EG4 6500 EX Under an Oscilloscope

I think too much emphasis is on the waveform and not the overall voltage stability. If the inverter is hunting, stepping up & down, over correcting, under reacting or delayed response only a high end oscilloscope can record the change and see the jump/change in scale. What your eyes are seeing is happening within a few cycles and a common scope is still processing the average. The overall shape is the same just the scale changes. This is a picture of a waveform of a cheap generator with terrible percent of distortion and you wouldn’t want it powering anything you really like other than power tools yet
there’s absolutely no flickering of any light, LED or not.

Skypower said:

I think too much emphasis is on the waveform and not the overall voltage stability.

Click to expand...

Agree; there are likely 4 intertwined issues: DC bus stability, L-L/L-N/N-G voltage stability, harmonics/waveform, and AC transient response, roughly in order. Without DC bus stability you have to work really hard to make any of the other things stable. Transient response requires everything else to be good. All of these issues are also dependent on PV, battery state of charge, and inverter load.

Shimmy said:

Agree; there are likely 4 intertwined issues: DC bus stability, L-L/L-N/N-G voltage stability, harmonics/waveform, and AC transient response, roughly in order. Without DC bus stability you have to work really hard to make any of the other things stable. Transient response requires everything else to be good. All of these issues are also dependent on PV, battery state of charge, and inverter load.

Click to expand...

Tha is why my thinking these days is tending towards adding some ultra capacitors to slow down rapid voltage fluctuations on the dc bus.

Skypower said:

I think too much emphasis is on the waveform and not the overall voltage stability.

Click to expand...

Not at all, the waveform is telling a story. It should be rock solid.
I scoped a Growatt 6000T and it had a perfect waveform, better than FPL.

That is not the issue.
It may look perfect with a steady constant load, but if a sudden high power load is applied (or suddenly removed) the inverter voltage regulation may not be quite fast enough. Especially if the incoming dc is going up and down significantly with load changes as well.

A rock steady dc, or one that only changes fairly slowly goes a very long way to fixing light flicker.
Its not always the fault of the inverter.

Shimmy said:

Agree; there are likely 4 intertwined issues: DC bus stability, L-L/L-N/N-G voltage stability, harmonics/waveform, and AC transient response, roughly in order. Without DC bus stability you have to work really hard to make any of the other things stable. Transient response requires everything else to be good. All of these issues are also dependent on PV, battery state of charge, and inverter load.

Click to expand...

Many inverters have an internal dc bus that runs at a different voltage than the battery. Unfortunately, we have no insight into what is happening on that bus but I suspect that is the best place for filters to clean up transients.

Anyone have a waveform for the Growatt 12k? I imagine it is similar to the 6k...

Skypower said:

Depending on the load on neutral (against L1&2) AND the distance from the N/G bond connection(main breaker box?) the greater the line loss differential can be seen to ground( that hopefully doesn’t have any line loss load). So that 7 Volt you see could be normal for your situation since now you removed a secondary bond point.

Click to expand...

There is some impedence due to the distance to the N-G bond. However, the normal voltage seen is less than 3V which would be correct with the distance on 4AWG. With the light flicker the voltage seen is 7V.

This video shows the AC waveform at 7V on Neutral to Ground at the loads panel during the flicker. When the additional light is turned on, the voltage drops to background noise.

solar8484 said:

It's plausible that there are design issues in these Voltronic inverters but yours look like they could have additional manufacturing defects.
Have you tried to exchange for new units? Frankly, even PowerJack inverters at much lower cost put out much cleaner and more stable waveforms than what you have.

Click to expand...

Currently moving thru RMA process. Will certainly be checking the waveforms when the replacements are installed.

Skypower said:

I think too much emphasis is on the waveform and not the overall voltage stability. If the inverter is hunting, stepping up & down, over correcting, under reacting or delayed response only a high end oscilloscope can record the change and see the jump/change in scale. What your eyes are seeing is happening within a few cycles and a common scope is still processing the average. The overall shape is the same just the scale changes. This is a picture of a waveform of a cheap generator with terrible percent of distortion and you wouldn’t want it powering anything you really like other than power tools yet
there’s absolutely no flickering of any light, LED or not.

View attachment 147089

Click to expand...

I had to manually adjust the trigger to hold steady waveforms as the spike at the midpoint switch would make the waveform unstable.

I own some higher priced scopes I use in the shop but don't have a high amps probe for them and second, I was a little leery of using a $14K scope on these inverters knowing there was a voltage fluctuation problem.

As shown in the second video in this post, the voltage peak to peak is fluctuating. https://diysolarforum.com/threads/eg4-6500-ex-under-an-oscilloscope.54931/post-770237

Warpspeed said:

That is not the issue.
It may look perfect with a steady constant load, but if a sudden high power load is applied (or suddenly removed) the inverter voltage regulation may not be quite fast enough. Especially if the incoming dc is going up and down significantly with load changes as well.

A rock steady dc, or one that only changes fairly slowly goes a very long way to fixing light flicker.
Its not always the fault of the inverter.

Click to expand...

I would expect some transient fluctuation when a large load is placed on the inverter, take my well pump kicking in that runs about 50A when it kicks in. One small blink and that is it. Inverter handles the load well, no flicker after the initial blink.

The light flicker will occur with light loads usually in the 600W to 800W range on the system. Here is a video on the inverter voltage fluctuating when the lights had flicker. The video was taken to show Solar Assistant had no interference with the inverter voltage output. You can see the inverter display reflects the varying output voltage and I have graphs from SA showing the fluctuations also.

Zwy said:

Currently moving thru RMA process. Will certainly be checking the waveforms when the replacements are installed.

Click to expand...

A common issue with these rebranded inverters is little or non-existent QA. The customers are the QA department.

Do we have any updates on this very important thread?

v_green57 said:

This is probably about average for a cheap AIO unit.

Bear in mind that there is no "truth in advertising" in regards to solar. You can say pretty much whatever you want as long as it sells stuff.

The only way I know of to create "pure sine wave" AC (< 1% THD, in my opinion) involves using a class AB amplifier or a high-quality Class C amplifier.

Neither of those approaches is very power efficient or cheap, so aren't used for inverters.

"PSW" is visibly better than "MSW" though, when the waveform is examined.

Various reactive components of your load tend to smooth the jaggies seen on that waveform. Do you know if whoever posted it had a "real-world" load attached? Or were they just looking at the unloaded output?

This reminds me of the "audio amplifier big lie of the 60's" where equipment mgfrs were yelling about "50 watts per channel!" for POS amplifiers. They were only capable of that as an instantaneous value at at 50% THD, under which conditions it was pretty much unlistenable.

The FTC finally stepped in and imposed the "% THD at a specified power level into a specified load" to stop the consumer being ripped off.

Click to expand...

Hmmm….I do own a recapped Pioneer SX1250. Love the tanks.

v_green57 said:

This is probably about average for a cheap AIO unit.

Bear in mind that there is no "truth in advertising" in regards to solar. You can say pretty much whatever you want as long as it sells stuff.

The only way I know of to create "pure sine wave" AC (< 1% THD, in my opinion) involves using a class AB amplifier or a high-quality Class C amplifier.
Neither of those approaches is very power efficient or cheap, so aren't used for inverters.

Click to expand...

There is another way to generate very low THD waveforms with high power efficiency, but its not cheap and not really cost competitive commercially.

Several individual square wave inverters driving individual transformers can have different voltage secondaries all connected in series, and is then able to generate any type of output waveform. Its a kind of high power direct digital to analog conversion method, at the multi kilowatt level (or any power level).
It produces a stepped output waveform, but the steps can be made very small.
Each inverter produces a square wave output with three possible logic levels, a negative voltage output, zero voltage, or a positive voltage output.
One inverter has 3 possible voltage steps. Two inverters 9 steps, three inverters 27 steps, and four inverters 81 steps.
Four inverters combined, can produce an unfiltered output of a measured 0.85% THD.
Its very power efficient because the square wave inverters are hard switched at low frequency. The largest inverter switches at only 50/60Hz.
The smaller inverters are slightly more frantic, but still only switch at relatively low audio rate.

My own "Warpverter" works on this principle. Its not commercially viable because of the cost of winding four transformers, but if you wind your own, its an excellent way of making a very high powered robust inverter with excellent performance for very reasonable cost.
The output waveforms across each secondary winding looks like this:

The inverters are hard switched, the difference in voltage is produced by the primary/secondary ratios of the transformers.
The waveforms below are shown with the correct relative amplitudes on the crappy difficult to photograph analog oscilloscope I had at the time.
Largest inverter +/-225v, second inverter +/-75v, third inverter +/-25v, fourth inverter +/-8.3v.
Peak amplitude 225 + 75 + 25 + 8.3 = 333.3v
RMS value 235.6v

When the voltages are combined by connecting the transformer secondaries in series, the output looks like this :
Total combined output easily <1% THD.
Highest harmonic is the third, and that is down by -40db on my spectrum analyser.
This is not theoretical, its actually measured from the inverter I am, running on right now, and have been for almost six years.

Warpspeed said:

There is another way to generate very low THD waveforms with high power efficiency, but its not cheap and not really cost competitive commercially.

Several separate square wave inverters driving individual transformers can have the different voltage secondaries all connected in series, and is able to generate any output waveform. Its a kind of high power direct digital to analog conversion at the multi kilowatt level (or any power level).
It produces a stepped output waveform, but the steps can be made very small.
Each inverter produces a square wave output with three possible logic levels, a negative voltage, zero voltage, or a positive voltage.
One inverter has 3 steps. Two inverters 9 steps, three inverters 27 steps, and four inverters 81 steps.
Four inverters can produce an unfiltered output of a measured 0.85% THD.
Its very power efficient because the square wave inverters are had switched at low frequency. The largest inverter switches at only 50/60Hz.

My own "Warpverter" works on this principle. Its not commercially viable because of the cost of winding four transformers, but if you wind your own, its an excellent way of making a very high powered inverter with excellent performance.
The output waveforms across each secondary winding looks like this:

The inverters are hard switched, the difference in voltage is produced by the primary/secondary ratios of the transformers.
The waveforms below are shown with the correct relative amplitudes on a crappy difficult to photograph analog oscilloscope.
Largest inverter +/-225v, second inverter +/-75v, third inverter +/-25v, fourth inverter +/-8.3v.
Peak amplitude 225 + 75 + 25 + 8.3 = 333.3v
RMS value 235.6v

Click to expand...

Is that the same as how Trace did it? I believe that their inverters actually contained three separate ones.

And if so, is this still how Schneider and Outback (the low frequency models) still work or did they change that?

Yes, Xantrex and Trace make inverters of this type, its nothing really new.
Theirs have three inverters, which is entirely adequate with probably around 3% THD.
I went a bit further with four inverters, but the principle is exactly the same.
Its not a method as widely known about as it should be.
These are all excellent inverters, but just not cost competitive with cheap Chinese.

Warpspeed said:

Yes, Xantrex and Trace make inverters of this type, its nothing really new.
Theirs have three inverters, which is entirely adequate with probably around 3% THD.
I went a bit further with four inverters, but the principle is exactly the same.
Its not a method as widely known about as it should be.
These are all excellent inverters, but just not cost competitive with cheap Chinese.

Click to expand...

I suspect Trace might have had a patent on it which was sold to Xantrex and by the time it was sold to Schneider, the patent had expired which allowed the OG Trace folks to start Outback with the same tech. That’s conjecture though on my part.

Edit: I think that the OG Magnum folks might have come from Trace too.

And of course MidNite though they are only currently producing their new high frequency model.

Warpspeed said:

Its very power efficient because the square wave inverters are had switched at low frequency. The largest inverter switches at only 50/60Hz.
The smaller inverters are slightly more frantic, but still only switch at relatively low audio rate.

Click to expand...

Yep, true old school LF inverters.

I believe all those companies have some kind of relationship as ownership changed hands.
I built my first inverter of this type quite independently over forty years ago, about the same time Trace built their first version, although I knew nothing about Trace at the time.
I quickly realised it was just not commercially viable, but over the years I have slowly incorporated new ideas of my own, and the Warpverter is a totally different and unique design in several ways.

For instance, my inverter does not use any form of voltage feedback.
It uses voltage (and current) feed forward to regulate the output voltage.
Its much faster to respond and unconditionally stable, much more than than voltage feedback.
As far as I know my Warpverter design is the only inverter to use feed forward instead of feedback.
That is another very powerful technique that still seems to be unknown in the inverter design world for some reason.

Warpspeed said:

I believe all those companies have some kind of relationship as ownership changed hands.
I built my first inverter of this type quite independently over forty years ago, about the same time Trace built their first version.
I quickly realised it was just not commercially viable, but over the years I have slowly incorporated new ideas of my own, and the Warpverter is totally different in many ways.

For instance my inverter does not use any form of voltage feedback. It uses voltage (and current) feed forward to regulate the output voltage.
Its much faster and unconditionally stable than any voltage feedback. As far as I know my Warpverter design is the only inverter to use feed forward instead of feedback. That is another very powerful technique that is still relatively unknown in the inverter world.

Click to expand...

You should get a patent on it. Might not be viable for home use, but might be for commercial applications. I.e. power plant internal backups.

I was basing my conjecture on the timeline and comparing to how Bose had/had utility patents on various subwoofer designs.

Its all free and open source.

If you are interested, go over to The Back Shed Forum. Many Warpverters over there, and a couple more in the final stages of assembly right now.

Click to expand...

There are now about twenty Warpverters that have been built successfully and running around the world that I know of. All are in the 5Kw to 8Kw class, but could very easily be scaled up to ANY power level. There are now at least three different sources of driver boards to generate the necessary waveforms, and by agreement are all plug and play with identical pin outs.

So its well and truly out loose in the public domain, no patent would be possible. But I prefer it that way. My contribution to solar.

Below about 4Kw, a PWM driven low frequency transformer inverter is definitely cheaper to make and makes a very robust inverter, but that topology does not scale up very easily to much higher power levels, especially for the inexperienced inverter builder.

The Warpverter, is also extremely robust, completely bi directional, and because it switches so slowly is very non critical of layout and component choice.
Very high power levels could be achieved by the home constructor with excellent potential reliability.

I am not trying to sell anyone anything.
Just letting people know that its all free open source technology and very accessible to any aspiring home inverter builders.

Arguably, the biggest advantage of a home built inverter is that it is quickly and easily repairable.
You know EXACTLY what every part does, and where to get a replacement.

If I had a lot more time I'd build one.

One of the joys of retirement is having plenty of time to research, plan and carry out projects such as this at minimal cost.

Dumpster diving and sourcing suitable scrap toroidal transformers to rewind can take some time, and the rewinding process is physically demanding (these transformers are large and heavy) but the whole thing need not end up being horribly expensive.
I suppose it depends on how you value your time...