View Single Post
Old 03-29-2010, 05:55 AM
witsend witsend is offline
Gold Member
Join Date: May 2009
Posts: 1,881
And here's a preliminary revised evaluation of the tests taken on Glen's replication

TEK00000 - 0.00 WATTS
TEK00002 + 3.59 WATTS
TEK00004 + 2.46 WATTS
TEK00006 - 0.87 WATTS
TEK00008 - 0.04 WATTS
TEK00010 + 4.91 WATTS
TEK00012 - 0.17 WATTS
TEK00014 - 0.09 WATTS
TEK00016 - 0.71 WATTS
TEK00018 - 0.42 WATTS
TEK00020 - 0.12 WATTS

average performance over the entire test period was, therefore 0.822 watts.
Heat dissipated was an average of 5.5 Watts.
THEREFORE COP > 6.23 OR 669%

No heat profile was conducted on the mosfet arrangement and the attached heat sink nor on the shunt resistor. Yet there was clear measured evidence of heat being disipated at both points. Temperature measured at both points was higher than evidenced at the load. If these values were also factored in as wattage dissipated at a conservative three quarters of the wattage at the load then the actual wattage dissipated as heat would have been 5.5 * 75% plus 5.5 being 9.62 Watts. This would place COP at anything between 6.23 and 11.7.

It is not an ideal test evaluation. Ideally the test should be run and data logged continuously over an extended period of time. But this was not possible given the equipment available. However, what is significant is that there were any records at all of negative wattage delivered. This should not be possible according to mainstream predictions. Yet it is clearly a predominant condition of this circuit. It is also evident in the occassional recharge of the battery which is not typically known to climb when run under load conditions.

What is signficant is that any value of wattage that is recorded to be greater than COP 1 is quite simply not expected within classical prediction on this circuitry. And measurement of temperature rise is widely considered to be sufficient proof of the wattage dissipated. Therefore, in all cases and under all possible conditions, this circuit is capable of exceeding classical prediction even under the most conservative assessment. And this using classical measurement protocols.

These measurements were confined to multiple waveform sample range which is required for 'proof' of data. Faster time scales or fewer samples fall outside the range of accuracy required by Tektronix.