oscilloscope debugging

[stopgo]

I'm using an HP54501A 100 MHz scope that I got off of ebay.  I'm relatively new to these things.  It's worked fine for several months.  I've always been surprised that it could work w/o a shared ground with the circuit... that doesn't seem possible.  But it did.

One day, I started it up, accidentally auto-scaled (something I hadn't done before), readjusted to my desired resolution, and then took a measurement.  Now everything reads as a very slight sine wave around 0 volts.  Even DC sources.

The internet says my oscilloscope's ground (which is connected to my wall socket) needs to be connected to my circuit's ground (which is connected to a battery).  That seems reasonable to me, but I have no idea how to accomplish that.  Could that be the problem?  If so, how do I connect those two grounds.  If not, does it just sound like a broken oscilloscope?

[hazzer123]

What is the frequency of the slight sine wave which you mention? If it is the same as the local mains power supply frequency then  there are some problems with the internal circuitry since it is allowing mains voltage to intefere with the input signals.

To connect the grounds together, you should be able to connect the negative terminal of the battery to either any metal part of the oscilloscope's casing, any metal pipes in the building or a ground terminal on the oscilloscope. Which ever is most convenient (i suspect the last one).

[stopgo]

I just connected the ground from the casing and now it works again!  Thanks hazzer123!

[Admin]

stopgo . . . Im trying to understand how you can use your oscope for months without using both connectors . . . to get a voltage signal, you actually need two connections (because a voltage is actually a difference in charges).

The (-) is typically ground, and (+) is your signal.

Read my notes on ground:
http://www.societyofrobots.com/robot_faq.shtml#circuit_debug

[stopgo]

Admin, I am equally confused.  It always absolutely shocked me that I was able to examine, say, PWM signals without connecting my scope's ground to the circuit - but it happened.  The only possible explanation I can think of is that my circuit's ground already happened to be at a voltage near the ground coming from my wall, but that seems very unlikely.

[Steel_monkey]

In most O-scopes, ground connector is already internally connected with ground from the plug. Of course, it only happens if socket have third ground connection. As I know, in most highly-developed countries sockets have 3 pin, so if you are using wall adapter with 3 pins to power your circuit, potential of ground lines of your circuit and o-scope will be equal. PC also should be grounded, perhaps you power from USB. But this is just an occasion.
P.S.
Ground-connected o-scopes can cause a lot of problems working with uncoupled power supplies ( socket is not coupled from supply circuit with transformer). "Ground" of power supply will have floating potential of one phase from your socket, and you connect it to actual ground... Don't try to test power supplies!