Always check your earthing
04/09/26 13:08 Filed in: Hardware
The modern DSO (Digital Storage Oscilloscope) is a powerful diagnostic instrument well suited for diagnosing hardware but there are traps for young players the most insidious of these is differing earth potentials. If you get this wrong you can easily release the magic smoke from a circuit and worse still put your life in potential danger.
This week's adventure started with a Rigol DHO924S a perfectly adequate 12bit 1.25GSa/S scope for most tasks (except for the one I really want to use it for) and just before connecting up the signal probe earth to the USB powered microcontroller circuit I did my usual check with a multimeter of the difference of the zero volt on the circuit under test and the signal probe ground wire. 88 volts. A scarily large number for a sensitive device under test. While in this case the connection would probably have been safe, good practice made me change the supply arrangement - just in case.
This week's adventure started with a Rigol DHO924S a perfectly adequate 12bit 1.25GSa/S scope for most tasks (except for the one I really want to use it for) and just before connecting up the signal probe earth to the USB powered microcontroller circuit I did my usual check with a multimeter of the difference of the zero volt on the circuit under test and the signal probe ground wire. 88 volts. A scarily large number for a sensitive device under test. While in this case the connection would probably have been safe, good practice made me change the supply arrangement - just in case.
So what is happening?
Power supplies can be made in two ways: floating or ground referenced. If you have a floating power supply it still provides the required voltage difference between its 0v and positive terminal, but nothing is known about those terminal voltages and any other voltage. In practice there is always some coupling to something via a high impedance route ie large voltage difference possible but only at trivial amounts of current. (While pure speculation an explanation would be the present of noise reduction capacitors in the power supplies and that the measurement was conducted in Australia which has 240V 3 phase power)
The Rigol scope is almost certainly NOT ground referenced. The big green and yellow wire with banana plugs on each end supplied with the scope is an excellent clue. This is probably means I was worrying about nothing and only trivial current would flow if the circuit and probe earths were connected.
However as 88V is a rather large difference, to be safe I put the USB powered circuit on a power supply where I could attach the green and yellow earth to the power supply 0v terminal measured my 0v ground terminal difference again at a satisfyingly small fraction of a volt.
Another alternative would have been to have used a differential probe.
Why all the fuss. Because:
- if the power supplies are not connected by a large impedance route real current can flow with damaging results.
- other scopes are earth referenced the moment you plug them in (the Rigol can float - you can run it from a USB battery pack) and on a large class of other scopes this would be a sure sign that you are about to cause damage
No circuits or test equipment were harmed in this adventure
So what is the problem with the Rigol DHO924S scope?
I happen to be measuring a Pulse Per Second (PPS) Signal. The scope will not trigger at intervals more the 20ms / div. It only provides free run mode at longer intervals. As a result you have to stop the waveform to capture it. No single shot. No stable display.
This makes taking measurements a lot less convenient than on other scopes. Unfortunately this failing in the Rigol is not obvious from the spec sheet.