CS trigger glitch vs. isolation

I solved a long standing glitch problem by connecting the CS power adapter via a two wire adapter to the AC---floating the earth ground wire.

I am using the CS to observe the output of a high gain amplifier (about 40,000). When the input of the amplifier is connected to a function generator, the CS was causing a large glitch, even with the function generator turned off. The bnc cable has a T at the function generator, with one branch of the T grounded, the other branch going to the function generator output. (I tried an old analog function generator, same problem)

If the amplifier input is grounded with a bnc short right on the input (no bnc cable),
there is no glitch.

The first picture shows the setup in which the glitch is observed.
Second picture: The rigol scope shows a big glitch every time the clever scope triggers.

Third picture If the CS is not triggered, or if the CS power ground wire is left floating, there is no glitch. The input referred voltage is only about 500uV. The trace width is due to 60Hz interference.

This probably could be fixed by a circuit change---something for the next hardware rev.

I solved a long standing glitch problem by connecting the CS power adapter via a two wire adapter to the AC---floating the earth ground wire. I am using the CS to observe the output of a high gain amplifier (about 40,000). When the input of the amplifier is connected to a function generator, the CS was causing a large glitch, even with the function generator turned off. The bnc cable has a T at the function generator, with one branch of the T grounded, the other branch going to the function generator output. (I tried an old analog function generator, same problem) If the amplifier input is grounded with a bnc short right on the input (no bnc cable), there is no glitch. The first picture shows the setup in which the glitch is observed. Second picture: The rigol scope shows a big glitch every time the clever scope triggers. Third picture If the CS is not triggered, or if the CS power ground wire is left floating, there is no glitch. The input referred voltage is only about 500uV. The trace width is due to 60Hz interference. This probably could be fixed by a circuit change---something for the next hardware rev.

Hello BNevins,
We will have to check this out. A possible reason is that after triggering, we disable the ADC until we next start it capturing again. This results in about a 300 mA load current change. The negative terminal of the power supply is connected to ground. When there is a load change, I guess this could be coupled into the ground system shared with your high gain amplifier, and it amplifies the ground current change. Any impedance in the ground, and you'll get some voltage. With 500 uV input, you don't need much impedance! So what you have done fixes the problem but increases the risk if you are working with mains voltages.

Tricky problem! Another possible solution is that we have a way for you to disable the ADC on/off when dealing with sensitive situations, though probably getting rid of the ground is the best way out. Note that you still need a ground reference somewhere - the best place probably being the copes BNC ground being connected to your own ground. Otherwise external signals such as the mains will get in.

Hello BNevins, We will have to check this out. A possible reason is that after triggering, we disable the ADC until we next start it capturing again. This results in about a 300 mA load current change. The negative terminal of the power supply is connected to ground. When there is a load change, I guess this could be coupled into the ground system shared with your high gain amplifier, and it amplifies the ground current change. Any impedance in the ground, and you'll get some voltage. With 500 uV input, you don't need much impedance! So what you have done fixes the problem but increases the risk if you are working with mains voltages. Tricky problem! Another possible solution is that we have a way for you to disable the ADC on/off when dealing with sensitive situations, though probably getting rid of the ground is the best way out. Note that you still need a ground reference somewhere - the best place probably being the copes BNC ground being connected to your own ground. Otherwise external signals such as the mains will get in.