Does a transformer make the waveform looks more like sine wave?

[BeBuLamar]

I want to test the output of a cheap inverter which produces what they call modified sine wave. I want to see how bad it is but I don't want to hook up my scope directly to 120VAC so I want to use a step down tranformer. What I wonder if the transformer makes the waveform looks better than it is?

[BillyO]

It depends on the transformer.  Transformers definitely have BW limitations and something like a 220V to 12V power supply transformer will do horrible things to sharp edged signals like square waves.  However, one designed for RF might do a better job, if it will handle the voltage.

I would put a mostly ohmic load on it (like an incandescent bulb) and use a few turns of wire around one lead going to the bulb.  A home-made HF transformer.

[Whales]

Yes it can.  Transformers + termination load have a frequency response.  If it attenuates higher frequencies then it can make the waveform look better.

Ordinary steel 240V transformers will likely be particularly distorting, they are only designed to work around 50Hz.  They might be OK up to audio frequencies, but I expect they will still make a 50Hz square or stepped ("modified") wave look smoother.  Also they behave nonlinearly sometimes, which will add extra strangeness.

Make sure to have a termination resistor on the output of the transformer.  If you just hook up a (1Megohm) oscilloscope probe then you will see lots of unwanted capacitively coupled higher frequency signals (and possibly have oscillation/damping issues?).  A 1k resistor might work, YMMV.

[thm_w]

What oscilloscope and what scope probe do you have?

As BillyO says, a step down transformer is not going to have a flat frequency response. I can't find a good reference but here is a power transformer that you can see goes all over the place after about 200Hz: https://netaworldjournal.org/advanced-transformer-diagnostics-sweep-frequency-response-analysis/

You can build a resistive divider, though, that is essentially what a 10x scope probe is already.

[BeBuLamar]

OK so I used a 10x probe and connect it directly. It's just a square wave with only about 70% duty cycle.

[BillyO]

OK so I used a 10x probe and connect it directly. It's just a square wave with only about 70% duty cycle.

That's quite the "modified" sine wave 

[edpalmer42]

OK so I used a 10x probe and connect it directly. It's just a square wave with only about 70% duty cycle.

That's quite the "modified" sine wave 

If you measure the RMS and peak voltage, the so-called modified sine wave is actually better than a pure square wave.  However, I always call it 'modified square wave' because that's what it is!  Modified sine wave is a marketing term that has no basis in reality.  Oh, wait ... I guess that's kind of a definition of marketing, isn't it?   

[CaptDon]

Modified sinewave inverters have a horrible output waveform!! They can damage older equipment with power transformers and solid state rectifiers, even more so if the rectifiers are low loss Shottkey due to reverse breakdown voltages being exceeded by ringing transients of the nearly square waves. If you think the inverter output looks bad with a resistive load put an inductive load on it and scope it!! I have a home made multiple section L C Filter inline with the output of my inverter. It is basically a low pass filter but it does reflect some bad stuff backwards into the inverter.

[bdunham7]

So I just happen to be replacing an old modified-sine Back-UPS.  Most moderate-sized mains transformers I've seen don't do much to filter out anything, especially under a few kHz.  Here is my mains vs the Back-UPS 500 "modified sine".  Transformer (magenta line) is a decent-quality desktop power brick from an old printer or something (several amps output @ 13.5VAC) and I used a 10x  probe, the direct connection (yellow line) I used a MicSig DP10013 differential probe at 500x.  There's no load on the UPS or the transformer, various loads might make a difference, especially something like a linear power supply.  I have no idea what PFC SMPS units do with that input.   

[edpalmer42]

Those waveforms look similar to ones I've seen.  Apparently, the zero-voltage sections reduce the RMS values to something closer to the value for a sine wave while still maintaining a reasonable peak-to-peak voltage.  But the fast rise and fall times still wreak havoc with a lot of equipment.  Something like CaptDon's filter would help knock down the harmonics that those fast times create.

Another interesting way to knock down the harmonics is with a ferroresonant voltage stabilizer, if you can find one.  It's an old technology that's big, heavy, and sometimes noisy (acoustic, not RF).  But if you come across one for a good price it's worth considering.

[Kleinstein]

Even the iron transformers are not that bad for at least the audio range. The get little distortion from the magnetizing current one should use less than the nominal voltage. So for the US, use a 230 V transformer to look at 110 V mains waveform.

The sharp transitions can still stress slow rectifiers and active PFC citcuits could act up.

[Aldo22]

I once had a look at this with my DSO-TC2 (you don't need a good scope for this. Probe in 10x).
The so-called “modified sine wave” can be dangerous for connected devices.

e-ast CL300-12:



e-ast CLS 600-12:

[janoc]

Hooking up that Fnirsi tester to 580-680Vpp takes balls, mate. 

How good is the insulation of your probe and that gizmo?

[Aldo22]

Hooking up that Fnirsi tester to 580-680Vpp takes balls, mate. 

How good is the insulation of your probe and that gizmo?

It is an inverter connected to a 12V battery.
The scope runs on rechargeable batteries.
Nothing is connected to earth/ground.
The probe is a P6100. 10x:600Vpk CAT II.
I haven't touched anything, just switched on the inverter and took a photo.
The risk is manageable.

[tggzzz]

Another interesting way to knock down the harmonics is with a ferroresonant voltage stabilizer, if you can find one.  It's an old technology that's big, heavy, and sometimes noisy (acoustic, not RF).  But if you come across one for a good price it's worth considering.

They also have the advantage of "keeping the power flowing" during short brownouts.

Equally they take a lot of power/time to "get going" in the first place, and that temporary "overload" can cause inverters to give up and protect themself.

[janoc]

Hooking up that Fnirsi tester to 580-680Vpp takes balls, mate. 

How good is the insulation of your probe and that gizmo?

It is an inverter connected to a 12V battery.
The scope runs on rechargeable batteries.
Nothing is connected to earth/ground.
The probe is a P6100. 10x:600Vpk CAT II.
I haven't touched anything, just switched on the inverter and took a photo.
The risk is manageable.

I would be worried that it explodes in my face due to vastly overstated/fake safety ratings and generally poor design (Fnirsi isn't exactly known for quality gear), not that I get electrocuted. And an inverter powered from a 12V battery has plenty enough energy to kill you. 

[Aldo22]

I would be worried that it explodes in my face due to vastly overstated/fake safety ratings and generally poor design (Fnirsi isn't exactly known for quality gear), not that I get electrocuted. And an inverter powered from a 12V battery has plenty enough energy to kill you. 

Here we go again. 
After the standard probe in 10x, it's 24Vrms.
The input has 1MΩ.
I'm sure it will explode and kill me. 

P.S. Please don't start another “Your DMM will kill you” thread.
This has been discussed ad nauseam.

[tggzzz]

I would be worried that it explodes in my face due to vastly overstated/fake safety ratings and generally poor design (Fnirsi isn't exactly known for quality gear), not that I get electrocuted. And an inverter powered from a 12V battery has plenty enough energy to kill you. 

Here we go again. 
Behind the standard probe in 10x, it's 24Vrms.
The input has 1MΩ.
I'm sure it will explode and kill me. 

P.S. Please don't start another “Your DMM will kill you” thread.
This has been discussed ad nauseam.

I'm glad that, unlike me, you've found a way to ensure it is always *10, and the switch will never slip into the *1 position.

I'm glad that, unlike me, you've found a way to guarantee that the probe shield (and hence scope case) will never touch something other than earth. SMPSs are high energy devices, and that always makes me wary that I may encounter some unexpected consequences.

I don't see any spec for how the voltage rating must be derated with frequency. While that's unlikely to be a problem at 50Hz, it doesn't engender a feeling that the probe is well specified, designed and manufactured.

[Aldo22]

 

[janoc]

P.S. Please don't start another “Your DMM will kill you” thread.
This has been discussed ad nauseam.

And yet people still do stupid unsafe shit like this, thinking they know better and that it can't happen to them. Nothing happened last time, so it has to be alright, right?

What to say. 

[tggzzz]

P.S. Please don't start another “Your DMM will kill you” thread.
This has been discussed ad nauseam.

And yet people still do stupid unsafe shit like this, thinking they know better and that it can't happen to them. Nothing happened last time, so it has to be alright, right?

What to say. 

Yup. It is also a lack of experience and imagination, along the lines of "I don't see how this can be dangerous" - because they haven't witnessed it and haven't thought sufficiently about the failure mechanisms.

Imagine telling children that "it is OK to run into the road without looking, because I've done it without consequences".

[BeBuLamar]

I once had a look at this with my DSO-TC2 (you don't need a good scope for this. Probe in 10x).
The so-called “modified sine wave” can be dangerous for connected devices.

e-ast CL300-12:

(Attachment Link)

e-ast CLS 600-12:

(Attachment Link)

Yes what I got looks like your waveform. I used an analog scope so I don't have screen capture. My digital scope is a USB one and I don't want to connect it directly to 120VAC.

[BillyO]

So I just happen to be replacing an old modified-sine Back-UPS.  Most moderate-sized mains transformers I've seen don't do much to filter out anything, especially under a few kHz.  Here is my mains vs the Back-UPS 500 "modified sine".  Transformer (magenta line) is a decent-quality desktop power brick from an old printer or something (several amps output @ 13.5VAC) and I used a 10x  probe, the direct connection (yellow line) I used a MicSig DP10013 differential probe at 500x.  There's no load on the UPS or the transformer, various loads might make a difference, especially something like a linear power supply.  I have no idea what PFC SMPS units do with that input.   

Looks like a very nice transformer.  I imagine there are frequencies in that signal that go into the 10s or 100s of kHz.

[bdunham7]

Looks like a very nice transformer.  I imagine there are frequencies in that signal that go into the 10s or 100s of kHz.

It is pretty decent, it's a few decades old and was made in Taiwan.  I can't remember what it originally powered but I keep it around for all sorts of testing.  It's heavy and probably has plenty of iron, but I think it just might be saturating at the end of the cycle when the UPS is going.  Or perhaps that is the UPS output transformer.  The UPS is also several decades old and quite heavily built.  Anyway, transformers like this often will work through most of the audio range with little distortion, especially if you operate them well below their rated voltage.  Maybe I'll characterize one someday just for fun.

[BillyO]

Maybe I'll characterize on someday just for fun.

Run a bode plot on it. 

[bdunham7]

And an inverter powered from a 12V battery has plenty enough energy to kill you. 

Anything with mains output voltage and more than a few mA can kill you by direct contact, but saying that a typical battery powered inverter poses a danger of making your instrument (however crappy) explode might be overstating the case just a wee bit. 

[bdunham7]

The so-called “modified sine wave” can be dangerous for connected devices.

The one you're showing is especially bad for some linear power supplies because it has too much duty cycle and not enough peak-to-peak voltage.  Properly done modified-sine setups have to start with an output equal to the peak of the sine they are replacing and then shorten the on-time to bring the RMS down to spec.  IDK what the best arrangment is for induction motors, but I imagine they'll run hotter in any case.  A non-PFC SMPS is probably fine either way.

[Aldo22]

The one you're showing is especially bad for some linear power supplies because it has too much duty cycle and not enough peak-to-peak voltage.  Properly done modified-sine setups have to start with an output equal to the peak of the sine they are replacing and then shorten the on-time to bring the RMS down to spec.  IDK what the best arrangment is for induction motors, but I imagine they'll run hotter in any case.  A non-PFC SMPS is probably fine either way.

With motors such as those in a fan, you can hear that something is wrong with the e-ast CL300-12. It sounds sick.
But what surprised me the most was that my Voltcraft Energy Monitor 4000 Pro emitted a little smoke when it was connected to the inverter.
There is no problem with the e-ast CLS 600-12.

[BeBuLamar]

I have a small step down transformer to step it down from 120V to 12V. The waveform looks the same as when I connected it directly. This time I use the digital scope so I can make some measurement. Compared to a sine wave with the same peak value it has the same RMS value as well.

[Vovk_Z]

Power transformers in general don't work as good low-frequency filters. They have limited BW of cause but it is still to wide to be used as good filter.
I mean you possibly want 100-150 Hz LF filter but typical power transformer is  more like 1.5-15 kHz LF filter (depends on load and exact transformer). So it doesn't filter most needed frequency.

[Zero999]

I have a small step down transformer to step it down from 120V to 12V. The waveform looks the same as when I connected it directly. This time I use the digital scope so I can make some measurement. Compared to a sine wave with the same peak value it has the same RMS value as well.

Yes, that's exactly what you should expect to happen.

In my experience small mains transformers have a fairly flat frequency response up to a few tens of kHz, then there's a peak at nearly 100kHz. They can even be used for audio, if the voltage is reduced to account for the minimum frequency being 20Hz, rather than 50Hz. It will do nothing to smooth your modified sine wave into anything remotely resembling a proper sine wave.

[jonpaul]

no see the freq resp of a power trsf and spectrum of the wave in question

j