Scope Loading?

[cswanson]

I have a circuit (5V DC) that takes a 1 Hz sine wave at 800 mVpp and amplifies it to 4.46 Vpp with a DC bias that puts the lowest swing just above +80mV. This works great.

I send this output to a MCP3008 ADC which is an SPI slave to an Arduino Uno. The master does nothing more than query the MCP3008 for the current value and sends it to the serial output. (Note: I am powering the arduino with the 5V DC supply and it is connected to the computer USB for serial output. ) The output is an increasing and decreasing list of integers mapping to the voltage swing.

This works great as well...

as long as the scope probe is connected to the circuit's ground rail and the analog input pin of the MCP3008.

If I disconnect the probe, I get a constant output of 868 or 869 on the arduino's serial output.

The scope doesn't have to be on, but the probe must be connected to the scope. ( Scope is a Rigol DS2072A )

I measured the resistance (1.2MOhm) and capacitance (3.3 uF) of the probe and tried attaching passive components to the MCP3008 input channel and circuit ground but it had no noticeable effect.

What am I not considering?

[Tac Eht Xilef]

Without seeing the actual circuit you've built, it's impossible to say.

I measured the resistance (1.2MOhm) and capacitance (3.3 uF) of the probe and tried attaching passive components to the MCP3008 input channel and circuit ground but it had no noticeable effect.

Again, without seeing the actual circuit that would've been the first thing I would've tried, but you've done it. But 3.3uF probe capacitance seems awful (as in "a few orders of magnitude") high...

[ludzinc]

"The scope doesn't have to be on, but the probe must be connected to the scope."

Do you mean as long as the probes ground wire is connected, or as long as the probe tip is connected to your circuit.

I've seen the former a lot - broken ground path, never the latter.

[cswanson]

@Tac Eht Xilef I was surprised too, I'll double check.

@ludzinc The probe must be connected to the circuit and plugged into the scope. If I disconnect the probe from the scope I get garbage.  Thanks for keeping me honest.

[w2aew]

Also sounds to me that the scope's ground lead is establishing a common ground between the Arduino and your ADC. Without the scope probe, one of them is floating with respect to ground.

[cswanson]

Cool.

I'm planning to integrate the atmega328 MCU onto the breadboard, and I am currently replacing the serial output with a 7 segment display.

So when I rid myself of the computer the problem should clear.

Thanks!

[T3sl4co1l]

Sounds like bad layout... pictures?  Schematic?

Tim