MPP Rev-D ‘Falsing’ on end of swing.

[waltr]

Read this thread:

http://www.geotech1.com/forums/showthread.php?13113-Designing-an-Integrator-circuit&highlight=purpose+integrator

[Goaty]

Thanks for your info waltr.


I feel I am closing in on what needs to be done but don’t quite have the knowledge about the functioning of the circuit at a component level to tackle it. I understand what this section does but not how it does it.


You said - “ re-check that the integrator Caps and resistors MATCH. That is the RC on the Op-amp Feed back (out to inverting in match the RC on the non-inverting in to ground” and this is the part that I am having trouble getting my head around.
I think I will have to do a bit more study on op-amps first!


I noticed that Qiaozhi has posted an LTSpice simulation of the MPP sampling integrator but have never used a circuit simulator and unfortunately don’t really have the time to get into it at the moment.

[waltr]

The Hammer Head I built has a pot to balance integrator inputs (pre-amp out to pot wiper, then each pot end into invert and non-invert integrator).
This is a touchy adjustment to get it working best so fairly critical. The pot has a much lower value than the integrator input Resistors so doesn't do much of a change.

If I remember correctly, to adjust this set sample delay out a bit (not minimum) then look at integrator output with O'scope.
The adjust so the two samples (main & EFE) give the same change but opposite direction.

May also be possible to look with scope on the sampling switch outputs and adjust so amplitude of each, when switch is on, to be the same but opposite direction.

These two samples are now Canceling each other so any EF affects should also cancel.

[green]

If Main & Sec samples widths are the same then all is good. Exact width is not that important.

There are discussions in the threads on this forum about what and how the integrator works. I just don't have links.

Since the main & sec sample widths are the same then re-check that the integrator Caps and resistors MATCH. That is the RC on the Op-amp Feed back (out to inverting in match the RC on the non-inverting in to ground). A mismatch of these components can cause the same effect as mismatched main vs sec sample widths. Again Exact values are not important but values Matching is very important.

Some detectors have a pot in series with input resistors to balance EF. Does anyone know how close(out to inverting in match the RC on the non-inverting in to ground)needs to be?

[waltr]

If Main & Sec samples widths are the same then all is good. Exact width is not that important.

There are discussions in the threads on this forum about what and how the integrator works. I just don't have links.

Since the main & sec sample widths are the same then re-check that the integrator Caps and resistors MATCH. That is the RC on the Op-amp Feed back (out to inverting in match the RC on the non-inverting in to ground). A mismatch of these components can cause the same effect as mismatched main vs sec sample widths. Again Exact values are not important but values Matching is very important.

[Goaty]

I have noticed is that the sample pulse widths are all the same as the Tx oscillator (58uS)

C11 (2n2) and R11 (22k) with U7 seem to set the pulse widths for the main & sec samples to 50uS.

What would be a suitable value?

Would this be enough to upset things?

[Goaty]

@ Qiaozhi
“Since you've checked the main sample and the EFE sample are the same width, also check the EFE delay time.”


Today I spent some time re-measuring the pulse signals up to, and including, the pulse generator just to check my original settings and have listed my measurements below-


Transmit Oscillator (TP1)
Pulse rate - 1030 pps
pulse width 58 uS


Receive Pre-amp (TP3)
Damped flyback with correct response to metal near coil


Sample Pulse Generator (TPs 5,6,7,
Minimum sample delay (TP5) 22uS
Secondary sample delay (TP7) 120uS after start of Main sample pulse.
All pulse widths 58 uS


I’ve also started checking a few components in the Sampling Integrator since the Sample pulse section looks OK.


R24 2k147 ohms
R25 2K110 ohms ‘in-circuit’ measurements.


TP8 (Q3 gate) -4.72V
TP7 (Q4 gate) -4.71V


U8
Pin 3 non-inv input (main) 92.7mV
Pin 2 inv input (Sec) 91.8mV
Pin 1 output -117.8mV


Well, this is as far as I have got.
At the moment the sampling integrator circuit has me a little bewildered as to how it functions so any explanations would be appreciated.
Also, if anyone has some suggestions as to how to go about testing this section I would be appreciative as my understanding of op-amps is limited to just the basics.


And, one last thing I have noticed is that the sample pulse widths are all the same as the Tx oscillator (58uS) and yet in the Sample Timing diagram (Fig 10) it shows the Tx pulse widths being wider than the Main Sample and Sec. Sample widths – is this correct? Measurements of the circuit also indicate they are the same but but I was concerned that I may have missed something.


Once I get my head around the Sampling Integrator I will probably try a trimpot to adjust the signal feed levels to FET switches Q3 and Q4 but other than that I’m starting to run out of ideas.

[qute1]

This detector building hobby is a bit of an addictive drug isn’t it!

I think you will find that we are all suffering from this particular affliction (i devote the majority of my free-time to the development, construction and testing of my mods and designs) drives the wife nuts there are literally MD's everywhere, running out of space to store even the regularly used ones never mind the collectors editions.

[Goaty]

Sorry about taking so long to reply but I have been in & out of hospital a couple of times since and am still trying to catch up.

I have not had time to make measurements on the board but have made a couple of quick replies to some of the comments -

@ soverey

your shielding has to be in the range of a few hundred ohms per square

The coil is wound as ‘self shielding’ and I don’t think the graphite layer was just an experimental ‘after thought’ - it probably does not contribute much to the overall shielding.






@ qute1

Are your coil bundles tightly bound and secured in the housing and are you checking outside away from emi

I tried a slow swing with a bump stop against my shoe and no falsing evident, so windings seem solid enough.




@ waltr

be sure nothing can move inside the coil housing and that the coil to box cable is not free to move about.

As above for windings and if I slightly unwind the coil cable from around the shaft and wriggle them there is no problem.


@ kt319

sorry but MPP is simple design for beginners for getting new experience and starting the hobby.
you have to start to do new project, hammerhead, with GB tuning advantage. do not worry, you went through the first stage and are ready
overcome next one. you are on right way amigo.

I fully agree – as a matter of fact I ordered a couple of the Mirage boards from Sven only a week or two ago. I believe these are based on the Hammerhead. This detector building hobby is a bit of an addictive drug isn’t it! And learning all the time.

Thanks for your cross references – I had found and read the first previously but not the second.

Well, enough for now,
I will report again ASAP

Thanks again for all the help

[Goaty]

Thanks everyone for your helpful suggestions.
I can't wait will work through them and respond when time permits in a few days time!

[kt315]

crossreferenced to
http://www.geotech1.com/forums/showthread.php?23898-False-beep-at-end-of-swing

http://www.geotech1.com/forums/showthread.php?24024-Does-single-coil-circuit-suffer-from-EFE-beep

[green]

I think you need to trim R24 or R25. Maybe lift ends going to R19, connect a 200 to 500 ohm pot ends to the resistors with wiper going to R19. Or connect a high resistance(500k, 1meg)pot across R24 or R25. Adjust pot for no indication when quickly lifting a ceramic magnet(fridge magnet) from the coil.

[kt315]

sorry but MPP is simple design for beginners for getting new experience and starting the hobby.
you have to start to do new project, hammerhead, with GB tuning advantage. do not worry, you went through the first stage and are ready
overcome next one. you are on right way amigo.

[waltr]

Are your coil bundles tightly bound

My thought too. Also be sure nothing can move inside the coil housing and that the coil to box cable is not free to move about.
Also ensure everything inside the box is secured.

I run a PI MD without shielded coil without issues. EFE pulses seem to work as I get no falsing normally.
However, I do get a false if a bump the coil on something or catch the coil cable on stuff. It only takes a very slight movement of coil internals or cable near the coil to false.

[qute1]

Are your coil bundles tightly bound and secured in the housing and are you checking outside away from emi (i had trouble near wifi signals and there are a lot of them where i live they propagate a lot further than you might think) just my four pennies worth so if you have checked this just ignore me.