PI noise

[mickstv]

OPA1611 OPA1612

Same as what I use in my detector. I also use the OPA1642 for the rest of the stages. No problems with noise out in the field.

[moodz]

Same as what I use in my detector. I also use the OPA1642 for the rest of the stages. No problems with noise out in the field.

Ti apparently acquired National Semis line of business in opamps and has been shutting down the main Fabs that make them ... hmm ... I have even been looking at "discrete opamps" like this open source Jensen design.

http://mnats.net/JE-990.html

Theres a commercial version of it here ..http://www.johnhardyco.com/pdf/990.pdf by JohnHardy

[greylourie]

moodz, have you looked at the discrete input of/on your rigol oscilloscope. Also appears to use a discrete differential arrangement following. Though seems like a lot of effort compared to just using factory made opamps.

http://www.eevblog.com/files/Rigol-D...c-FrontEnd.pdf

http://www.eevblog.com/files/Rigol-D...ic-DiffAmp.pdf

[greylourie]

Thank you Eric ,its good to learn something.

I echo your sentiments, 6666.

Now I have learnt about a new beast (cross coupled stage/differential), and the potential benefits.

[moodz]

moodz, have you looked at the discrete input of/on your rigol oscilloscope. Also appears to use a discrete differential arrangement following. Though seems like a lot of effort compared to just using factory made opamps.

http://www.eevblog.com/files/Rigol-D...c-FrontEnd.pdf

http://www.eevblog.com/files/Rigol-D...ic-DiffAmp.pdf

Yes I did see those on eevblog .. they may have use discretes to save every last dollar as labour inputs are probably much cheaper than inventory inputs in their cost model for those boxes. There is also the possibility that the frontends are more robust using discrete transistors as the rigol scopes are targeted heavily at the education and repair bench market where all sorts of abuse occurs :-)

[green]

Monolith suggested using a pendulum to swing the target in another thread(Amplifier frequency response). A test swinging a US nickel 8, 10, 12 and 14 inches above the coil on a pendulum(about 76 inches in length, .36Hz). The target was pulled back 17 to 18 inches and let go and caught on the return. Should be about 1 meter/second peak velocity. Coil, Rx(two eight inch round coils, figure eight) Tx(oval coil surrounding Rx). Nickel centered over point where two coils touch, swings end to end. Loose about 2/3 of the signal for each two inch increase in distance. Eventually you run into noise level. The reason I wonder what is causing the noise increase when sampling. I would be interested if some one would do a test. Look for noise at post amplifier out on a scope, normal and with a added capacitor across the fdbk resistor of the last stage preamp (fdbk RC=200usec to 500usec). I have been getting about 1/4 the peak noise with the added capacitor. The short glitches happen when the control one shots switch and are there even if the scope ground and probe are connected to common. Thought I had solved the glitch problem, evidently not.

[Monolith]

Monolith suggested using a pendulum to swing the target in another thread(Amplifier frequency response). A test swinging a US nickel 8, 10, 12 and 14 inches above the coil on a pendulum(about 76 inches in length, .36Hz). The target was pulled back 17 to 18 inches and let go and caught on the return. Should be about 1 meter/second peak velocity. Coil, Rx(two eight inch round coils, figure eight) Tx(oval coil surrounding Rx). Nickel centered over point where two coils touch, swings end to end. Loose about 2/3 of the signal for each two inch increase in distance. Eventually you run into noise level. The reason I wonder what is causing the noise increase when sampling. I would be interested if some one would do a test. Look for noise at post amplifier out on a scope, normal and with a added capacitor across the fdbk resistor of the last stage preamp (fdbk RC=200usec to 500usec). I have been getting about 1/4 the peak noise with the added capacitor. The short glitches happen when the control one shots switch and are there even if the scope ground and probe are connected to common. Thought I had solved the glitch problem, evidently not.

Still a very good signal S/N, at 14inches, specially when considering an 8 inch coil. 1.75 times diameter of the TX coil is excellent.
The glitches seem to be inherent to our quality level Digital Storage scopes. I have also observed various aliasing errors on my scope.

There is a tiny noise sinewave of about 2K frequency visible on the 14 inch picture. Is that your PP rate?

The analog switches add some noise. I add a discrete LP filter at the output of the switch that seems to help.

Limiting the bandwidth to the desired one seems to be a good idea for every opamp circuit, but maybe it would even be a good idea to reduce the pass-band frequency for the initial stages to the absolute minimum in the intention of reducing as much noise as possible before further amplification. In the case of your last stage preamp, I would consider 5Hz as low, but I have not measured the actual target response time of an 8 inch coil. At 1m/s, the time of the target above the 8 inch coil is 200ms = 5Hz, or should we call it 10Hz, as the response is more like a half sine wave? However, at a distance, the detection field is narrower than the coil, so the frequency increases. A common frequency used is about 15Hz.

I found that for the pendulum, filling a glass bottle with water, so that the pendulum is really heavy, helps keeping the pendulum swinging for quite a while, giving me time to look at the scope and saving traces.

[greylourie]

green, I recall seeing "1 metre/second" mentioned on other forums..... and what Monolith says about limiting bandwidth ties in with that maybe. Possibly half or even a little more than less than half of what you are using currently...

[green]

Still a very good signal S/N, at 14inches, specially when considering an 8 inch coil. 1.75 times diameter of the TX coil is excellent.
The glitches seem to be inherent to our quality level Digital Storage scopes. I have also observed various aliasing errors on my scope.

There is a tiny noise sinewave of about 2K frequency visible on the 14 inch picture. Is that your PP rate?

The analog switches add some noise. I add a discrete LP filter at the output of the switch that seems to help.

Limiting the bandwidth to the desired one seems to be a good idea for every opamp circuit, but maybe it would even be a good idea to reduce the pass-band frequency for the initial stages to the absolute minimum in the intention of reducing as much noise as possible before further amplification. In the case of your last stage preamp, I would consider 5Hz as low, but I have not measured the actual target response time of an 8 inch coil. At 1m/s, the time of the target above the 8 inch coil is 200ms = 5Hz, or should we call it 10Hz, as the response is more like a half sine wave? However, at a distance, the detection field is narrower than the coil, so the frequency increases. A common frequency used is about 15Hz.

I found that for the pendulum, filling a glass bottle with water, so that the pendulum is really heavy, helps keeping the pendulum swinging for quite a while, giving me time to look at the scope and saving traces.

Not sure what the 2Hz oscillation is. It's not always there, comes and goes. The butterworth filter overshoots at about that frequency with a step input. Next order I'll add a Bessel filter and try it. Sampling at 1kHz, about 10usec target and EF sample.

The noise is a lot lower with the integrator input connected to common, so I'm thinking the switches aren't causing the problem.

More scope traces comparing the different output stages in(Amplifier frequency response)thread. Look at the enclosed schematic. Integrator cutoff 16Hz low pass, two more stages with a 16Hz low pass cutoff. More on the added filtering in(Amplifier frequency response). I don't know what the best frequency cutoff is. A scope trace of target signal and noise should give a good indication.

I use a bottle of water for the pendulum. I caught it on the return swing so at least half of the trace was noise.

When I first joined this site I read a reply in a thread that stated if you can't detect a US nickel at 15 inches keep looking. From my testing a mono coil gives the best signal but so far I haven't been able to get the best signal to noise using a mono coil. Could probably get there using what I have by increasing coil diameter and peak current. Or by lowering the noise.

Stated the Rx coils were 8 inch. They are spider web(flat basket)wound, 8 inch I.D. so the mean diameter is about 8.3 inches

[Monolith]

Not sure what the 2Hz oscillation is. It's not always there, comes and goes. The butterworth filter overshoots at about that frequency with a step input. Next order I'll add a Bessel filter and try it. Sampling at 1kHz, about 10usec target and EF sample.

The noise is a lot lower with the integrator input connected to common, so I'm thinking the switches aren't causing the problem.

More scope traces comparing the different output stages in(Amplifier frequency response)thread. Look at the enclosed schematic. Integrator cutoff 16Hz low pass, two more stages with a 16Hz low pass cutoff. More on the added filtering in(Amplifier frequency response). I don't know what the best frequency cutoff is. A scope trace of target signal and noise should give a good indication.

I use a bottle of water for the pendulum. I caught it on the return swing so at least half of the trace was noise.

When I first joined this site I read a reply in a thread that stated if you can't detect a US nickel at 15 inches keep looking. From my testing a mono coil gives the best signal but so far I haven't been able to get the best signal to noise using a mono coil. Could probably get there using what I have by increasing coil diameter and peak current. Or by lowering the noise.

Stated the Rx coils were 8 inch. They are spider web(flat basket)wound, 8 inch I.D. so the mean diameter is about 8.3 inches

Eric's coil chart.pdf

Attached is Eric's coil chart. This will give you an idea for choosing the best coil diameter for the Nickel.

Are you using the 2 RX coils in differential mode?

[green]

[ATTACH]36592[/ATTACH]

Attached is Eric's coil chart. This will give you an idea for choosing the best coil diameter for the Nickel.

Are you using the 2 RX coils in differential mode?

If I read Eric's chart correctly a 30 inch diameter coil would give the best signal for any target at 15 inches. Maybe a diameter between 20 and 30 inches for the best S/N.

The Rx coils are in differential mode.

What is the answer to the ring at 20 inches with a 10 inch coil question?

[Monolith]

coil calculations

If I read Eric's chart correctly a 30 inch diameter coil would give the best signal for any target at 15 inches. Maybe a diameter between 20 and 30 inches for the best S/N.

The Rx coils are in differential mode.

What is the answer to the ring at 20 inches with a 10 inch coil question?

Attached is a .pdf with links to calculators that are very useful for all sorts of calculations for metal detectors. Also screenshots with some calculations for an 8inch coil and a 14inch coil

[green]

[ATTACH]36592[/ATTACH]

Attached is Eric's coil chart. This will give you an idea for choosing the best coil diameter for the Nickel.

Are you using the 2 RX coils in differential mode?

I had seen Eric's chart before but I don't remember reading the explanation with it. I ran a test with three different size coils, 200mm/1.5, 200mm and 200mm x 1.5 that I posted in another thread awhile back. Same current profile for all three coils. Instead of same number of turns I multiplied amplitude reading by 300uH/coil inductance for each coil. Same inductance instead of same turns. The results look similar except signal out is the same at a distance half way between their radius. 200mm coil, 300mm coil((100+150)/2=125mm))crossing.

[green]

HI Eric,


I have tried several lme9990 and found that they all take quite a bit longer to settle(think it was up to 5 or 7us) than the AD797. Though a mate of mine was able to use the lme49990 and get it to settle earlier than I could with the same input impedance and gain, go figure! Though he was using a different coil...
When I did noise tests, I could not see any notable difference between the 797 and the lme. Same goes for using an lm394/5534 preamp. This settles as fast as the 797 and there seems to be no notable output noise difference. Though one thing I am still to try, from one of my initial tests, it looks as though the 797 is not as sensitive as the lm394/5534 preamp, however some other things were set up differently. Currently changing back those differences and will test on the weekend. I will know more after the weekend..

Cheers Mick

Hi Mick

Did you learn anything over the weekend? I tried the mono coil and amplifier again. Having a hard time lowering the noise with a mono coil. A mono coil would be a lot easer to make than the one I use with the differential amplifier.

green

[Mechanic]

Hi Green,

After changing things back to the way they were meant to be, the 797 seems comparable to the lm394/5534 preamp.

If the preamp noise is high, and you also use 1k input impedance, the output noise from the detector will always be high/ bad signal to noise ratio. Also, you need to make sure you are using a fast opamp otherwise it takes too long(longer than what it could be) to settle after the tx, thus delaying when you can take your first sample.

I once did an experiment and replaced the front end fets(ML) with a 1k resistor and clamping diodes and the jump in output noise was amazing! I could have lowered the gain of the preamp to get the noise back to the same level, but in doing so depth and sensitivity would be compromised.

Testing for noise with a mono coil(or a DD for that matter) in suburbia will always be too difficult due to the electrical interference. The best way is to tackle the system noise and then when you are away from interference, this is the best case of how the detector will run.
The reason a DD coil will seem quieter is that it has less receive area for a mono of the same size, ie the receive of the DD is half the size of the coil.
However, there is the case where something in the tx cct is not quite right and the receive picks this up with a mono coil as the receive is directly coupled to the tx cct, whereas a DD the receive is isolated from the tx cct.

Cheers Mick