Oscilloscope as a DMM

[zorthgo]

Hi Guys,

      I've got a question for you. I was playing around with some voltage division circuits (to refresh my memory from circuits I), and I was trying to measure the voltages with my oscilloscope (because I only have a cheap $20 radio shack multimeter I thought it would be more accurate). Well, the problem is that when I did all my calculations, the %error was pretty high @ 52%. So, I took out the radio shack multimeter and I measured everything all over again, and now the %error between my calculated values and my measured values went down to 17%. Does this mean that I should not rely on my oscilloscope for voltage measuring? Which would mean that a DMM is essential, and I should pick on up?

Thanks for the help!

[c4757p]

How large were the resistors? The scope has only one tenth the input impedance of the multimeter, so the readings will drop out with resistors roughly a tenth the size. Try using a 10x probe.

BTW, oscilloscopes are for seeing wave shapes, not making accurate measurements. Don't trust it for that - though it should be better than this.

[alm]

A scope should do a decent job as a voltmeter, but will be less accurate than a DMM unless it's way off. Expect the accuracy to be about 2%. They should be in the same ballpark, however, unless you're using a 1x probe/simple wire and the 1 MOhm input impedance becomes an issue, as c4757p already suggested. If your circuit was ground referenced, than clipping the ground clip to the circuit may have shorted parts of the circuit, and affected your measurements. Also when measuring AC, pay attention to the difference between peak, peak-to-peak and RMS voltage. If it is a digital scope, then it should have the RMS voltage as one of the automated measurements. Even then the DMM and scope will disagree if the signal is not sinusoidal and the DMM is not true RMS.

A DMM is essential, however. A scope is quite inaccurate as a voltmeter, and will not do current, resistance or continuity without external components. It's also ground referenced, while a DMM can easily measure between any two points in the circuit. A decent DMM will also have a wider range without switching probes, auto-ranging, better input protection and a bunch of other features.

A DMM is by far the most used test instrument for most electronics work. I would definitely buy a decent DMM before buying a scope.

[Paul Price]

My older Tek 475B analog oscilloscope was an excellent voltmeter that offered with 10x probe isolation very quick and reliable DC and AC voltage and current readings straight from the graticule that just about matched the accuracy and beat the settling time of  my 1.999 Full Scale count Fluke DVM. The 10x probes assured me that I wasn't upsetting the circuit I was attempting to measure, unlike the unshielded probes of the DVM which did.

In my opinion, a well functioning analog oscilloscope is the ideal universal measurement device and should give precise, repeatable and accurate DC readings and also accurate readings of measurements of changing voltage waveforms versus time.

A digital scope adds the  the possibility of recording capability, but my analog oscilloscope would never be just thought of as some occasionally used piece of equipment considered to be inaccurate and just needed to view "wave shapes."

My Fluke DVM offered portability, best for safe mains voltage isolation and a good  way to measure low-impedance floating voltages and is an essential tool to measure resistance and even test semiconductors and circuits for short-circuits.

A DVM and an analog scope are the two most essential electronic test bench and circuit development measuring and testing tools.

[zorthgo]

@c4757p, the resistors weren't very high. One was a 2K and the other was a 100k. Very simple circuit!

@alm, I was using a 10X probe, and had configured the oscilloscope accordingly.

Should I have been getting better readings from the oscilloscope or is this normal?

Thanks guys!

[c4757p]

Yes, if it has a history of accuracy. My Tek 2445A is calibrated in both time and voltage to within the limits of my eye to read from it, including both graticule and digital cursors. But I calibrated it myself, I've used it for a long time, and I have reason to believe that it's accurate. If all you can compare it against is a $20 Rat Shack multimeter, you don't know how accurate it is.

Should I have been getting better readings from the oscilloscope or is this normal?

You should have read much closer. Check the VAR / vernier knobs - they may be out of the set position, and a lot of times people don't think to check them.

[zorthgo]

@Paul Price, at school we used tektronix oscilloscope and they were fairly accurate, but the one I have is an Owon DS8102V so I don't know if it is as accurate as my school's!

[c4757p]

A new DSO should be quite accurate, even though an Owon's no Tek. Check all your settings and check your probes.

When I warned you not to trust the voltage reading, I didn't know whether you had a brand new DSO or an old clunker analog CRO. You should absolutely be able to trust the voltage and time readings from a new digital scope. (Just remember that the ADC is only 8-bit, meaning it's only a "256-count" voltmeter. Your $20 meter is undoubtedly a 2000 count meter. So it's accurate, but not precise.)

[Paul Price]

Zorthgo, in one word, yes!

Your readings should not have been anywhere so far out of the ballpark.

[zorthgo]

I might be reading the wrong value on the oscilloscope. There are quite a bit of different voltage values. I was reading the pk-pk voltage. But the Vmin seams to give me a closer value to what the DMM is giving me. Should I be looking at Vmin for a simple DC voltage division circuit? 

[c4757p]

Yes. Peak-peak reads from the lowest value to the highest. For pure DC that should be zero. In reality you'll pick up some noise and it will measure that. Use Vaverage.

[zorthgo]

I don't have V-average, the only one I have are: V-pk-pk, Vmean, Vmax, Vmin, Vtop, Vbase, Vamp. Vmin seams to be closer to what the DMM is giving me. But at school, we usually looked at Vmax. But here, Vmax is also off by quite a bit.   

[c4757p]

Average = Mean.

 

If Vmax is significantly out from Vmin, you've got a good amount of signal noise that you're picking up. Some of the Owon scopes are infamous for it. Look for an averaging function (for the waveform, not the measurement) if it bothers you. Also, consider what I said before about the 8-bit ADC. This means that if you're set to 0.5V per division, with 10 divisions (so 5V across the screen), the smallest voltage difference that can be resolved is 5/256 = 0.02V. Don't expect to get any closer than that.

[zorthgo]

The Vmax is reading 400mV while the Vmin is reading 240mV which is exactly what the calculated voltage is. The oscilloscope is set to 2V per division. Is there a way to eliminate the signal noise?

Thanks!

[c4757p]

Disconnect your probe and zoom in as far as you can to whatever you see on the screen, and post a screenshot. Do it with both channels. That is a lot of noise, and there's always the chance that it's faulty.

It could also just be a limitation of using measurement functions that expect a substantial AC waveform with DC, but it's probably best to make sure.

[zorthgo]

@c4757p, none of my thumb drives works with this oscilloscope so I had to take a picture with my phone.

Both leads are measuring the same thing. But giving me different readings!

[c4757p]

Ch2 is a bit yucky. You did disconnect the probes completely, right? I'd say Ch1 is perfectly normal, though. There are a few threads here about either this scope or a related one, why don't you go read up on it? I seem to recall noise issues.

[IanB]

The Vmax is reading 400mV while the Vmin is reading 240mV which is exactly what the calculated voltage is. The oscilloscope is set to 2V per division.

Wait a minute, what? 

Elementary instrument theory says that for maximum accuracy you should measure as close to the full range of the instrument as possible.

You can't measure 240 mV with the scope set to 2 V per division. That would be what, 16 V or something full scale deflection? Set the scope to a scale compatible with your measurement range.

[c4757p]

Nice catch, I missed that. I still think the point about the noise stands, though - look at that picture.

[zorthgo]

@c4757p. My bad, That was a picture of the circuit I was talking about. This is a picture of the oscilloscope without any probes connected to it.

[zorthgo]

@IanB, I don't quite know what "full scale deflection" means! Are you saying that I should go up or down in voltage per division?

[IanB]

An oscilloscope has a beam that is deflected by an input signal to produce a trace on the screen. In a traditional analog scope this is a real electron beam deflected by high voltage plates in the CRT, but in a digital scope the beam is emulated using digital signal processing.

Nevertheless, it helps to think of the basic operation of a digital scope in the same way as an analog scope.

So the beam with no signal has a baseline on the screen. When there is an input signal the beam is deflected up or down to produce a signal trace.

To get the maximum resolution with a positive going signal (i.e. upwards), you would position the baseline near the bottom of the screen and then adjust the vertical scale so the maximum upward deflection is near the top of the screen.

[zorthgo]

Thanks guys, for the help! I guess I am going to be buying a better DMM so that I can compare measurements with the oscilloscope (at least in the beginning while I am getting the hand of how to use the o-scope).


Thanks!

[AndrejaKo]

I have the 100 MHz model of that scope, and here are a few tips I hope you will find useful:

If you want to do averaging on that scope, press acquire button, H1 for acquire mode, F3 for average and use the multipurpose know to set how many samples you want to average.  The more samples you use, the more time it takes for the values to settle down, so this will keep noise out. Use acquire, H1, F1 to go back to sample mode. If you're looking for quickly changing signals, you can use H1, F2 to go to peak-detect mode. With it, the scope will display the most extreme values it captures on the screen.

Also keep in mind that you can have up to 8 measurements at the same time, although it looks as if only 4 will fit the screen.

Another thing to keep in mind is that your scope does all measurements and cursor calculations on the displayed data and not on the captured data. So what does that in practice mean? Well it means that if you can't see it nice, it can't be measured nice. For example period measurements and rise times will provide incorrect results at too low sec/div values, voltages will be incorrect if volts/div is set too high and so on. Try to have the actual wave fill as much of the screen as possible.

To get rid of that "1: None is saved" screen at the scope, press the save button, then H3 button to go to Object and show and finally F2 button to set Show to off.


About precision:
Keep in mind that the scope has 3 measurement ranges which it can use. At 10X settings their borders are at 2 V and 20 V. Try to measure your signal with the lowest range that will fit on screen.

There's one more thing that you can do to increase the accuracy and that's self-calibration. Turn on the scope and keep it running for at least 30 minutes, so that it warms up. Then disconnect the probes and press utility button and then H1. Use the multipurpose knob to go to adjust function then press H2, pause for a second (a text screen should pop out now) and then press it again. You should get a screen with a progress bar. Wait for it to finish. Also keep in mind that you need to do this whenever the room temperature changes significantly, since measurements are affected by temperature. If you haven't done this in a while, make sure you do it before you continue to use scope voltage measurements.

My scope's Vk (cycrms in the measurement menu) usually very close to the reading I get from multimeter for DC and low frequency signals.

[ptricks]

I have a Tektronix 2246 and the voltage accuracy is listed at:
+- (.5% of reading + 2% of the SEC/DIV division + 250 uv)