Claimed OU circuit of Rosemary Ainslie

[MileHigh]

Poynt and Glen:

It is highly significant that your two sample waveforms are nearly identical.  Glen is using a breadboard and Poynt is using the Groundloop PCB.  The fact that your two sets of data are a very close match, and from an engineering perspective are essentially identical, merits a flood of positive emoticons and beer mugs clicking and all that good stuff.  Let the bells ring!

It's actually pretty amazing that your waveforms match and you have different layouts.  You are in a position to replicate each other's data if you choose to.  It's.... un-beelleeebbibble...  I am choked up!  lol

Glen you get a gross of Brownie Points for doing a clean build that looks as good as a real PCB.

MileHigh

[poynt99]

The test setup.

I will be getting a power supply tomorrow that will allow me to test with 24V.

.99

[MileHigh]

Poynt:

Thank ya v'ry much.

Glen:

Let me try to help you and others with the thermal analysis.  In looking at your data, you could see that the temp was nearly always about the same.  Let me explain the nuts and bolts of it and give you some suggestions.

Firstly, let's just eyeball your load resistor and make a guesstimate as to how long it would take the temperature to stabilize.  Just from real life experience, a life of experiencing hot coffee cups and turkeys and pots on the stove, and looking at the material it is made of, it's shape, and how it is designed to conduct heat away from itself by being vertically oriented, I will guess that it would take about 10 minutes for the temperature to stabilize.

Going back to your posting on EF, #2875, you see the increasing exponential waveform for the blue Channel 1 trace?  That's a standard exponential curve.  The temperature of the resistor follows that same curve, but just stretched out over about 10 minutes (we guess).

So the trick is to know how long it takes for the temperature to stabilize.  More precisely, we make a compromise and say when it is at 99% of it's maximum temperature because the last 1% takes a very long time.

Do you have a variable power supply?  I hope so.  You make a precise measurement of the load resistor with your best multimeter, and then connect it across the variable power supply and dial the voltage so that it dissipates say, 30 watts.  The wattage does not matter at all, but 30 watts sounds good.   Measure the temp of the body of the resistor in the center of the shaft and always measure in the same place.  You don't need to look for the hottest spot on the resistor, just use the same spot.  You make an initial temp measurement on the center of the body of the resistor before you apply the power.  This is your true "ambient" temperature.  All of the temperature increases are relative to this temperature.  Recording the room ambient temperature would not hurt also, but the true reference temperature is on the body of the resistor before any power is applied.

Then apply the 30 watts of power and note the temperature every minute or so.  You only objective is to get a handle on when you are at 99% of the max temp.  Here is a scenario:  You apply the power and see it start leveling off after around 8 minutes.  At 9 minutes you are at 130 F.  At 10 minutes you are at 130.3 F.  At 25 minutes you are at 131 F.  So you look at that and decide that after 9 minutes you are close enough to the final temperature to use that data.

So, now that you are armed with this information, for the rest of your testing, all that you have to do is two things to be in good stead, 1) record the resistor body "ambient" temperature, and 2) just wait 9 minutes and record your final temperature and you are done and can move on!

The big caveat that is easy to abide by is that all of this is based on keeping the resistor suspended in air, you can't change your setup.

Now it is easy to do the thermal profiling.  Aaron did his as temperature vs. DC voltage, and you could see a slight curve in his trend line which makes sense because power is proportional to the square of the voltage.  Personally, I would do a temperature vs. power graph, and it looks like it should be a nearly straight line.

So, knowing your precise resistance of your inductive resistor, you can easily calculate what the precise power supply voltage should be for 10 watts, 20 watts, 30 watts, etc.   Every time you change the power supply voltage, you know that you only have to wait 9 minutes to get a very accurate final temperature.  If my guesstimate of 9 minutes is right, then in a few hours you will have a deadly accurate thermal power dissipation profile graph for the rest of the testing.  Every time you try a new waveform to heat up the load resistor, just wait 9 minutes and get the final temp to get the thermal power dissipation.  In the mean time you could be dumping your DSO waveforms into your spreadsheet and crunching the power levels there also.

Once you have that "delta temperature relative to initial body temperature" vs. power dissipation graph loaded into Excel and then printed out on a paper for you, you would be able to go gangbusters.  You don't always have to record the initial resistor body temperature for the start of each experiment, you could simply record it at the start and then use that as the reference point for a series of runs, one after the other.  The assumption is that you have normal air circulation in the room where you are doing your testing so the ambient temperature of the air in the room itself is not going to be affected by your presence (60 watts) and the lighting and the action that you have going on with the coil resistor.

If you became a well oiled machine, you could start a test and then get the waveform dump during the 9 minutes that you are waiting for the temperature to stabilize.  Once you get the data, start another test right away, change the waveform, supply voltage, whatever you want to do and fire it up and wait 9 minutes again and record the temp.

Now to relax a bit, in the real world you should be able to make two or three measurements per hour where you can crank out a real-world COP measurement, the real thing.  The key is to have your temperature vs. power graph first, and then you can go completely crazy.

MileHigh

[MileHigh]

Just a few more comments for the sake of completeness.

I am also assuming that the weather is stable while the testing is done, no big warm or cold fronts changing the air temperature too much.

Also, the Holy Grail is the suspended load resistor and the ambient air currents around the load resistor.  You want these to always be nominal and normal.  That means you can't decide to switch on a fan by your bench because you are getting hot, etc.

The thermal resistance is modeled by the load resistor floating in a sea of air that is slowly sauntering by with a certain average velocity.  This velocity should stay the same because it directly affects the "heat peel" off of the body of the resistor.  Changing the "heat peel" changes the effective thermal resistance to the outside world.  You don't want the thermal resistance to change at all.   This might affect the final temp by plus or minus a few degrees.

A more drastic and obvious change would be to block the lower opening of the "smokestack" preventing air funneling up and through the resistor tube.  This would change the effective thermal resistance drastically and the final temp would probably be 20 or more degrees higher as compared to the standard setup.

This may all sound like I am splitting hairs, but it should in fact be easy to keep the ambient conditions very stable for everybody.  The important thing is to be aware of the issues.

MileHigh

[Rosemary Ainslie]

Rosemary:

Keep hoping for that COP!  My eighth impression is that everything looks normal.  In other words, it was startling to see the big reverse current spike lined up with the big positive voltage spike in Glen's waveforms, it looks like that should not be happening.  So I had to iterate over and over in my head to finally figure it out.  That's all part of the fun!

MileHigh

Hello MileHigh.     5 almost consecutive posts and climbing?     There are those of us on both sides of this aetheric pond  who speculated that you'd be turned on by that resistor arrangement.       Personally I was only disappointed in one relatively small aspect of your reaction. 

Rosie   

EDIT BTW I read through your advice to Fuzzy on thermal analysis.  I think Fuzzy is looking to the battery draw down rate as part of the required analysis.