Thoughts and questions about the "High Voltage DC Bus" inside of AIOs

[fmeili1]

I've read in some threads about the error "HV DC BUS voltage too high" (sometimes also a too low fault) and I try to understand why it may happen and what to do to prevent it. This problems seems to target many different brands and models of AIOs.

Luckily I saw this error on my EG4-6500EX only once, right at the beginning of my solar project while in testing phase and it never happened again after the AIOs went into production more than a year ago.

As far as I understand, the following principle schematic shows where the high voltage DC Bus is located inside a non-insulated AIO.


I guess, the reason for such a DC Bus is to "normalize" the different voltages to a value in a "common" range to let the inverter (DC/AC converter) do it's PWM job to generate the pure AC output sine wave. For a 120VAC inverter the DC bus voltage should or must be at least 340VDC (2 x 120 x √2) to cover the full min to max sine wave delta. I think because of losses, it should be at least a bit higher.

I understand why a "HV DC Bus voltage too low" error can occur if both PV voltage (dark) and battery voltage (empty) are too low and the DC/DC boost converters are not able to step up the remaining voltage to the required DC Bus voltage (with the given requested load which the inverter is asking for).

From my understanding a "HV DC Bus voltage too high" error can only occur if the "consumers" of the DC Bus suddenly no longer requesting power while PV still produces power. The PV DC/DC boost converter may not able to react fast enough to the suddenly rising DC Bus voltage (e.g. when the battery suddenly stops charging and the inverters load is switched off e.g. buy switching off the AC-out breakers.

These are just my assumptions when looking at this schema above, but I'm not sure if I'm right here (I'm not an AIO schematic designer).

So I have a couple of questions in my mind:

• Is my assumption correct about the minimum HV DC Bus voltage of 340VDC for an 120VAC inverter (inside an AIO)?
• If yes, all HV DC Bus voltages over different brands and models should be comparable
• What is a typical max. HV DC Bus voltage? At which value is the voltage DC too high that a PWM inverter is not able to generate a clean 120VAC sine wave?

I try to answer my initial questions with my assumptions partly by myself.

To prevent a "HV DC Bus too high" error, the inverter should always be running and should have at least a bit of load to "consume" energy from the HV DC Bus to prevent it from over voltage situations (I hope that in typical setups, the self consumption and losses of the inverter module is enough to not let the HV DC Bus voltage overshoot, even if PV delivers high voltage and the battery is full).

A connected battery with at least a bit of remaining "room" for energy may be helpful to smooth down an overshooting HV DC Bus voltage (but this is just speculation).

I think the risk for such an high voltage problem is the highest if the inverter module have a relatively low load while the PV produces a huge amount of energy to charge the battery but the battery is nearly full. If the batteries energy consumption for charging is dropping very fast (because the max. charging voltage has been reached) there may be too much energy stored in the HV DC Bus (via Caps) and with sudden reduced requested amps and maybe a slow reacting PV DC/DC boost converter, the HV DC Bus voltage may spike over a critical value.

Will the risk of such an over voltage problem rise in summer if the solar system is massive over paneled with PV modules (some are doing this to cover more load in winter)?

But these are just hypothesis.

What I know for sure is how the HV DC Bus voltage in our solar system behaves. Here are the diagrams over the last 7 days".

HV DC Bus voltage:

The SOC diagram:


Battery voltage:


PV power:


Load:


In my case it could be clearly seen that the HV DC Bus voltage reaches it's maximum values if the batteries stop charging because they are full. At this time the PV production is usually at or near "full power" before PV is clipping.

But I have no idea if my max. seen HV DC Bus voltage of about 460VDC is close to an over voltage problem or if these values in a range between 360-460VDC are "normal". I think so because everything is working without problems.

Maybe other can share their DC bus diagrams for other AIO brands and models to be able to compare and better understand which behavior is "normal" and when it may starts to develop a problem.

 

[42OhmsPA]

Past week hasn't been the best for production...
SRNE ASPs in parallel, I've had 0 issues.

 

[timselectric]

Yeah, I've seen the DC bus voltage too high error, many, many times.
With my early (don't do it) learning to AC couple a hybrid system to an off grid system.
And can confirm that it happens when the loads are almost nonexistent, and the PV rises to near VOC.

 

[fmeili1]

Past week hasn't been the best for production...
SRNE ASPs in parallel, I've had 0 issues.
View attachment 301308
View attachment 301309

Your Bus voltage has a different range, interesting!
Is this a 240VAC inverter? I'm asking because your Bus voltage is much higher compared to mine, which are just 120VAC models.

But if my guess about the minimum required DC Bus voltage is the voltage delta between min an max of the sine wave, a 240VAC inverter would need at least 680VDC... so this could not be the reason.

It looks like my assumption that AIOs with the same output voltage may have the same DC Bus voltage range. At least yours are a lot higher compared to mine for an 120VAC inverter.

 

[zanydroid]

Doesn’t a 240VAC inverter still only need 340VDC?

The PWM would force one leg at 0 while the other would be at 340 at max voltage in the AC waveform

 

[42OhmsPA]

Your Bus voltage has a different range, interesting!
Is this a 240VAC inverter? I'm asking because your Bus voltage is much higher compared to mine, which are just 120VAC models.

But if my guess about the minimum required DC Bus voltage is the voltage delta between min an max of the sine wave, a 240VAC inverter would need at least 680VDC... so this could not be the reason.

It looks like my assumption that AIOs with the same output voltage may have the same DC Bus voltage range. At least yours are a lot higher compared to mine for an 120VAC inverter.

Yes, they are 240VAC, essentially two 120V 5kW inverters in the same box.

I have 1 string ~360VOC, 1 string ~280VOC, and one ~400VOC.

 

[zanydroid]

To prevent a "HV DC Bus too high" error, the inverter should always be running and should have at least a bit of load to "consume" energy from the HV DC Bus to prevent it from over voltage situations (I hope that in typical setups, the self consumption and losses of the inverter module is enough to not let the HV DC Bus voltage overshoot, even if PV delivers high voltage and the battery is full).

Maybe the DC voltage is too high for the PWM frequency + output circuitry to have enough of a safety margin to keep it within the acceptable voltage range. IE, there is a minimum pulse width determined by the output transistors used in the PWM, which means there’s a limit to how short the pulse can be to try to compensate for the unexpected high voltage.

In that case, the fail safe response should be to disconnect the output instead of emit an out of spec voltage

(I don’t think there needs to be a buck stage in the MPPT, the inverter PWM does a bucking already)

 

[DIYrich]

The 340vdc is needed to make 240v split phase ac, not just 120v ac.

The high voltage bus is also another reason for high voltage batteries of commercial inverters and some growatts.

 

[fmeili1]

I'm not sure how a 240VAC split phase AIOs H-bridge is designed. If it's possible to create the split output with just one H-bridge (I think it is at least possible with a huge output transformer), the minimum required DC bus voltage would be the double compared to the 120VAC inverter, resulting in at least 340VDC. In case 2 separate H-bridges are used inside a split phase AIO to create the 2x120VAC a minimum bus voltage of 170VDC would be enough to create both.

As others mentioned, the DC bus voltage will be (or can be) higher if higher PV MPPT voltages are supported by an AIO which would make totally sense. As @zanydroid mentioned, it would not make much sense to reduce the PV voltage with a buck converter to a lower DC bus voltage because the PWM of the inverters H-bridge already does this. As far as I know the PV entries are going through a boost converter and this is only able to lift the voltage but not reduce it.

As a result, perhaps the following statements are generally valid ?

- The max. DC bus voltage must be at least as high as the maximum allowed MPPT input voltage.
- The min. DC bus voltage must be at least >170VDC to be able to create 120VAC (or maybe >340VDC to create a split with just one H-bridge)

So far we have two real life DC bus voltage ranges for specific used max. Voc string voltages:

AIO	AIO type	max. string Voc connected	measured HV DC bus voltages
SRNE ASP (@42OhmsPA )	HF, 120V/240V, split	400	510-550
EG4-6500EX (me)	HF, 120VAC, single	305	355-455

Maybe others can share their values to be able to extend these table to get a more clear picture about the relationship between different brands/models, max used Voc and DC bus voltage range.

It would be also interesting if someone who hit a "DC bus voltage too high" fault would provide the bus voltage when this error occurred (if historic data is available, maybe via SA).

 

[timselectric]

It would be also interesting if someone who hit a "DC bus voltage too high" fault would provide the bus voltage when this error occurred (if historic data is available, maybe via SA).

My high DC bus voltage errors at 500v.
I try to keep it below 480v, for a little buffer.

 

[timselectric]

Here's a high DC bus voltage event.

 

[fmeili1]

update

AIO	AIO type	max. string Voc connected	usual measured HV DC bus voltage range	DC Bus high voltage fault limit
SRNE ASP (@42OhmsPA)	HF, 120V/240V, split	400	510-550	?
EG4-6500EX (me)	HF, 120VAC, single	305	355-455	?
Growatt SPF-5000-ES (@timselectric )	HF, 230VAC, single	480 ?	300-480	500

 

[fmeili1]

My high DC bus voltage errors at 500v.
I try to keep it below 480v, for a little buffer.

How exactly do you try do keep the bus voltage below 480V?

The PV strings are usually fix installed and your max. Voc is like you've organized your strings. Also you does not have very much control over your load - or do you add additional load in case the DC bus voltage climbs to critical values?

 

[timselectric]

update

AIO	AIO type	max. string Voc connected	usual measured HV DC bus voltage range	DC Bus high voltage fault limit
SRNE ASP (@42OhmsPA)	HF, 120V/240V, split	400	510-550	?
EG4-6500EX (me)	HF, 120VAC, single	305	355-455	?
Growatt SPF-5000-ES (@timselectric )	HF, 230VAC, single	480 ?	300-480	500

Usual bus voltage 420-480
Max PV VOC 450

 

[fmeili1]

update

AIO	AIO type	max. string Voc connected	usual measured HV DC bus voltage range	DC Bus high voltage fault limit
SRNE ASP (@42OhmsPA)	HF, 120V/240V, split	400	510-550	?
EG4-6500EX (me)	HF, 120VAC, single	305	355-455	?
Growatt SPF-5000-ES (@timselectric )	HF, 230VAC, single	450	420-480	500

 

[timselectric]

How exactly do you try do keep the bus voltage below 480V?

The PV strings are usually fix installed and your max. Voc is like you've organized your strings. Also you does not have very much control over your load - or do you add additional load in case the DC bus voltage climbs to critical values?

I have my garage (hybrid) system, AC coupled to my house (off grid) system.
If the hybrid exports too much, it drives the bus voltage too high.
This only happens when I am testing things and not paying attention.
I'm still working out the kinks.

Disclaimer: don't AC couple to an off-grid AIO. They aren't designed to handle it.

 

[42OhmsPA]

Some new charts, sun was playing hide and seek with the clouds this afternoon.

 

[fmeili1]

Some new charts, sun was playing hide and seek with the clouds this afternoon.

View attachment 301604
View attachment 301605
View attachment 301606

Same here yesterday, we hit 10 times cloud edge effects which up to +5% over STC (about 10 minutes over STC all together over the day)

Bus voltage jumped around a lot, but stays in it's usual limits (like yours).


solar power:


load:

 

[timselectric]

It must be pretty boring, living in the normal world. lol

 

[fmeili1]

It must be pretty boring, living in the normal world. lol

I'm not quite sure how to understand this. My English is still not good enough to quite grasp all the subtle nuances of your humor
Are you referring to the charts or the thread in general? I'm just curious. Give me a hint Btw. My world doesn't feel boring at all and I feel comfortable in it. LOL

 

[timselectric]

I'm not quite sure how to understand this. My English is still not good enough to quite grasp all the subtle nuances of your humor
Are you referring to the charts or the thread in general? I'm just curious. Give me a hint Btw. My world doesn't feel boring at all and I feel comfortable in it. LOL

I was referring to the mostly normal bus voltage stability.
Don't mind me. My humor can be dry at times.
And I had just completed 15 hours on the road.

 

[fmeili1]

I was referring to the mostly normal bus voltage stability.

yes, is a boring bus voltage
But a boring bus voltage is a good bus voltage

Don't mind me. My humor can be dry at times.

I like your humor and mostly I'm able to get it

And I had just completed 15 hours on the road.

sounds like no fun...

 

[wpns]

FWIW, my EG4 18Kpv inverters over the last 4 days:


Where the bus voltages are Vbus1, Vbus2, and VbusP (half-bus voltage, 200-ish volts, might be used for 120V?)

I'm told the bus voltages have been tweaked in the latest firmware, but with no idea why, what problem they were trying to solve, or what the other unintended consequences might be, I'm in no hurry to 'upgrade'.

FWIW:

MariaDB [Monitoring]> select max(EG4$Alpha$Vbus1), max(EG4$Alpha$Vbus2), max(EG4$Alpha$VbusP),max(EG4$Beta$Vbus1), max(EG4$Beta$Vbus2), max(EG4$Beta$VbusP) from EG4;
+----------------------+----------------------+----------------------+---------------------+---------------------+---------------------+
| max(EG4$Alpha$Vbus1) | max(EG4$Alpha$Vbus2) | max(EG4$Alpha$VbusP) | max(EG4$Beta$Vbus1) | max(EG4$Beta$Vbus2) | max(EG4$Beta$VbusP) |
+----------------------+----------------------+----------------------+---------------------+---------------------+---------------------+
|            540.30000 |            351.00000 |            270.50000 |           455.40000 |           350.80000 |           227.70000 |
+----------------------+----------------------+----------------------+---------------------+---------------------+---------------------+
1 row in set (6.200 sec)

So my high bus voltage alarm is at least over 540V, as I don't think I've ever seen one.

 

[fmeili1]

update

AIO	AIO type	max. string Voc connected	usual measured HV DC bus voltage range	DC Bus high voltage fault limit
SRNE ASP (@42OhmsPA)	HF, 120V/240V, split phase, off-grid	400	510-550	?
EG4-6500EX (me)	HF, 120VAC, single phase, off-grid	305	355-455	?
Growatt SPF-5000-ES (@timselectric)	HF, 230VAC, single phase, off-grid	450	420-480	500
EG4-18kPV (@wpns)	HF, 120V/240V, split phase, hybrid	537	227-540	?

 

[wpns]

MariaDB [Monitoring]> select max(EG4$Alpha$Vpv1), max(EG4$Alpha$Vpv2), max(EG4$Alpha$Vpv3),max(EG4$Beta$Vpv1), max(EG4$Beta$Vpv2), max(EG4$Beta$Vpv3) from EG4;
+---------------------+---------------------+---------------------+--------------------+--------------------+--------------------+
| max(EG4$Alpha$Vpv1) | max(EG4$Alpha$Vpv2) | max(EG4$Alpha$Vpv3) | max(EG4$Beta$Vpv1) | max(EG4$Beta$Vpv2) | max(EG4$Beta$Vpv3) |
+---------------------+---------------------+---------------------+--------------------+--------------------+--------------------+
|           536.70000 |           537.20000 |           346.50000 |          360.50000 |          323.60000 |          359.60000 |
+---------------------+---------------------+---------------------+--------------------+--------------------+--------------------+
1 row in set (4.176 sec)