Colpits frequency deviation

[porkluvr]

Hello all, this is a question mainly for Tesoro IB owners. I don't have the luxury of possessing a frequency counter (at this time), so please indulge me. This question applies exclusively to those IB detectors that use the TX coil to determine output frequency. Tesoro models with colpitts oscillators come to mind, but I know there are others. Metallic objects interacting with the EM field will have an effect on the operating freqency.

I am working up a Tesoro-type detector that derives audio by zero-beating the oscillator output of CMOS voltage converter (LT1046 etc.) with the metal detector transmit frequency, resulting in baseband audio. Using an active low pass filter and squarer - and with the proper selection of the LT1046 oscillator frequency, after down-converting I should have an audible frequency deviation, correct?

So my question is, how far above and below nominal will the output frequency change for different types of targets i.e. iron, silver, brass, black sand, landmines; big, small, or WHATEVER?

I'm mainly concerned with 10~18KHz but I'll be grateful for any and all data.
The picture I posted shows the predicted result of mixing a steady 9750Hz tone with 9900Hz~10500Hz, but I don't know if this is even remotely practical. What is a practical range?

In the screen-shot image, V2 represents the LT1046 pin2, and A1 is supposed to be a Bandito at roughly 10.2KHz +/- 300Hz. Thank you for your time. I've had trouble in the past with maintaining white-space in this forum, so I've placed smileys to try and break thinks up.

[Max]

Hello all, this is a question mainly for Tesoro IB owners. I don't have the luxury of possessing a frequency counter (at this time), so please indulge me. This question applies exclusively to those IB detectors that use the TX coil to determine output frequency. Tesoro models with colpitts oscillators come to mind, but I know there are others. Metallic objects interacting with the EM field will have an effect on the operating freqency.

I am working up a Tesoro-type detector that derives audio by zero-beating the oscillator output of CMOS voltage converter (LT1046 etc.) with the metal detector transmit frequency, resulting in baseband audio. Using an active low pass filter and squarer - and with the proper selection of the LT1046 oscillator frequency, after down-converting I should have an audible frequency deviation, correct?

So my question is, how far above and below nominal will the output frequency change for different types of targets i.e. iron, silver, brass, black sand, landmines; big, small, or WHATEVER?

I'm mainly concerned with 10~18KHz but I'll be grateful for any and all data.
The picture I posted shows the predicted result of mixing a steady 9750Hz tone with 9900Hz~10500Hz, but I don't know if this is even remotely practical. What is a practical range?

In the screen-shot image, V2 represents the LT1046 pin2, and A1 is supposed to be a Bandito at roughly 10.2KHz +/- 300Hz. Thank you for your time. I've had trouble in the past with maintaining white-space in this forum, so I've placed smileys to try and break thinks up.

Hi,
yes, frequency shifts occours easy on these designs. Colpitts osc are affected by this but also other configurations too. This is because inductance vary with flux chaining due to targets (not considering capacitive effects e.g. in unshielded coils).
Frenquency tend to increase or decrease related to targets. Ferrous and non-ferrous targets gives different behaviour. The principle it's vastly used in old fashioned VLF-PLL machines that uses often monocoils or ib configuration too...but mainly for pinpointing. One example is the french project that is from something I think 1980 or 1981 magazine. There are many others, often magazine ones that used same principle. In these circuits the tuned osc frequency is compared with a stable reference like a crystal osc frequency to give informations about disc.

Sorry , but I can't measure exact freq. shift for bandido...and don't know exactly what they are related to composition, size, etc.
I know that are small changes...something from 0.1 Hz to some e.g. 100Hz.

Best regards,
Max

[Carl-NC]

PL,

The only info I can offer is from the old BFO detectors. They usually ran around 100kHz and would frequency shift about 1kHz, or 1%. So you might expect a 10kHz Colpitts to shift 100 Hz or so.

The only way to know for sure is to do the test. You'll need a zero-motion model like the Bandido. Freq meters are cheap.

- Carl

[Geo]

[quote=Carl-NC;54027]PL,

The only info I can offer is from the old BFO detectors. They usually ran around 100kHz and would frequency shift about 1kHz, or 1%. So you might expect a 10kHz Colpitts to shift 100 Hz or so.

The only way to know for sure is to do the test. You'll need a zero-motion model like the Bandido. Freq meters are cheap.

- Carl[/quote

Generally with good quality components you may construct an Colpits oscillator with frequency shift of 0.1% or less . For me 1% frequency shift is a little big (remember at am radios 1% at freq 1Mhz is 10 Khz shift. ..... it is big)

[Max]

Hello all, this is a question mainly for Tesoro IB owners. I don't have the luxury of possessing a frequency counter (at this time), so please indulge me. This question applies exclusively to those IB detectors that use the TX coil to determine output frequency. Tesoro models with colpitts oscillators come to mind, but I know there are others. Metallic objects interacting with the EM field will have an effect on the operating freqency.

I am working up a Tesoro-type detector that derives audio by zero-beating the oscillator output of CMOS voltage converter (LT1046 etc.) with the metal detector transmit frequency, resulting in baseband audio. Using an active low pass filter and squarer - and with the proper selection of the LT1046 oscillator frequency, after down-converting I should have an audible frequency deviation, correct?

So my question is, how far above and below nominal will the output frequency change for different types of targets i.e. iron, silver, brass, black sand, landmines; big, small, or WHATEVER?

I'm mainly concerned with 10~18KHz but I'll be grateful for any and all data.
The picture I posted shows the predicted result of mixing a steady 9750Hz tone with 9900Hz~10500Hz, but I don't know if this is even remotely practical. What is a practical range?

In the screen-shot image, V2 represents the LT1046 pin2, and A1 is supposed to be a Bandito at roughly 10.2KHz +/- 300Hz. Thank you for your time. I've had trouble in the past with maintaining white-space in this forum, so I've placed smileys to try and break thinks up.

Hi,
another good example is on Electronique Pratique #183 (think was something 90s project), that unfortunately I haven't...maybe someone has it to post.
This project were copied by other sister magazines (and not only) around the Europe in 90s...and is a pure PLL that uses freq. deviations to do disc.
It used a 4046 as PLL if I remember well...can't find the schematic now !!!


Best regards,
Max

[Carl-NC]

Generally with good quality components you may construct an Colpits oscillator with frequency shift of 0.1% or less . For me 1% frequency shift is a little big (remember at am radios 1% at freq 1Mhz is 10 Khz shift. ..... it is big)

In a metal detector, the Colpitts oscillator will have a frequency shift from metal targets that cause the coil inductance to change. It has nothing to do with the quality of the components. BFO detectors need this shift to work, and for synchronous TR detectors it doesn't really matter.

- Carl

[Geo]

In a metal detector, the Colpitts oscillator will have a frequency shift from metal targets that cause the coil inductance to change. It has nothing to do with the quality of the components. BFO detectors need this shift to work, and for synchronous TR detectors it doesn't really matter.

- Carl

I meaned the frequency shift from the component "heat" without any metal target near the coil. Other way the shift must me from very little to 5%. As i remember before 30 years i made a simple bfo at 500... Khz and when i had a very big metal target near the coil then the output was out of audio rate...... i had a frequency shift >20 kHz

[Max]

I meaned the frequency shift from the component "heat" without any metal target near the coil. Other way the shift must me from very little to 5%. As i remember before 30 years i made a simple bfo at 500... Khz and when i had a very big metal target near the coil then the output was out of audio rate...... i had a frequency shift >20 kHz

Hi Carl and Geo,
both of you are right about deviation.
It occourrs like Geo said due to thermal effects e.g. components warming expecially capacitors on the osc section , indepently if is colpitts configuration or not. That's why sometimes wima special series are used or also matched temperature coeff. in components is required. Such deviations could be around 1% of nominal frequency. As Carl said in sync detectors (tr ones) it's not necessary having a perfect frequency stability...cause phase is involved and then a correct phase analysis can be performed without care on frequency shifts...but only considering phase shifts.
In BFO and VLF-PLL (that use too frequency shifts as disc parameter) the discrimination is based on freq. shifts...
in common disc BFO we have just increase or decrease output tone (and sometimes visual indication too) to signal ferrous vs non-ferrous targets...where in PLL output can be a simple buzzer or a speaker with a double (multiple) tone driver etc.

Anyway, in both (BFO and freq-VLF-PLL) frequency shifts are detected by comparing ext osc signal (coil osc signal) with an internal osc (that can be a simple transistor osc...a crystal osc...or also a cmos/integrated rc osc etc) to give a signal due to the difference between. Here freq. shifts are related to inductance variations...due to flux chaining ...though there are always frequency "drifts" due to thermal effects...
Sometimes (for cold climate) some coils integrates a compensation system for thermal effects...to avoid continuos recalibration of center frequency point...but often it's not required to do this in normal conditions.

The EP project that I've mentioned hasn't a compensation for thermal...and it's based on frequency shifts...detected by diff from a crystal osc...and all is built around a 4046 and some op. amps (lm324 if I remember well).

Best regards,
Max

[porkluvr]

I finally purchased a frequency counter (combined with cheap LCR meter) and measured the output frequency of my Super Sabre 2 using different targets. The results I got were not what I expected. Even worse, the results were not what I was hoping for.

By that, I mean that I expected (or HOPED, that is) that the direction of frequency change would be different for iron vs. colored metal. I had this grand notion that I could somehow use an audible frequency change - as translated from the transmitter oscillator via direct conversion - in order to discriminate while operating in an all metal mode. I wanted a Silver Sabre with BFO audio discrimination.

But, as luck would have it: a screwdriver caused about the same direction and magnitude of frequency change as a silver spoon. SS2 nominal output was about 12020Hz. Placing either the spoon or the screwdriver next to the search head caused frequency to increase by about 200Hz. My measuring instrument's frequency resolution is limited to tens, so "about" is as exact as I can get -but there it was. This kind of throws all my confabulation about all-metal mode discrimination into a fiddle. I don't have a lot of space to experiment, and so I have to be left to wonder: I always thought that Iron would would cause a decrease in transmitted frequency, while silver would cause an increase. Why was I so wrong?

[Max]

I finally purchased a frequency counter (combined with cheap LCR meter) and measured the output frequency of my Super Sabre 2 using different targets. The results I got were not what I expected. Even worse, the results were not what I was hoping for.

By that, I mean that I expected (or HOPED, that is) that the direction of frequency change would be different for iron vs. colored metal. I had this grand notion that I could somehow use an audible frequency change - as translated from the transmitter oscillator via direct conversion - in order to discriminate while operating in an all metal mode. I wanted a Silver Sabre with BFO audio discrimination.

But, as luck would have it: a screwdriver caused about the same direction and magnitude of frequency change as a silver spoon. SS2 nominal output was about 12020Hz. Placing either the spoon or the screwdriver next to the search head caused frequency to increase by about 200Hz. My measuring instrument's frequency resolution is limited to tens, so "about" is as exact as I can get -but there it was. This kind of throws all my confabulation about all-metal mode discrimination into a fiddle. I don't have a lot of space to experiment, and so I have to be left to wonder: I always thought that Iron would would cause a decrease in transmitted frequency, while silver would cause an increase. Why was I so wrong?

Hi,
it depends on magnetic properties of material you use in tests.

Iron and ferrous-like items are ferromagnetic and thus concentrate field lines inside , like happens in transformers and other magnetic-circuits. This way you get an increase in inductance of tx coil so a decrease of frequency.

This happens in e.g. radio receivers with moving-core-tuning section, when core(s) are totally inserted inside the coil(s) structure the inductance increase.

Paramagnetic behaviour you can see e.g. in aluminium and cause similar behaviour, though not so strong like with ferromagnetic materials.

Diamagnetic materials have opposite behaviour of ferro- and para-magnetic materials. They tend to distorce/deflect field lines from their inside... so if you put a diamagnetic core inside a coil you'll get a small reduction of inductance, and frequency increase.

So the question is why you get both times an increase ?

Silver is diamagnetic, so the increase is ok.

But what about the screwdriver ???

MY GUESS:

Is it made of iron? No. Never.

Often they are made of special steel, then sometimes are also plated with other metals or alloys to prevent e.g. corrosion.

Effects could be strange.

I think your screwdriver have something that makes increase your frequency.

You have to try with just iron. Use e.g. an iron hammer.

For example I have a total disc on also big iron hammers but weak disc on some kind of steel and e.g. tools that I expect to be ignored by MD in disc mode. Even if disc is phase based is strange that a positive phase shift appear for a supposed ferromagnetic item.

Also found screws made of iron and steel BUT plated by other metals in disc mode.

So it's the sum of effects that make one behaviour win on the others in composite materials items.

Kind regards,
Max

[porkluvr]

I do have a hammer....

Hi,
it depends on magnetic properties of material you use in tests.
Max

I do have several large hammers. I have nails (where....?). OK I will redo simple tests. I live in a shoe, but the shoe is not so small that I cannot do such testing. It is uncomfortable. Ummh. I hate that I am not only lazy, but also that I am also allergic to pain. I will do it anyway, retest with a hammer, nails (16 penny if I can find one). Boat anchor? I do not own a boat, or an anchor.

Car? Not likely. Not allowed in the house. Not even the old Ford... .

My last "test" was hastily conceived, Not much planning went into it.

Re-test?? Not today... soon. Man, I'll be honest with you, I waved my search head around the computer case etc.etc., and not one single time did I get any indication that frequency was going to dip. BUT, I will do it some more. That's the scientific method..... try it again. But I'm scared.

I have a 8# sledge hammer. If THAT doesn't make the frequency drop, then I must live in a black hole, and, but I just thought It was a shoe. In any event, I will get back to you in a couple of days.

I have three racks of "Tennessee Style Barbeque" ribs to roast today, so the "sledge hammer" event will have to wait until Monday.

Thanx, Max
I love my baby back, baby back, baby back - I love.... my...

TIME OUT

[Max]

Hi,
I've tested that on my BandidoII-clone.
Results are as expected...

About on surface of the coil...

-200Hz less using a small hammer made of iron, not steel
-about 120Hz more with a large silver coin

I've used concentric coplanar original by Tesoro (9x8'') to test this.
Don't understand why you get different, always reading increase.

Kind regards,
Max