Tweet
Hi all !
Today I want to show one interesting idea that I found when I finished my previous design . I mean that recuperative PI device , published here - http://www.geotech1.com/forums/showt...y-recuperation As you remember , the main idea of that detector was not to stop the coil current after the charge interval , but to reverse it and store the current in the coil , keeping it constant for some time . This solution has some benefits - we can receive a double target response with the same coil current ( due to reversing the current ) , and another bonus - we can restore this coil current after finishing all measurements . So after the measurement interval we can return almost all the energy stored in the coil to the battery , during a special recuperation interval . And the target information we can get not by measuring the VOLTAGE on the coil loaded to the resistor , but measuring the CURRENT in the shorted coil . The whole idea looks quite attractive and works good enough ... but one day another strange idea came to my head
. I began to thinking how I can increase the operating frequency ( Pulse Repetition Rate ) of the device , and what is the limit ? As everyone can see , in a conventional PI device we does have two intervals - a charge interval when the energy being stored in the coil , and a flyback interval - with exponential current decay and sharp flyback peak - when all stored energy heats the damping resistor And when the coil voltage ( and current ) drops enough - we can strobe the coil voltage to get the useful information about the presence of the metal targets near the coil . When we can start a new measuring cycle ? Of course , we need to wait the time when all the current is totally dissipated ( and we get all the info from the coil ) , and then we can start a new charging interval , and so on . So the maximum PRR is limited by the duration of the charge interval - we need a time to pump the current to the coil , anyhow .
But in my device I need to wait even more - I need to wait when my recuperation interval is finished . So , for example , charge interval is about 150 uS ( A-B interval on the timing diagram ) , then 8 uS flyback ( B-C ) , then 100 uS measuring interval ( C-D ) , and then 150 uS to return the power back to the battery ( D-E ) , so total time is about 400 uS , and maximum PRR must be about 2,5 Khz . Of course , my device is far from this problem - it has PRR of 320 Hz , but what if I wanna make the device for some special purpose , when a high PRR is need ?
And suddenly I found a very beautiful solution that completely eliminates all this problem and allows almost unlimited PRR ... speaking more precisely , it's limited now only by a flyback pulse duration ( 8 uS with my coil , for instance ) . If I can , indeed , easily reverse the current before my measuring interval , why I cannot do it just after it ? What if I reverse the current , measure my target response , and then don't recuperate the energy but reverse it again ? Of course , when I wait 100 uS after reverse , I receive all the current decaying processes in the target , but when those processes are finished - it just mean that the target is properly "magnetized" for the next cycle , isn't it ? So why I need to recuperate the energy and pump it to the coil again , if I can do a very simple thing - reverse the current again and again , ad infinitum
And utilizing this technique I easily can push the frequency up to 50 kHz and even more , for instance , and what is really beautiful - device will consume almost the same current as on 5 kHz PRR .
So , when I use this idea , every current interval does have a double purpose . At the same time I do two things - receive the incoming signal from the target ( decaying exponent function ) , and prepare the target to the next measuring interval , magnetizing it in the opposite polarity . And what is interesting - as I have TWO flyback pulses during the period ( positive and negative ) , I am doubling the equivalent PRR of the whole device . In the other words , this device running at 10 Khz PRR does work completely equal the conventional PI device at 20 Khz PRR , consuming very little energy . You see , in the usual PI device all the coil energy is wasted in every cycle , but in this new detector the current is only reversing back and forth And by the way , bipolar operation does have some advantages itself - we know some examples of conventional bipolar PI's .
The coil current in this device will have the shape of continious square wave , and the target signal is just being mixed with this wave . So all what we need is to use the current transformer ( just like I did it in my previous design ) to pick up the signal from the coil , and perform some operations to separate this continious idle current wave and the incoming signal . My theoretical analysis showed that this principle is quite powerful , and we can do the signal processing using some different algorithms - some of them are simple , and some more complicated .... and now I am finishing the working device using one of this ways . I decided to publish all the circuits little by little , because it's not so easy to draw all the papers at once . And because the entire device isn't finished yet - some circuits can be changed .... Here is the first circuit - the initial power chain concept .
Today I want to show one interesting idea that I found when I finished my previous design . I mean that recuperative PI device , published here - http://www.geotech1.com/forums/showt...y-recuperation As you remember , the main idea of that detector was not to stop the coil current after the charge interval , but to reverse it and store the current in the coil , keeping it constant for some time . This solution has some benefits - we can receive a double target response with the same coil current ( due to reversing the current ) , and another bonus - we can restore this coil current after finishing all measurements . So after the measurement interval we can return almost all the energy stored in the coil to the battery , during a special recuperation interval . And the target information we can get not by measuring the VOLTAGE on the coil loaded to the resistor , but measuring the CURRENT in the shorted coil . The whole idea looks quite attractive and works good enough ... but one day another strange idea came to my head
. I began to thinking how I can increase the operating frequency ( Pulse Repetition Rate ) of the device , and what is the limit ? As everyone can see , in a conventional PI device we does have two intervals - a charge interval when the energy being stored in the coil , and a flyback interval - with exponential current decay and sharp flyback peak - when all stored energy heats the damping resistor And when the coil voltage ( and current ) drops enough - we can strobe the coil voltage to get the useful information about the presence of the metal targets near the coil . When we can start a new measuring cycle ? Of course , we need to wait the time when all the current is totally dissipated ( and we get all the info from the coil ) , and then we can start a new charging interval , and so on . So the maximum PRR is limited by the duration of the charge interval - we need a time to pump the current to the coil , anyhow . But in my device I need to wait even more - I need to wait when my recuperation interval is finished . So , for example , charge interval is about 150 uS ( A-B interval on the timing diagram ) , then 8 uS flyback ( B-C ) , then 100 uS measuring interval ( C-D ) , and then 150 uS to return the power back to the battery ( D-E ) , so total time is about 400 uS , and maximum PRR must be about 2,5 Khz . Of course , my device is far from this problem - it has PRR of 320 Hz , but what if I wanna make the device for some special purpose , when a high PRR is need ?
And suddenly I found a very beautiful solution that completely eliminates all this problem and allows almost unlimited PRR ... speaking more precisely , it's limited now only by a flyback pulse duration ( 8 uS with my coil , for instance ) . If I can , indeed , easily reverse the current before my measuring interval , why I cannot do it just after it ? What if I reverse the current , measure my target response , and then don't recuperate the energy but reverse it again ? Of course , when I wait 100 uS after reverse , I receive all the current decaying processes in the target , but when those processes are finished - it just mean that the target is properly "magnetized" for the next cycle , isn't it ? So why I need to recuperate the energy and pump it to the coil again , if I can do a very simple thing - reverse the current again and again , ad infinitum
And utilizing this technique I easily can push the frequency up to 50 kHz and even more , for instance , and what is really beautiful - device will consume almost the same current as on 5 kHz PRR . So , when I use this idea , every current interval does have a double purpose . At the same time I do two things - receive the incoming signal from the target ( decaying exponent function ) , and prepare the target to the next measuring interval , magnetizing it in the opposite polarity . And what is interesting - as I have TWO flyback pulses during the period ( positive and negative ) , I am doubling the equivalent PRR of the whole device . In the other words , this device running at 10 Khz PRR does work completely equal the conventional PI device at 20 Khz PRR , consuming very little energy . You see , in the usual PI device all the coil energy is wasted in every cycle , but in this new detector the current is only reversing back and forth
And by the way , bipolar operation does have some advantages itself - we know some examples of conventional bipolar PI's . The coil current in this device will have the shape of continious square wave , and the target signal is just being mixed with this wave . So all what we need is to use the current transformer ( just like I did it in my previous design ) to pick up the signal from the coil , and perform some operations to separate this continious idle current wave and the incoming signal . My theoretical analysis showed that this principle is quite powerful , and we can do the signal processing using some different algorithms - some of them are simple , and some more complicated .... and now I am finishing the working device using one of this ways . I decided to publish all the circuits little by little , because it's not so easy to draw all the papers at once . And because the entire device isn't finished yet - some circuits can be changed .... Here is the first circuit - the initial power chain concept .
But now I prefer the traditional name of the topic - let's see , where it will lead us .
. I can be caused by eddy currents in the coil wire , dielectric losses in the capacitor , etc . That is why I added a special adaptive compensation circuit in that device ( what was a little tricky indeed
Comment