PI never gets old - constant power drive with added bonuses.

[Tinkerer]

Current flowing in the coil wires causes the parallel wires to attract / repel. These wire oscillations cause electrical noise.

[Teleno]

Current flowing in the coil wires causes the parallel wires to attract / repel. These wire oscillations cause electrical noise.

Fascinating. I would never have thought of that. If this were the cause then replacing the finger with a plastic clamp would work. I'll try it out.

[Teleno]

Current flowing in the coil wires causes the parallel wires to attract / repel. These wire oscillations cause electrical noise.

.

I've tried it out but it doesn't help. The palm of the hand held at about 1 cm from the coil does help without touching so I've discarded the mechanical cause.

EDIT: Placing a fast target such as a 1 cm2 piece of aluminium foil on the coil's edge stopped the ringing !

I believe it's in the feedback to the virtual ground, the hand could be seen as a salty target.

EDIT 2: The alu foil worked only because I was holding it with my hand

[Qiaozhi]

Examining your scope shots it looks to me as if the noise on the floating ground is similar to the preamp output noise. When you place your hand/finger on the coil you're also affecting the floating ground, which is directly connected to the non-inverting input of U3. Have you considered that this may be the result of a PCB layout problem causing positive feedback at the preamp? What happens if you place your finger on the non-inverting pin of the preamp?

[green]

Have you tried shielding the coil? I need to shield my coils when trying to operate with a low micro second first delay.

[Teleno]

Can you rotate the coil to different positions? Usually if there is a specific EMI source you can find a null. Also, make sure there are no other coils nearby. "Nearby" can be 1-2 meters. This one gets me a lot.

The oscillation doesn't change with the coil's orientation. I took the whole set up outside far from appliances but and it stayed the same.
So the next suspects are the feedback circuit, the PCB layout and the coil itself (see more on this below).

Have you tried shielding the coil? I need to shield my coils when trying to operate with a low micro second first delay.

How do you shield a planar coil without the shield looking as a target? I can think of graphite coating or a mesh made from high resistivity conductive material.

Examining your scope shots it looks to me as if the noise on the floating ground is similar to the preamp output noise. When you place your hand/finger on the coil you're also affecting the floating ground, which is directly connected to the non-inverting input of U3. Have you considered that this may be the result of a PCB layout problem causing positive feedback at the preamp? What happens if you place your finger on the non-inverting pin of the preamp?

The PCB layout is always a suspect but a finger on the non-inverting pin didn't affect the ringing. A finger on Rdamp does though. I suspect the true cause is a secondary resonance mode in the coil, but first I had to test the feedback theory. If true, then the 12 MHz ringing should be gone if the wideband Op Amps are replaced by slower types. I used the TLC081 ( 10 MHz GWB) as UC1 in the schematic with the following results:

Preamp output (b in the schematic):



Preamp output with hand next to the coil:




Even without "handing" all the periodic noise is practically gone. The preamp is still fast enough to sample at 2.5 us. Responsive to my 0.2 g test gold nuggets.

From this it would seem that the feedback theory was correct... but then I looked at the virtual GND during flyback and I caught the 12 MHz oscillation still hanging in there!

Virtual GND during flyback:




The bandwidth of the TLC081 is unlikely to afford a 12 MHz ringing, but just to make sure I tried a TL071 (5.25 MHz GWB) with similar results at the virtual GND.
Again, the hand makes it go away.

Virtual GND during flyback (hand next to the coil):





From all of the above the following causes can be discarded: EMI from nearby appliances and positive feedback in the preamp.

It is apparent to me that the 12 MHz ringing is coming directly from the coil either as EMI fro a transmitter (unlikely) or as a secondary resonance mode of the coil.
The latter seems the real culprit because if you look at the back side of the coil it looks like this:




I guess all those crisss-crossing strides of different lengths are behaving as coils within the coil causing higher resonance modes (additional poles in the impedance spectrum).

However, I still haven't discarded remote EMI sources by looking at the effects of shileding the coil (as per green's suggestion).

I have to decide now whether to further pursue the earliest sampling possible, scrambling for the last 0.5 us with high BW Op Amps and lots of HF noise, or just settle for the TLC081 which gets rid of all the noise with no significant penalty delay wise.

[Altra]

Could be how you are grounding your scope? This was discussed on the forum a while back.

https://www.youtube.com/watch?v=zodp..._channel=w2aew

[Tinkerer]

Could be how you are grounding your scope? This was discussed on the forum a while back.

https://www.youtube.com/watch?v=zodp..._channel=w2aew

Thank you for the link to the video. Very interesting.

[Qiaozhi]

Thank you for the link to the video. Very interesting.

Yes, very interesting ... but I found it really irritating how many times he says "OK".

[Tinkerer]

Looking at your coil, I see that your using plain wire for the coil leads or cable.
How long is the cable? Could the the 12Mhz come from the cable?
Did you try to twist the cable?
Did you try with coax? single coax, or one lead of the coil using a coax, the inner wire being the coil lead, the outer wire connected to the shield. The other lead, inner wire as lead, outer part of the coax connected on the one side to the ground on the TX board, the other end left free, thus being connected as a shielded wire.

For the shielding I like to use 3mm to 7mm PVC foam sheet, painted with graphite paint. The thickness of the shield serves as spacer between the coil wires and the graphite shielding. A thin wire embedded in the graphite paint connects to the outer coax.

[Teleno]

Looking at your coil, I see that your using plain wire for the coil leads or cable.
How long is the cable? Could the the 12Mhz come from the cable?

Did you try to twist the cable?

It's a twisted cable indeed. Touching the cable doesn't affect the ringing.

Did you try with coax? single coax, or one lead of the coil using a coax, the inner wire being the coil lead, the outer wire connected to the shield. The other lead, inner wire as lead, outer part of the coax connected on the one side to the ground on the TX board, the other end left free, thus being connected as a shielded wire. .

I may try that if the shielding doesn't solve the issue. Thank you for the tip.

For the shielding I like to use 3mm to 7mm PVC foam sheet, painted with graphite paint. The thickness of the shield serves as spacer between the coil wires and the graphite shielding. A thin wire embedded in the graphite paint connects to the outer coax.

If I undestand you right the foam is painted on one side opposite the wires, correct?

[Tinkerer]

Correct. 3mm is the minimum distance, more is better.

My reasoning is that the graphite particles are very small and the binder prevents from being a perfect contact between particles, therefore the paint ends up having a relatively high resistance. The eddy currents generated in the shield will therefore dissipate before travelling a large distance.
I see the shielding as an electrostatic shield to prevent E fields from interfering with the RX coil.
I insulate the shield with several layers of Epoxy resins.
This shielding greatly reduces the influence of the hand, leaving only the target response of the conductivity of the hand itself.

[Tinkerer]

Did you try adding a ferrite on the cable? I would suggest a ferrite of the type that becomes resistive / dissipative with a frequency above 1.5Mhz.

[green]

How do you shield a planar coil without the shield looking as a target? I can think of graphite coating or a mesh made from high resistivity conductive material.

Similar to Tinkerer, add 3/16 foam board painted with graphite paint each side of coil. Last coil I made, tried reducing the shield time constant. Stretched AWG32 bare wire across board in a few places, tape wires down with 1/8inch masking tape before painting, removed tape after drying. Used one of the wires each side as shield return. Should reduce grahite time constant. Coil seemed to work fine. Trying to sell my house, need to do more testing after I move.

[Teleno]

Did you try adding a ferrite on the cable? I would suggest a ferrite of the type that becomes resistive / dissipative with a frequency above 1.5Mhz.

Eureka! That did the trck!

I placed one of those ferrites seen on USB cables and the ripple is practically gone.

I guess it was a shielding issue in the end.