EG4 6500 EX Under an Oscilloscope

[ckanu]

If people want to run inductive loads, I tell them to buy a transformer based inverter like victron or sigineer. For most people HF works great. Especially if you have an inverter based heat pump or resistive loads.

From what I found out Victron does not have a "transformer based inverters" - or maybe I missed something. They have separate transformers.

 

[Zwy]

Has any one placed the EG4 6500EX inverters under an oscilloscope to examine it's waveform to see how clean the power is? Found a very short video on YT with a guy testing the SunGold Power 6548, ( which is similar to the EG4 6500EX) and the waveform was not as clean as I would've expected.

Yes, I have. I can say the waveforms posted in this thread in later posts are by members who don't quite understand aliasing, bandwidth and data collection.

Here is one glitch.

Here is one where lights flicker, voltage is not constant. The units that have the flicker problem can not hold a steady output voltage.

I have more.......

 

[Zwy]

The PWM L-C filter is a compromise across load range on inverter. Filter response depends on output load. Unloaded and maximum loading are normally worse looking waveforms.

Sharp, small time period glitches are normally caused by dead zone timing gap where the output IGBT switching devices are switching H-bridge polarity. The dead zone is a purposely designed safety gap to avoid any cross conduction between upper IGBT and lower IGBT to prevent a momentary shorting across power supply. Too much dead zone gap causes stronger glitches. Any actual cross conduction is deadly to devices so needs to be avoided across all operating load and temp ranges.

The required dead zone gap depends on given device, load current, and temperature of devices. The way the Chinese constantly substitute parts between manufacturing builds without much or any performance evaluations, it is likely the optimum gap has changed from original design.

I've looked at this scope capture I've made countless times of the AC output on the EG4 6500EX. It's a little long but shows the distortion with full PV power coming in. This was taken under light load, intermittent full sun with clouds coming and going. I noticed the switching on the H bridge and why I had to manually set the trigger high or the scope trace would jump all over due to the spikes at the switch, some are over 10V.

I would appreciate any input you have about this and other scope waveforms I've linked.

 

[Quattrohead]

Those little glitches look like timing errors where the two inverters aren't quite getting their sine waves joined up together 100%.

 

[Zwy]

Those little glitches look like timing errors where the two inverters aren't quite getting their sine waves joined up together 100%.

I have an idea how these work after some research with all the "free time" I had the last few days. When you really zoom on a waveform at 200us you can see the pulsing of the high voltage bus, then you get to where the switch has to occur on the H bridge. It needs the delay as @RV10flyer has stated when it switches. Why the spikes, I can't say.

Like I said, I'm a tech not an engineer and I only need in many cases to know whether the waveform looks correct (dropouts, spikes, frequency, holds to ground when a driver grounds a circuit for examples) and have a good grasp on how the circuit works for troubleshooting.

 

[Quattrohead]

I have a strong feeling they cheaped out on the output filtering components, and I cooberate this by the fact that some people have put capacitors on the output and resolved their issues.

 

[Warpspeed]

I would not worry too much about the actual waveform, the grid can look pretty awful at times too, and it still works just fine.

The problems start with highly non linear and reactive loads, as well as short term massive current surges.
If you have really nasty hostile loads, a solid LF transformer inverter is going to be a lot more robust and more likely to survive for longer.

The trouble is, a decently robust inverter does not come cheap, and its going to be large and heavy, as well as expensive.
You get what you pay for (sometimes).

 

[Zwy]

I would not worry too much about the actual waveform, the grid can look pretty awful at times too, and it still works just fine.

The problems start with highly non linear and reactive loads, as well as short term massive current surges.
If you have really nasty hostile loads, a solid LF transformer inverter is going to be a lot more robust and more likely to survive for longer.

These flicker lights on light loads without any inductive load. You do get dirty grid power, but do you see the voltage jumping from 111V to 127V repeatably? I don't.

The trouble is, a decently robust inverter does not come cheap, and its going to be large and heavy, as well as expensive.
You get what you pay for (sometimes).

I like the 6500EX, it's a good unit is many aspects. But the voltage regulation on the 2 I have doesn't work.

 

[Warpspeed]

Light flicker, as you say, is due to a sluggish to respond voltage regulation.
That is another problem quite separate from all of the other inverter limitations listed above.

Most of it is usually due to very sudden changes of dc voltage at the input of the inverter under step load change, rather than any weakness in the inverter power stage.
A larger battery with a much lower source impedance can help a lot with light flicker.
Failing that, some ultra capacitors across the dc supply right at the inverter can also make a big improvement to the light flicker problem, by slowing down the short term dc voltage fluctuations.
If the dc voltage changes are more gradual, the sluggish inverter feedback can then cope a lot better.

 

[Zwy]

Light flicker, as you say, is due to a sluggish to respond voltage regulation.
That is another problem quite separate from all of the other inverter limitations listed above.

Most of it is usually due to very sudden changes of dc voltage at the input of the inverter under step load change, rather than any weakness in the inverter power stage.
A larger battery with a much lower source impedance can help a lot with light flicker.

My bank is 55Kwh, I think it is big enough. Usually in the evening with 80 to 90% charge I see the light flicker.

Failing that, some ultra capacitors across the dc supply right at the inverter can also make a big improvement to the light flicker problem, by slowing down the short term dc voltage fluctuations.
If the dc voltage changes are more gradual, the sluggish inverter feedback can then cope a lot better.

Explain this, if lights are flickering and you turn on more lights the flicker stops. Turn off the extra lights, flicker returns.

Only loads are my base load of 150W and the lights.

Meanwhile output voltage on the display is ramping up and down between 111V and 128V.

Do you still believe it is a DC issue?

And how would this explain the increase on neutral to ground voltage seen when the lights flicker to 7.0V? I do have a video of this occurring both on a multimeter and scope. Lights not flickering and all you see on the scope is noise in the millivolt range. Where is this AC voltage coming from? It is AC voltage, scope shows it as such.

Removal of the bonding screw in the inverter and firmware update made the flicker worse, that is when I found the 7.0V on the neutral. It wasn't there before. N-G are bonded at the service panel at the meter and I did check the connection. The firmware update changes the inverter to a common neutral and is to rely on the N-G bond at the service panel. But where is the 7.0V from?

 

[Zwy]

I have a strong feeling they cheaped out on the output filtering components, and I cooberate this by the fact that some people have put capacitors on the output and resolved their issues.

I've thought about it. I thought about adding N-G bond at the loads panel to see if it changes anything. Removal of the neutral on the ac input. Experimental only.

 

[Warpspeed]

@ Zwy
I have observed this very same effect myself, it only does it sometimes on very light loads when solar input exactly matches the inverter load.

If something changes, where inverter load is either increased or decreased very slightly, the flickering stops.
Also if solar changes slightly due to variations in cloud density, the flickering stops.
It comes and goes, but only seems to do it around sunrise and sunset, and only for a very short time.

As far as I can see, its due to the voltage characteristics of the battery. If a battery is just sitting with nothing connected to it, there will be a certain measurable voltage.
As soon as you start charging, even at very low current, the voltage jumps significantly upwards.
Likewise if you connect a discharge load, even a very small load, the battery voltage jumps downwards.

If you are running where charging current and load current are equal, the battery may be right at the point of either charge/discharge, and the battery voltage can jump up and down by enough to cause light flicker.

It never happens during the day while charging, or at night when there is only discharge.
I have only seen this effect at dawn and dusk where we are transitioning between day and night.
Sometimes its only one or two flickers, sometimes it can go on flickering for several minutes.

Not sure what to do, but will very soon be greatly increasing my battery capacity from 5Kwh to 28Kwh and hope that fixes it.

If it persists, I might try some ultracapacitors. As it is now its tolerable, as the point at which it occurs I really do not need lights switched on.
But I would still like to understand better the exact cause.

 

[PreppenWolf]

That sounds like a crappy BMS issue than a battery issue?

The flickering at 80-90% charge also sounds like a battery issue where your BMS starts balancing a cell? Maybe throttling power and causing the inverter to have a hard time adapting?

I've often wondered if my bank being so large is the reason I don't have some of these issues. (8) banks of 105kWh using (8) 100A 16s Overkills.

 

[Zwy]

@ Zwy
I have observed this very same effect myself, it only does it sometimes on very light loads when solar input exactly matches the inverter load.

Mine never does it with solar input that I have noticed. Always with no charging taking place.

If something changes, where inverter load is either increased or decreased very slightly, the flickering stops.

Noticed the same.

Also if solar changes slightly due to variations in cloud density, the flickering stops.
It comes and goes, but only seems to do it around sunrise and sunset, and only for a very short time.

Only at night for my units, no PV input.

As far as I can see, its due to the voltage characteristics of the battery. If a battery is just sitting with nothing connected to it, there will be a certain measurable voltage.
As soon as you start charging, even at very low current, the voltage jumps significantly upwards.
Likewise if you connect a discharge load, even a very small load, the battery voltage jumps downwards.

If you are running where charging current and load current are equal, the battery may be right at the point of either charge/discharge, and the battery voltage can jump up and down by enough to cause light flicker.

I've seen it at the flat area of the discharge curve where battery bank voltage doesn't change any.

It never happens during the day while charging, or at night when there is only discharge.
I have only seen this effect at dawn and dusk where we are transitioning between day and night.
Sometimes its only one or two flickers, sometimes it can go on flickering for several minutes.

Not sure what to do, but will very soon be greatly increasing my battery capacity from 5Kwh to 28Kwh and hope that fixes it.

Large bank might fix yours, it certainly won't on mine.

If it persists, I might try some ultracapacitors. As it is now its tolerable, as the point at which it occurs I really do not need lights switched on.

But I would still like to understand better the exact cause.

This, I'm looking for answers but not finding any other than the inverter is not regulating output voltage.

 

[Zwy]

That sounds like a crappy BMS issue than a battery issue?

The flickering at 80-90% charge also sounds like a battery issue where your BMS starts balancing a cell? Maybe throttling power and causing the inverter to have a hard time adapting?

I've often wondered if my bank being so large is the reason I don't have some of these issues. (8) banks of 105kWh using (8) 100A 16s Overkills.

Are you referring to my system? I run a Batrium, none of the current goes thru the BMS.

55Kwh isn't enough?

 

[PreppenWolf]

I don't know what Batrium is, but your batteries aren't on a BMS?!

 

[Warpspeed]

I don't have any answers, but the first step might be to monitor the dc voltage right at the inverter while its in flicker mode.

 

[Shimmy]

Kind of a related question: what frequency do the IGBTs switch at on these units? If you have fast enough switching and good controls a step load/transient response is not a problem with a high frequency inverter. But the difference between 1kHz and 7kHz is huge. Unfortunately higher switching frequency has an efficiency penalty.

 

[Shimmy]

I don't have any answers, but the first step might be to monitor the dc voltage right at the inverter while its in flicker mode.

That is where a scope is really needed. Are you in closed-loop communications?

 

[Zwy]

I don't know what Batrium is, but your batteries aren't on a BMS?!

I have a BMS, it is a Batrium. And none of the current goes thru the BMS.

Search results for query: batrium

diysolarforum.com

 

[Sanwizard]

Has any one placed the EG4 6500EX inverters under an oscilloscope to examine it's waveform to see how clean the power is? Found a very short video on YT with a guy testing the SunGold Power 6548, ( which is similar to the EG4 6500EX) and the waveform was not as clean as I would've expected.

Wow, my Genmax generator puts out a cleaner waveform than that, unless his scope had issues.

 

[Zwy]

Ran across this post in a SP6548 thread. Confirms what I'm seeing. https://diysolarforum.com/threads/s...ng-lights-after-new-install.43573/post-637264

 

[Skypower]

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.

 

[solar8484]

I've looked at this scope capture I've made countless times of the AC output on the EG4 6500EX. It's a little long but shows the distortion with full PV power coming in. This was taken under light load, intermittent full sun with clouds coming and going. I noticed the switching on the H bridge and why I had to manually set the trigger high or the scope trace would jump all over due to the spikes at the switch, some are over 10V.

I would appreciate any input you have about this and other scope waveforms I've linked.

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.

 

[Skypower]

My inverters are MPP LV6548 purchased back in Sept 2021 from Watts247 and the software is original. There is no flickering unless there is very little on and the Keurig coffee maker is brewing because it’s high current element is regulated by quickly pulsing on/off and my inverters are playing chase the load. About 20% of the time, I may see subtle single flicker when the HVAC starts or the clothes dryer starts, but I almost have to be anticipating it to not have it go unnoticed. A rock solid DC supply mandatory; cable sizing, contact or mosfet and battery capacity. I can assume you got that down.
I would expect that the EG4 version should have the same performance if not better than mine, especially if its software has been updated. If it’s anything less I’d be demanding that the seller/distributor make it right.