Tweet
Hi Friends,
Due to Tony's chimera, progress on the GPZ coil has been very slow coming, but after over a year I've now completed two coils.
I've yet to complete the DOD that I originally set out to make. I decided to make a concentric first as this seemed more straight forward, so I have ended up with two of those, a 10 inch and a 17 inch. The 10 inch is extremely sensitive, air detecting pieces around 0.05 of a gram at about 3 to 4 inches.
I had to construct a vacuum former to make the coil shells. Many thanks to Don Bowers for the many tips he has posted here and elsewhere. Due to my Daughter suffering terribly from allergies, along with endocrine disrupting thermoplastic fumes, I decided to use a 2.2kWh 4 bar radiant heater as the heat source instead of the kitchen oven. I picked one up new very cheaply. I built an aluminum foil lined box and placed the heater at the bottom and the plastic holder fits snuggly over the top. There are air gaps around the bottom of the box to ensure the heater doesn't overheat. I used a 2kW household vac and with this can easily pull 2mm HIPs and 1.5mm ABS, I suspect I could also pull 2mm ABS if I got it hot enough.
The next problem:
Many here familiar with Minelab PIs will be well aware of the exact requirements of inductance and resistance to achieve a reasonably functional coil. My first experiments didn't work out very well in this regard with the Q being way off my objective of 5. I have a 19" DOD with a Q of 5.4 and it's a noisy old deadweight that's so talkative many deep nuggets need to be so large that the stock 14" coil would respond to them anyway.
Fortunately from Stefan's many comments over the years I knew what I had to do, make some of my very own pseudo-Litz wire.
For reference here's one such comment "As you know I make my own litz wire out of 0.2mm enameled wire. The criterion I use is that my range of coils should have the same Q to match the detector. Also the coils I make are to be used on highly mineralised ground as found in the Pilbara region.
When you start to make your own litz wire you realize that the number of strands in the bunch say for a 12 inch coil will be different that say a 21 inch coil to meet the above criterion. I use 30 strands for a 12 inch compared to 34 for a 21 inch as an example. For an elliptical the number of strands will increase as you require more turns for 300 uh. The resistance increases and the Q drops.
If you have good ground then the difference in Q if you use a standard litz wire bunch will not be an issue. If you are going to use a manufactured litz then I would start with this premise, build the coils and see how the coils perform. Don't use a solid single conductor as you are wasting your time." Stefan.
Because the GPZ 7000 is so sensitive I needed to use wire much smaller than 0.2mm to ensure residual currents dissipated fast enough. This required building a wire winder, but first I needed to make a bobbin loader.
Here's a couple of videos showing the bobbin Loader and wire Twister, they're a bit fuzzy, I need to upgrade my phone. They are not quite finished yet I need to install the controllers in boxes, not easy using a QE single handed and the software needs tweaking a bit, but that's for a future rainy day.
I had a bit of trouble getting the tensioning to work on the Twister so bypassed it for now, but the resulting wire seems to work just fine anyway, with a differential strand length of 20 cms over 30 metres, or just over half a percent. I assume given the bandwidth any effects will be minimal.
The Twister motor rotation speed is monitored by a hall sensor and another six pulse per rotation Hall provides rate of travel, this enables fairly precise control of the wire pitch. To digress here, I found that a bucking coil wound with a tighter pitch requires less turns and seems to require much less effort acquiring a fine null point.
https://vimeo.com/620390674
https://vimeo.com/620396557
I started by winding a dummy coil on the former to achieve the target inductance, this gave me the rough length. I could then calculate how many conductors I would need to achieve the correct resistance to give me my target Q of 5. But it is imperative at this point to include the connecting cable as the resistance is not trivial when the total target R is only 0.376 Ohms.
So following on from Stefan's advice I roughly calculated how many conductors I would need to wind the coil and connecting cable, but then tweak it so I would end up a little below the target resistance, then the connecting cables can be fine tuned to achieve a dead on R.
The Twister only has 14 bobbins, but with patience very close cables can be made. For instance if a 20 meter cable of 100 conductors is required, wind 200 metres onto 10 bobbins and twist that together. Next cut the 10 ply into 20 metre lengths and load the 10 bobbins again with that and twist together. It seems like a lot of work but when what you require isn't readily available there's no alternative. Plus with each coil cost 1.5 thousand dollars....
With the TX coil there is an option to buy a readymade pseudo-Litz that gives a lower resistance than required and then the connecting cable can be constructed to achieve the desired resistance. This has the advantage in that the inductance of the wound coil is definite and so the R can be accurately determined to achieve the target Q.
Even the 17" settles at 6 microseconds with a 1k damping resistor, I'm impressed at how fast pseudo-Litz is.
It can be done, although some of my previous assumptions were incorrect, it just takes time.
Cheers Kev
Due to Tony's chimera, progress on the GPZ coil has been very slow coming, but after over a year I've now completed two coils.
I've yet to complete the DOD that I originally set out to make. I decided to make a concentric first as this seemed more straight forward, so I have ended up with two of those, a 10 inch and a 17 inch. The 10 inch is extremely sensitive, air detecting pieces around 0.05 of a gram at about 3 to 4 inches.
I had to construct a vacuum former to make the coil shells. Many thanks to Don Bowers for the many tips he has posted here and elsewhere. Due to my Daughter suffering terribly from allergies, along with endocrine disrupting thermoplastic fumes, I decided to use a 2.2kWh 4 bar radiant heater as the heat source instead of the kitchen oven. I picked one up new very cheaply. I built an aluminum foil lined box and placed the heater at the bottom and the plastic holder fits snuggly over the top. There are air gaps around the bottom of the box to ensure the heater doesn't overheat. I used a 2kW household vac and with this can easily pull 2mm HIPs and 1.5mm ABS, I suspect I could also pull 2mm ABS if I got it hot enough.
The next problem:
Many here familiar with Minelab PIs will be well aware of the exact requirements of inductance and resistance to achieve a reasonably functional coil. My first experiments didn't work out very well in this regard with the Q being way off my objective of 5. I have a 19" DOD with a Q of 5.4 and it's a noisy old deadweight that's so talkative many deep nuggets need to be so large that the stock 14" coil would respond to them anyway.
Fortunately from Stefan's many comments over the years I knew what I had to do, make some of my very own pseudo-Litz wire.
For reference here's one such comment "As you know I make my own litz wire out of 0.2mm enameled wire. The criterion I use is that my range of coils should have the same Q to match the detector. Also the coils I make are to be used on highly mineralised ground as found in the Pilbara region.
When you start to make your own litz wire you realize that the number of strands in the bunch say for a 12 inch coil will be different that say a 21 inch coil to meet the above criterion. I use 30 strands for a 12 inch compared to 34 for a 21 inch as an example. For an elliptical the number of strands will increase as you require more turns for 300 uh. The resistance increases and the Q drops.
If you have good ground then the difference in Q if you use a standard litz wire bunch will not be an issue. If you are going to use a manufactured litz then I would start with this premise, build the coils and see how the coils perform. Don't use a solid single conductor as you are wasting your time." Stefan.
Because the GPZ 7000 is so sensitive I needed to use wire much smaller than 0.2mm to ensure residual currents dissipated fast enough. This required building a wire winder, but first I needed to make a bobbin loader.
Here's a couple of videos showing the bobbin Loader and wire Twister, they're a bit fuzzy, I need to upgrade my phone. They are not quite finished yet I need to install the controllers in boxes, not easy using a QE single handed and the software needs tweaking a bit, but that's for a future rainy day.
I had a bit of trouble getting the tensioning to work on the Twister so bypassed it for now, but the resulting wire seems to work just fine anyway, with a differential strand length of 20 cms over 30 metres, or just over half a percent. I assume given the bandwidth any effects will be minimal.
The Twister motor rotation speed is monitored by a hall sensor and another six pulse per rotation Hall provides rate of travel, this enables fairly precise control of the wire pitch. To digress here, I found that a bucking coil wound with a tighter pitch requires less turns and seems to require much less effort acquiring a fine null point.
https://vimeo.com/620390674
https://vimeo.com/620396557
I started by winding a dummy coil on the former to achieve the target inductance, this gave me the rough length. I could then calculate how many conductors I would need to achieve the correct resistance to give me my target Q of 5. But it is imperative at this point to include the connecting cable as the resistance is not trivial when the total target R is only 0.376 Ohms.
So following on from Stefan's advice I roughly calculated how many conductors I would need to wind the coil and connecting cable, but then tweak it so I would end up a little below the target resistance, then the connecting cables can be fine tuned to achieve a dead on R.
The Twister only has 14 bobbins, but with patience very close cables can be made. For instance if a 20 meter cable of 100 conductors is required, wind 200 metres onto 10 bobbins and twist that together. Next cut the 10 ply into 20 metre lengths and load the 10 bobbins again with that and twist together. It seems like a lot of work but when what you require isn't readily available there's no alternative. Plus with each coil cost 1.5 thousand dollars....
With the TX coil there is an option to buy a readymade pseudo-Litz that gives a lower resistance than required and then the connecting cable can be constructed to achieve the desired resistance. This has the advantage in that the inductance of the wound coil is definite and so the R can be accurately determined to achieve the target Q.
Even the 17" settles at 6 microseconds with a 1k damping resistor, I'm impressed at how fast pseudo-Litz is.
It can be done, although some of my previous assumptions were incorrect, it just takes time.
Cheers Kev
Comment