Menu
What's new
Find's Treasure Forums

Welcome to Find's Treasure Forums, Guests!

You are viewing this forums as a guest which limits you to read only status.

Only registered members may post stories, questions, classifieds, reply to other posts, contact other members using built in messaging and use many other features found on these forums.

Why not register and join us today? It's free! (We don't share your email addresses with anyone.) We keep email addresses of our users to protect them and others from bad people posting things they shouldn't.

Click here to register!

Need Support Help?

Cannot log in?, click here to have new password emailed to you

Question for the tech heads

A

Anonymous

Guest
Hi all,
I am playing here and wanted to run a question by those in the know. I have a conventional PI "transmit" circuit whereby I have 24V being switched via an FET to a 7inch coil of 1ohm with a 300ohm damping resistor across it. I realize that the coil and FET will get hot because I'm pulsing so much current through it however I'm not sure why the damping resistor is getting hot. I will have to compromise somewhere because they go into meltdown! I can't leave it like this for very long. I know in previous posts Eric commented that he uses a 1W metal film for the damping resistor and it gets quite warm. My calculations using ohms law and the power formulas show that there should be 80ma through that resistor for a small amount of time yet the 2W resistor gets very hot. Is this do to the 500V inductive kickback across it? I intend to replace the small coil with a large (40"X40") and use this to find large deep objects. What should be my targets (coil pulse width, wire size, #of turns, etc) for getting the most depth I can? Any ideas, suggestions, and advice is welcomed.
Thanks,
FJ
 
Hi FJ,
Several things may cause the symptoms you are seeing - The FET "Going into meltdown" worries me slightly - it implies that you are not fully enhancing it, and that therefore there is significant power loss in the FET itself.
What actual circuit are you using to drive the FET? What FET is it, what Vgs are you using, what pulse width? How fast is the FET being turned on/off. Is the FET really turned off, or is there leakage resulting in DC heating? (use a 'scope across the resistor to watch the current flow). Have you a protective diode on the FET gate to prevent damage to the drivers, or do they have their own protection? Is their any ringing or oscillation taking place? etc. etc.
BTW. Most of the "techies" tend to hang out on the Tech Forum (http://www.thunting.com/cgi-bin/geotech/forum/view.cgi?action=index&forum=tech).
HTH,
Nicko
 
Thanks for the response Nicko. I posted the question here because the PI experts are over here such as Eric Foster, Dave Johnson, Dave Emery, etc but none of them seem to post much anymore.
The circuit is basically the one Eric posted here once before. The only difference I've made is that the circuit is being driven by a PIC chip. When I say "meltdown" its because I've increased the specs and am pushing it further than its probably capable of handling. For instance, increasing the on time for the coil from 100us to 200us and increasing the pulse frequency. Therefore I expect the FET and coil to get warm, I'm just not sure why the damping resistor is warm. I suspect from the inductive kickback but I'm not sure. If that isn't the reason its getting warm, maybe there is something wrong. I figure the experience of some experts here might point me in the right direction. I'm not an expert, still learning.
Thanks,
FJ
 
It may not be necessary to increase the on time as you have done. The key as I understand it from previous posts is that you only need to supply current to the coil until the coil resistance limits the current flow to some maximum level. Beyond that you are wasting power and no appreciable improvement in performance will be obtained. A scope with a current probe is a handy way to measure it,, though measuring the voltage drop across the damping resistor with a scope should also tell you. A shorter transmit time will certainly reduce component heating.
 
Hi Russ,
I have been pouring over old posts, especially the echange regarding flyback pulses kicking the target (or not) and it would almost seem from what Dave Johnson has to say would re-enforce what you are saying. However, in all the posts before and since that exchance, Eric always makes the statement that a 50us to 100us pulse time is sufficient for small, low conductivity targets but for larger more conductive targets as would be for a 1meter coil detector you would want to increase the coil ontime. He states he used a 5ms on time for Mel Fisher's unit. I was thinking quite a bit shorter because of the heat and current draw from the battery pack.
I guess I'm trying to figure out how to get the maximum depth I can for a large coil detector, utilizing a 24V battery while keeping the overall current consumption at an amp or less. Maybe I'm asking for too much, its all about compromises right?
Am I better off utilizing more turns on the coil than trying to pump more current through it? Am I better off using a small wire guage to get the resistance up or using a larger wire gauge resulting in a lower resistance and putting a current limiting resistor in series? These are the things I'm tryint to get a feel for here before I have to reinvent the wheel.
Thanks,
FJ
 
Thanks for the response Nicko. I posted the question here because the PI experts are over here such as Eric Foster, Dave Johnson, Dave Emery, etc but none of them seem to post much anymore.
The circuit is basically the one Eric posted here once before. The only difference I've made is that the circuit is being driven by a PIC chip. When I say "meltdown" its because I've increased the specs and am pushing it further than its probably capable of handling. For instance, increasing the on time for the coil from 100us to 200us and increasing the pulse frequency. Therefore I expect the FET and coil to get warm, I'm just not sure why the damping resistor is warm. I suspect from the inductive kickback but I'm not sure. If that isn't the reason its getting warm, maybe there is something wrong. I figure the experience of some experts here might point me in the right direction. I'm not an expert, still learning.
Thanks,
FJ
 
Hi FJ,
Yes, the damping resistor will get warm because of the flyback or spike dissipation as well as the coil ON time. The damping resistor is the primary resistor that discharges the coil current. So, what current or energy is in the coil at the time of the shutoff also flows in the damping resistor.
I would expect the damping resistor to get warm. In your case where you have increased the voltage and the ON time, you might want to go to a little higher wattage damping resistor to be safe.
As far as whether you have the ON time too long, just look at the spike signal at the damping resistor with a scope and note the waveform of the FET ON signal. You will see this voltage step to some low voltage and then this voltage will decay. Once this signal levels out and becomes a flat line, the coil is saturated. After that, the current is just heating the coil.
Since you can adjust the ON time, you might do some simple air tests on an iron, silver, or copper object to see what advantage if any you can note between different pulse widths. This will give you some indication of what to expect from buried objects of similar composition.
Personally, I wouldn't use any more on time than necessary. Again, I would do some air tests of objects to see just what differences the longer ON time makes.
I would be curious as to what you find.
I hope this helps.
Reg
 
Thanks Reg, for testing purposes, I have been using a simple 7" coil of 27 turns of wire strapped to a wooden yard stick. I was getting 24 to 25inches on a coke can before I started trying to increase the PPS or the coil on time. I was getting a few more inches for a few seconds and then it decreased considerably, I'm sure because the FET is getting so hot (I called it going into meltdown) and its performance began falling off pretty quick. I don't know if this is a valid test for having the ultimate goal of really using a 1meter coil. I really don't have anything to compare against to know if this range I'm getting is good or if I'm just hitting the limits of this setup. Another words, I shouldn't expect more out of the thing without going to a bigger FET, a better heatsink, maybe 2 FETs in parallel to distribute the load, etc. I guess I'm just looking for some suggestions around this and also suggestions on the coil. Is it better or does it matter if I use more turns of wire on the coil, should I use larger wire with a lower coil resistance and use a series resistor to limit the current or a smaller wire gauge to increase the resistance and avoid using a series resistor? I was hoping someone here has done these experiments already. Thanks to all for your help,
FJ
 
Hi FJ
There are several reasons why you might see an increase in distance besides the increase in current in the coil. One very important thng is the noise factor, which by itself can cause a significant depth loss or increase if the internal noise level is changed. Increasing the voltage has a tendency to increase internal noise levels.
Also, when you increase the current on time, you also will change the decay time, which will then change the sampling time with respect to the final decay. So, the sampling may occur sooner with respect to when the voltage decays to 0V. This can be very important because the sampling now may be such after increasing the current that the sample is actually taken too soon and as the result, the differential amp is in or near saturation, which will clearly cause an apparent reduction in depth capabilities.
If I remember correctly, FET's have a positive temperature coefficient, meaning they will increase in resistance when heated. This will reduce the current flow. Just how much of a change this may be, I cannot answer. I am not sure this is causing the depth loss you are experiencing.
Don't forget that if you are using a battery, the internal resistance of the battery becomes quite important when trying to draw a lot of current. Now, if you are using several batteries in series, then the maximum current flow is really limited by the weakest link, meaning the weakest battery.
You have asked a lot of questions I cannot answer from experience. I can speculate however.
Now, I would like to note that one of my DD coils is about the same size as your test coil and I get greater depth capabilities on a can than you are presently getting. In fact, my distances are in the 30" to 36" range and this is with less than a 50 usec on time.
I am only using 12 V and way less than 1 amp of current to accomplish this. So, as you can tell, there is more to fine tuning a PI than just increasing the current into the coil.
I am mentioning this because I am just trying to point out that there are many aspects that have to be looked at besides just increasing the current if you are looking for more depth. I for one, began reducing the time before the sample is taken, this really helps more than some people realize. I also concentrated on noise reduction and am still working greatly on that one.
From your questions, I would have to guess that you have about the same knowledge as I had when I first got started trying to build a PI. It is fun, yet frustrating because there are more unknowns than what is known, which makes decision making very difficult. I will also openly admit that today, I am no where close to having the knowledge of Eric on PI's, so I am not the most knowledgeable in answering all your questions. However, my knowledge has increased to the point that I am more confident in my statements.
When I started trying to increase depth capabilities, I first tried increasing the current in the coil, much like you. I found this led to other problems, so that is why I began to look at the delay factor more closely as well as reducing the noise. As a result, I have been able to stretch the depth capabilities considerably with the same current.
If I had to make a recommendation, I would recommend you maximize the depth using a 12V battery, and then slowly bring up the voltage if you still feel this is necessary. However, before increasing the battery voltage, I would experiment with the delay time to find the optimum point.
Do you have a scope so you can look at all the signals? This really helps. BTW, is your PI a personal design?
Finally, the speculation. The selection of the number of turns and the wire size for the coil is normally determined by the first selecting the desired inductance and then the current to be drawn. So, to answer your question about increasing the wire size, number of turns, etc, you have to know a lot of other factors. All of this will affect the minimum allowable delay time you will be able to use and this is very important. My guess is that if you alter the number of turns or wire size, you will see a change in depth capabilities, but more than likely this will be the result of the secondary effects and not just because of the power applied.
In a nutshell, it takes a whole lot more power to increase the depth 1" if all other factors are held constant. It doesn't take but a few additional inches of potential depth capabilities before you will be dragging a car battery along behind you just to power the thing.
I wouldn't complicate things by trying to use 2 FET's, but I would try a very good heatsink for the FET used.
Reg
 
Thanks for the help again Reg.
"If I remember correctly, FET's have a positive temperature coefficient, meaning they will increase in resistance when heated. This will reduce the current flow. Just how much of a change this may be, I cannot answer. I am not sure this is causing the depth loss you are experiencing. "
You are correct, the Rds on resistance increases significantly as the FET heats up. With the FET getting so much hotter after a few seconds, I speculated that this was the reason that it worked better for a short time and then diminished quickly.
"Don't forget that if you are using a battery, the internal resistance of the battery becomes quite important when trying to draw a lot of current. Now, if you are using several batteries in series, then the maximum current flow is really limited by the weakest link, meaning the weakest battery. "
I'll be using a battery, the first time I was playing with this for a period of time, I realized that I was losing my rechargable battery pack and connected it up to a power supply and am running it from that so that I am assured of constant clean power.
"Now, I would like to note that one of my DD coils is about the same size as your test coil and I get greater depth capabilities on a can than you are presently getting. In fact, my distances are in the 30" to 36" range and this is with less than a 50 usec on time.
I am only using 12 V and way less than 1 amp of current to accomplish this. So, as you can tell, there is more to fine tuning a PI than just increasing the current into the coil. "
That sounds great...for you, not me of course <IMG SRC="/forums/images/smile.gif" BORDER=0 ALT=":)"> Can you tell me what coil you are using? I'd like to buy one for testing purposes.
"If I had to make a recommendation, I would recommend you maximize the depth using a 12V battery, and then slowly bring up the voltage if you still feel this is necessary. However, before increasing the battery voltage, I would experiment with the delay time to find the optimum point."
This is a good point that I will do. By the way, I have it adjustable from 30 to 250us pulse delay.
"Do you have a scope so you can look at all the signals? This really helps. BTW, is your PI a personal design? "
I do have a scope. I started with the front end circuit that Eric once posted here. I am running everything with a PIC so I have had to do a fair amount of mods but I believe that the heart of the detector is still unaltered, just the timing being provided by the PIC as well as level shifting to make that happen. Obviously I'm doing something wrong though if I'm not getting the depth you and Eric are on a coke can.
"I wouldn't complicate things by trying to use 2 FET's, but I would try a very good heatsink for the FET used. "
I looked up the characteristics of the heatsink I'm using on the Thermalloy website and saw it has a thermal resistance of around 17. I am going to order some more efficient heatsinks. I know from RC speed controllers(one of my other hobbies) that paralleling FETs is used to increase the current handling capability and to distribute the load to keep the heat down on one device. I also read a post while going through the archives about minelab using 2 FETs in parallel.
I really thank you for the time you are taking to help me out. Also, I'm serious about buying a coil if you would tell me which one you are using.
Thanks,
FJ
 
It's the meltdown of the 300 ohm damping resistor
that gets me. If the fet doesn't avalanche then
it would be wise to clamp the flyback with a zener to around 210 volts and see how you go with
out the high rise time high voltage spikes which
are probably invisible on your cro and are most
likely very much higher than 500v.
 
Hi FJ,
First I need to apologize and revise my depth capabilities downward for my smaller coil. I was able to get 30" to 36" when using the 11" coil, and I thought I had gotten the same distance with the smaller coil. Now, I don't think that was the case.
I just took my smaller coil outside in the local snow storm and I was only able to obtain about 27" to 30" max on the air tests on a pop can. Noise made the difference between the two distances. Most of the time, I was limited to about 27" to 28".
At 24" the signal was still quite respectable, but the signal became very weak very quickly and after about 27", the signal was extremely weak and difficult to hear. I had to minimize external noise by turning off my TV inside to reduce the ambient noise outside to obtain any further increase and even then the noise from other sources was sufficient to I could not hear the very weak signal some of the time.
I guess what I am trying to say is your detector is not too far out of range with the size of coil you have.
My coil I used for testing is a DD design using approximately 8" diameter coils I made myself for my application. Also, my coil will not hold up under the current conditions you are using, so it is not really one that can be purchased or even used in your application.
I would recommend you try building an 11" coil, which is more the standard and then install the coil in a housing so all tests can be made outside away from some of the big electrical noise problems. With this size coil, you should be able to obtain 30" and up to 36" air test on a can. If you cannot, then you need to look at several things as to why you cannot.
It could be because you simply do not have enough amplification throughout the rest of the circuit to provide a signal loud enough to be heard. It could be because the noise is so great that the signal is simply buried. The most logical reason would be a combination of both.
I should qualify something about the 30" to 36" depth on a can even with a 11" coil. The signal from such a target is not an ear blasting signal. Instead, it is quite subdued and is subdued for a distance.
I am going out on a limb on this one, but I suspect that with a 11" coil, you will be able to obtain a typical depth of something like 30" to 36" with 1 amp coil current. Now if this were to be increased to say 4 amps, you might see another inch or two max, and it would take maybe 16 amps to add another inch or two. Those are just guesses and are assuming everything else is held constant.
Hopefully, someone will jump in with the calculations to determine actual distances possible.
Reg
 
Hi Carl,
I had an IRF840LC in there and switched it out with an IRFP32N50K. It didn't seem to make any difference. This FET has a higher current rating and much lower RDSon. Thats when I started to investigate my heat sink and found that it was less than optimal and thought it might help me to keep the FET cooler with a better heatsink.
Thanks,
Fred
 
Hi Robby,
I didn't reach avalanche with the IRF840LC but I do with the heftier FET. I think Reg is correct in his earlier post as to why the damping resistor gets hot. He says that the coil is dumping the flyback into the damping resistor. This must be pretty hard on it.
Thanks,
FJ
 
Hi FJ,
Sorry for not coming in on this discussion, but I am very busy putting the finishing touches to my new workshop at Brixham, on Devon's south coast. Much better location for developing beach and water hunting detectors. Reg is doing a good job, though, in sorting out what is going on with your setup.
Be back very soon.
Eric.
 
The damping resistor sees the flyback energy in
any design. All I was saying was to try clipping
those very high frequency tops off at say 210v
or before avalanche occurs.
As for the fet, Reg is right. A coil that works at
50us may not like 500us and at the same or higher
frequency may tend to cook. If you are using
heavy gauge single strand wire then the f/b pulse
will take longer to settle and this will result
in more average current through the resistor. The
fet can be turned on slowly but not to slow and
it must be slammed off. Most heating occurs during
transission time where the fet is acting as a
resistor.
So my 2 cents.............
Unsuitable coil for pulse length, pps, p/supply.
Driving the coil from the wrong supply rail making
it struggle to turn on or maybe even off.
Switching it on and off too slowly.
A fet with high on resistance.
A damping resistor made from the wrong material.
And all the above are just guesses.
 
Maybe just guesses but some good points to pursue. Especially the coil, I'm going to play with that next. Thanks for the input!
FJ
 
FJ, Check to make sure that your FET gate drive switches all the way down to -V. You may be running the FET in linear mode. The damping resistor is ALWAYS going to get hot as most of the energy you apply to the coil is wasted in it during the flyback period. This is why so many engineers are trying to recover the flyback energy. Recovering the flyback energy = smaller batteries.
Sadly, the PIC chip, especially a little one like the PIC16F84 is a lousy choice for a PI. WHY??? Because you cannot resolve the timing close enough. You need to be able to set the first sample timing in increments of 1uS to get maximum sensitivity to small low conductive targets. The PIC is simply too slow to do this. My suggestion is for you to acquire a learning curve by building Carl Moreland's excellent little PI design. This will give you a good foundation from which you can launch your own ideas. Good Luck and keep at it, Dave. * * *
 
Thanks for the reply Dave. I am using a 20Mhz pic and I'm generating the PPS via interrupts. This occurs by setting a countdown timer in the chip. Since all instructions take 1 or 2 clock cycles this should be .2uS accuracy. My resolution may not be small enough but my accuracy should be very good. Remember, my goal is not to have sensitivity to small low conductive targets but to pick up larger conductive targets at maximum depth. I currently allow the PIC to change the pulse delay from 20uS to 250uS in 10uS increments. I would expect to change that to 30uS on the low end when I connect it to a large 1Meter coil. In the ISR I am waiting the pulse delay amount after pulsing the coil, then I use a time delay loop to time the sample gates. This all occurs before the next interrupt is generated so the accuracy should be at least as good as using a 555 timer, especially since the system is using a crystal to generate accurate timing.
I'm no expert so feel free to point out anything I might be missing. From all that I've heard so far it would seem that pulsing a 1ohm coil with 24V is going to cause the excessive heat I'm seeing on the FET and the flyback on the damping resistor. All feedback and critisisms are welcome :)
Thanks,
FJ
 
You must log in or register to reply here.
Top