Acceptable Ripple

Don North north at alum.mit.edu
Sun Dec 7 17:12:18 CST 2014


On 12/6/2014 10:26 AM, Robert Jarratt wrote:
> Just checking a PSU in a MicroVAX 3100 Model 80. I am measuring about 10mV
> ripple on the 5V supply and 20mV ripple on the 12V supply. Does that seem
> reasonable?

Those are very reasonable ripple voltage levels on those supplies.

> I am not sure if it is a scope artefact, but there appear to be large spikes
> at the top and bottom of each sawtooth. The machine appears to partially
> work, it seems to be getting some errors during the memory test, but I don't
> know if that is the spikes I am seeing or just a bad component somewhere.

Could be either ... depends on your scope probe setup. If you are using a 
passive probe with a long flying ground lead, you can easily cause your scope to 
display much larger spikes than exist in the circuit, due to the added 
inductance of the flying ground lead.

The best possible probing can be had using a direct 50ohm coax wired across a 
decoupling cap (or other convenient power/ground spot) into your scope as a 
50ohm input. The probing ground loop will be very small and the scope will very 
accurately reflect on its display what is happening in your circuit. It takes a 
bit more work to set it up, but you will get a better view of what the board 
power rail looks like.

HOWEVER you must be careful not to overload the power input of your 50ohm scope 
termination.

Your scope must be able to handle V^2/R watts on its input, or 0.5W at 5V, and 
3W at 12V. Probably OK at 5V, but not at 12V.
The solution is to add a series resistor on the center coax lead of for example 
150ohm, so the scope only sees 50/(50+150) = 1/4 the input voltage.
At 12V then the scope will see only 3V, and the input scope termination must 
only handle 200mW at 12V power rail.
Of course you must multiply the scope display by 4X to get the real voltage display.

Don



More information about the cctalk mailing list