2 hours ago, ClauBebe said:

this is the graphic on dark night.

Yeah, I've been thinking that a current threshold of 0.05 A is too low. I think I'll try 0.20 A on the next patch.

2 hours ago, Coulomb said:

Yeah, I've been thinking that a current threshold of 0.05 A is too low. I think I'll try 0.20 A on the next patch.

Very low 🔅, rain condition 

Condition verry low sun, and rain condition 

 

Pic 1 : Voltage MPPT 1/2

Pic 2 : Curent MPPT 1/2

PIC 3 : Power MPPT 1/2

Pic 4 : Solax Inverter with same PV         Power but a little more badly oriented about 5 pieces

L.E. : I confirm the MPPT don't stuck for now, I don't have to reset manual and the production is "wow" for this condition when the MPPT are stuck to 90V !!! The my manually reset will last 20 second and then the MPPT are stuck again to 90V !!!

This is a werry improvement in the low sun, the only thing left would be the "production per night" !

 

TX for the hard work, Coulomb !

 

 

 

Edited December 24, 20223 yr by ClauBebe

6 hours ago, Coulomb said:

His is an Infini V IV, which looks the same as an Axpert MKS IV, and very similar to an Axpert VM IV. The Infinis have different settings, and even equivalent settings like battery type are merely in a different order. Just use setting 05 (battery type) on an Axpert; the last setting option should be protocol LIC, just after SOL and LIb. I believe that all OG10s should have this option.

Yes mine does and I have played around with pin outs on the LIC setting and still no joy. 
i will try and get hold of the CAN to RS485 converter from FreedomWon and try that out…

Heh. In these very low light conditions, the every 5 minute resets to 200 V are actually costing PV power. I think the resets need to be conditional on a certain minimum panel current.

17 hours ago, Coulomb said:

Here is patched firmware version 556.05 for the Axpert Max II 10kW with the older '2809 DSP processor.

It implements:

* Fixes the premature float bug.

* Displays as version 556.05; I think I messed up the version number for what was supposed to be 456.05.

* Implements Georg's current threshold patch, 0.50 A -> 0.05 A.

* No longer stops setting the direction flag when the current is low; I decided it would make no difference.

* Every 5 minutes now (was 1 minute), it checks each MPPT voltage. If below 175.0 V, the MPPT (1 and 2) is "reset" by changing the direction flag to 1 (was to 0 in 456.05) and sets the MPPT voltage threshold to 200.0 V. Setting the flag to 1 should hopefuly fix the "glitchiness" of the PV voltage.

* MPPT1 and MPPT2 are handled in separate patches now, in case they get out of sync, which could have been another source of glitchiness.

* Fixed the bug where I was testing the solar voltage reference against 175.0V, instead of the latest actual voltage measurement.

Use at your own risk.

OG-10 DSP Reflash_556.05 patched.zip 1.61 MB · 3 downloads

 

6 hours ago, Coulomb said:

Yeah, I've been thinking that a current threshold of 0.05 A is too low. I think I'll try 0.20 A on the next patch.

The Inverter 3 has the latest firmware (VERFW:00556.05. At the moment I am running 4 different variants of the firmware and have not lost track yet...
There are no major changes to the previous variant. The two MPPTs regulate in the opposite direction (both brown). In the current weather conditions, no regulation would be necessary.
What is very strange is that when the voltage is increased, the reading for PV1 drops briefly and there is a short power peak for PV2. So there must be a fault here!
The original firmware (red) does not regulate and only provides 30 to 50 % of the power.
I'm beginning to think it's a hardware fault in inverter 3! I must test next!
Inverter 5 has the previous software version. Yesterday I connected another field to I5. And lo and behold, there are no fluctuations in the upper range here - both MPPTs regulate perfectly!

As with Georg's firmware, starting at low current leads to an early yield. So don't change it!

The fluctuations in the lower range could possibly be reduced by not setting a fixed voltage, but only adding the offset (if V < Vmin then V := V + Voffset). I hope this is enough to get out of the "clamping zone".

52 minutes ago, Coulomb said:

Heh. In these very low light conditions, the every 5 minute resets to 200 V are actually costing PV power. I think the resets need to be conditional on a certain minimum panel current.

bearing in mind that with the original firmware it was stuck at 90V in bad lighting conditions and had to be manually reset or exceed the current from the panels by a minimum of 0.5 - 1 A per string to start producing again and jump from 90V to ~320V, we were losing a lot of production per day sometimes on cloudy days and up to 60%-70% compared to SOLAX which I monitor and which has the panels a little worse arranged, 5 are on the west and the rest on the south and for me they are all oriented to the south and without obstacles for now I say it is a huge improvement! thank you !

 

Inverter 5 QVFW (VERFW:00465.05 is perfect!

Important note: Yes, the inverter 3 has problems, the fluctuations in the upper range (from 90 V) are not caused by the firmware changes!

After switching to another inverter, the fluctuations also occurred. They were also present earlier, when the original firmware was still installed. Since the identically constructed twin fields do not show this effect, I do not yet have an explanation.

Edited December 25, 20223 yr by Volker

The PV fields of I3 are now connected to I2 and I2 had the latest firmware. We were waiting for the next sunrise.
Now it's time to celebrate Christmas with the family.

Merry Christmas to all!

Hi everyone and Merry Christmas I am back with graphs related to QFW 556.05 for the last 2 days, yesterday being dark, a rainy day and today being predominantly sunny.  I saw that I no longer have voltage during the night at MPPTs.

I also put a graph with the comparison between the original firmware 56.02 and the custom firmware 556.05

I continue monitoring custom firmware . Thank you all for the hard work done to obtain the best possible firmware.

There still seems be be a bit of a 90 V shelf at the start and end of the day. I can't tell what the current is when it jumps from 90 V to ~350 V. Perhaps we need a shady day before we decide that 556.05 is good enough.

If you guys have a feel for whether 1 min or 5 min or other value would be best for the time between resets, please say so. Too frequently and those spikes start adding up; too infrequently and you might get stuck at low voltage too long in certain conditions.

Thanks for the feedback.

here are the charts from last night and this morning and and the comparison between Solax and Aexpert

I don't have that much time today. Therefore, I have not yet been able to examine the fluctuations (now it`s I2, so it is not a hardware error and does not depend on the firmware version).

The following is a diagram of 4 PV arrays with 10 modules each. Since the light was very dim, they are comparable.

Where do the gaps in the morning and evening come from?

Why are the curves not more even?
(you can see the original firmware getting stuck at 90V in the voltage curve)

here are the graphs in the early hours of the morning and in the evening

 

I apologize for so many graphics, but I want Master Coulomb to see in as much detail as possible what is happening with the MPPTs 

today was a sunny day, tomorrow is expected to be darker and in the following days it seems that it will also rain and then we will see better how it behave

this morning it seems that on MPPT1 it was 0.500 A to increase from 90V to 357V and on MPPT2 it was 0.200 A to jump from 90V to 347V, both identical strings

in the evening it dropped from ~240V to 90V to a current of 0.100A on MPPT1 and from ~240V to 90V to a current of 0.100A on MPPT2. So it seems that it remains stuck at 90V when it no longer has current, so 0 power

 

Edited December 25, 20223 yr by ClauBebe

16 minutes ago, Volker said:

Where do the gaps in the morning and evening come from?

I assume it's due to the lack of amperage, at least that's what the graphs show to me, please look at them

and this is how the production looks to me today, a mostly sunny day

 

 

Edited December 25, 20223 yr by ClauBebe

12 hours ago, ClauBebe said:

12 hours ago, Volker said:

Where do the gaps in the morning and evening come from?

I assume it's due to the lack of amperage,

I thought maybe it's the fact that when the sun is really weak, the open circuit voltage is not much, if any, over 200 V, so that power at 200 V is actually less than at 90 V. So the logic is, going up in voltage didn't work, let's go down instead. But it should go down from 200 V, which means it should surely find a peak above 90 V, I would think. Certainly that very sharp increase in voltage and hence power can't be optimal.

Maybe the problem is that at that time of the morning, the current is increasing so fast, that even though the firmware takes the voltage down by say 5 V, the increase in current more than makes up for the loss of voltage. So it rewards this voltage decreasing behaviour by keeping on going till it gets to 90 V again. That's a weakness of the hill climbing algorithm: even though power went up by decreasing voltage, it don't realise that the power would have increased much more by increasing the voltage.

The same thing could happen at the end of the day when the current is rocketing down. Even though increasing the voltage loses power, it doesn't realise that increasing the voltage would have lost less power.

If this rate of current change theory is true, it's hard to think of a simple work around.

 

On 2022/12/23 at 6:27 PM, Coulomb said:

Here is patched firmware version 556.05 for the Axpert Max II 10kW with the older '2809 DSP processor.

It implements:

* Fixes the premature float bug.

* Displays as version 556.05; I think I messed up the version number for what was supposed to be 456.05.

* Implements Georg's current threshold patch, 0.50 A -> 0.05 A.

* No longer stops setting the direction flag when the current is low; I decided it would make no difference.

* Every 5 minutes now (was 1 minute), it checks each MPPT voltage. If below 175.0 V, the MPPT (1 and 2) is "reset" by changing the direction flag to 1 (was to 0 in 456.05) and sets the MPPT voltage threshold to 200.0 V. Setting the flag to 1 should hopefuly fix the "glitchiness" of the PV voltage.

* MPPT1 and MPPT2 are handled in separate patches now, in case they get out of sync, which could have been another source of glitchiness.

* Fixed the bug where I was testing the solar voltage reference against 175.0V, instead of the latest actual voltage measurement.

Use at your own risk.

OG-10 DSP Reflash_556.05 patched.zip 1.61 MB · 3 downloads

 

Version 556.05 always makes it last and gives up first at the end of the day. The previous version 456.05 works better!

 

Edited December 26, 20223 yr by Volker

On 2022/12/23 at 6:27 PM, Coulomb said:

Here is patched firmware version 556.05 for the Axpert Max II 10kW with the older '2809 DSP processor.

It implements:

* Fixes the premature float bug.

* Displays as version 556.05; I think I messed up the version number for what was supposed to be 456.05.

* Implements Georg's current threshold patch, 0.50 A -> 0.05 A.

* No longer stops setting the direction flag when the current is low; I decided it would make no difference.

* Every 5 minutes now (was 1 minute), it checks each MPPT voltage. If below 175.0 V, the MPPT (1 and 2) is "reset" by changing the direction flag to 1 (was to 0 in 456.05) and sets the MPPT voltage threshold to 200.0 V. Setting the flag to 1 should hopefuly fix the "glitchiness" of the PV voltage.

* MPPT1 and MPPT2 are handled in separate patches now, in case they get out of sync, which could have been another source of glitchiness.

* Fixed the bug where I was testing the solar voltage reference against 175.0V, instead of the latest actual voltage measurement.

Use at your own risk.

OG-10 DSP Reflash_556.05 patched.zip 1.61 MB · 3 downloads

„Georg's current threshold patch, 0.50 A -> 0.05 A."

OK: Produces a yield even in poor light conditions.

 

„5 minutes“

Could be enough, not sure yet. At the moment the light conditions are changing slowly, maybe a minute is better....

 

„Fixed the bug where I was testing the solar voltage reference against 175.0V, instead of the latest actual voltage measurement.“

Unfortunately, this makes it worse again.

I recommend: if Va < Vmin then Va := Va + Voffset (Voffset = 25 V)

I hope this is enough to get out of the "clamping zone".

Edited December 26, 20223 yr by Volker

23 minutes ago, Volker said:

The previous version 456.05 works better!

Huh. I'll have a careful think about this.

8 hours ago, Coulomb said:

I thought maybe it's the fact that when the sun is really weak, the open circuit voltage is not much, if any, over 200 V, so that power at 200 V is actually less than at 90 V. So the logic is, going up in voltage didn't work, let's go down instead. But it should go down from 200 V, which means it should surely find a peak above 90 V, I would think. Certainly that very sharp increase in voltage and hence power can't be optimal.

Maybe the problem is that at that time of the morning, the current is increasing so fast, that even though the firmware takes the voltage down by say 5 V, the increase in current more than makes up for the loss of voltage. So it rewards this voltage decreasing behaviour by keeping on going till it gets to 90 V again. That's a weakness of the hill climbing algorithm: even though power went up by decreasing voltage, it don't realise that the power would have increased much more by increasing the voltage.

The same thing could happen at the end of the day when the current is rocketing down. Even though increasing the voltage loses power, it doesn't realise that increasing the voltage would have lost less power.

If this rate of current change theory is true, it's hard to think of a simple work around.

The following diagram shows the voltage and reading (FW version 456.05). The voltage rises to 150V in the morning (with interruptions). Power and current are not yet present (only 3 very very small spikes). The MPPT controller is therefore still asleep.

Then the MPPT seems to wake up and fully expands the PWM ratio (too high current!). The voltage collapses to 0 V (short circuit, therefore no power).

When the PV modules supply enough current (approx. 0.15 A), the first usable power is generated. Unfortunately, the voltage is then held at 90 V. Normally, the voltage should rise quickly (see the characteristic curves of the modules). Later, setting the voltage to 200 V achieves the release of the clamping.

Normally, the voltage should rise continuously (!) and quickly, as it does at the beginning. But we don't have an analogue circuit, only a switch and a duty cycle!

Unfortunately, I have too little experience with inverters. I should look for curves from other manufacturers...

Edited December 26, 20223 yr by Volker

3 hours ago, Volker said:

Version 556.05 always makes it last and gives up first at the end of the day. The previous version 456.05 works better!

I will also upload the previous version 456.05 tonight and I will come back tomorrow with the observations today it was cloudy in the morning and it seems that it stayed longer at 90v

 

For now it seems that it still has to pass 0.6A to be able to jump from 90v to ~ 320v

in the evening it seems to be a little better, it freezes at 90V when the amperage drops to 0.200A

 

3 hours ago, Volker said:

I recommend: if Va < Vmin then Va := Va + Voffset (Voffset = 25 V)

I hope this is enough to get out of the "clamping zone".

maybe it's not a bad idea if it could be implemented

hello everyone, I managed to write the QFW 456.05 firmware and I have back with the comparative graphs between QFW 556.05 and QFW 456.05 in the morning . QFW 456.05 seems to open faster in the morning...

the first 2 days. (26/27) are approximately the same cloudy lighting conditions and the third (28)day being sunny. 

in the evening it seems to be about the same thing

 

maybe the time should be changed to one minute in QFW556.05 to see how it behaves

On 2022/12/26 at 3:37 PM, Volker said:

Version 556.05 always makes it last and gives up first at the end of the day. The previous version 456.05 works better!

it seems Volker is right
 

13 hours ago, ClauBebe said:

maybe the time should be changed to one minute in QFW556.05 to see how it behaves

Good idea, and easy to implement. It eliminates one change.

This is patched firmware version 656.05 (yes, I had to use a different method to make the version number come out > 655.36, the limit for 16 bits). It's identical to 556.05, except that the resets are considered every minute instead of every 5 minutes. In the mornings and afternoons, 5 minutes is a long way up or down the ramp.

Use at your own risk.

OG-10 DSP Reflash_656.05 patched.zip

On 2022/12/29 at 8:39 AM, Coulomb said:

This is patched firmware version 656.05 (yes, I had to use a different method to make the version number come out > 655.36, the limit for 16 bits). It's identical to 556.05, except that the resets are considered every minute instead of every 5 minutes. In the mornings and afternoons, 5 minutes is a long way up or down the ramp.

With QFW 656.05 I caught a cloudy and rainy day with very weak sunlight
Compared to the Solax that I monitor and which works at maximum capacity, which today produced only 0.80kWh all day, my system produced 0.90kWh, something that never happened to me with the original firmware, which remained stuck in 90v for a day like this, and it didn't produce anything only if I keep resetting it manually
I am attaching the comparative graphs between the 3 QFW 456.05, 556.05, 656.05

and here the whole chart for today, cloudy and rainy day with QFW 656.05

 

On 2022/12/26 at 3:37 PM, Volker said:

Version 556.05 always makes it last and gives up first at the end of the day. The previous version 456.05 works better!

 

I'm curious to see Volker's graphs, which can track in real time the behavior of the 2 firmwares QFW456.05 and QFW656.05

Edited December 30, 20223 yr by ClauBebe