I'm I facing with a bad ftdi card?

Spent some time googling about my problem, but I didn't find anything useful for my problem.

That's my first project with arduino and is a very simple one. I just want to detect a small current through an SCT-013-030 and then activate a relay. To do so, having problems of space, I decided to use an arduino pro-mini board (a clone bought from aliexpress) and an fdti card to upload the sketch and acquire the results during the development.
No problem to upload the program, but my readings were very unstable.
Deepening the problem I found that I had a something more than 100 mV garbage on ground and 5 volts lines.
At first I was using the usb 5V to power up the ftdi interface, arduino an my circuitry:
| tried to improve the power supply filtering and then to use a 7805 regulator to get rid of the USB supply for both the arduinoi and my circuitry with no effect. I have been able with the oscilloscope to determine that the main frequency of the noise is around 500 MHz.
If I remove the ftdi board then I have clean power lines (~20 mV noise on the power lines).
Am I facing with a bad FTDI board, despite it's communicating correctly, or that's normal?
Any trick to reduce this noise (the ftdi board already has a 10 µF and I guess a 100nF on its VCC lines) or a suggested vendor for an ftdi board that does not have this problem?.

USB power is very noisy, although I doubt that the main frequency is 500 MHz. Use a different power supply. Batteries are extremely “quiet”.

Post a link/picture of your ftdi card. 500Mhz sounds like an onboard switching supply.

You said small current. How small.
Is it mains AC. Is your code reading it as AC, not as DC.
Did you float the sensor "mid-voltage"
Are you using an amplifier (opamp) between the sensor and Arduino.
Because, according to the datasheet, this transformer outputs 5mv (~TWO digital steps for AC) per 10Amp!
Leo..

Thank you foe your answers.
I don’t see any switching power supply on the board. The strange thing is that I get the noise even if I remove the usb cable and the Vcc pin from the FTDI card, thus leaving it with no power at all. Despite this, I still see the leds on the FTDI card lighted.
About the circuitry, yes I have an OP237 that amplifies the signal by 10 and the second half of it as an emitter follower in series to the inverting input. Not having a voltage reference pin I compare the measurements to the 2.5 Volts I get from the partitors and this doubles the noise I get.
Having to detect currents as low as 1 Amp or something similar I have to deal with signals of about 30 mV, which can be a limit of my circuitry in case of noisy Power supply. As far as I can see no perceptible noise is added by the amplifier.
At present I’m not using the sensor to test the circuitry but a DC voltage obtained with 2x56K resistors plus a 10K trimmer.
The attached schema is not complete as I am still experimenting, but gives an idea of what I am doing.

http://www.ebay.it/itm/FT232RL-FTDI-USB-a-TTL-Conversor-Serie-3-3-5V-Arduino-Pro-Mini-by-ElectroHobby-/271815647483?pt=LH_DefaultDomain_186&hash=item3f4979a0fb

Current sensor.pdf (61.1 KB)

USB 2 uses 480Mbit rate, which would account for approx 500MHz signal.

Any USB connected device is going to produce substantial noise.

However that sort of frequency should be easy to knock down with good
ceramic decoupling - you'll find that unless you are using a low-impedance
low-inductance 'scope probe most of what you see on a good scope is
pick-up from the probe's ground lead forming a loop-antenna.

Your problem with ADC readings is down to lower frequency noise
directly from the USB 5V rail - it should be reduced a lot if you replace
the 5V with a 7805.

The 480MHz signal on the oscilloscope is almost certainly mainly probe-pick
up and not present on the power rails (assuming a decent amount of decoupling
is present). If your 'scope can set a different bandwidth limit, try adjusting
that to see what is present in the DC-->1MHz range.

MarkT:
USB 2 uses 480Mbit rate, which would account for approx 500MHz signal.

Any USB connected device is going to produce substantial noise.

However that sort of frequency should be easy to knock down with good
ceramic decoupling - you'll find that unless you are using a low-impedance
low-inductance 'scope probe most of what you see on a good scope is
pick-up from the probe's ground lead forming a loop-antenna.

Your problem with ADC readings is down to lower frequency noise
directly from the USB 5V rail - it should be reduced a lot if you replace
the 5V with a 7805.

The 480MHz signal on the oscilloscope is almost certainly mainly probe-pick
up and not present on the power rails (assuming a decent amount of decoupling
is present). If your 'scope can set a different bandwidth limit, try adjusting
that to see what is present in the DC-->1MHz range.

I already did that and even isolated the scope from ground, but the noise is still there. Even the addition of a 7805 did not sort any effect, as it didn't taking away any power supply from the fdti card itself, both those coming from arduino and the USB. Only completely removing the fdti card seems to have effect, but in this case I am not able to acquire any value.
Iḿ trying to upload the image of the ftdi card but it looks like the board does not accept links from dropbox. About limiting the bandwidth of the scope, the lower limit is 20 MHz, that is not useful in this case.

20MHz is lower than 480MHz.

I already did that

What? It can't be limiting the bandwidth to 1MHz so I am not sure what you
tried.

I repeat the noise on the scope is an artifact of using the wrong probe, not
necessarily real noise on the line. Does the noise appear to analogRead()
when the Arduino is powered from a 7805?

480MHz noise is invisible to the ADC, its far beyond what the sample/hold
circuit can track.

Sorry, I wrote 500MHz but it was 500KHz, and that's what I had in mind. My scope is limited to 300MHz.
In any case it is definitively the FTDI card.
I remembered I had a 4pin rs232 ttl card, so I connected it to arduino and I am back to roughly 40 mV noise with the scope connected to ground, so it's probably collected noise.
In any case now the readings are much more stable: making an average of 5 consecutive readings, I get 39 to 45 from Arduino with an input voltage of 20.7 mV and the instability remains the same even for lower values.
Now my question is I have a defective FTDI board or I should look for a different board? If yes, which-one?

So are you saying that with the ftdi card connected to arduino by gnd, tx, rx, rst but not 5V, you see the noise?

Then is the card connected to PC when it is connected to arduino? Computer power supplies usually have switching supplies in hundred of kHz switching frequency.

You are likely seeing USB signal switching frequency noise, getting coupled into the power supply line, either internally to the FTDI chip, or from the short trace on the board, or maybe in the USB cable itself.
Try a better quality cable, like one designed for USB3. Try adding a small value cap across the 5V/Gnd pins at the FTDI board output connector, down in the pF range;
try a different FTDI Module,
perhaps from www.tinyosshop.com ($6.90 plus shipping) or from mouser.com (MIKROE-483, $9).

You could have a measuring groundloop.
If your scope is grounded to the mains.
And you use a grounded PC.
Then you have a groundloop, and you see all the hash from the PC and it’s supply on the scope.
If you use a laptop, same story.
Even a non-grounded laptop supply has laptop ground connected to the mains wiring via X/Y safety capacitors.
If you use a laptop, try unplugging the brick, and run it on battery power.
Leo…

liudr:
So are you saying that with the ftdi card connected to arduino by gnd, tx, rx, rst but not 5V, you see the noise?

Then is the card connected to PC when it is connected to arduino? Computer power supplies usually have switching supplies in hundred of kHz switching frequency.

I guess it is the ftdi chip itself generating the noise, as I have the noise even if I unplug the USB cable and disconnect the Vcc pin on the ftdi board.
About ground loop, yes I had one during my last measurement as I didn't isolate my scope from ground, but I did see the noise even with the scope Isolated from ground.
Regarding the cable, it is a micro or mini USB one (the one small that does not fit in cellular phones) and I am not aware of USB 3 cables in that standard, but the one that I am using, judging on its thickness compared to others I have should be shielded.
I guess that, if adding a 50 pF capacitor (I should have some somewhere) will have no effect, I'll look for another FTDI card. Can anyone post the image of a card that is not affected by the problem, just to stay on the safe side with my next purchase?

Another question. Even though I did not make any test with it yet, examining my stc-013-030 current sensor I saw that one of the two signal wires is shorted to the shield. I understand that the sensor itself is a good antenna, but do you think it's worth to separate the shield and connect it to the mains ground?