Hi. I have the moku go and have a question about the earth connection. I only have continuity between earth and bnc ground when the unit is off. Ones I turn the moku go on this continuity is lost. Is this normal or is there anything wrong with my unit?
When i use my scope a measure the output from the moku go waveform generator 75mV sinus 1khz the signal is shaky and noisy. If I measure a similar signal from another signal generator I have with the exact same setting on the scope the signal looks perfect.
This was the start of thinking something is wrong with the moku go and I then found the lack of earth connection on the bnc ground.
Do you have continuity between earth and bnc ground when your unit is on?
The ground of the power supply and the BNC’s are connected, they are not isolated or floating. The earth of the power supply is only connected to ground for safety and ESD compliance, it is not expected to have a direct connection between the power supply earth and the Moku:Go ground. For a stable signal, the oscilloscope ground needs to be connected to the Moku ground. Relying on the oscilloscope power supply and moku power supply earth to provide a common reference will result in shaky signals.
I too am noticing noise issues (10 mV rms) when I connect the Moku:GO output to a low noise current driver for external modulation (using a BNC cable). This noise level is much lower (1 mV rms) when connect another function generator (RIGOL). In this context, I have a bit of confusion understanding the statement made by Nadia:
As I see in my MokuGO, the outer shield of the BNC (ground) is connected to the input power supply earth. In this case, the way is to have a common ground reference point among Moku:Go and any other device is via the BNC cable outer shield. However, this still is 10x noisy as compared to other function generators (as mentioned above). Is there any suggestion to resolve this issue?
Thank you for reaching out to Liquid Instruments! When you mention 10 mV of noise, is this the DC offset of your signal? How are you measuring the noise of the Moku:Go and other function generator? Once I know a bit more about your setup I can try to help and optimize the system. Looking forward to hearing from you!
Thanks for getting back. The 10 mV noise level was measured as follows:
MokuGo (or another function generator) output BNC yerminal was connected to the input of an oscilloscope (1 Mohm, AC coupling). The noise appears like a large spike (-ve pulse) appearing every now and then, riding on top of smaller fluctuations spread over time.using the FFT feature, I see that there is some component at 20 kHz.
I can measure it more carefully on a spectrum analyzer and share the results.
The oscilloscope noise floor itself with it’s input shorted to ground was about 0.5 mV RMS.
Both MokuGo and the oscilloscope are powered through the same AC mains using an extension board.
I see some improvement when I connect the ground of the oscilloscope with the ground of MokuGo power DC supply output using a banana plug and thick wire (16Amp).
My questions specifically are:
What is the voltage noise (in nV/sqrtHz) of the MokuGo output?
If this is a ground loop issue then what is the best way to tie together the grounds to a common reference?
( other devices usually have a ground connection on the back panel).
Hi. I resolved the issue by replacing the power supply and building my own. I added an on/off switch, which I always missed in the original design, and a DC switch that allows the DC negative to be connected to PE or left floating.
After changing the power supply, all issues with the Moku Go disappeared, and the previously unstable signals are now completely stable.
Assuming your oscilloscope is properly grounded, the best way to ensure the oscilloscope ground and Moku:Go ground are connected is to directly connect a BNC from the Moku:Go output and the oscilloscope input. This will connect the Moku:Go and oscilloscope ground resulting in a stable signal.
We were able to resolve the noise issue finally by operating the Moku with an isolated power input (similar to that suggested in Tom’s post). Having the devices connected by the grounds shield of the BNC cables was not sufficient in our case.
