Tweet
There are hundreds of posts on damping a coil and I have read many of them.....even ones imbedded in other topics. I'm sure this has already been contemplated but I can't find much on the concept of shunting the back EMF with a semiconductor to bring it down fast. ex; A diode in parrellel with the damping resistor. Something similar is done to relay coils to make them respond faster to a change in state. Can anybody elaborate? Good/bad idea? If it is being done, what component best suited for this?
-
-
As far I remember this was discussed by Thinkerer and others (maybe in Aziz thread about using PC as MD). Do some search.Originally posted by Ism View Post -
The purpose of the damping resistor is to provide critical damping of the back emf from the coil. If this resistor is too large a value, the signal will be under-damped and will ring. If the value is too low, the signal will be over-damped, which will also reduce the weak target signal that you are trying to detect. Maximum sensitivity is achieved when you find the Goldylocks resistor - i.e. the one that is just right.Originally posted by Ism View Post
The best way to determine the correct value of the damping resistor, for your particular coil, is to use the arrangement shown in bbsailor's article, which can be found here -> http://www.geotech1.com/pages/metdet...s/FastCoil.pdf (see Figure 11).Comment
-
Thankyou for your replies.
Qiaozi, I know about critical damping with the resistor and have saved and read bbsailor's article. WM6, I will look up PC as MD, I always do a word search and read before asking questions but much of the acronyms are cryptic to me. Actually I have read about using some components to quickly dampen the ringing, but that seemed to be conversations on designing the MD's front end, not at the coil. Aparently the topic is covered somewhere in the forum and I will dig deeper.Comment
-
I've looked at various 2-tau methods, where a fast tau gets the flyback down quickly, then a normal tau switches in for critical damping. But I've never had much success with it.Comment
-
I presume you are looking for gold nuggets. Is that correct?Originally posted by Ism View PostComment
-
Yes. Winter is upon me and I have a couple coil projects in the works.Originally posted by Qiaozhi View PostComment
-
coil damping
hello ism , you mention using a diode to damp the coil , big no no, tried this.
kills the flyback spike to , depending on diode , about 0.7V above rail , end up with an extended duration and less "kick" to the taget.
also when "amped" the deviation for given target is a hell of a lot less,might detect a fridge at 6 inches.
however , i have seen a schematic on here of a "new" detector think it was called a CHANCE PI , which "apears" to use a mosfet to damp the coil controlled by ecu,
but just as to when it damps the coil , i don't know top / middle / bottom of curve.
on the face of it though , looked very interesting , mosfet used as a variable resistor that can be used to damp the coil as and when , poss , ecu looks at curve and then turns on mosfet to damp acordingly , so this would be the ultimate "goldylocks" damping resistor , not based on one resistance that's near enough / compromise , but can be automaticaly altered every pulse by ecu as / when required.
got to go , off to "THE BENCH" TO PLAY WITH 2 MOSFETS !!!Comment
-
Thanks for the feedback. I tried a couple configurations and found the effects you described.Comment
-
more on dampling
on the face of it , damping with a resistor is unsuitable due to the large voltage variation , and also it would apear that damping is only required / benificial right at the end of the decay to stop oscillations , here is my argument :
right , say we have a resistor accross the coil , firstly it can reduce the peak flyback by loading it / slowing it down , as far as i'm aware , what kicks the target is a very fast ,large voltage change , so by adding a perminent resistor , we are not helping ourselves.
second , a resistor slows down the decay slope , so if effort has been put into making the fastest coil possible , then some of your work is being undone by the resistor slowing it down.
third ,adding a resistor accross a signal / voltage /potential of limited current will reduce the voltage deviation we see , so will reduce sensitivity.
forth , all though quite minor powering up a resistor as well as a coil wastes energy , not allot , but it wastes some.
and lastly , current accross a resistor varies with voltage , so most of the damping is taking place at the highest voltage , and less damping at low voltages , where we need it to stop oscillations.so if we half the resistance to stop osillation at the bottom we load the top / peak voltage by a factor of 50 (if 600V peak)
we only need a resistor right at the end to stop oscillations , not all the way through the signal life.
so conclusion is , resistor is a big no no , it's just the easy way.
WHAT IF :
we use a mosfet on its own , or with a fixed resistor , to only switch on to stop oscillation happening then switch it off.
this could be done by monitering voltage when it drops below say 24V.
as long as we damp it so that the first rebound doesnt go below the top rail then switch off, it should decay without oscillation.
OR:
we could damp to a fixed voltage above top rail , say 1V above , then switch off , and the bit we want to see below 0.7V is left alone.
OR : we could switch a mosfet in with a fixed resistor at a fixed voltage point above top rail to just damp the last bit.
many ways it could be done ,but as carl said , it sounds as though it could be tough to get it right , but we should only need to "damp the bend" at the bottom then leave well alone.
not thinking of doing this using ecu's and the like , could all be done using descrets , this way reaction time would be faster , more reliable , and self correcting as conditions alter.
only a thought.
Comment
-
hello to all,
ive just come accross an apparent comercial schematic on the internet for a walk through scanner , that's not the interesting bit.....
on the schematic instead of a damping resistor there are 3x 4.7V zeners pointing down from + to mosfet drain and a 4002 pointing up from drain joining onto the 3 zenners in series.
above this it says "zeners to reduce decay time"
so zenners add upto 14.1 V and poss 0.7 V for the 4002 totals 14.8 V
so this would limit flyback peak to 14.8v above rail which was stated to be 15 V
WHY ???
i could poss understand it if there was a resistor in series so that it would damp from flyback peak all the way down to 14.8 v above rail then left well alone , but limiting flyback would extend the delay time.
also should there not have been a "light" resistor across coil the stop oscillations ??
as it stood , it would have a peak of 14.8v above rail , long decay followed by oscillations.
oh , and the other thing was , the irf150 was driven by a 555 set for only 19.5hz with 7.671% on duty , 19.5 hz , thats a bit low ???
irf150 has a very low 0.055 ohm on resistance apparantly and good for 40A continuous,
60A pulsed but only 100v max , that's a bit low , still zeners would ensure it would never see this.
ANY THOUGHTS ???
Comment
-
Where's the drawing?Comment
-
-
"I see", said the blind man.
Yes, I am familiar with that design and the explanation is quite simple.
They have deliberately slowed down and the flyback impulse so that Grandpa's pacemaker wouldn't fry which could result in fatality, and the designer also did not want the metal plate in my cousin Freddy's head to warp, causing him a massive headache?
That is a logical explanation.
Seriously.
Of course you might ask "well why bother zapping them in the first place if you can't zap 'em good?
I agree.Comment
-
more on damping
hello , mikebg , diagram correct for what i saw .
porkluver , funny as hell , made me laugh
so it was done that way for saftey , and not for depth and penitration , so not allot of good for us then , all makes sense now , still dont know how they got a useable signal out of the oscillations at the end , a damping resistor would have made all the differance.
unless they were going to look at the slope somehow and ignore the bit at the end.
eitherway , i rekon its easyer metal detecting a captive target with a coil all the way round , they should try a single sided coil and see how difficult that is , the cheats!!!Comment
Comment