Not so DC waveform of minus 5 volt rail at the ne5534 - could this be cause of noise problems

[green]

I've taken it off the bench now, but it was only a few milli-volts if I remember correctly.
Given that the amplification level through the whole chain from coil input to speaker is huge, it should be no surprise that the audio threshold randomly wobbles around somewhat. No doubt this noise level could be reduced by some redesign, but then it would no longer be a Baracuda. That's why I resisted making too many changes, and tried to keep to the original design as much as possible while fixing the unreliablity issues.

Not suggesting a redesign. The question, is the noise level what is expected. There were some suggestions earlier in the thread on how to fix the problem. Do we know if Polymer's detector has a problem? You got a few mv at U7 pin7 where Polymer got 300mv p-p with coil connected. Maybe he does have a problem.

[Qiaozhi]

That all sounds about right and similar to the Hammer Head (also a fairly simple design).
If I turn the threshold too low then the audio is wobbly - so turn up the threshold so the audio is just quiet. Still very good sensitivity.

The Hammerhead works best with the VCO option.

[Qiaozhi]

Do you remember what the p-p noise level was at U7 pin7?

I've taken it off the bench now, but it was only a few milli-volts if I remember correctly.
Given that the amplification level through the whole chain from coil input to speaker is huge, it should be no surprise that the audio threshold randomly wobbles around somewhat. No doubt this noise level could be reduced by some redesign, but then it would no longer be a Baracuda. That's why I resisted making too many changes, and tried to keep to the original design as much as possible while fixing the unreliablity issues.

[waltr]

That all sounds about right and similar to the Hammer Head (also a fairly simple design).
If I turn the threshold too low then the audio is wobbly - so turn up the threshold so the audio is just quiet. Still very good sensitivity.

[green]

I put the Geotech Baracuda back on the bench and checked the audio threshold level. Yes, there is a random variation of the audio, but this is a consequence of the Baracuda design. The audio tone is generated from the TX oscillator, and is gated by the strength of the target signal.

These are the measurements I made on the circuit with the scope:
1. PL4, pin2 - audio signal shows some random instability.
2. Junction of R35 (100k) and D9 ( 1N4148 ). i.e. U8 base - signal amplitude variation detected.
3. Junction of R24 (2M2) and C20 (100nF). i.e. D4 gate - signal is rock solid.
4. R34 (22k) pin connected to U7, pin 7 - signal amplitude variation detected.
5. Junction of R29 (100k) and C9 (470nF) - small amplitude variation detected. Note: Following stage amplifies this variation by 100x.
6. Junction of R31 (33k) and R33 (1k) - minimal signal amplitude variation.

This appears to be normal behaviour for this particular design where the audio tone increases in amplitude as the target signal gets stronger.

The Geotech Baracuda version does not change any of this original behaviour. The purpose of this revision was to fix some design issues which were causing unreliability and early component failure. As a consequence there was a increase in sensitivity due to some of these changes, but nothing that would introduce an audio instability problem. The bottom line is that the Baracuda design is somewhat "cheap and cheerful", but it does the job and gives reasonable results.

If you want a detector that is more stable, then select one which has a VCO type audio output, such as the MiniPulse Plus (MPP). In that particular design the audio is generated by a separate oscillator, and increases in frequency as the target signal gets stronger. Since human hearing is more sensitive to changes in tone than to amplitude variation, this can provide more sensitivity to weak signals.

I'm interested in Polymer's noise problem because I have the same problem with my detector. Different circuits, amplitudes not the same but same general shape(random noise?) I started a thread awhile back on the noise problem and had some suggestions on how to improve noise level such as shielding the circuit. Seems like the place to start is to calculate amplifier out noise with coil input shorted and compare with measured amplifier noise to see if noise level could be improved. Polymer disconnected the amplifier from the integrator and verified the noise was coming from the amplifier. Detection distance is controlled by S/N ratio, reducing noise in half should be same the same as doubling coil current. Could someone calculate p-p amplifier noise out with the coil shorted for the Geotech Baracuda?

Do you remember what the p-p noise level was at U7 pin7?

[Qiaozhi]

I put the Geotech Baracuda back on the bench and checked the audio threshold level. Yes, there is a random variation of the audio, but this is a consequence of the Baracuda design. The audio tone is generated from the TX oscillator, and is gated by the strength of the target signal.

These are the measurements I made on the circuit with the scope:
1. PL4, pin2 - audio signal shows some random instability.
2. Junction of R35 (100k) and D9 ( 1N4148 ). i.e. U8 base - signal amplitude variation detected.
3. Junction of R24 (2M2) and C20 (100nF). i.e. D4 gate - signal is rock solid.
4. R34 (22k) pin connected to U7, pin 7 - signal amplitude variation detected.
5. Junction of R29 (100k) and C9 (470nF) - small amplitude variation detected. Note: Following stage amplifies this variation by 100x.
6. Junction of R31 (33k) and R33 (1k) - minimal signal amplitude variation.

This appears to be normal behaviour for this particular design where the audio tone increases in amplitude as the target signal gets stronger.

The Geotech Baracuda version does not change any of this original behaviour. The purpose of this revision was to fix some design issues which were causing unreliability and early component failure. As a consequence there was a increase in sensitivity due to some of these changes, but nothing that would introduce an audio instability problem. The bottom line is that the Baracuda design is somewhat "cheap and cheerful", but it does the job and gives reasonable results.

If you want a detector that is more stable, then select one which has a VCO type audio output, such as the MiniPulse Plus (MPP). In that particular design the audio is generated by a separate oscillator, and increases in frequency as the target signal gets stronger. Since human hearing is more sensitive to changes in tone than to amplitude variation, this can provide more sensitivity to weak signals.

[green]

Aye, that would be great - a figure 8 coil with less noise than no coil I get your point here! Thank you.

I did make two scope pics at 1msec & erronously at 2msec instead of 10msec. If you need the 10ms one, I will hook up everything again.
So I am posting 1ms & 2ms. Coil is disconnected and there is no damping resistor either.

[ATTACH]42783[/ATTACH] [ATTACH]42784[/ATTACH]

Here two scope pics of Pin 7 of U7(LF412) @ 200ms.
One with C15 diconnected and one with C15 connected again at same scope range.
From what I can tell the signal without C15 is dead quiet ... the threshhold sounds super solid here.

[ATTACH]42785[/ATTACH] [ATTACH]42786[/ATTACH]

Thought it might be interesting to note: When I do my noise "measurements" out in the wild to
see what may be interfering or not, I have no oscilloscope with me, as this might be causing noise too.

So I am just listening to the threshold, hearing if the wah-wah of the threshold wobble is still there and how strong it is.
Not a precise measurement by any means, but good enough for me if I can get that wah-wah amplitude down enough.

Thanks for the measurements. My error on the amplifier out measurement. I'm interested in the noise level at amplifier out, forgot about Tx causing a signal with coil disconnected. I'm expecting maybe 5mv p-p noise at amplifier out with coil disconnected and Tx off. I disable Tx when measuring amplifier out noise with my circuit. Maybe not needed with your circuit. 1ms time base is probably good, maybe try 200usec. Play with the scope settings and see what looks best to you trying to see p-p noise level. Disconnecting C15 verified the noise is coming from the amplifier and coil. Don't know if my estimate of noise level at amplifier out is correct with coil disconnected. Maybe someone could calculate what is expected. Don't know how to calculate expected noise level at U7 pin7 knowing amplifier out noise. Maybe someone could suggest a method.

Maybe you could try shorting D1, D2 when measuring amplifier out noise with coil disconnected to see what effect removing resistor noise has.

You need the damping resistor when making the noise measurements with coil disconnected.

Might have to adjust null offset when measuring amplifier out noise with coil disconnected.

[Polymer]

Some thoughts, figure8 coil makes a difference for me with my detector. Your noise without the coil is about half the noise with the coil. Guessing noise with coil would be about noise without coil if noise without coil was zero. http://www.geotech1.com/forums/attac...2&d=1523301124 http://www.geotech1.com/forums/attac...4&d=1523301171. from reply #21
The noise with the figure8 coil isn't going to be less than noise with no coil. Could you record a scope picture of amplifier out(U2 pin6) with coil disconnected at maybe 1msec/div and 10msec/div? Would be interested a scope picture of Pin 7 of U7 at 200msec/div with R16 or C15 removed.

Aye, that would be great - a figure 8 coil with less noise than no coil I get your point here! Thank you.

I did make two scope pics at 1msec & erronously at 2msec instead of 10msec. If you need the 10ms one, I will hook up everything again.
So I am posting 1ms & 2ms. Coil is disconnected and there is no damping resistor either.



Here two scope pics of Pin 7 of U7(LF412) @ 200ms.
One with C15 diconnected and one with C15 connected again at same scope range.
From what I can tell the signal without C15 is dead quiet ... the threshhold sounds super solid here.



Thought it might be interesting to note: When I do my noise "measurements" out in the wild to
see what may be interfering or not, I have no oscilloscope with me, as this might be causing noise too.

So I am just listening to the threshold, hearing if the wah-wah of the threshold wobble is still there and how strong it is.
Not a precise measurement by any means, but good enough for me if I can get that wah-wah amplitude down enough.

[green]

Thought I would put up a summary of this problem and what I have tried up to now.
Thank you very much to everyone for participating in this hunt so far and hopefully we will find solutions.

But first a description of the problem for everyone not reading the entire thread.

I have a strong wobble/jitter/pulsating/warble of the threshold.
The frequency of this "wobble" is in the low Hz (anywhere from 0.5 Hz to 5 Hz).

Lowering the threshold enough to not hear it causes the detector to have
detecting "holes" (aka not detecting) when the amplitude of the wavering threshold is at a low.

What I have done so far as to coils/location. Results:

- Changing coil position vertical/horizontal makes no difference
- figure 8 coil makes no difference
- went into valley with no cellphone reception, no mains, problem persists
- coil in 3ft of salt water, problem still there

-> What I still want to do is going into a cave with lots of rock around to exclude radar pulsing or similar stuff.

What have I done with the pcb:

- running on batteries makes no difference to benchtop power supply
- tossed out the lt1054 inverter and ran off pos+neg battery pack. LT1054 is NOT source of problem.

-> I will be changing the 5534's, as SVEN1 mentions in Post#6. It is plausible to me. I do have TI 5534's.
-> I have read many little tidbits of info here and elsewhere, will be doing these changes, will see what happens, too many to mention.

For now it looks like it has to do with components/combinations or a very strong signal I cannot get away from. Cave will answer that.

Have you had this problem? Any other ideas or theories ... you're welcome.

Some thoughts, figure8 coil makes a difference for me with my detector. Your noise without the coil is about half the noise with the coil. Guessing noise with coil would be about noise without coil if noise without coil was zero. http://www.geotech1.com/forums/attac...2&d=1523301124 http://www.geotech1.com/forums/attac...4&d=1523301171. from reply #21
The noise with the figure8 coil isn't going to be less than noise with no coil. Could you record a scope picture of amplifier out(U2 pin6) with coil disconnected at maybe 1msec/div and 10msec/div? Would be interested a scope picture of Pin 7 of U7 at 200msec/div with R16 or C15 removed.

[6666]

Does anything alter if you put a 100 PF cap across the preamp input pins ?

[Polymer]

Recently rec'd some knock off Texas Instruments NE5534's, they give off a rather wonky wave pattern, not even close to what it normally looks like. The detector still functioned.
As soon as I changed out to a real TI NE5534, the wave form was normal.

Hello again SVEN1,
Do you by any chance have scope pics of the wonky wave pattern?
Perhaps one of the "normal" waveform as a comparison?
Thank You!

[Polymer]

Thought I would put up a summary of this problem and what I have tried up to now.
Thank you very much to everyone for participating in this hunt so far and hopefully we will find solutions.

But first a description of the problem for everyone not reading the entire thread.

I have a strong wobble/jitter/pulsating/warble of the threshold.
The frequency of this "wobble" is in the low Hz (anywhere from 0.5 Hz to 5 Hz).

Lowering the threshold enough to not hear it causes the detector to have
detecting "holes" (aka not detecting) when the amplitude of the wavering threshold is at a low.

What I have done so far as to coils/location. Results:

- Changing coil position vertical/horizontal makes no difference
- figure 8 coil makes no difference
- went into valley with no cellphone reception, no mains, problem persists
- coil in 3ft of salt water, problem still there

-> What I still want to do is going into a cave with lots of rock around to exclude radar pulsing or similar stuff.

What have I done with the pcb:

- running on batteries makes no difference to benchtop power supply
- tossed out the lt1054 inverter and ran off pos+neg battery pack. LT1054 is NOT source of problem.

-> I will be changing the 5534's, as SVEN1 mentions in Post#6. It is plausible to me. I do have TI 5534's.
-> I have read many little tidbits of info here and elsewhere, will be doing these changes, will see what happens, too many to mention.

For now it looks like it has to do with components/combinations or a very strong signal I cannot get away from. Cave will answer that.

Have you had this problem? Any other ideas or theories ... you're welcome.

[Polymer]

Don't know if there is a problem. Amplifier and resistor noise are expected, also noise from coil pickup. I think what is needed is a test method to determine if noise level is higher than expected. Would be nice if someone would post a scope picture of U7 pin7 of a good bara.

I absolutely agree with you green! A good reference scope shot for a "normal" noise level would be great. For abnormal wobble noise, see my scope pics

I will put up a summary tonight of tests done so far inside and outside and what else I will be doing
(like taking a closer look at components, like waltr has suggested here) to hopefully solve this.

[Polymer]

Been thinking of what might be the problem.
One thing is the possibility that the AC coupling Cap from Pre-amp to Integrator is causing the issue. Maybe this cap is the wrong type, quality.
With many, many cap types the capacitance changes with Voltage and we know the Voltage at the pre-amp output goes rail to rail.

Exactly what cap is used here?

The C15 on the board is is a 100nF WIMA MKS 2 (100v).

I think the quality is ok. Here's a part of the datasheet:

I will put up a summary tonight of tests done so far inside and outside and what else I will be doing
(like taking a closer look at components, like you have suggested here) to solve this.

[green]

interesting thing i though... i did bara board to decise all problem of people... but sametime George also run in the process...
strange enough that problems are still not fixed and only growing up )))

Don't know if there is a problem. Amplifier and resistor noise are expected, also noise from coil pickup. I think what is needed is a test method to determine if noise level is higher than expected. Would be nice if someone would post a scope picture of U7 pin7 of a good bara.