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

[waltr]

Please ignore the revD comment- I mixed your posts with someone else.

[green]

Bara_Geo_pin7_lf412_persistance 5s.png from reply #14

Wondering if high frequency signal matters. Looks like low frequency signal is more than 6 times high frequency signal. Could you repeat scope picture with persistence off and 200msec/div instead of 200usec/div? Channel 1 only.

[Qiaozhi]

Bara_Geo_pin7_lf412_persistance 5s.png from reply #14

Wondering if high frequency signal matters. Looks like low frequency signal is more than 6 times high frequency signal. Could you repeat scope picture with persistence off and 200msec/div instead of 200usec/div? Channel 1 only.

I'm sure it doesn't matter. On my Geotech Baracuda board I shorted out R22, and reduced the 4mV ripple to zero. There was no detectable difference in the preamp output. This is exactly the same power supply filtering arrangement as the Hammerhead design, and there's no problem there either.

Polymer appears to have an error on his board somewhere.

[green]

I'm sure it doesn't matter. On my Geotech Baracuda board I shorted out R22, and reduced the 4mV ripple to zero. There was no detectable difference in the preamp output. This is exactly the same power supply filtering arrangement as the Hammerhead design, and there's no problem there either.

Polymer appears to have an error on his board somewhere.

Could you post a scope picture like I asked Polymer in reply #17 to compare with Polymer's picture. I'm interested in what the low frequency noise signal looks like, less than 10Hz since that is what we are trying to detect.

[Polymer]

Sounds like either R22 (27 Ohm) value is not correct as has been suggested (did you check?) or the 47uF cap is NOT a Cap. Adding a 100uF seems to have fixed this so that should be good enough.

Your scope capture of U7a Output (pin 1) looks like the Offset from the Pre-amp is not adjusted correctly- but the pre-amp is AC coupled (cap) so can not be the case here.
OR - the second Sampling switch (Q2) is not working correctly.
Re-check the second sample pulse and Q2 is switching. Scope probe on Q2-R26 connection should show a change of Voltage level when Q2 switches on.
Look at Q3 output and the RevD to see what this point should look like.

We were suggesting using a second battery instead of the LT1054 Voltage inverter. To do this remove the LT1054 and connect a 7-15V battery into pin 5 (B-) and pin 3 (B+) to power the 79L05 (negative regulator). This second battery's Positive goes to 0V (Ground).
The 12V batter still connects to +VB & 0V to power the 78L05 (positive regulator).

I did check the 27 Ohm resistor. It is so.
There is one thing I noticed when taking the temporary 100uF bypass cap off and measuring again with only the 47uF in place.
That is, that the funny signal on the -5v at pin 4 of the 5534 was still gone, although only the 47uF was now in place.
You having mentioned it maybe not being a capacitor had me poking around quite a bit.

From what I found, one of the damping resistor wires had a faulty, intermittent connection.
When the damping resistor is not connected, the funny signal on the -5v line of the 5534 is hefty, like in the scope picture.
Oddly enough the added 100uF capacitor mitigated this effect.
Is the damping resistor in place as it should be, then that funny signal is very small, like in Qiaozhi's measurement.
Sorry, my mistake ...

Is the entire middle text part invalid or just the reference to "RevD"? Thanks for correcting this, I have not looked at this yet.

Finally the knot in my head has unravelled. Your battery connecting information did it Yippee! THANK YOU!

I will be grabbing a bag of batteries tomorrow and powering with that. Don't have enough new ones around right now.

Here a picture of how I understand it, just to be sure:

[Polymer]

Quote from green post #17
"Wondering if high frequency signal matters. Looks like low frequency signal is more than 6 times high frequency signal. Could you repeat scope picture with persistence off and 200msec/div instead of 200usec/div? Channel 1 only."

Yes, of course! here you go:




Both are without trace persistance.

[Polymer]

I'm sure it doesn't matter. On my Geotech Baracuda board I shorted out R22, and reduced the 4mV ripple to zero. There was no detectable difference in the preamp output. This is exactly the same power supply filtering arrangement as the Hammerhead design, and there's no problem there either.

Polymer appears to have an error on his board somewhere.

You're right, I had an intermittent Rdamp connection causing the excessive -5v noise at the 5534. The wonkiness is still there though ... grumble ...
The hunt shall continue ...

Thank you for putting the GeoBara on the bench!

Could you, please, please, make a scope shot like green is suggesting in post #17 !!!!!

Now where's the emoticon "pretty please with a cherry on top" ...

[waltr]

Is the damping resistor in place as it should be, then that funny signal is very small, like in Qiaozhi's measurement.
Sorry, my mistake ...

Is the entire middle text part invalid or just the reference to "RevD"? Thanks for correcting this, I have not looked at this yet.

Finally the knot in my head has unravelled. Your battery connecting information did it Yippee! THANK YOU!

I will be grabbing a bag of batteries tomorrow and powering with that. Don't have enough new ones around right now.

Here a picture of how I understand it, just to be sure:

[ATTACH]42721[/ATTACH]

Ahhh...if Damping R is not connected then the pre-amp was probably getting overloaded for too long.
Good find and fix.

Only my reference to rev D (was thinking of the MPP rev E thread). The rest is valid to be checked.

Battery connection diagram looks good. You now see how to get a Negative Voltage using two batteries.

[Polymer]

Ahhh...if Damping R is not connected then the pre-amp was probably getting overloaded for too long.
Good find and fix.

Only my reference to rev D (was thinking of the MPP rev E thread). The rest is valid to be checked.

Battery connection diagram looks good. You now see how to get a Negative Voltage using two batteries.

Thank You!

I ran the detector with batteries only without the lt1054.
The detector works just like before, wobblenoise included!

So for now the lt1054 is not part of my threshold jumpiness problem.
Maybe I can't see what other noise it may be causing at the moment, but that's something for another time.

I probed the Q2 jfet like you advised. I also probed Q3 as a comparison.

Both gate signals look good to me coming from pin 10 & pin 12 of the 40106.

The output of Q2 (second sample) does not look good to me.
It's in the right position timewise though. But who am I to say ... Here's the pic:



Here the screenshot of the measurement of the main sample at Q3.
Looks somewhat "better", but the signal shape is odd to me nonetheless.
I have no reference scope pics of them working at "nominal" value.



Just a somewhat wild guess here from what I have learnt so far about jfets:

- They may be ESD partially damaged
- The wide range of Upinch in these components may be causing problems

What is your take on these signal shapes?

I did read "My take on the HHII". It's a veeery long read and a lot to digest for me.
Chock full of interesting knowledge. Those posts I will be reading again. Learning ....

[waltr]

The second sampling looks pretty noisy compared to the first sample.
It does seem the two samples are doing what they should in the scope trace on the Integrator output in Post #14. !st sample increase signal, 2nd sample decrease signal.

Now trying to not get your project confused with someone else's (the MPP rev E) but could have a similar problem.

Possible the wobbling is due the AC mains interference. Have you taken the detector far from the house and AC mains?
If the wobbling goes away then that is the problem. My HH detector does this when the heater kicks on a few feet away in my shop but is fine when away from the house.

Another possibly thing to try is increase the delay a few usec. You do have the 1st sample a bit early and sampling a little later may help.

Also possible one or both the JFETs are not working properly OR U7 needs better caps on its power pins (adding caps may help).

[green]

Quote from green post #17
"Wondering if high frequency signal matters. Looks like low frequency signal is more than 6 times high frequency signal. Could you repeat scope picture with persistence off and 200msec/div instead of 200usec/div? Channel 1 only."

Yes, of course! here you go:

[ATTACH]42722[/ATTACH]
[ATTACH]42724[/ATTACH]

Both are without trace persistance.

What is the best test procedure to determine the cause of noise and how to minimize it? My thoughts, maybe someone that knows could suggest the correct procedure.
Pin 7 of U7 is the important test point to monitor, 200msec/div. Noise pickup from coil would vary with location so testing with coil disconnected would be best. Lifting one end of R16 or C15 would remove amplifier noise. Record p-p noise at U2 pin6 to see if noise is near what expected could be a place to start.

[kt315]

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 )))

[waltr]

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?

[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.

[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.