PI noise

[moodz]

The circuit was put together as a learning platform. Didn't have SMD adapters for the control circuit and coil driver. They have made it a lot easier to wire the circuit.

...going back to this post ... where you have a picture of the circuit boards built on veroboard (?) ... the main problem you have with noise is 80 percent due to this type of layout ... you could alleviate it possibly by placing all the boards on wire / capacitor standoffs on a blank sheet of PCB ... solder wire standoffs to ground and capacitor standoffs to power rails. The PCB forms a groundplane. The standoffs should not exceed 10mm in height.

The other 20% reason you have excessive noise is probably due to low CMMR on your diff integrator ... this can be fixed using the advice as per attached picture ...

[green]

...going back to this post ... where you have a picture of the circuit boards built on veroboard (?) ... the main problem you have with noise is 80 percent due to this type of layout ... you could alleviate it possibly by placing all the boards on wire / capacitor standoffs on a blank sheet of PCB ... solder wire standoffs to ground and capacitor standoffs to power rails. The PCB forms a groundplane. The standoffs should not exceed 10mm in height.

The other 20% reason you have excessive noise is probably due to low CMMR on your diff integrator ... this can be fixed using the advice as per attached picture ...[ATTACH]36755[/ATTACH]

Thanks for the suggestions. The circuit boards are copper clad on one side and used as the - rail for the control circuits, common for the analog card. The circuit was built on the un clad side when I didn't have SMD adaptors. With the adaptors I can build on either or both sides. Was thinking the adaptors were acting as standoffs. Maybe not. Been using punched boards because that is what I had and the holes allow connecting circuits from side to side. I use a 1C integrator.

[green]

The diff amp has less noise than the resistor input inverting input preamp for me. Requires a separate Rx, Tx IB coil. I had made a DOD coil awhile back. Connected it to the detector, the noise looked higher than I remembered the figure 8 Rx oval Tx looking. Ran a noise test comparing four different coils. Each connected to the diff amp, all circuits working except Tx command off. GEB switched off. Recorded 5 scans for each coil. Posting the lowest and highest p-p noise scan for each coil. Including some information I posted awhile back for the coils + the noise data from today. The 8 inch o0o is what I called figure 8 Rx oval Tx when I made the chart.

[green]

[Tried the switch both places with the 1C integrator and didn't see a difference in the noise but wondered if I was missing something.]

I didn't notice a difference before. Yesterday I disconnected the leads from preamp out to the switch inputs and connected them to common to test integrator noise. Noticed if I touched common or either of the power supply pins I had a small change in output. Rewired the circuit with the switches to integrator -input, resistors to preamp. That problem fixed, don't think it changed the operating noise level. Ordered a different amplifier and switches to try to reduce noise level.

Using a dual transistor matched pair for the preamp input. Tried four different transistor pairs. I had tried the BCM 857 vs the 2N3906 in LT spice, the 857 was better. Wondering if I might be doing something wrong. Scope pictures of post amplifier out, preamp input shorted, target and EF sample(no ground sample). Recorded 5 scans with each transistor pair. Posting the lowest p-p(top row) and the highest p-p(bottom row) of the 5 scans for each pair. I had tried a different npn pair vs the 2n3904 awhile back that didn't test as good, don't have the data. Haven't tried the new switches yet. The test pendulum_9 reply #51 shows need S/N increase of about 3 times to gain a detection distance of 2 inches. Not much difference with 2N3904, 2N3906 and SSM2220 with the SSM being best.

[green]

Lately I've been using an integrator and two post amp stages with a .01usec feed back TC each stage followed by a negative absolute value and a 2.3Hz 5pole butter worth filter. Recorded some scope pictures with the integrator and post amp feedback TC changed to .033usec without the abs. value and 5 pole filter. R7 and C4 are the feedback RC for the integrator, 10k input resistors in series with input switches(1C integrator). A photo cell was mounted 8 inches above the coil, a 200mm diameter disk was glued to the bottom of the water bottle swinging from a pendulum, US nickel stuck to the disk. 200mm/.2second=1meter/second. The nickel swings across one end or from end to end across both. Bottle was pulled back about 18 inches and let go then caught on the rebound, crosses coil twice, I think I like the circuit without the absolute value and 5 pole better.

reply#91

Some more scope pictures with 2N3906 matched pair. Integrator and post amplifier same as above. Nickel swinging end to end across both coils(Rx two 8 inch round coils figure 8, Tx oval surrounding Rx). Photo cell to show when target crosses coil. Target and EF sample, no ground sample.

[Orbit]

Hi Eric !
Which logically brings less noise integrator ?

[BASEL]

see this video maybe can help you to figure where is noise come from
https://youtu.be/BFLZm4LbzQU

[green]

Can you draw a timing diagram that starts with the TX pulse (coil on) and then all sample pulses, delays and widths from that point. Just a typical setup will do. I want to see the relationship of each part of the cycle along to the next TX pulse, particularly how far out your GEB sample is from the 'off' edge of the TX.

Eric.

Any comment on the timing? I've posting things in the past that made sense at the time but looking at them latter I had to think for awhile to understand them. I could try to draw the timing if the scope pictures aren't clear or. Timing sequence: coil on 160usec, target delay 5.7usec,(sw#1) target sample 9.6usec, GEB delay 5.4usec, (sw#2)GEB sample 100usec, EF delay 604.3usec(1kHz sample rate changes with sample rate), (sw#2)EF sample 9.6usec, EF delay 5.4usec, (sw#1)EF sample 100usec. Repeat. (1C integrator, timing that I was using at the time, probably not optimum.) sw#1 preamp out normal, sw#2 inverted out

[green]

Any comment on the timing? I've posting things in the past that made sense at the time but looking at them latter I had to think for awhile to understand them. I could try to draw the timing if the scope pictures aren't clear or. Timing sequence: coil on 160usec, target delay 5.7usec,(sw#1) target sample 9.6usec, GEB delay 5.4usec, (sw#2)GEB sample 100usec, EF delay 604.3usec(1kHz sample rate changes with sample rate), (sw#2)EF sample 9.6usec, EF delay 5.4usec, (sw#1)EF sample 100usec. Repeat. (1C integrator, timing that I was using at the time, probably not optimum.) sw#1 preamp out normal, sw#2 inverted out

Some test data. Recorded integrator out change for different targets(target distance, 2 inches for the nuggets and 4 inches for the coins) with and without GEB. Target sample time was adjusted for no change in integrator out when a quart zip lock bag about 6x6x1.5 inches of California ground was placed flat on one of the figure 8 Rx coils. Post amplifier out was monitored with a scope, targets were lowered on a pendulum until a definite change could be seen in the noise trace. Integrator out mv and detection distance for each target with and without GEB was charted. Preamp feedback TC=.2usec. The nuggets might chart a little different with a different feedback TC. The US nickel looses some distance with GEB. US quarter shows an increase in signal(opposite polarity) with GEB but the noise is higher so the detection distance stayed the same. Not a big difference in detection distance except for the nuggets with the change in target delay time. Another test could be leaving the target sample at 10usec and adjusting GEB sample to balance ground or repeating above test at different GEB delay times. After looking at the test results don't think it would get much if any better. The shorter target delay time is needed for the 4 grain nugget and is a little better for the coin targets.

[green]

Ground is a large noise source. A test comparing some different ground and ground simulation targets. Ca. ground from California gold field, Al. ground from my back yard. Each in a quart zip lock bag. Hot rock from California, Brick I was using for another project. Charted change in integrator out(no target vs target placed on one of the figure eight Rx coils). The two grounds were close, GEB reduced the hot rock and brick but didn't cancel. Are test results what I should expect?

[green]

Can someone calculate the ground decay slope using the delay and sample times that cancelled ground in the test (target delay_1)?

[Ferric Toes]

Can someone calculate the ground decay slope using the delay and sample times that cancelled ground in the test (target delay_1)?

The best thing I can do at the moment is to find my decay plots for various coins and also iron mineralised ground from various places, including Ca and Va.

Iron mineralised material always has a similar decay, so if you take some ground from California, Virginia, England, France, Italy, Cambodia, Australia, or indeed any country, and plot its decay, it is always about the same. Large or small quantity makes no difference. What does change dramatically is the amplitude, and the ground in parts of Australia can be more that 10x that of bad ground in California. Fired earthenware pottery and housebricks have the same decay

Metal targets are quite different and the decays depends on the conductivity, surface area, and mass.

I will post some plots shortly.

Eric.

[green]

Still don't know how to calculate ground slope using the times in (target delay_1). For a ground slope of -1, I think the formula for GEB sample time is(GEB sample time=GEB sample start time/target sample start time*target sample time). (21.2/6*10.2=36.04), (29.4/10*14.4=42.34), (36.6/14*17.6=46). Needed 100 for my bench setup. Modified the controller so I could control which switches turned on mainly to see how much target signal there is when taking the EF sample. The EF signal is a little high because I didn't see a easy way to turn both EF switches on(just the 100usec time)should be GEB sample time-target sample time. Didn't expect the ferrite bead(1in O.D. 5/8in I.D. 1/2in high)readings. The targets weren't at the same distance from the coil in (GEB_on_off_2), just shows relative readings for each target.

[green]

Decay charts for the Ca ground, US nickel, US quarter and the 1 troy oz 99.9% copper coin. Tx=160usec constant rate(6250 amps/sec). The nickel decayed close to a straight line on a lin-log chart. The others close to straight on a log-log chart. Charted target recording-no target recording.

[Ferric Toes]

I finally found some of the plots I did in 2013. First is a Nickel test that I did with 1 coin, then a second laid on top, and then a third. This demonstrated an interesting effect that three identical objects sitting on top of one another have a longer decay. This is true even if there is no electrical contact between them. Try it with cling film between coins and the result is the same. Linear time and amplitude followed by linear time/ log amplitude for this one shows a good single exponential fit.


Next is a non-conductive volcanic ash, so the signal is purely a viscous magnetic decay. This is plotted on linear time/amplitude and log time/log amplitude scales and has a slope close to the theoretical t^-1.


Any other viscous material from fired brick, volcanic basalt, to Australian ironstone, has a similar decay and when plotted will run substantially parallel to the Tiva Canyon material. Soil from Chico in California or Red Hill Virginia, is the same. The only major difference is the amplitude, and somewhere I have a plot of six or more samples from these different sources on the same graph. I will post this shortly.

Eric.