Testing AC Line Quality with an Oscilloscope - Generator power quality

[Apothecary]

I have an older analog Hitachi Oscilloscope V-1050F.

I want to see the sinusoidal waveform for several power generators I have.
I've watched probably a dozen YouTube videos on this but am still not sure.

If I use an isolation step transformer, will that change the waveform or just the voltage?
If I must use a transformer, can you suggest one appropriate for this?

If my probes have a 10X feature, why can't I just set the probes to 10x and check the voltage directly at the AC outlets?

This is directly from the User Manual for this Oscilloscope.....
Do not apply an excessive voltage. The input withstand voltage of each in put connector and probe input is as follows. Never apply a voltage higher than specified.
INPUT direct 250V(DC + AC peak at 1 kHz)
When probe is used 500V(DC + AC peak at 1 kHz)
EXT TRIG INPUT 250V(DC + AC peak)
EXT BLANKING 20V(DC + AC peak)


( If I do not connect anything to the ground terminal, should it be safe to just use the Hot and Neutral ? )

 

[Nigel Goodwin]

This is a VERY controversial subject on these forums

The best answer, for you, is that is you have to ask you shouldn't be doing it.

No matter what the solution it's important that you understand exactly what you're doing, and WHY!.

The 'safest' solution, in your case, is probably to feed the scope via a mains isolation transformer, and use a x10 probe - and be VERY, VERY aware that the probes, scope, and any connected metalwork may all be at full live mains potential!

 

[Apothecary]

OK. Thanks. It's been a few year since I dusted off the old Oscilloscope and used it but I'm pretty darn sure I used it to test AC line quality before with no problems.
But again, it's been a while and I'm not certain.

I know even less about paragliding, but I'm asking tons of questions about that too
I've felt 120v AC many times. Never seemed enough to do much harm. Just startles you a bit.

I have no plans to mess with 240v tho.

 

[Ylli]

These generators are portables? Not wired into any switchgear? OK.

Leave the scope grounded. Connect the scope probe ground to the generator chassis. Probe the 'Neutral' on the generator output, and you should see almost nothing. We are just checking to see if the generator has the neutral connected to chassis ground. Assuming it does, carefully probe the hot side of the output. Do use the 10:1 probe, and make sure the scope is set to accept the expected voltage. 20 V/div or perhaps 50 V/div would be good.

Alternately you could connect a filament transformer to the output and then scope the output of that transformer. A transformer will not appreciably affect the waveform.

 

[Apothecary]

Thanks YLLi. Will try that.

 

[Apothecary]

What about this? My Oscilloscope is a 4 channel unit......

 

[Ylli]

If the output of the generator is not ground(chassis) referenced, it should be. If not though, I would go with using the filament transformer rather than trying to do a differential measurement.

 

[Apothecary]

If the output of the generator is not ground(chassis) referenced, it should be. If not though, I would go with using the filament transformer rather than trying to do a differential measurement.

Any filament transformer will do?

This One?
CHICAGO STANDARD P-6469 FILAMENT TRANSFORMER 115 VAC INPUT 25.2 VAC @ 1 AMP OUTP

 

[Apothecary]

These generators are portables? Not wired into any switchgear? OK.

Leave the scope grounded. Connect the scope probe ground to the generator chassis. Probe the 'Neutral' on the generator output, and you should see almost nothing. We are just checking to see if the generator has the neutral connected to chassis ground. Assuming it does, carefully probe the hot side of the output. Do use the 10:1 probe, and make sure the scope is set to accept the expected voltage. 20 V/div or perhaps 50 V/div would be good.

Alternately you could connect a filament transformer to the output and then scope the output of that transformer. A transformer will not appreciably affect the waveform.

Just to be sure......
Could I just as easily use a VOM to do this test for chassis ground? This test is done while the generator is NOT running correct?

 

[Apothecary]

Ok, after checking the generator(s) in question............(VOM, generator NOT running)

They ALL have connection between the chassis and the GROUND PIN of the outlets only.
None of them have any continuity between the chassis and Neutral or Hot

These are all portable generators not connected to any other equipment.

 

[Ylli]

So it sounds like the output is floating and not ground referenced at all. Just to keep things as safe as possilbe, use the filament transformer. The one you referenced is fine.

 

[KeepItSimpleStupid]

If a generator is feeding a house, then the neutral-bond is made inside the house (at the house ground rod)

if it's a stand-alone generator, then the neutral-ground bind should be made at the generator.

In both cases, you would want a fuse to blow if ground is connected to hot.

A GFCI would allow an ungrounded source. You can in the US have outles with 3 prongs and no earth connectos IF they are labeled properly (No Earth ground) and connected to a GFCI.

 

[Reloadron]

I can give you the same answer here as the other forum. You can do a differential measurement as suggested using two channels and two 10:1 probes. Since the generator(s) are stand alone isolated and if your scope is running off and grounded to AC mains power of another source you can also place scope ground on neutral and the vertical input on the line side (hot). The differential method insures a better margin of safety.

Keep in mind while a scope will display a sine wave just fine it is not really designed to measure power line or in your case generator power quality. Distortion of the signal consist of harmonics, hum and noise with the latter two showing up on the peaks and valleys. Because of the V/Div settings of the vertical input channels it becomes hard to see anything since you are using the gain to display the entire signal.

While some newer scopes offer a FFT (Fast Fourier Transform) feature most older scopes do not.

As to the use of a transformer to reduce the signal you plan to measure. The transformer likely will not pass on harmonic distortion if the frequencies exceed the transformers intended use or frequency response. The use of a transformer can also add noise so it has a few things going against it.

There is an old trick which is sometimes used. Using a cheap DMM which is not a true RMS responding meter measure the output voltage. Then measure it again using a quality true RMS responding meter. The cheap DMM will actually be an average responding RMS indicating meter so it will only be accurate on a nice true sine wave. A quality RMS responding RMS indicating meter will afford a true RMS reading regardless of input waveform shape. Subtract at cheap meter reading from the quality meter reading and the greater the delta the more distorted the waveform. Not great but it works.

The ideal solution to measuring power quality is to use an instrument like a line analyzer designed for the purpose.

Finally as I mentioned in the other forum when looking at generator output the best results will be obtained when the source is under a load and the load should be about at least 75% of the generator's rated power output.

Ron

 

[Apothecary]

Here is the result.
A few notes.
1). I attempted to check the house AC first.
2). This connection was to a multi outlet (cheap) adapter (Not directly to a home outlet)
3). The probes did not have to be unsheathed....this signal was generated when the probe was placed within 1 mm of the hot or neutral.
4). The probes were set to 10X

 

[Reloadron]

Same response I left you on the other forum.

Placing the probe in close proximity to the AC source is not good enough. The probe needs to be connected directly to the source. You can use either method suggested as to using a single or both channels. You can get the same results by just placing a finger tip on the probe tip and increasing the vertical gain of the scope. You don't mention the scope's time base settings or the vertical gain setting but what you see pretty much looks like general 60 Hz noise.

Ron

 

[ronsimpson]

the probe was placed within 1 mm of the hot or neutral

The 1mm air gap makes a fractional pF capacitor. So you are ac coupled. It is likely that high frequencies come through stronger than low 50/60hz frequencies. It is hard to know from here, but likely the high frequency noise you see on the 60hz is not as strong as it appears.
I have some 100:1 probes that are good for 1500 volts.

 

[ronsimpson]

If money was not a problem:

There are a number of diff probes made for measuring signals like yours.

I have a battery powered, isolated scope. Not like this one in the picture but close. It is safe to connect the scope's ground to the power line because the entire scope floats. The scope's case is plastic so it is safe to have the "ground" at a voltage of 1000 volts. I often use it for power line problems.

 

[Reloadron]

Here are a few examples of using a scope, single channel and looking at AC mains: 60 Hz Sine Wave

V/Div 50 (Using 10:1 Probe)
Time Base 5.0 mSec/Div
Measured True RMS voltage was 123 VAC

Example of placing probe tip in close proximity to AC Hot but not connected:

V/Div 2 (Using 10:1 Probe)
Time Base 50 mSec/Div

Just a noisy garbage signal and not a true representation or actual power quality.

Ron

 

[The Electrician]

Ylli is right on about the use of a filament transformer, and Reloadron is overly pessimistic.

If a typical filament transformer is used with no load other than a 10x scope probe on the secondary, the frequency response is more than adequate. When there is a resistor load on the secondary the leakage inductance forms an LR low pass filter which rolls off the response (just like an RC low pass would).

Here are some scope captures showing just how good a filament transformer can be in this situation. I used an ordinary dimmer with an incandescent light bulb to create a distorted grid waveform. The primary of a filament transformer was also connected in parallel with the light bulb. A 100 MHz bandwidth isolated differential probe like the one ronsimpson showed in post #12 was connected to the light bulb/filament transformer primary, and that signal is shown as the yellow trace on the scope. The waveform is somewhat asymmetrical, apparently the fault of the triac in the dimmer:

Here is the output of the filament transformer shown as the green trace. I've offset the green trace to better show each waveform separately:

Here I've increased the sweep speed and removed the offset. The two waveforms are almost perfectly superimposed now--they appear to be essentially identical.

 

[The Electrician]

Here i've increased the sweep speed even more to zoom in on the fast rising portion of the waveforms. Again, they're essentially identical:

Now, to show the effect of a load on the secondary creating an LR low pass, I've connected a 100 ohm resistor to the secondary. Notice how the rise time of the green trace is now slower. The bandwidth of the transformer has been reduced enough to cause the secondary waveform (green) to not be essentially identical to the waveform from the isolated differential probe:

I used the FFT function of the scope to look at the harmonics of the yellow (wide band) trace, and compare them to the harmonics of the green (filament transformer) trace. They are essentially identical out to the 40th harmonic as displayed by the FFT, and probably even higher.

The lesson here is that a filament transformer is more than good enough to accurately examine the grid waveform. I did this test on 5 different filament transformers, all less than 50 VA, with the same result. It's unlikely that there will ever be distortions with significant harmonic components beyond several kHz on a typical grid (or generator) waveform, and an unloaded filament transformer can handle that.

I would hope that it goes without saying that if there is noise with components at high audio frequencies and above, this technique won't accurately show them. This method is mainly to see gross deviations from a 50/60 Hz sine waveform.