Hammerhead Info

[Judge]

I have been constructing the HH, and have reached a point where I have a few questions that may have answers that will help others. Most may seem pretty simple, but maybe I am not alone with these difficulties. Anyway, I figure others may benefit too from this. Here are my questions.

When everyone talks about delay times before the first sampling for a fast coil. Is the correct measure of the delay time for a coil the time from the end of the TX pulse to the zero point on the flyback pulse? Should I be trying to reduce the time between the switch off of TX and the start of the flyback pulse? It currently is about 5 usec. How is this done? Should I adjust the Tx pulse width to reduce the width of the flyback pulse?, or should build a different coil?
How do I use the measured resonance frequency of a coil to calculate its inductance in microHenries?

On a scope I can see a change in the output at TP9, but this does not register as anything coming from the speaker. Is this to do with the threshold voltage setting? I only get a sound out of the speaker/headphones if the signal saturates at TP9 (hits the ceiling).

In the construction and calibration section of Carl's paper on the HH it says that the output from the NE5534AP should be adjusted with R18 to be 0.0. Why is this done?

I have used 30 micron Al foil for a shield and could see no effect from this? How do I know if my foil shield is working correctly?

Is there such a thing as ground balance for PI like the HH?

Well, enough of my ramblings.


I would like to thank Carl. The HH is a brilliant learning tool, and I have enjoyed every moment working on it.

[Reg]

Hi Judge,

You have posted some good questions and I will try to give you my explanation of some of them.

First, I refer to the delay time as the time from when the transmit pulse is turned off to the time when I actually sample the receive signal. One can measure this several ways and get different answers so the minimal delay setting may vary a microsecond or two (usec) depending upon how you measure it. A good technique is to simply measure the pulse to the FET gate with one probe and the output of the preamp with the other, providing you can see the slight nitch caused by the sampling pulse. Another technique is to look at the transmit pulse and and the output of the preamp and simply measure from when the transmit is turned off to when the sample is taken.

When trying to obtain a very short delay time, one should realize that the signal at the output of the preamp should have leveled off at 0V when the sample is taken, or very close to it. Sampling while the signal is leveling off can be done but may saturate the next stage and cause the detector to be much more ground sensitive. The big problem is the saturation of the next stage. If this happens, then any signal response you get will be extremely weak if noticeable at all.

It doesn't really matter just how you measure this delay time as long as one knows the trick to finding lower conductive objects is to shorten the time. If you can detect your desired targets then you have reached your goal.

The key to shortening the overall time is to figure out ways to shorten all things involved. The most common thing to try to shorten the the decay time of the coil. This is the time it takes for the flyback pulse to decay to 0V. Several techniques are involved including coil winding techniques, type of wire used, total inductance, damping resistance, and even changing the FET. If you reduce the inductance, you can reduce the delay time also, but will reduce the sensitivity somewhat. My recommendation if you are trying to build a PI with a shorter delay is to keep the delay between 200 uh and maybe 350 uh. If you make the inductance 300 uh, then you probably can use any Minelab PI coil also. They are expensive but available. Personally, I prefer to make my own.

If you are getting the delay down to 15 usec or so, then you are doing fine. Reducing the delay below this point becomes much more difficult, especially on the Hammerhead which is a higher powerered PI.

One way to measure the resonant frequency of a coil is to take a coil to be measured and connect a scope across the coil leads. Then move the coil close to your operating PI that has a different coil on it. YOu can also look at the output of the preamp by bringing this second coil close also. When you do this, you will see an oscillation which is the result of the second coil characteristics. Now, you can measure the time between peaks on the sinewave to get an approximation of the frequency. The trick is to reduce this time between peaks.

The reason you adjust for 0V at the output of the preamp (NE5534A) is to allow for maximum swing of the signal. If the signal is too far either way, the next stage can saturate quite easily. So, keeping the voltage at 0V is the best.

I am not sure just why you are not getting any signal out of the speaker even though you see a change at TP 9. I am not sure whether you built ver 1 or 2, which will make a difference. In either case, you might carefully check all connections and make sure you have all parts in correctly. My guess is there is something wrong with one of the transistors, but that is just a wild guess at this time.

The use of shielding is somewhat difficult to see unless you are looking very carefully at all things. Shielding can reduce the oscillation of a coil if done correctly. This will allow for a little better or faster coil. Now, too heavy of a shielding will affect the decay curve because the shielding is "detected" as a target, so care has to be used as to just what is used as a shield.

One of the main reasons for a shield is to reduce what is commonly called the ground capacitance. This capacitance will cause more of a dramatic ground signal as the coil is lowered towards the ground. The shield will also help reduce any static problems which can commonly cause false signals when the coil bumps grass or other vegatation.

Now, if the delay is quite long, you can minimize or even eliminate the shielding since the ground signal will be gone. However, it is a good idea to use a shielding, at least to begin with.

Is there a form of ground balance available for the Hammerhead (HH)? Nope, not at this time, but one can build a form of it if they want to. What is needed is to take a later sample, maybe at 100 usec or so and then amplify this second sample and subtract it from the main sample at some later stage after the integrators (then next stage after the preamp is the integrator stage). You can look at that and get the general idea by looking at the Goldscan schematic posted on this site in another thread.

You can also modify the HH so a DD coil can be used and build a DD coil. There are some posts on how to do this also. A DD coil will reduce the ground signal quite a bit also.

Well, I haven't answered all your questions, but have hit most of them. One question I didn't answer is how to calculate the inductance of a coil. I prefer to use an LCR meter and simply measure it. That is because I am lazy, among other things.

Hope this helps.

Reg

[Judge]

Reg,

Thanks for all the great info. It has given me a good definition of delay time, and given me some idea of what to search for with my audio weakness. I will have to borrow or buy an LCR meter to measure inductance. I will look at whether the stage following the NE5534 is being saturated, and check the transistors.



I have one other question for the moment.



When I am looking at the flyback pulse I get a very different trace with the 1x and 10x on my oscilloscope. For 1x I see a large negative segement of the flyback pulse after the positive spike. With the 10x the trace looks like the trace in Carl’s article. Should I be using the 1x or 10x probe to look at the flyback pulse?



Thanks again,

Judge

[Reg]

Hi Judge,

You will get a different signal when using a x1 probe than a x10 probe for two reasons. First, is the sensitiivty and the second is the circuit loading. When looking at the Xmit pulse, it is a good idea to use the 10x probe just to minimize problems.

Now, you didn't ask about how to look for the saturation of a stage, but I will jump in anyway. If the power supplies are +5 and -5, the general saturation voltage of many opamps that are not rail to rail output is about 4.2V or so. So, if you are very close to 4V or more, then you have something wrong or the signal is simply too much coming in. It is best to work with the voltage much lower when possible.

Reg

[Judge]

Reg,

Thanks for the additional info. I will use the 10x on the scope, and your tip on saturation of the op amps should help me track my problem. I built option 1 VCO.

I borrowed an LCR meter and my test coils are 500 and 300uH, with delay times of 15usec and 7usec.

Thanks again, I will let you know how I progress.

Judge