microvolt amp for scope

[ZeroResistance]

I trying to build a microvolt preamp for my usb oscilloscope. I have a Picoscope pen style scope.

I am using a INA118 dip package and am currently testing on a breadboard.

My gain resistor is around 470 ohm which gives a gain of approx 110 as per the datasheet.

I have built my circuit as per the reference circuit in the datasheet. Currently I have just shorted the input terminals of the INA and this the waveform at the output of the INA I am getting (attached)

As you see its quite noisy. My input voltages would be in the 10's of microvolt it could be either polarity.
Currently I have powered the INA with a dual supply +5V / -5V, reference pin is grounded.

[StillTrying]

Your 4 pulses each look near identical, so I would think it's a 21kHz SMPS somewhere.

[schmitt trigger]

just to make sure, that the noise is not power supply induced, try to fully power the circuit from batteries.

[ZeroResistance]

Your 4 pulses each look near identical, so I would think it's a 21kHz SMPS somewhere.

I have powered the circuit from a linear power supply apart from that there are computers and fluorescent tube lights on and running and the usb scope is connected to the computer.
I have also decoupled the INA supply rails by 0.1 uF ceramic capacitors.

PS  The fluorescent tube lights are running on electronic ballast.

[ZeroResistance]

just to make sure, that the noise is not power supply induced, try to fully power the circuit from batteries.

Excellent idea I could try that. But what about the usb connected scope would that still be ok with the battery powering the circuit?

[bson]

What if you just ground the USB scope probe by hooking it to the ground clip?  Do that and check that it's not created by the USB scope itself.

If you see it still it's coming from scope or USB port's power supply/management.  (You don't say how the scope is powered.) If it's from the computer port supply, consider a USB decrapifier.

[Ian.M]

Its pointless trying to breadboard microvolt level stuff^*. There's just too much noise pickup. A well designed and decoupled  minimum stray wire matrix board layout in a screening can has a reasonable chance of working OK, so dig out an Altoids tin, some feedthrough caps for the power leads (but typically you solder Ground direct to the tin), and suitable bulkhead mount connectors to get the signal into the tin and out to the scope probe via a BNC adaptor or similar *without* using the probe's ground clip lead.

* unless its a narrow-band tuned circuit or low-pass filtered to near DC

[ZeroResistance]

What if you just ground the USB scope probe by hooking it to the ground clip?  Do that and check that it's not created by the USB scope itself.

If you see it still it's coming from scope or USB port's power supply/management.  (You don't say how the scope is powered.) If it's from the computer port supply, consider a USB decrapifier.

What you are trying to say is connect the oscilloscope earth alligator clip to the probe tip and so this way the probe tip is shorted to ground and now measure the signal level.

That seems at good idea. The oscilloscope has no seperate power it is usb powered.

[rstofer]

What you are trying to say is connect the oscilloscope earth alligator clip to the probe tip and so this way the probe tip is shorted to ground and now measure the signal level.

That seems at good idea. The oscilloscope has no seperate power it is usb powered.

I would imagine the test lead still works like an antenna.  Better, perhaps, but not a dead short.

Why not put a 50 ohm terminator on the scope input?  Lacking a terminator, how about shorting the BNC connector with a very short piece of wire.

[ZeroResistance]

Its pointless trying to breadboard microvolt level stuff.  A well designed and decoupled  minimum stray wire matrix board layout in a screening can has a reasonable chance of working OK, so dig out an Altoids tin, some feedthrough caps for the power leads, and suitable bulkhead mount connectors to get the signal into the tin and out to the scope probe via a BNC adaptor or similar *without* using the probe's ground clip lead.

These are excellent insights... many thanks

[ZeroResistance]

I would imagine the test lead still works like an antenna.  Better, perhaps, but not a dead short.

Why not put a 50 ohm terminator on the scope input?  Lacking a terminator, how about shorting the BNC connector with a very short piece of wire.

This is as pen style oscilloscope there is no BNC connector just a usb cable that directly goes into a pen style blob that can be hand held like a pencil.
I can do what you say using another oscilloscope though.