The individual coils can be pretty far. Remember that Coils in series are accumulative in their inductance. But again I would not try to go beyond a 1:4 ratio. Meaning that if your total inductance is 100uH ... The sense coil should be no less than 25uH while the 'other' coil should be no more than 75uH ... If you go too much out of bounds, then the sensitivity diminishes to a point that it is difficult to see a signal.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Yes I see coils in my sleep. I think I got it figured out. I used two coils in series. One larger (25 uH) and one small (15 uH). The smaller was on the tip of the probe. I am making different configurations of coils to experiment with. I'll let you know how it turns out. Thanks, Curtis
Thinking about your coil winding... You mentioned a 2mm size (ID or OD length? number of turns &c). Anyway, how about using 2 continuous rotation servos. The first (A axis) turns a sewing machine needle (with the eye used to hold your wire [noparse][[/noparse]and a sleeve to provide diameter]) and the second (X axis) is used to provide the feedrate (coil turn spacing). Such a small size should not need a very complex fixture to hold it together.
Also, the recently re-released Servo Driver Board would give you some enormous flexibility. Combined with the article (in this forum) of using a secondary servo as an input (learning) device could allow you to "teach" it the appropriate winding specifications (ratios). Once the sequence is learnt, you could drive pairs of servos to make mutiple coils if needed - or mod the code to automate different coil specifications for testing and the like.
That sounds like a good plan. I was thinking of using servos or stepper motors. I want to make the ID of the coil 2mm and my final inductance would be 40 uH. I am trying Beau's suggestion of spliting the inductance with a small 15 uH coil (probe) and a larger coil of 25 uH in series. I like you setup though. Could you go more into detail? I have the serial version of the servo controller on hand. Thanks, Curtis
The latest Servo board was here: http://forums.parallax.com/showthread.php?p=834730
There was an article by either Chris Savage or Mr Schwabe (sp?) both Parallax people that described how to use a servo as an input device. Of course, now I can't find it but I'll keep looking and post it when I can.· Essentially, I was thinking to use these combined ideas as a learning "method" to "teach" the two servos used to wind the coils.· Using a lathe as a reference, the spindle (A Axis) has a sewing machine needle as a mandrel and the main carriage slide (The X Axis) provides the winding pitch for the coils.· I only chose a sewing machine needle because of it's diameter and the eye that can be used as an anchor for your (very) fine wire (I envisioned you are using).· Additional diameter can easily be added, cigarette papers might work or whatever, of course if you have a lathe you can make whatever is needed.
Call the A axis SRV0, the Z axis SRV1 and the two "learning servos" SRV2 and SVR3 (See sketch).· Using the learning servos and the input routines to "teach" the A and Z axes their relational movements.· Wind SRV2 and oscillate SVR3 to give the desired rotation/pitch.·
I am asuming the coils have multi-layer - this is why the oscillate (L to R) motion provided by you via SVR3 to be used as "playback" for SRV1.
When you have the basic plots, you can fine tune them.· Basically, this is a simple 2 axis CNC "machine".· I figure the short distance for the coil pitch/length could be handled by the length of the servo horn.· The picture attached, shows a simple oscillating mechanism I am building for the daughter for use in a yarn winder/counter.· This "machine" will wind spun yarn (she does this as a hobby) onto a spool while counting the length.· Basically, this is really a "coil winder" (if I think Electronics ;-)·whereby the spun yarn is counted (by an indexed wheel of 32mm diameter (1 rev = ~100mm)), wound onto a takeup spool and oscillated from end to end while doing so.· This is a work in progress and my first Propeller project (as a pre-cursor to the router table I plan to make).
Of course, just program control could be developed for the A and Z axes but I was thinking if your coils are special or have specific movements in them then the "teaching" ideas would also be fun as well as practical for your needs.
Also, if I am correct in my thinking of what you are doing:· I suppose you are using resonant frequencies to determine iregularities in the metals you are testing using the coils as part of a tuned circuit.· If true, then you could also use variable capacitors in your tuned curcuits to provide additional flexibility and spectrum.· Either parallel, series or a combination of tuned curcuits may be useful.· You might find circuits about "Grid Dip Oscillators" useful for this kind of work.· There are lots of referneces to circuts, formulae and the like the Internet.· The ARRL publishes some excellent reference materials as well.
Sorry about the typos. Similar "mechanisms" and ideas could also be employed to provide an X-Y scanning grid of your target if used to hold your "probe". But I guess that's somewhat evident from the single-axis photograph There is certainly enough horsepower/pins available on a Propeller chip to do this.
If you are not familiar with G-Code (used with CNC controls), check it out; there is a specific G-code that provides a "probe" function. This function builds a "point cloud" that provides an X,Y,Z; 3-D representation of the object being probed and can then be used to re-create an object of (probed) dimensions. Although the function and their controls are primarily designed to use a "touch probe" that uses physical contact (either switches in the probe or conduction between the probe and object) it should be possible to realise the same function using "proximity sensing" - especially if you are using RF (amplitude, FM or other) to sense irregularities in your target object.
Hummm, an "active filter design" also using a programmable resistive component in the tuned circuit could be an interesting multi-pass option too [noparse][[/noparse]especially if the 'scanning' is automated]
but the idea is multiple frequency scanning to generate overlayable mathematical models (way out of my area).
(Sorry, need to ask someone else, it's been way too long since I actually did any of this; just getting back into it after a "few" years.
But wow, what a great construction and programming project!
All of the information is great. I have a three day weekend this week and will work on all of this. My ultimate goal is to have an object that will use the Eddy Current Analyzer. I will post everything for anyone who is interested when I finish documenting it.
I'll read these post. I have attached a spin file that does:
1. Excites the coil in my eddy current analyzer and the results are read by a ad0831. The code works fine.
What I want to do now is to sweep a frequency range and find the Resonant Frequency by getting the highest value from the adc and then storing it in a variable (NewFreq) and set the frequency to the value using Synth() in the code.
I tried an If statement, but my trouble is where to put it. The routine should only run once at bootup. My mind is not seeing it and if anyone can help, I would appreciate it.
I am a new SPIN user so my coding skills with this (clever and unusual ;-) language are somewhat limited; I've seen some clever and unusual code fragments that made me think "gee, that's clever, unusual but clever" so I won't try to provide any serious SPIN code.· There are others in this Forum who would have a much better grasp of SPIN efficiencies...
What I notice however is perhaps a cross between what I see you say and what I extrapolate based on my (incorrect?) assumptions on what you are doing and the way in which you do this.· If I take the following as example:
1. You have a UUT (Unit Under Test) [noparse][[/noparse]the metal sample being "scanned"]
You can either:
CASE_A:
2a. Start a resonance frequency scan test from startFREQ to endFREQ using stepVALUE
2b. Record the value returned at each stepVALUE and if the next value is higher, replace the current
2c. The value returned by each·next stepVALUE, when lower or == to the current terminates the scan
or,
CASE_B:
3a. You already have the (general) resonant frequency of the UUT
3b. You want to scan between a range that is (perhaps) some range of +/-·of this.
In either case, you will need a loop structure that:
4a. Sets up the frequency for the initial, final and step values
4b. Reads the value and stores it (for a subsequent compare)
4c. Inc/dec the frequence by the step value
4d. Read and compare the new value
4e. Detemine if the new returned value is larger (less &c) and take appropriate action.
Depending on the intent, where to do the resonance test becomes part of the choices.·
In CASE_A, the loop structure might be well suited in the Synth method (using AcquireValue) to get the best return.· In CASE_B, the loop structure could be implemented in the main PUB method, calling the Synth method on each pass.
Finally, if performance/speed if not really the intent - especially if using new UUT's, coils or other test parameters/criteria, a combination of both might be in order.· This technique could also provide coarse and fine "tuning"; coarse in the Synth method and fine in the caller.· For a coarse method, a higher stepFREQ would be used with a successively reduced step rate to get "closer" quicker.· Maybe more code but it should run and find the resonant truth faster.
· if (val_2 < val_1) break ' It doesn't get better than this ;-)
[noparse][[/noparse]/code]
+
To be sure, it might be useful to repeat the initial testing from lower to higher and then from higher to lower, recording the "best" return on each pass; this could then be used as the "base" especially if using a coarse/fine technique.· Harmonics can be "interesting"·so this·might help some·too.
Anyway, I hope this makes sense, if not let me know.· Let us know how things pan out - looking forward to any updates or queries.
Drat, I omitted to replace the val_1 value the new (higher) val_2 value for the next pass (on the line with the "keep going"). But I hope it makes sense that you'd have to do that
I'm new to Parallax but I'm an NDT tech with certification in Eddy Current inspections. I find this project very interesting.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
2) What material are you expecting to test?
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
5) How are you displaying your output?
I'm interested in your project. It sounds like it has good potential and I look forward to seeing more about it.
Last March I went to the ASNDT conference in St. Louis. It was great.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
Answer: I want to get the frequency of the probe. I was hoping to have a autocalibrate for each coil or zero the coupling before I scan. Also maybe you would know, probes are rated for a frequency range. How do they find this range?
2) What material are you expecting to test?
Answer: Any metal that artifacts are made of.
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
Answer: It is an absolute probe, but I experimenting with other types.
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
Answer: I agree, but my VB 6.0 software finds the little changes in ferrous materials.
5) How are you displaying your output?
Answer: A Visual Basic program and an LCD from Parallax.
Last March I went to the ASNDT conference in St. Louis. It was great.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
Answer: I want to get the frequency of the probe. I was hoping to have a autocalibrate for each coil or zero the coupling before I scan. Also maybe you would know, probes are rated for a frequency range. How do they find this range?
Probes are designed for a specific resonant frequency by varying R, XL and XC. They also have a bandwidth to allow frequency variation to match the probe to the material under test. The designed frequency will be the peak, and the bandwidth will be a usable range on each side of the peak. Once the probe is coupled to the material surface, the test instrument circuitry is used to balance the complete circuit.
2) What material are you expecting to test?
Answer: Any metal that artifacts are made of.
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
Answer: It is an absolute probe, but I experimenting with other types. Perhaps you want to check out a Cross point probe or a Plus point probe. It will increase the resolution of your probe by adding a linear·direction to your indication.
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
Answer: I agree, but my VB 6.0 software finds the little changes in ferrous materials.
5) How are you displaying your output?
Answer: A Visual Basic program and an LCD from Parallax. Is this a Complex Impedance Display? or strip chart? or other
WRT: "Also maybe you would know, probes are rated for a frequency range. How do they find this range?"
Coil/tank circuits have formulae for resonant frequencies, scratching my head I remember f=1/2pi(LC) - if I recall, this is for a series LC circuit but it might be the parallel one... The GDO (Grid Dip Oscillator) mentioned previously is a test piece for determining the Q of coils in such circuits - it works (as I recall) by showing the absorbtion of a LC network as a representation of the Q of the coil - Q being the bandpass(?) curve and therefore the attenuation points (in dB from resonant centre freq) but it's been some 30 years since I played with that stuff so I could be off base here - someone else care to correct or clarify? If I had access to my old ARRL radio books, I could look it up for you but alas...
Yes, my pseudo code fragment/comments were intended to auto-tune the UUT (not specifically the probe but I guess both could apply); then provide a sweep of the UUT. I don't know how you are scanning the UUT but I guessed that once you have the resonant frequency of the UUT (I assume each UUT is different), deformation in the UUT would show as a "glitch" when passed over. I was supposing that slighty off-resonant frequencies could provide a 3-D mathematical model to give either more detail or less false positives (the overlay stuff I mentioned) but this is way out or my area - just thinking "aloud"
I've been thinking about the resonance &c. There should be nothing to prevent the use of another COG, pin(s) and ADC; connected to a sense circuit included with your primary excitation circuit that can be used to set a global vaiable that the excitation circuit uses for the drive (freq) value. Unless you've run out of pins...
ie: Like previous concepts, you are seraching for resonance and how you do this is unknown (specifically) but, if a sense pickup point is available from the excitation circuit and fed through an ADC back into another COG; running code to determine the value, it could be used to set the value also used by the excitation circuit. A closed loop arrangement using a global variable.
I'm basing this idea on the operation of the GDO mentioned; ie: the indication of resonance is provided by the _dip_ in the meter of a GDO. By comparison, this is showing the best I/E/P transfer from the excitation circuit to the UUT (or probe; or both). I can't see why a sense circuit (I or E as appropriate) would not work as part of a feedback loop in a close loop model.
I have finally finished the PCB design for the Eddy Current Analyzer. I have attached an ExpressPCB file of the finished circuit using a 40 pin Prop. I ready to etch it and I would appreciate if someone would check it for me. The software is almost complete and I will post the complete project as soon as I build the instrument from the attached PCB. Included in the complete project is about 20 scans of different metals. Thanks again and any suggestion or question are appreciated.
Hi Curtis;
I do not have the expertise to help with the circuit but I would like to know what it is. Could you upload these files in a non-proprietary format (e.g. PDFs)?
Cheers!
Paul Rowntree
A few thing I noticed. First, it's good practice to put small ceramic capacitors between the power and ground pins of every digital IC as close to the IC as practical. Digital IC's have lot's of high frequency switching internally and this capacitor keeps the power to the IC clean and helps trap noise generated by the internal switching of the IC. Second, Power traces are usually made several times wider than signal traces to minimize resistive losses. Third, it looks like you're using the bottom copper layer as a silk screen layer? (silk screen is most often colored yellow in PCB programs) I know ExpressPCB's proto-board service has no silk screen, but it's still usually more useful to use both copper layers to route signals. A 1:1 printout of your design can then be taped to the top of the board to replace the missing silk screen layer. (two routing layers would also allow an unbroken ground plane, but your ground plane forms a loop right now so it should work plenty well.) Fourth, I'd suggest double checking the pin-out of your regulators just to be safe. I've been bitten by linear regulators that mix and match the pin-out. [noparse]:)[/noparse]
Lawson
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Lunch cures all problems! have you had lunch?
Comments
The individual coils can be pretty far. Remember that Coils in series are accumulative in their inductance. But again I would not try to go beyond a 1:4 ratio. Meaning that if your total inductance is 100uH ... The sense coil should be no less than 25uH while the 'other' coil should be no more than 75uH ... If you go too much out of bounds, then the sensitivity diminishes to a point that it is difficult to see a signal.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Curtis
Still counting windings?
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.:
YADRO
Yes I see coils in my sleep. I think I got it figured out. I used two coils in series. One larger (25 uH) and one small (15 uH). The smaller was on the tip of the probe. I am making different configurations of coils to experiment with. I'll let you know how it turns out. Thanks, Curtis
Thinking about your coil winding... You mentioned a 2mm size (ID or OD length? number of turns &c). Anyway, how about using 2 continuous rotation servos. The first (A axis) turns a sewing machine needle (with the eye used to hold your wire [noparse][[/noparse]and a sleeve to provide diameter]) and the second (X axis) is used to provide the feedrate (coil turn spacing). Such a small size should not need a very complex fixture to hold it together.
Also, the recently re-released Servo Driver Board would give you some enormous flexibility. Combined with the article (in this forum) of using a secondary servo as an input (learning) device could allow you to "teach" it the appropriate winding specifications (ratios). Once the sequence is learnt, you could drive pairs of servos to make mutiple coils if needed - or mod the code to automate different coil specifications for testing and the like.
Cheers,
Harry E.
That sounds like a good plan. I was thinking of using servos or stepper motors. I want to make the ID of the coil 2mm and my final inductance would be 40 uH. I am trying Beau's suggestion of spliting the inductance with a small 15 uH coil (probe) and a larger coil of 25 uH in series. I like you setup though. Could you go more into detail? I have the serial version of the servo controller on hand. Thanks, Curtis
The latest Servo board was here: http://forums.parallax.com/showthread.php?p=834730
There was an article by either Chris Savage or Mr Schwabe (sp?) both Parallax people that described how to use a servo as an input device. Of course, now I can't find it but I'll keep looking and post it when I can.· Essentially, I was thinking to use these combined ideas as a learning "method" to "teach" the two servos used to wind the coils.· Using a lathe as a reference, the spindle (A Axis) has a sewing machine needle as a mandrel and the main carriage slide (The X Axis) provides the winding pitch for the coils.· I only chose a sewing machine needle because of it's diameter and the eye that can be used as an anchor for your (very) fine wire (I envisioned you are using).· Additional diameter can easily be added, cigarette papers might work or whatever, of course if you have a lathe you can make whatever is needed.
Call the A axis SRV0, the Z axis SRV1 and the two "learning servos" SRV2 and SVR3 (See sketch).· Using the learning servos and the input routines to "teach" the A and Z axes their relational movements.· Wind SRV2 and oscillate SVR3 to give the desired rotation/pitch.·
I am asuming the coils have multi-layer - this is why the oscillate (L to R) motion provided by you via SVR3 to be used as "playback" for SRV1.
When you have the basic plots, you can fine tune them.· Basically, this is a simple 2 axis CNC "machine".· I figure the short distance for the coil pitch/length could be handled by the length of the servo horn.· The picture attached, shows a simple oscillating mechanism I am building for the daughter for use in a yarn winder/counter.· This "machine" will wind spun yarn (she does this as a hobby) onto a spool while counting the length.· Basically, this is really a "coil winder" (if I think Electronics ;-)·whereby the spun yarn is counted (by an indexed wheel of 32mm diameter (1 rev = ~100mm)), wound onto a takeup spool and oscillated from end to end while doing so.· This is a work in progress and my first Propeller project (as a pre-cursor to the router table I plan to make).
Of course, just program control could be developed for the A and Z axes but I was thinking if your coils are special or have specific movements in them then the "teaching" ideas would also be fun as well as practical for your needs.
Also, if I am correct in my thinking of what you are doing:· I suppose you are using resonant frequencies to determine iregularities in the metals you are testing using the coils as part of a tuned circuit.· If true, then you could also use variable capacitors in your tuned curcuits to provide additional flexibility and spectrum.· Either parallel, series or a combination of tuned curcuits may be useful.· You might find circuits about "Grid Dip Oscillators" useful for this kind of work.· There are lots of referneces to circuts, formulae and the like the Internet.· The ARRL publishes some excellent reference materials as well.
Cheers,
Harry E.
If you are not familiar with G-Code (used with CNC controls), check it out; there is a specific G-code that provides a "probe" function. This function builds a "point cloud" that provides an X,Y,Z; 3-D representation of the object being probed and can then be used to re-create an object of (probed) dimensions. Although the function and their controls are primarily designed to use a "touch probe" that uses physical contact (either switches in the probe or conduction between the probe and object) it should be possible to realise the same function using "proximity sensing" - especially if you are using RF (amplitude, FM or other) to sense irregularities in your target object.
Hummm, an "active filter design" also using a programmable resistive component in the tuned circuit could be an interesting multi-pass option too [noparse][[/noparse]especially if the 'scanning' is automated]
---CAP---V_Resistor---(more_than_zero)
}coil =============================Resonance == Base_Reference
Componets
(less_than_zero)
but the idea is multiple frequency scanning to generate overlayable mathematical models (way out of my area).
(Sorry, need to ask someone else, it's been way too long since I actually did any of this; just getting back into it after a "few" years.
But wow, what a great construction and programming project!
FWIW
Cheers,
All of the information is great. I have a three day weekend this week and will work on all of this. My ultimate goal is to have an object that will use the Eddy Current Analyzer. I will post everything for anyone who is interested when I finish documenting it.
Thanks,
Curtis
http://forums.parallax.com/forums/default.aspx?f=25&m=121278
via:
http://forums.parallax.com/forums/default.aspx?f=25&m=148376
"Peripherals and interfacing:"
Propeller Application: Proportional feedback from a Standard Hobby Servo (Updated video clip)
http://forums.parallax.com/forums/default.aspx?f=25&m=121278
I'll read these post. I have attached a spin file that does:
1. Excites the coil in my eddy current analyzer and the results are read by a ad0831. The code works fine.
What I want to do now is to sweep a frequency range and find the Resonant Frequency by getting the highest value from the adc and then storing it in a variable (NewFreq) and set the frequency to the value using Synth() in the code.
I tried an If statement, but my trouble is where to put it. The routine should only run once at bootup. My mind is not seeing it and if anyone can help, I would appreciate it.
Curtis
I am a new SPIN user so my coding skills with this (clever and unusual ;-) language are somewhat limited; I've seen some clever and unusual code fragments that made me think "gee, that's clever, unusual but clever" so I won't try to provide any serious SPIN code.· There are others in this Forum who would have a much better grasp of SPIN efficiencies...
What I notice however is perhaps a cross between what I see you say and what I extrapolate based on my (incorrect?) assumptions on what you are doing and the way in which you do this.· If I take the following as example:
1. You have a UUT (Unit Under Test) [noparse][[/noparse]the metal sample being "scanned"]
You can either:
CASE_A:
2a. Start a resonance frequency scan test from startFREQ to endFREQ using stepVALUE
2b. Record the value returned at each stepVALUE and if the next value is higher, replace the current
2c. The value returned by each·next stepVALUE, when lower or == to the current terminates the scan
or,
CASE_B:
3a. You already have the (general) resonant frequency of the UUT
3b. You want to scan between a range that is (perhaps) some range of +/-·of this.
In either case, you will need a loop structure that:
4a. Sets up the frequency for the initial, final and step values
4b. Reads the value and stores it (for a subsequent compare)
4c. Inc/dec the frequence by the step value
4d. Read and compare the new value
4e. Detemine if the new returned value is larger (less &c) and take appropriate action.
Depending on the intent, where to do the resonance test becomes part of the choices.·
In CASE_A, the loop structure might be well suited in the Synth method (using AcquireValue) to get the best return.· In CASE_B, the loop structure could be implemented in the main PUB method, calling the Synth method on each pass.
Finally, if performance/speed if not really the intent - especially if using new UUT's, coils or other test parameters/criteria, a combination of both might be in order.· This technique could also provide coarse and fine "tuning"; coarse in the Synth method and fine in the caller.· For a coarse method, a higher stepFREQ would be used with a successively reduced step rate to get "closer" quicker.· Maybe more code but it should run and find the resonant truth faster.
+
_rough_ example:
[noparse][[/noparse]code]
Synth(Pin, startFREQ)
val_1 := AcquireValue
repeat NewFREQ from (startFREQ + step_FREQ)·to endFREQ step stepFREQ ' already done 1st
· Synth(Pin, NewFREQ)
· val_2 := AcquireValue
· if (val_2 > val_1) keep going
· if (val_2 < val_1) break ' It doesn't get better than this ;-)
[noparse][[/noparse]/code]
+
To be sure, it might be useful to repeat the initial testing from lower to higher and then from higher to lower, recording the "best" return on each pass; this could then be used as the "base" especially if using a coarse/fine technique.· Harmonics can be "interesting"·so this·might help some·too.
Anyway, I hope this makes sense, if not let me know.· Let us know how things pan out - looking forward to any updates or queries.
Cheers,
Harry E.
I'm new to Parallax but I'm an NDT tech with certification in Eddy Current inspections. I find this project very interesting.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
2) What material are you expecting to test?
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
5) How are you displaying your output?
I'm interested in your project. It sounds like it has good potential and I look forward to seeing more about it.
Tom Lewis
·
Why don't you run your resonate frequency sweep in a cog by itself at start-up, store the results and kill the cog?
rs_jim
Last March I went to the ASNDT conference in St. Louis. It was great.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
Answer: I want to get the frequency of the probe. I was hoping to have a autocalibrate for each coil or zero the coupling before I scan. Also maybe you would know, probes are rated for a frequency range. How do they find this range?
2) What material are you expecting to test?
Answer: Any metal that artifacts are made of.
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
Answer: It is an absolute probe, but I experimenting with other types.
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
Answer: I agree, but my VB 6.0 software finds the little changes in ferrous materials.
5) How are you displaying your output?
Answer: A Visual Basic program and an LCD from Parallax.
Thanks for your interest Tom.
Hi Rs_Jim,
Great idea, I will try it.
Thanks,
Curtis
Last March I went to the ASNDT conference in St. Louis. It was great.
1) Your question above, with respect to sweeping a range of freq to find the Rresonant frequency, is to find the resonant frequency of the probe, is it not? Perhaps it should be a standalone program? The resonant freq wouldn't change once the probe is constructed.
Answer: I want to get the frequency of the probe. I was hoping to have a autocalibrate for each coil or zero the coupling before I scan. Also maybe you would know, probes are rated for a frequency range. How do they find this range?
Probes are designed for a specific resonant frequency by varying R, XL and XC. They also have a bandwidth to allow frequency variation to match the probe to the material under test. The designed frequency will be the peak, and the bandwidth will be a usable range on each side of the peak. Once the probe is coupled to the material surface, the test instrument circuitry is used to balance the complete circuit.
2) What material are you expecting to test?
Answer: Any metal that artifacts are made of.
3) Your probe... is it a straight 'absolute' coil, or are you winding it in another configuration?
Answer: It is an absolute probe, but I experimenting with other types.
Perhaps you want to check out a Cross point probe or a Plus point probe. It will increase the resolution of your probe by adding a linear·direction to your indication.
4) Have you tested it on ferrous materials yet? Variable permiability in ferromagnetic material affects eddy current signals, but I don't know if it will hide the indications you are seeking. Usually, the affect of variable permiability and·large probe size prevents eddy current inspection from reading serial numbers filed off guns or engine blocks. (A 1mm probe is teeny-tiny in the ECT world. Most people find them too delicate and too expensive for normal use.)
Answer: I agree, but my VB 6.0 software finds the little changes in ferrous materials.
5) How are you displaying your output?
Answer: A Visual Basic program and an LCD from Parallax.
Is this a Complex Impedance Display? or strip chart? or other
Thanks for your interest Tom.
Hi Rs_Jim,
Great idea, I will try it.
Thanks,
Curtis
Is this a Complex Impedance Display? or strip chart? or other
The display is both. I use an XY plot for the amplitudes of the excite frequency and the amplitude of the voltage
measured between the coil and capacitor. I use Viewport for a strip chart also.
Thanks for the info,
Curtis
Good to hear things make sense.
WRT: "Also maybe you would know, probes are rated for a frequency range. How do they find this range?"
Coil/tank circuits have formulae for resonant frequencies, scratching my head I remember f=1/2pi(LC) - if I recall, this is for a series LC circuit but it might be the parallel one... The GDO (Grid Dip Oscillator) mentioned previously is a test piece for determining the Q of coils in such circuits - it works (as I recall) by showing the absorbtion of a LC network as a representation of the Q of the coil - Q being the bandpass(?) curve and therefore the attenuation points (in dB from resonant centre freq) but it's been some 30 years since I played with that stuff so I could be off base here - someone else care to correct or clarify? If I had access to my old ARRL radio books, I could look it up for you but alas...
Yes, my pseudo code fragment/comments were intended to auto-tune the UUT (not specifically the probe but I guess both could apply); then provide a sweep of the UUT. I don't know how you are scanning the UUT but I guessed that once you have the resonant frequency of the UUT (I assume each UUT is different), deformation in the UUT would show as a "glitch" when passed over. I was supposing that slighty off-resonant frequencies could provide a 3-D mathematical model to give either more detail or less false positives (the overlay stuff I mentioned) but this is way out or my area - just thinking "aloud"
Cheers,
I've been thinking about the resonance &c. There should be nothing to prevent the use of another COG, pin(s) and ADC; connected to a sense circuit included with your primary excitation circuit that can be used to set a global vaiable that the excitation circuit uses for the drive (freq) value. Unless you've run out of pins...
ie: Like previous concepts, you are seraching for resonance and how you do this is unknown (specifically) but, if a sense pickup point is available from the excitation circuit and fed through an ADC back into another COG; running code to determine the value, it could be used to set the value also used by the excitation circuit. A closed loop arrangement using a global variable.
I'm basing this idea on the operation of the GDO mentioned; ie: the indication of resonance is provided by the _dip_ in the meter of a GDO. By comparison, this is showing the best I/E/P transfer from the excitation circuit to the UUT (or probe; or both). I can't see why a sense circuit (I or E as appropriate) would not work as part of a feedback loop in a close loop model.
Cheers,
I have finally finished the PCB design for the Eddy Current Analyzer. I have attached an ExpressPCB file of the finished circuit using a 40 pin Prop. I ready to etch it and I would appreciate if someone would check it for me. The software is almost complete and I will post the complete project as soon as I build the instrument from the attached PCB. Included in the complete project is about 20 scans of different metals. Thanks again and any suggestion or question are appreciated.
Curtis
I do not have the expertise to help with the circuit but I would like to know what it is. Could you upload these files in a non-proprietary format (e.g. PDFs)?
Cheers!
Paul Rowntree
This is the Eddy Current Analyzer and its in ExpressPCB format (free software). I have attached an image as you requested.
Thanks,
Curtis
Lawson
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Lunch cures all problems! have you had lunch?