests, an equally big hurdle is to make a truly accurate measurement of the power consumption of the device under test so that you know precisely what load resistor to use for the control test.  That is another big +/-X% variable that could easily muck up the whole works.  You really have to work hard on that one.

MileHigh

[jibbguy]

Hehehe, i never said you were incompetent, although i have often thought you were married to a real "bitch" (...meaning conventional mainstream science)

Regarding the "triple nickles", What you can do if you try hard enough, is to not properly use them (...not use the correct value ranges in the support caps & resistors). But that's not the same thing  

They could possibly have been made more robust within the last few years... But i doubt it. IC's tend to just go on as ever, because engineers are a Conservative lot and don't like their stuff being screwed around with; as it can cause lots of woe with the older circuits still out there, often in very unexpected ways. Usually another ID number will be used if specs are significantly altered; and there are several similar packages of oscillator/single-shots on the market with slightly different specs & ID no's.

[MileHigh]

TK:

You don't need integration for this test.  The DSO does the weighted I*V waveform dump.  The only thing that you have to do in Excel is calculate the average waveform value and then multiply it by the conversion factor to turn it into the true average power consumption over X cycles.

jibbguy:

Its also possible that when the first field tests on transistors were done and the first real-world circuits designed around them, done in secret by contractors to the US Defense Department in the 1950's and early 60's long before they were available to the public (...they had them for nearly 10 years before the general public was allowed to), that this effect was noted then... And sat on; either ignored or suppressed.  In fact, these early days would probably be the ONLY time it would be noted and properly studied in all these years.

You are dreaming here.  Engineering academics would certainly be more comfortable than front line engineers in explaining these phenomena.  If you look hard enough you will find hard core analog engineers that understand this stuff also.

You are burying analog oscillation effects in mysticism, and nothing could be further from the truth.  The reasons that you get spurious oscillations is that the "s-plane transfer function" for a given analog circuit has "poles" that lie in "the right half of the s-plane."  There is no point in saying any more but rest assured this kind of stuff is understood inside-out.

I fully agree with you though that conventional analog engineers had their heyday in the 40s, 50s, 60s, and to a certain extent the 70s.  On the other hand, the people that design high-speed computer boards and design ASICs have in a sense become microwave/analog engineers.  They are much different from "old school" analog engineers, but high-speed digital and ASIC design is all about PCB transmission lines and whatnot because the frequencies are so high now.

MileHigh

[Hoppy]

IMO there is very little chance that Aaron will use the full capability of his DSO at the real risk of learning the truth about the real power in his Ainslie spikes. Anyway I'm signing off for a week so I hope that on my return that this Ainslie COP17 Heater fiasco will be cooling off.

Hoppy