500 nano second sampling

[Tinkerer]

Hi Tinkerer

Could you post the power supply and transmitter circuit diagram? You may wish to protect it for commercial reasons. If so I understand.

Chet

Send me your email by PM and I will send you the schematics. This goes also for Green and for everybody who gives valuable feedback and advice.
Attached is today's noise picture, #2. Same picture as the first noise picture, same target, same distance. I think you can see significant improvements.

[CAS]

Not having seen your circuit, just a small tip I learnt early on when we were building military spec power supplies is to have a .1 uf cap and 10uf cap (in parallel) as close as possible to any regulator ie input to gnd, out to gnd. Also any other ics always have a .1uf from the supply pins to gnd as close as possible to the ic itself. I have noticed these days that even in LED and plasma tv's that all the ic's have a .1 uf cap on supply pic to gnd at the ic.

[Chet]

Hi Tinkerer


Looking at the power noise it has a very high oscillation frequency of around 15 MHz. This could be related to small diameter wires and the layout on your proto board. It might be worth a try to take the MOSFET circuity off of the proto board and wire it directly using heavy short conductors. The addition of .1 uf capacitors such as suggested above by CAS is standard practice. In this case for 15 MHz you might try smaller values such as .001 uf.


The 3.5 to 6 mHz oscillation in the receiver may go away when the 15 mHz is cured.


Thinking ahead the layout of the PCB should have heavy runs and as much ground plane as possible for the transmit area.


Have another good day,
Chet

[DOOLEY]

Erm,
Just to rule something out,
Could be barking up the wrong tree ,
But worth a try.........

Assume tinkerer that your using a mosfet for
The switching,
And at half a micro second it's switching very fast,
Try a 220 ohm or so between gate and whatever your
Driving it with.

Just thinking a while back I had a noise issue
When driving a coil // mosfet to fast resulted in
Noise at the peak of the flyback,
A 220 ohm was just enough to stop it.

Not sure about what circuit your using but wondering
If noise could be migrating down and causing a problem.

Loads // high voltage on a mosfet do affect the gate.

Just a thought.

[Carl-NC]

T, have you looked at the coil itself? This looks a lot like issues I've dealt with when trying to make an IB coil that's really fast. If you haven't potted the coil, even the wires can vibrate enough to cause problems. Also, make sure there are no other stray coils anywhere in the vicinity. You might want to replace the TX coil with a (physically) much smaller coil with the same inductance.

[Tinkerer]

Thank you for all the valuable advice.
Seeing the great response, I am posting the schematic and PCB here. I drew these some time ago, but then decided to investigate all sources of possible problems fist, trying to eliminate them, as the making of the PCB represents a lot of work for me.

On this schematic, I added a lot of ferrites, with the intent to reduce HF oscillations. I have not actually tried these on the perforated prototype board. On the board they can easily be removed or shorted.

OK, tried to upload the DESIGNSPARK 7 files but the forum does not accept them. Will try later again zipping the files.

All the best and thanks for the help.

[crane]

Make sure you don't have any ferrite beads in the rx circuit, you will get oscillations if you do.

ML have a large ferrite core in the coil tx circuit but both wires must pass thru it, you will get saturation, which will then result in an offset if you only pass one tx lead thru the core. It is mainly to stop out going noise though.

[Tinkerer]

T, have you looked at the coil itself? This looks a lot like issues I've dealt with when trying to make an IB coil that's really fast. If you haven't potted the coil, even the wires can vibrate enough to cause problems. Also, make sure there are no other stray coils anywhere in the vicinity. You might want to replace the TX coil with a (physically) much smaller coil with the same inductance.

My lab is in the attic, with all sorts of copper sheets nearby on the roof. The space is also very narrow, so there is a huge amount of clutter, equipment, computer, tools, rolls of wire, half finished coils and such. To run a 1 meter diameter TX coil in this mess is really not easy.
I started with small coils first, but to attain the design depth of 1 to 2 meters, I see no other way than to use a large coil.
However, I would like to hear any suggestions of how to tackle the challenge.

The circuit itself has been tested in the lab and outside. For this purpose it needs a balance control, to adjust the induction balance for the great difference in the environment inside and outside. This also worked, but I am in the process of making the balance control more robust.

The noise:
At present we are using an ADC that samples a sample width of about 2.5us. The noise is not visible at all after integration. However, the plan is to sample with 500ns, so I need to eliminate the noise.

There are many sources of noise in this circuit. My goal is to tackle them one by one, starting with the switching noise caused by the TX board. So I will disconnect all other parts and show only this noise.
The next thing I will try, is a snubber designed for the frequency in question. HELP WITH THE CALCULATION OF THE SNUBBER PARAMETERS would be very welcome.

Ah, the coil assembly itself. No, this coil is not yet cast or shielded. Been trying to decide if I add a turn or take a turn off, but otherwise it is ready for sealing. Also the coil cables are the full length, twisted pairs, without shield. Might want to shield them though.

Thanks for all the help.

[Tinkerer]

this picture shows:

blue trace, no TX running, PSU and analog board connected. All 3 traces look the same low noise.

Red trace, RX coil input, disconnected from board, with PSUand analog board disconnected.
Yellow trace, input to TX board, with PSU and analog boards disconnected.

It looks like I opened a can of worms.

[Chet]

this picture shows:

blue trace, no TX running, PSU and analog board connected. All 3 traces look the same low noise.

Red trace, RX coil input, disconnected from board, with PSUand analog board disconnected.
Yellow trace, input to TX board, with PSU and analog boards disconnected.

It looks like I opened a can of worms.

Hi Tinkerer

I am trying to understand your findings here? Does this test indicate the noise is from the power supply or from external noise in your work area?

Chet

[Chet]

Hi Tinkerer

In reviewing the data sheet for the TPS25940ARVCR it is critical to have short leads and a proper layout. Maybe some of the noise will go away when it is installed on a well laid out PCB. Is it possible to bypass it for bench testing to see if the noise quiets down?

Have a good day,
Chet

[Davor]

...and maybe a ferrite bead on source lead?

[scrungy_doolittle]

...and maybe a ferrite bead on source lead?

I'm no hardware guru. but we had a similar problem on out medical device. The flyback from the transformer created enough noise that coupled to an i/o line that it was triggering and simulating a push button. I.E. it was enough that the processor would see it as a button press. The solution was a cap and resistor to ground to make a filter. Perhaps a small inductor and cap and resistor in the power supply line, creating a low pass filter with a cut off in the khz range would block the noise?

I have managed to get the STM32F411 development platform running, and have compiled and debugged a test program. Currently I am working on the ADC and will be building the LCD display as well. I've been going through a personal crisis the last two weeks. Looks like I will be permanently batching it, the wife left two weeks ago. Right now I am scrambling to get everything done. Way more work than normal there....

[Tinkerer]

Hi Tinkerer

I am trying to understand your findings here? Does this test indicate the noise is from the power supply or from external noise in your work area?

Chet

Thanks for the feedback. So far I can not see any noise from external sources. The noise from post 39 comes from the LPC2148 that we use for the ADC, timing and writing to the SD card. This board also used most of the power.

[Tinkerer]

Hi Tinkerer

In reviewing the data sheet for the TPS25940ARVCR it is critical to have short leads and a proper layout. Maybe some of the noise will go away when it is installed on a well laid out PCB. Is it possible to bypass it for bench testing to see if the noise quiets down?

Have a good day,
Chet

Sorry about the mix up. The schematic I posted is the way the final PCB is projected. The prototype perforated board is obviously not the same. It also does not include the TPS25940. The work done on the board today proves that the bad layout of the perf-board is the cause of most of the switching noise.