Target response tester

[Skippy]

VLF maths:

tan (phi) = 2*pi*f_detector * tau

tau is the target time-constant.
phi is phase lag .... IF you say that a zero-time-constant target has zero lag. Unfortunately is seems commonplace to specify Ferrite as zero degrees, then all non-ferrous start at 90 degrees, up to 180 degrees.

So for example if I measured a US 5c 'nickel' as having a phase lag = 129 degrees = (90+39) degrees with a 13kHz VLF,
then I would calculate target freq f_target = 16.05kHz; tau = 9.9 microsecs, approx.
Similarly, a 'wheat cent' may measure 164.6 degrees lag, so f_target is about 3.6kHz, tau about 44 usec.

...................................

I had an attempt at measuring the %IACS of some lengths of my copper wire this afternoon. The fancy HP34401A multimeter disappointed badly with its 4-wire measurements, jumpy readings, and more annoying, steady readings, that were just plain wrong. The problem is noise or thermocouple issues, made much worse due to the truly miserable test current of one whole enormous milliamp.
So I abandoned that, and did it old-skool, sticking a sensible constant current from a bench PSU through the test sample, and measuring the millivolts drop on a marked length of the sample, and accurately measured the test current (200mA up to 1A as appropriate).
I tested the 0.83mm and 1.32mm wires, about a 1 metre length, and in addition, tested the 3mm (2.99) copper rod that I was unsure about, as it may not have been electrical grade.
Results:
Taken at 24C temperature:
0.83mm wire: C = 58.26 x 10⁶
1.32mm wire: C = 58.65 x 10⁶
'3mm' rod: C = 58.8 x 10⁶
These respectively work out as IACS figures of 101.8% ; 102.5%; and 102.8% after temperature correction.
So it appears what they say is true - normally better than 101%, often exceeding 102% IACS.
It also clears up the 3mm rod mystery, which after some investigation, turned out to be RS components stock number 684-248 , described as very pure, and evidently true.

So that suggests I can knock another 1% off my discrepancy between measured and theory 20mm rings.
I believe the L calculation produces results slightly lower than actual, by maybe one or two %, but pinning that down is pretty tricky.
......................

And incidentally, I have 'discovered' that copper tube as typically used for water pipe, is normally specified as 'C106' grade, Phosphorus de-oxidised. This makes it better for processes like brazing and welding. But .... the %IACS figure is typically 85% - 90%. I found an old post where you tested a square of it flattened out , and got poor results which I picked up on. Maybe this is the reason why?

[green]

And here's two graphs for the '20mm' diameter copper rings. One shows all of them, clearly showing what I assume is skin-effect ( 13kHz VLF used ), and one showing the smaller gauge wires in more detail, to show the decent straight-line.
Twelve wire diameters tested in total:
0.46; 0.57; 0.73; 0.83; 1.02; 1.10; 1.23; 1.35; 1.58; 1.73; 2.01 and 2.44mm.

All were bare copper electrical grade wire, the two smallest were enamelled.
The 'theoretical' figures assume conductivity is 100% at 20C ( 58 x 10⁶ ). The tests were done at a temp of about 17C.

The slope for the smaller gauge wires is about 1.065, ie. 6.5% above theory. 1% of this is down to temperature, so it's really 5.5% above theory.
The other 'errors' I've discounted. The solder joint is maybe 0.1mm thick, roughly equivalent to 1mm of copper, so the loop resistance is increased by (63+1)/63 = +1.5%. But if this is wire with IACS of 101 - 102%, that roughly cancels the solder-joint resistance.
Theoretical inductance values are estimates, but seemingly pretty good ones ... I'll post more about this later.

Wire time constant varied from 8us to 100us. Tried a simulation in spice. I'm guessing time constant is calculated from phase shift. Does the simulation make sense? What is the formula to convert phase shift to time constant?

[green]

So here is the 19mm long wire test graph. Wire diameters used were:
1.10; 1.23; 1.35; 1.58; 1.73; 2.01 and 2.44mm
As the test was done at a temperature of 17C, I used a figure of 101% IACS in the maths. I assumed the wire was 100% IACS at 20C. ( I will have a go at testing some of the wire samples, to see if they really are 101 - 102% as some sources say )
Gotta say, it's the straightest graph I've plotted for a while , and I was expecting the fat 2.44mm wire to show some skin effect and spoil it.
Still, the end result is K = 0.029.

Got some larger diameter wire today. Charted the TC. I think I always get a little higher TC than you. Think it might be because I'm measuring decay Tx off and you are measuring TC Tx on. I need to figure why. Included the charts showing what I think is skin effect, one or two TC before straight line decay. I'll try to make some rings next.

[Skippy]

And here's two graphs for the '20mm' diameter copper rings. One shows all of them, clearly showing what I assume is skin-effect ( 13kHz VLF used ), and one showing the smaller gauge wires in more detail, to show the decent straight-line.
Twelve wire diameters tested in total:
0.46; 0.57; 0.73; 0.83; 1.02; 1.10; 1.23; 1.35; 1.58; 1.73; 2.01 and 2.44mm.

All were bare copper electrical grade wire, the two smallest were enamelled.
The 'theoretical' figures assume conductivity is 100% at 20C ( 58 x 10⁶ ). The tests were done at a temp of about 17C.

The slope for the smaller gauge wires is about 1.065, ie. 6.5% above theory. 1% of this is down to temperature, so it's really 5.5% above theory.
The other 'errors' I've discounted. The solder joint is maybe 0.1mm thick, roughly equivalent to 1mm of copper, so the loop resistance is increased by (63+1)/63 = +1.5%. But if this is wire with IACS of 101 - 102%, that roughly cancels the solder-joint resistance.
Theoretical inductance values are estimates, but seemingly pretty good ones ... I'll post more about this later.

[Skippy]

So here is the 19mm long wire test graph. Wire diameters used were:
1.10; 1.23; 1.35; 1.58; 1.73; 2.01 and 2.44mm
As the test was done at a temperature of 17C, I used a figure of 101% IACS in the maths. I assumed the wire was 100% IACS at 20C. ( I will have a go at testing some of the wire samples, to see if they really are 101 - 102% as some sources say )
Gotta say, it's the straightest graph I've plotted for a while , and I was expecting the fat 2.44mm wire to show some skin effect and spoil it.
Still, the end result is K = 0.029.

[Skippy]

"try more wires side by side ?"
No, I wanted consistency - all my 19mm lengths are 19.0mm with filed flat ends. Besides, the smaller ones give little useful data, bearing in mind I'm plotting Diam² , all the sub-1mm wires just crowd into the bottom end of the line. If anything, more wires in the 1.9mm to 2.3mm range would be more useful.

It also occured to me that consistency may be useful if they were needed for other tests, eg. signal strength measurements. I could also borrow a better meter from work if I wanted to measure the small signal better ... or knock up a x10 etc pre-amp.

[green]

[QUOTE=Skippy;252776]
Yes, I made some 19mm long straight wires. The bad news is the thinner wire ones give too little signal to get any accurate measurement. I was also expecting the thickest ones to also give misleading readings due to skin-effect. However, I did get 4 points on a straight-line graph, and you won't believe this , but they match the equation:

TC = 0.029 x %IACS x Diameter²

/QUOTE]

Have you tried more wires side by side for a larger signal? Doesn't seem to effect time constant with my tester and amplitude adds.

[Skippy]

I don't have an actual 50 usec ring, that was just an estimate of the point where measurements start to deviate from a straight-line, and it's probably more like 45 usec.
The actual readings near that value are for wires of diameter: 1.23mm, 1.35mm and 1.58mm, measured and calculated values for these were:
D = 1.23mm ; Theory TC = 39.15us ; Measured TC = 41.05us
D = 1.35mm : Theory TC = 46.1us ; Measured TC = 47.7us
D = 1.58mm : Theory TC = 58.9us ; Measured TC = 59.05us

The best-fit line for the smaller-diameter wires ( up to 40us TC ) shows: Measured TC = 1.06 x Theory TC. This could be 1.05 if test temperature ( 17C ) was taken into account. But you can see the 1.35mm wire is dropping off, being 1.035 x Theory, and for the largest 2.44mm wire measured about 0.84 x theory ( 99.6us vs. 119us ).

I only have one really thick wire loop, the 2.44mm, the next one down is 2.01mm, the other 10 continue down to 0.46mm.

[green]

I haven't completed my promised plots, as I wanted to get some extra wire sizes to fill in some gaps, particularly for larger diameter gauges. I have been successful , I've made three new rings and four 19mm lengths, and they are great fits, with the exception of the 3.00mm diameter '19mm' which is odd, more investigation needed on that one, but I don't think its pure copper, it seemed a bit 'springy'.

But the '19mm' TC values are excellent, a cracking straight-line, showing TC = 0.029 D² %IACS , not a hint of skin-effect.

The rings also chart great up to about TC = 50 usec , where measured drops down vs. theory. I assume this is mostly caused by Resistance increasing due to skin effect, though Inductance is also slightly affected by it.

Looks like you are getting good results. 2mm wire is the largest I've tested, will get some larger wire next time to store.

What size wire were you using for the 50usec TC ring? http://www.geotech1.com/forums/attac...2&d=1547342322 chart I posted above, AWG#16(1.29mm diameter)tested about 50usec for me and showed some skin effect at start of decay.

[Skippy]

I haven't completed my promised plots, as I wanted to get some extra wire sizes to fill in some gaps, particularly for larger diameter gauges. I have been successful , I've made three new rings and four 19mm lengths, and they are great fits, with the exception of the 3.00mm diameter '19mm' which is odd, more investigation needed on that one, but I don't think its pure copper, it seemed a bit 'springy'.

But the '19mm' TC values are excellent, a cracking straight-line, showing TC = 0.029 D² %IACS , not a hint of skin-effect.

The rings also chart great up to about TC = 50 usec , where measured drops down vs. theory. I assume this is mostly caused by Resistance increasing due to skin effect, though Inductance is also slightly affected by it.

[Skippy]

Some web info on copper wire:
In 1913, the International Electrotechnical Commission defined the conductivity of commercially pure copper in its International Annealed Copper Standard as 100% IACS. Furthermore, since 1913, processing technology has improved, to the point where copper conductors in electrical wire applications can exceed a conductivity of 100% IACS, even routinely reaching 102% IACS."

"The most common high conductivity copper for conducting electricity via wire, cables and busbars is Cu-ETP (Electrolytic-Tough-Pitch).
It has a minimum conductivity rating of 100% IACS, although most Cu-ETP sold today meets or exceeds the 101% IACS specification."

[Skippy]

And a quick straight-line plot of the temperature change on your test plate shows a decent fit, with the exception of the 61C point, which appears to match an actual temperature of 72C.
(Plotting TC^-1 against degrees C , straight-line at correct slope rate)

[Skippy]

All my rings were made by measuring the correct straight length, then forming it around a plastic rod, then fine-tuning with pliers to get the ends matching. With the exception of the 2.44mm wire one, which was made 7mm longer on each end, the cut points scribed on. It was then bent on the former, and then cut /filed to the marks. The last 5mm or so of a loop is the hardest to bend, making it 'over-long' gets round this.
I've since made two additional loops from smaller diameter enamelled wire, to fill in the gap at the bottom end. I'll try and locate some heavy gauge enamelled, it's in a box somewhere ....
Yes, I made some 19mm long straight wires. The bad news is the thinner wire ones give too little signal to get any accurate measurement. I was also expecting the thickest ones to also give misleading readings due to skin-effect. However, I did get 4 points on a straight-line graph, and you won't believe this , but they match the equation:

TC = 0.029 x %IACS x Diameter²

Assuming Cu wire = 100 IACS.
But I would prefer more points on the graph, and maybe if I used the 5" coil I could get better accuracy, so don't take the 0.029 figure as final.

The two additional thinner rings behaved as expected, plotting all but the two thickest rings gave a measured value 7% over theoretical value.
But..... I used a conductivity figure of 58 x 10⁶ = resistivity of 17.24 x 10sup]-9[/sup]
This is for annealed Cu at 20C . If I tested at 17C that makes theory TC increase 1.2%. And is Cu wire 100% IACS? It could be 101%, maybe higher? That would also increase theory TC by 1% . I will have to try out that fancy meter on a length of Cu wire.

I'll post up the plots etc tomorrow.

[green]

Temperature effect on target TC. Cut a piece 45mm square from one of the copper sheets reply#138. Covered one side with masking tape to aim infrared thermometer at for recording temperature. Heated target with a heat gun, placed target above coil and recorded four temperatures as the target cooled.

Skippy, have you done any straight wire tests yet?

[green]

For straight wire tests I was thinking, 19mm(3/4inch)