The Rangertell Examiner Project - Report 5

Apr 8, 2010 - Deluxe Examiner Calculator testing: TI-36X Solar model

The Examiner Deluxe model came with an upgraded Texas Instruments TI-36X Solar calculator. This calculator has more function buttons and a larger display with more digits than the calculator that comes with the T-G 8.08B model. But best of all, it does not automatically shut down after a few minutes of running. The instructions for the Deluxe model are also changed to use some of the keys that are not available on the calculator for the T-G 8.08B model.

I used the same Tectronics 442 oscilloscope to take measurements on this TI-36X solar calculator as the previous calculator. The first electronic measurements I made were the same as I did on the previous 8.08B Examiner, connecting the oscilloscope to a spiral coil identical to the spiral pickup coil in the Examiner. I used this spiral coil to scan around the calculator body for signals. I found the signal from this solar calculator are considerably weaker than the signal I picked up on the no-name calculator that came with the previous Examiner 8.08B model. I had to use special noise cancelling and methods and ground myself to the oscilloscope chassis in order to find a reasonably stable signal. See below for the signal I picked up from the calculator with the spiral coil resting on the back of the calculator at the same location as the Examiner pickup coil.

The large rising voltage at the left is a strong pulse signal which I used to trigger the the oscilloscope screen, and the smaller pulses to the right are a lower voltage pulse train which repeats itself at the 263 Hz frequency shown. This signal was seen when the spiral coil probe was placed at the back of the calculator in the same location where the Examiner spiral coil is located. The only difference was my coil was closer, touching against the plastic on the calculator, while the Examiner coil is nearly a centimeter farther away when the calculator is attached to the Examiner. I noted that I could pick up this signal with a spiral coil made as an exact duplicate of the Examiner coil, or I could use the plain alligator clip on my probe. The Alligator clip gave a stronger signal, but I used the coil so I could maintain the same kind of pickup that is used in the Examiner. The fact that an alligator clip picks up the signal through the air convinced me the signal is not inductively coupled to the probe, and is probably not inductively coupled to the Examiner either. It appears the coupling is capacitive considering I use high impedance probes to sense millivolt signals with an alligator clip. I later found that the weak low voltage electric field of this keypad is detectable at up to six inches when using only the amplifiers on my old Tek scope and when using good grounding and shielding practice for the signals measured at the probe ends. When using the spiral coil I observed a typical calculator pulse train at the keyboard that runs at a fixed frequency, and the remnants of electronic noise from the air that cannot be fully removed. I also observed the exact same signal when I shorted the ends of the spiral coil, or replaced it with a quarter, or various pieces of aluminum foil to pick up the signal. It appears the reason for using the flat coil shape is to approximate a capacitor plate shape to pick up the weak voltage signal from the calculator.

Since this signal did not show anything except a constant frequency pulse train, I decided to open the calculator and get a cleaner signal from the circuit board. Maybe I would be able to find other signals that are not visible from the air when I connect directly. When I opened the calculator I could see the TI-36X circuit board was similar to the previous calculator with a few differences. There was no battery, only a solar panel to provide power. In place of a battery, there was a small capacitor to store a charge for a short while in case the light is interrupted. There was also a red LED used as overvoltage protection soldered across the positive and negative supply conductors on the back of the board. The rest of the calculator was similar the other calculator on the 8.08B model, except it had a larger display with more characters on it.

The conductors on the back of the keypad showed a clean signal free of background noise that I could measure without any special shielding or noise reduction, while using any scale on the oscilloscope. We are looking at the pulses measured on the keypad conductors just beneath the Examiner pickup coil. These signals are sent to the keypad conductor from the processor (located under the black drop of epoxy). Each of the keypad conductors in the vicinity of the Examiner pickup coil has an identical pulse train like we see in this photo with the exception of the positive and negative supply conductors. Pressing keys did not change the frequency on any of the keypad conductors except for a momentary blanking signal while the key is held down. The frequency did not change after performing arithmetic operations or other functions, or the pressing the equal key. I did notice that whenever a function key was pressed it could take from a quarter second up to a half second for the calculation to finish and show the result on the display. When a new number appeared on the display, the signal I measured was not changed at any of the keypad conductors.

But I did observe some changes in the frequencies at the back of the keypad. At first I wondered if my test equipment was out of calibration. Then I noticed the frequency of the pulses on the keypad conductors was changing when the intensity of light falling on the solar panel changed. As the light became dimmer, the frequency of the pulses at the keypad conductors slowed down. When I adjusted the window blinds from the bright sunlight to a partially shaded room, the frequency could drop to half. And if I kept the calculator in a darker location long enough, it would drop to one third of it's frequency, then eventually it would stop sending internal pulses. This explained why I measured a different frequency than I did earlier from outside of the calculator. The light at the workbench was brighter than when I measured it earlier, so the frequenct was increased. The frequency variations seemed to be more noticeable on overcast days that had less sunlight, or at sunset time when the sky becomes dimmer. It seems like this calculator's frequency is dependent on the voltage supplied by it's solar panel. Like most pocket calculators, there is no crystal reference frequency, so the clock is free to drift with temperature and voltage variations. This is not a problem for making calculations because the refresh rate is fast enough to provide calculations in a fraction of a second regardless of the lighting condition. The impact of these frequency variations is negligable for making calculations because you won't notice a lag unless it is too dark to use the calculator.

Apr 4, 2011 - More Deluxe Examiner calculator testing: TI-36X Solar model

On April 3, 2011, I received an email request to take some measurements to solve some online forum debating about whether the TI solar calculator on the Examiner sends out various frequencies from the keypad conductors, or if these keypad frequencies are constant. Since I already knew the answer from my testing, I decided to photograph some measurements of the frequency on these conductors to show exactly what frequencies are present. I could also show how pressing keys or making calculations does not cause any of the frequencies to change, but changing the light conditions on the solar panel will cause the frequency to change.

Calculator frequency changes when light changes
I connected the oscilloscope and a Fluke 187 meter to the TI-36X solar calculator circuit board and recorded the voltage and pulse timing as I varied the lighting conditions. As the voltage changed in different lighting conditions, I photographed the signal displayed on the screen. In these images I am measuring the time duration between the large pulses that start each pulse train. As the frequency of the signal drops in dim light, the second pulse moves off the screen, so I extrapolated the period from the smaller pulses which can still be seen when the frequency drops (I also confirmed these frequencies by checking at different time sweeps where I could see the second pulse). The voltages seen at the oscilloscope trace are slightly lower than the actual voltage because the voltage trimmer was turned down slightly. But I used the Fluke 187 meter to to monitor the voltage at the the power leads for the solar panel and I watched this meter to see when the voltage reached the exact value I needed to take a picture of the signal. These voltages measured with the meter are the voltages you see printed on the images.

As you can see, there is quite a variation in frequency depending on how you adjust the light falling on the calculator. Notice that all the pulses on the screen showed dropping frequencies as the voltage dropped. I repeated these measurements on all the other keypad conductors and found the frequency of the signals were the same, and they varied the same way as the signals we see above, with the exception of the positive and negative supply conductors. During these measurements nothing was changed except adjusting the light falling on the calculator solar panel by adjusting shades to partially bkock out various amounts of light. When the voltage reached an even number I took a photo, then repeated the procedure for other voltages. Pressing keys had no influence on the signal at all except when the voltage was becoming low in the vicinity of 1 volt. At this level and below, pressing keys tended to cause the voltage to drop more rapidly and shut down the calculator. Somewhere around 0.67 v is where the calculator would suddenly stop producing a signal, and would not recover until it got a shot of bright light on the solar panel to wake it up again.

From the images and voltmeter readings above, we can chart the following graph that can used to predict the frequency of this TI-36X solar calculator. While we can't know the internal voltage while carrying the calculator across a field, we can have a rough idea of the keypad frequency depending on whether we are in bright sunlight, under the open shade of trees, or in a dark corner. This chart shows both the frequency of the individual pulses sent to each conductor, and the higher sequencing frequency which is a frequency 8 times higher.

    1.83v   288 Hz Maximum voltage attainable in any light
    1.80v   285 Hz Bright sunlight
    1.70v   270 Hz
    1.60v   260 Hz Various shaded areas
    1.50v   248 Hz
    1.40v   227 Hz
    1.30v   200 Hz dark shadows, sunset
    1.20v   174 Hz
    1.10v   152 Hz
    1.00v   124 Hz Very dark corner, display fades away
    0.90v    96 Hz No numbers visible
    0.80v    84 Hz No numbers visible
    0.70v    18 Hz triangle wave pulses breaking down
    0.67v      0 Hz No signal

When we look at the 263 Hz and 280 Hz I measured back in 2010, we can check this chart
to determine that the supply voltages were 1.63 and 1.77 volts on those two occasions.

Other calculator signal details

I also made quite a number of other measurements at the back of this calculator which are too numerous to make a detailed discussion from. One of the notable points I observed is there are four classes of conductors on the calculator circuit board: 1) Conductors for rows and columns that lead to the keys, and 2) conductors that have positive and negative supply power on them, and 3) there is a single conductor that has a square wave clocking signal at a peripheral location. And 4) there are also a group of conductors leading from the The processor to the display above it. (See the photo above showing the TI-36X Solar calculator circuit board). When I tested all these signals for their strength through the air, I found the signals at the display are the strongest. But these conductors are too far away to be detected from the pickup coil from a side angle. And the single clocking signal is also too far away to be detected from the pickup coil location. The only conductor signals that are picked up at the spiral coil are the keypad conductors near the coil, and the positive and negative supply wires also near it. The supply wires don't actually have any signal on them except for an ocasional over voltage flow of current through the LED, which is not seen as a frequency, but as a DC current that slowly increases or decreases. So the supply conductors really send no signal to the pickup coil. But the few keypad conductors in this location do send a signal that can be sensed at the spiral coil location. The keypad construction uses standard flexible conductive membrane which works by shorting two conductors together when a key is pressed. There is no decoder IC on this board, and no serial data stream. All the keypad conductors lead into the epoxy drop where the processor is located, as it is common for modern pocket calculators to do all the decoding inside the processor chip.

Two frequencies at the keypad
I also noted each of the small pulses is positioned at the same time slot where we see a large pulse firing at one of the neighboring keypad conductors. I confirmed this with a second probe channel to watch each of the keypad conductors fire a large pulse where one of the small pulses appears. Each of these conductors shows an identical signal as all the others, including the series of small pulses. It becomes apparent that the small pulses are only noise artifacts leaking in from the large pulses at the other neighboring conductors. Because these small noise pulses are far below the half-voltage level, they are not interfering with the functioning of the large pulse intended for the keypad button we are measuring.

Channel 2 shows a second nearby keypad conductor is pulsed 1.4 ms after channel 1, and at the same timing where we see a small pulse from the conductor at channel 1. I found each of the 8 rows of keys had a conductor that was pulsed at one of the 8 timing positions where we see a small pulse. This is a normal timing arrangement for a pocket calculator to pulse each of the keypad row conductors in a never-ending sequence of pulses.

We can now see that the small pulses are simply crosstalk artifacts from the other rows of keys. But it does not matter whether these small pulses are crosstalk noise or whether they were intended. They still make up part of the signal measured at the keypad conductors where the pickup coil is located, and are one of the two frequencies that are sent to the Examiner circuit. So we have two main frequencies found on the calculator keypad conductors: The keypad scanning frequency (2240 Hz small signal), and the row frequency which fires for each row of keys at 1/8 of the scan frequency (280 Hz large signal at supply voltage). It is important to note that these frequencies do not change after a key is pressed, but both the large pulses and the small pulse frequency will increase or decrease depending on the lighting conditions at the TI-36X solar calculator.

A third frequency at the keypad
During these measurements, I noticed that when I held a key pressed down, the signal on my screen was interrupted by a square wave which went high for about 4ms to obscure half of the keypad pulses. It was a 67 hz square wave that allowed the keypad to pulse only on the low half of the wave. The 67 Hz square wave disappeared and the keypad signal returned to normal after releasing the key. This square wave is a momentary interruption of the fixed frequency coming from the calculator keypad that I could see only during the time when the key is held down. I discovered most of the keys caused the exact same square wave to appear when they were held down. There were exceptions with the function keys, which caused a high signal to obscure all the keypad pulses for 1/4 second to 1/2 second. Then, after the function calculation was completed and a number appeared in the display, the 67 Hz square wave would again appear for as long as the key is held down. These momentary interruptions of the normal keypad frequency only occurred during the time while the keypad buttons were held down. When the key is released, the signal always returned to the same signal as we saw before any keys were pressed. This square wave interruption of the keypad frequency is common with many calculators when a key is held down. But for different calculator models, the frequencies and time durations that we see when a key is held down may vary.

A different frequency for different Examiners
The previous Examiner Model 8.08B used the same interanl parts as the Deluxe model, but had a different calculator which ran at a different frequency than this calculator.
Let's take a look at the signal we measured on the previous Examiner calculator and compare it to this signal.

We see the old calculator was sending out 30 Hz and 575 Hz signals, while this calculator sends out 280 Hz and 2240 Hz signals. None of these signals change after pressing calculator keys or completing calculations. So we have different Examiners with different calculators that produce different frequencies instead of sending the "resonant frequency" we are told will be produced by pressing the specified sequence of keys. The question now becomes "How do we set the calculator keypad frequency to the resonant frequency we want when it does not change after pressing keys"? And this raises another question: "how can this signal passing into the Examiner circuitry be at the resonant frequency of a substance that we are looking to find? The only frequency the calculator can have is a frequency its clock generates, or whatever frequency it drifts to as the supply voltage drifts. This can mean anywhere from 30 HZ to over 1 KHz depending on what calculator came with your Examiner. Further investigation shows that pocket calculators run at frequencies which will drift as the battery or solar panel voltage changes. Even two identical battery power calculators will not have the same exact frequency because the battery charge is not exactly the same, as well as the variations in the components that generate these frequencies inside the processor. I measured two new Casio fx-3000ES calculators and found they were different by 2% when measured under the same conditions. This is a normal state of affairs for pocket calculators that have no crystal or other stabilizing frequency reference. This amount of deviation causes no problems when performing calculations, but it can be disasterous if we needed to keep all the calculators running at the same frequency. But wait, there is a temperature drift that causes the calculators to run slower as the temperature rises. How much temperature drift? I measured 5% higher frequency when the calculator was at 0 C than when it was 25 C. You can verify this by measuring any pocket calculator frequency when it is warm, then place it in a freezer for a half hour and measure it again. These are the tempreatures we might expect to find in a real treasure hunting scenario for a climate where it snows in the winter, and becomes hot in the summer. As the temperature changes from cold winter day a to a hot summer day, we see the frequency drop 5%. This means if we had a frequency of 2500 HZ in the winter, it is now 2400 Hz on the hot summer day. All cheap pocket calculators see this temperature drift. You would need a more stable clock, maybe with a crystal to keep the frequency stable over that temperature swing.

How far does the calculator signal reach?
The distance you can detect the calculator signal depends on what equipment you are using to detect the signal. A more important question is "How far can you detect a useful signal?" The small signal that you can detect from a calculator at between 1 inch and 4 inches has a similar strength to the ambient noise in the air. My observations show that when you measure the signal at the back of the keypad in the exact same location as the Examiner coil is located, you will find a signal about 1-2 mv depending on what calculator you are using. The background noise in the air is usually less, so you can see this signal clearly. As you move the calculator away from the pickup coil, the signal becomes weaker until about 2-3 inches it has the same strengths as the noise in the air. By 4 inches, the signal is of little use because the stronger broadband noise in the air overpowers it. This noise factor can change dramatically at different locations and different times of day. Fortunately the Examiner pickup coil is less than a half inch distance from the calculator, where the calculator signal is stronger than the noise in the air.

But you could detect a useful calculator signal from a longer distance if you take special precautions to reduce noise and ground everything in the area to the chassis of your meauring equipment. It also helps to ground yourself to the chassis of your measuring instruments so you don't act as an antenna collecting stray noise to send to your test probes. I was able to observe the calculator signal on the oscilloscope up to six inches distance before it became so weak it could no longer be seen in the background noise. But when I removed the ground connections, the best I could get was maybe two inches. If I wanted to detect a calculaor signal at a longer distance than 6 inches, I could use a more sensitive oscilloscope and take more precautions to remove background noise from the air. But for our purposes, we don't need to see the signal any farther. We only removed the noise to get a clean signal from the air so we could see what the calculator sends to the Examiner pickup coil.

I should also point out there are other stronger signals that can be measured at the calculator. Another signal detectable at the LCD display is three times stronger than the keypad signal when measured at the exterior surface of the calculator. This signal becomes weaker as we move away, so we lose it in the noise at one extra inch farther than the keypad signal, But the Examiner pickup coil does not receive this signal at all. It is located so far away from the display conductors that the Examiner could not pick up the display signal from a side angle. The Examiner coil only picks up the the signals from the keypad conductors directly adjacent to it.

Calculator interface to the Examiner
But here is the interesting part. The signal that I can pick up using only a spiral coil with a capacitor at the end of it is much stronger than the signal I measure the Examiner receiveing. When I connect a probe directly to the Examiner antenna and the handle conductor, The calculator signal can barely be seen in the noise. It appears the Examiner is attenuating the calculator signal, and picking up additional noise from the air which is added to whatever signal noise comes from the calculator. Since the Examiner antenna is mostly placed in the air away from the calculator, it has plenty of surface and length to collect air noise, while only a small coil the size of a coin is near the calculator to pick up a calculator signal. Secondly, there is a problem with the way the pickup coil is wired. When we look at the internal Examiner circuit, we see there are two basic circuits combined together. One is a circuit from the ground, through the treasure hunter's body to the plastic handle, then a capacitive couple to a brass rod at the center of the handle, up through a 5 turn coil, a diode, and to the antenna. From the antenna, any signal can radiate to the atmosphere where some current can presumably return by ground path. (See the blue part of the circuit diagram below).

The second circuit is is an resonant circuit comprising the spiral coil, a 5.5 turn coil, two capacitors that result in about 5nF and a 4.4 MHz crystal. There is also a 5k series Potentiometer called sensitivity, which adds resistance to the circuit. It is hard to imagine this loop as a resonant circuit, unless we consider the loop as a simple tuner that is tuned to the frequency of the crystal which is wired in series. It could theoretically resonate if the circuit components were tuned to the same frequency as the crystal. When we check the values of the coils and capacitors, we see it is an exact match for the 4.4 MHz crystal when the trimmer is adjusted to 5.1 nF in the middle of its range. The only difficulty is the trimmer requires a very fine hair-trigger adjustment to get the frequency exact. Still, this is the first clue that this circuit was designed for a resonant frequency at 4.4 MHz.

But back to the problem...
The calculator frequency is nowhere near 4 MHz. The highest frequency it sends to the spiral coil is under 3 KHz. This frequency is easily attenuated by the 4.4 MHz circuit, both in theory and in the evidence shown when the signal was measured at the pickup coil on the Examiner. The 2260/288 Hz leaking out of the calculator sees the spiral coil only as a length of wire connecting it to the antenna. The calculator signal sees some resistance, (or more correctly impedance) to weaken it as it passes through the spiral coil on its journey to the antenna -- It does not find any resonance that would tend to strengthen it. When I measured the signal that I found at the Examiner spiral coil connection to the antenna, I saw it is about 1/3 the strength as I can measure it when only using a probe with a spiral coil clone connected to it.
Further investigation of this spiral coil revealed that it is not inductively coupled to the calculator signal. The Examiner spiral coil shows the same exact calculator signal strength when we short the leads together, or leave them untouched. Since the shape of the spiral coil approximates a capacitor plate, it can act as a capacitor to pick up the electric field of the calculator conductors as they clock on and off to scan the keypad. This is exactly what we found with the clone spiral coil. It acts like a capacitor plate, not an inductive pickup at these frequencies found on the calculator. Apparently the capacitive coupling is causing the entire Examiner circuit to charge with a very small amount of the calculator signal. This pickup coil could actually be substituted for a metal disc the size of a US nickel to make a capacitor plate which acts the same as the spiral coil to pick up a miniscule charge from the closest part of the circuit board. The appearance is the spiral coil is acting as a capacitor plate which picks up a small signal from the calculator and connects it to the antenna where it is mixed with other air noise. Then, remember my previous calculator signal measurements... I could get a good signal only when I took adequate sheilding and grounding precautions to keep the air noise away. I had to ground my own body to the chassis of my insruments in order to keep from acting like an airborne noise collector when I was handling probes and the calculator. Yet, when tresure hunting, there is no shielding. The person holding the Examiner is acting as an antenna probably better than the telescoping antenna. The idea that bio signals are coming from his body to interact with the Examiner signal seems ludicrous. Any bio signals could only be measured after connecting a closed circuit measuring instrument which is not influenced by atmospheric noise. Yet this is not a closed circuit for the Examiner. We see the treasure hunter standing on the ground acting as an antenna to help collect atmospheric charge and noise while holding the Examiner handle, which is made from plastic that is isolated from the examiner circuitry. In real treasure hunting conditions, any bio signals are easily swamped in noise that a person collects from the atmosphere while acting as an antenna. And a very small part of the AC noise he collects can be expected to enter into the Examiner circuitry through the handle capacitance. But what do these noise signals do after they enter the Examiner handle? They must pass through a small coil and a diode that is soldered to the antenna, which is also picking up air noise. So there is no measurable voltage differential between the noise coming to the antenna directly from the air or through the handle from the treasure hunter. The treasure hunter becomes, in essence an extension to the antenna that has a diode and small coil inline. Does the silicon diode do anything? No. Not with the weak microvolt signal coming in at equal strengths from both sides of it. Maybe if we consider the treasure hunter part of a dipole opposite the aerial, we could speculate that it can tune some VHF broadcast. But the Examiner has no circuit to tune anything in the VHF range.

But what about the "resonant circuit"?
The 4.4 MHz resonant circuit does not appear to be resonating at 4.4 MHz because of the placement of components in this circuit. In theory, it could resonate, but in practice it does not, as confirmed by testing. Since the microvolt levels of these atmospheric noise signals are too low to cause any serious current to flow in this circuit, we don't expect to have any measureable signal other than the same noise which you can find by holding a wire in the air. In fact, this is what it appears the examiner is collecting - airbound noise. I have been able to find no special tuning of any frequency at the internal conductors of the Examiner, nor on the antenna or handle or other conductors exposed at the exterior. It appears that all of the metal parts in the Examiner are acting as a single antenna which will pick up any electronic noise, including atmospheric noise, calculator clocking noise, RF broadcasts, or other noises such as sparks from a battery or welder. It collects airborne noise with the same efficiency as a coat hanger wire.

It should be noted that when I measured the calculator signal I could find from the Examiner antenna, the calculator signal became weaker as I tested at points along the antenna which were farther away from the calculator. It appears that the calculator signal I measure at the antenna is largely picked up by the clip at the end of the test probe, capturing the calculator signal directly through the air, And this secondary path becomes weaker as the probe is moved down the antenna away from the calculator. The weak calculator signal at the end of the antenna suggests that only a small fraction of the calculator signal strength is actually influencing the antenna which the pickup coil is soldered to. And it also suggests that the Examiner internal cirucitry attenuates whatever calculator signal it receives, rather than to strengthen it by means of a resonant circuit.

About power
The strongest signals I was able to measure inside the Examiner circuitry was less than 2 microvolts unless I intentionally held the Examiner near an electric power circuit. For normal treasure hunting, I expect somewhere between 0.25 and 1 uV strength of noise signals in the internal circuits, based on what I observed in outdoor tests. The question arises: "How does a microvolt signal in the internal wiring cause a substantial rotatational force in the Examiner?"
Electrically a microvolt signal is not capable of causing the Examiner to swivel against the force of gravity. It would take substantially more power and some actuating mechanism to cause the examiner to swivel. The Examiner does not have any swivel actuating mechanisms inside, or any power supply capable of moving the Examiner if it did have a suitable actuating mechanism. The only obvious force strong enough to move the Examiner against the force of gravity is the user's hand muscles.

So what can this circuit do?
1. The calculator signal stops at the antenna. It cannot resonate in the 4.4 MHz loop so it mixes with antenna noise collected from the air. Some of it is radiated from the antenna and metal parts of the Examiner to where it can be detected up to a few cm distance from the metal parts. The strongest signal from the calculator can actually be measured above the calculator, rather than at the Examiner antenna or other metal parts of the Examiner.

2. The resonant 4 MHz loop circuit could possibly resonate if there was an electric field or a magnetic field at that same frequency nearby of sufficient strength to be tuned and properly coupled to the circuit. There is no 4 MHz from the calculator. It does not clock anywhere near that high, But it is possible there could be 4.433619 MHz noise in the air at times. This is the same frequency used in video game monitors. I suppose you might find that frequency in the air if there was some video game hardware nearby. And there is also a chance of random noise in the air that may occasionally come at that frequency. And if it did... What would happen? If we are lucky and this circuit is working, then the loop will resonate in tune with the incoming 4.433619 MHz, and reinforce the magnetic waves at the spiral coil and helical coil to cause this frequency to be stronger locally in this loop. Thats all. There is no amplifier connected, so thats all it does. The other circuit with the operator and antenna is not actually connected except by a single solder joint. There is no circuit between the two, just a common connection. They appear to be operating independently. Signal tests have shown that there is no 4.4 MHz signal anywhere in the Examiner circuit or near it in the air that is detectable. So the theorized resonance is only a theory based on the calculated values of the components. This loop is not in fact resonating with enough strength to measure on the sub microvolt scales of an oscilloscope. The signal found at the loop is only noise, modified by the presence of the 1000 uF capacitor. It looks nearly the same as the noise coming from the air at the antenna, and the noise I can see if I connect the probe to a coat hanger and hold it in the air. I have established with measurements that the claculator is sending out 288 Hz and 2280 Hz signals and the Examiner circuit shows these signals weaker while it shows the air noise stronger, and no resonant frequencies changing after numbers are put into the calculator. But I did not establish anything about the other circuit from the operator through the handle and to the antenna except it looks about the same as air noise. It could be that this is all it is. But it could be something different. This is something I can't know, because I did not do an extensive test on this circuit. It is possible there is some treasure locating method connected with this second circuit, even if it does not appear to be likely. But I can say with certainty the 4.4 MHz circuit is not resonating at it's resonant frequency enough to measure. And I can say with certainty the two frequencies coming from the calculator are the same before and after any sequence of keys are pressed. And I can say with certainty that no signal from the calculator is resonating with the internal circuits of the Examiner. The calculator frequencies are attenuated by the internal circuits of the Examiner.

Calculator signal is coupled to the Examiner circuit by capacitive coupling
After spending hours detecting the calculator signal with different kinds of probes, it became very apparent that this signal is coupled to the Examiner through capacitive coupling rather than the inductive coupling that was claimed to transfer the signal. The calculator signal moves from the calculator to the spiral coil inside the Examiner. Testing has shown this coil is picking up the electric field which can be detected when conductors are pulsed at audio frequencies with less than 3 volts inside the calculator. Magnetic inductance is ruled out because there is no magnetic primary coil in the calculator to induce a magnetic field. We can also rule out RF tuning, because the pickup circuit is soldered to a circuit which is tuned to 4.4 MHz, not an audio frequency. we are left with capacitive coupling causing the transfer of the calculator signal. This capacitive coupling was confirmed by picking up an idential signal to the spiral coil signal from a flat plate or coin the same size as the coil. We expect to pick up an AC component of one or more pulsing frequencies from the calculator keypad between 30 Hz and 3000 Hz, which can vary between calculator models. This signal enters directly into the 4.4 MHz tuned circuit where it is virtually dissipated due to the manner in which it is wired. But more importantly, the calculator signal is not capable of finding resonance with this tuned curcuit even with proper wiring, because it is not a 4.4 MHz signal. Since there is no 4.4 Mhz signal picked up from the calculator, the calculator's audio frequencies arrive at a circuit where they have no influence. But the calculator signal also goes to the antenna to mix with other noise that was picked up, and it is free to travel out into the air for a few inches past the antenna until it becomes lost in the noise so it can no longer be seen. Signal testing has confirmed that the calculator signal can be measured on the Examiner antenna. But this signal is attenuated to be much weaker than if you measure it with only a spiral coil or small foil plate held at the back of the calculator. It appears that the internal examiner circuit attenuates this signal.

This calculator shows signals that are expected from any modern pocket calculator. It would be highly unusual if we found any frequencies changing after pressing keys or after performing calculations. This is because the electronic design is basically the same for all pocket calculators that use a chip-on-board processor. The only things that change are details such as batteries and solar power arrangements, and how many keys and digits are on the display. Like most pocket calculators, this calculator has no crystal or other reference to set its clock to. Whatever frequency the clock drifts to is unimportant as long as it remains slow enough to keep the power consumption low, but not so slow to cause an excessive lag in performing calculations. All of the calculating and decoding is done in the processor under the drop of epoxy. None of the other parts of the calculator have any electronics that can perform any calculating or change any frequencies. The keys have no electronics other than a switch that can short two conductors together underneath it. All the wires leading to the calculator keys are pulsed at a fixed frequency determined by the main clock. This pulsing is done in order to check to see if a key has caused any of the conductors to connect together. While most calculators maintain a relatively constant clock frequency, the TI-36x solar calculator has large frequency swings because there is no battery to hold the supply voltage constant when the light gets dim.

This is only a test report describing what I observed when I tested the signals for the Deluxe Examiner. You don't need to take my word for the test results. Everything I measured at the calculator keypad is repeatable so you can measure it yourself to see the same signals as I did.

Summary of test results:

Here are a few things I measured on the Examiner calculator:
1. The calculator signal is an audio frequency clocking signal which radiates out as an electric field several inches until it is no longer detectable above the electronic noise in the air.
2. The spiral coil picks up the signal from the calculator by capacitive coupling, as evidenced by an identical signal received using a flat plate the same size as the coil. The calculator signal is attenuated by the Examiner circuit.
3. The frequency that was picked up from the Examiner calculators did not change after pressing any combination of keys to put different numbers in the display. There is no way to adjust the calculator clock frequency other than allowing it to drift with variations of temperature or supply voltage.
4. The solar calculator frequency dropped to one third of its frequency when it was moved from bright sunlight to a dark corner.
5. The frequency of the calculator can change by a significant amount when the temperature changes in a treasure hunting field.
6. Various calculators installed on different models of the Examiner produce differnt fixed frequencies (The calculator on the deluxe model produced a weaker signal that is not the same frequency as the calculator on the T-G 808B model).
7. The same model calculator seldom has the exact same frequency due to variations in the components and aging of the battery, or variations of the light strength for solar models. This inherent clock frequency is not influenced by any numbers that are entered into the display.

Test conclusions

1. This testing wouldn't be worth the trouble, except it answered some nagging questions of "What kind of signal is sent from that calculator?" As it turned out, The signal sent is a keypad-scanning signal which can be displayed as a pulse train.

2. It seems the people who thought a calculator could not transfer a signal to a conductor nearby were wrong.
We see clearly the signal did reach through more a centimeter of air and plastic to arrive at the oscilloscope probe and show itself on the screen. I was able to detect this signal at up to six inches (15 cm) before it became too small for me to see it in the oscilloscope screen. Signal measurements taken directly on the circuit board showed that no part of this signal was missed when detecting it through the air.

3. Test results show the signal does not transfer by inductive means. It behaves as capacitive coupling as evidenced by better transfer to an alligator clip, coins and several scraps of aluminum foil than to a spiral coil identical to the pickup coil in the Examiner. The spiral coil picked up the identical signal whether it was shorted, open, or wrapped in aluminum foil. This result also dispells notions of resonant magnetic coupling.

4. People who thought that the frequency of the keypad conductors would change after pressing calculator keys or performing arithmeic operations were were also wrong. The evidence shows that no matter what keys are pressed, the frequency on the conductors is only momentarily inturrupted while the key is held down. When any combination of keys are released, the signal coming from the keypad returns the same as it was before the keys were pressed.

5. The solar calculator tested showed a constant fixed frequency as long as the light level remained constant. But the frequency showed some heavy drifting when the light level changed.

6. The test results showed the frequency of the no-name calculator used on the previous Examiner is different than the frequency on TI-36X calculator. We also saw that two identical model Casio calculators (Non-RangerTell) do not run at the exact same frequency.

7. The test results raise questions of whether there is any point in pressing keypad numbers or function keys when we see the the calculator continues to send the same signal as it did before any keys were pressed. And it seems more pointless when we see how far the keypad frequency can drift without pressing any keys at all.

8. The test results also show the signal coming from this calculator can be transfered and detected up to six inches distance in the air. But this does not answer the question of how the intended frequency can arrive at the receiving circuit when the frequency cannot be adjusted at the keypad. This spoils any theories of resonant frequencies coming from a calculator, especially when the solar calculator's fixed frequency drifts heavily as the light changes. Even identical calculators without solar panels show variations in frequency that can become significant as the battery ages.

9. These are not small details. These discreptancies pertain to the basic premise that a calculator sends out a resonant frequency to match a treasure frequency, and which can be adjusted by pressing calculator keys. These details lead us to believe that if the Examiner works at all, then it is not working by means a programmable calculator frequency.