180mV or more Vpp on Rigol scope, hooked up to a battery

[arcfault]

I recently purchased this used Rigol DS1052E. I'm pretty perplexed at what it reads. Consistently shows 180mV Vpp or more (or a little less) when hooked up to a battery (have tried two). AC coupled shows a similar number. Has done this with another probe as well. Have done a couple of self calibrations. Have tried different timescales, voltage scales, etc. I'm also not very familiar with scopes, so might be doing something stupid.

Even putting the probe on ground, I was getting a reading of 50mV Vpp. I think channel 2 behaves similarly.

Also attached a square wave for reference, in case it's useful.

Would appreciate any advice you  can offer. I was hoping to check for ripple from different solar charge controllers, but it seems impossible to do so when there's so much ripple being shown on just a battery.

Thank you!

[AG6QR]

Your red/black wires, combined with the battery, make a huge loop, which will act as an antenna if there are any nearby changing magnetic fields to pick up.  Take the battery out of the picture and put an ordinary wire between those red and black wires, keeping the loop area approximately the same size as you have it now (of course be careful not to short out the battery!).  Do you see any spikes?  Are they at all comparable to the spikes you saw when probing the battery?

If you do see voltage spikes when measuring the voltage across the wire, try re-arranging the wires so that the loop area is as small as you can make it.  Does that change things?  For fun, try making the loop as big of a circle as you can.  Does that change things?

Try switching on the 20MHz selectable bandwidth limit.  I can't remember where it is in the menus, but look around the vertical menu.  It might attenuate the spikes a bit.  Or not.

[jwet]

O'Scopes are very wideband devices and noise is everywhere and dictated by physics for high impedance sources like a scope probe.   Use a real 10x scope probe if you can, minimize any loop area as AG6QR noted.  The "ripple" on a pure DC source should be as good as it gets- a test is to short the probe with the shortest path possible, limit the BW to 20 MHz and look at the results- you shouldn't see more than 10 mV.  The noise inherent in a 10 M-Ohm resistive source is about 1.8 mV in a 20 MHz bandwidth.  This is RMS, the peaks you might see are less than 12 mV.   This is Johnson noise and is an inescapable fact of life.  The 1052E uses an 8 bit analog to digital converter- this gives a basic (quantization error- physics limited) noise level in dB of 6.02B + 1.76.  For 8 bits, this is  ~50 dB or about +-.3%- this gets added to the Johnson noise above.  If you're on a 100 mV AC scale, with BW limited to 20 MHz, you should see something like 1.8 mV from Johnson and .3 mV from quantization- peak to peak could be 15 mV.   If you don't limit BW, the 1052 would have 50 Mhz of BW that would see about 3mV RMS of Johnson noise.  The peaks of this noise are almost always less than 6x RMS or 20 mV - either way, you shouldn't be seeing noise approaching 30 mV if you have a decent connection to the source without loops.  There are some other minor error sources (sin(x)/x interpolation and excess noise) but that's about it.  You should be able to resolve 30 mV easily with a reasonable setup.  Rigol doesn't specify a noise value but you can get a pretty good handle from first principles.  A good test to see if the scope is right it to put a terminator on the input- 50 ohms doesn't produce much Johnson noise, this should be pretty quiet.

[arcfault]

Thanks for the feedback!

I live in a rural area, so interference should be at a minimum. My "lab" is a ways from the house.

I tested with a 10x probe, clamped to itself, with nothing on in the room. Still saw 160mVpp.

I did a primitive test with a AM radio and found the scope was fairly "noisy" when running. Not sure if that might indicate anything.

I should probably get a probe terminator to test the scope more, but I am definitely having my doubts about it.

[BrokenYugo]

Try going in the channel menu and setting the coupling to "GND" (ground) rather than AC or DC, this disconnects the input internally and grounds the amplifier input, what you see  then will be noise introduced by the scope.

The alligator clip ground wire that comes with a scope probe makes a great antenna, not for sensitive measurements or high frequency work. That's what the little spring looking ground probe that clips on right at the tip is for.

[jwet]

I actually looked more closely at your pictures and understand them a bit better.  You need shielding all the way out to the source.  Look up Dave's notes on measuring power supply ripple- all the same issues apply but its worse a battery, the terminals are farther apart. 

You're just picking up every radio signal and noise source between 0 and 50+ MHz.  If you have an FFT mode, you might be able to identify a dominant frequency- maybe a radio station or some man made noise from a lamp or motor inverter, etc. 

Once you're shielded and grounded and BW limited your setup, my Johnson noise and Quant noise will set a lower limit- that's the best you can do.  If you do the ground test either at the input relay as BrokenYugo mentioned or with a 50 ohm terminator and it fails, look at the scope.  It could be that someone went in to service it and didn't get the shields back in place over the front end.  Watch Dave's tear down and see what's supposed to be there.  There are usually matchbox sized tin half boxes over the input circuits- soldered down along their edges.  Check that they're present and soldered well.  Also make sure that the chassis and cover are properly grounded.  These scopes use switching power supplies and their own EMI can get into them if they're not put together properly.

Finally, big batteries aren't a low impedance source.  At RF frequencies, they're an antenna. You might have to resort to putting the battery in a shielded and grounded cabinet (old tool box? or something made of PCB? that is grounded)

Noise is everywhere...

Good Luck.

[arcfault]

Thank you all for the replies!

I haven't had a chance to test the ground mode or anything else yet, but I did pick up a Tektronix 2225. Beautiful scope!

Took me a minute to figure out how to work it. It reads the battery voltage completely stable, practically without any interference. As soon as I connect it to my high frequency variable power supply, I see a very distinct ripple of roughly 0.4Vpp under no load.

Using the same probes and all. I think the Rigol is definitely challenged in some way.

Does this make sense? Am I in the right direction?

Thanks!

[jwet]

Its very possible that your Rigol is noisy for one reason or another.  Do the test with the internal termination, etc to verify.  Like I said above, it might be missing a shield if you're lucky or could have something more complex if you're not...  Good luck.

I had a 2225 back in the 80's, bought it brand new after getting some kind of bonus at work.  These were Tek trying to keep up with low cost import scopes- they used to advertise in the hobby magazines for $999- I loved that thing.  THe 2225 was kind of the Rigol of its day.