Patched firmware based on version 73.00

[Coulomb]

The beta versions of patched firmware 73.00a are here: LFP,  and other chemistries.

It's split into two posts only because of a problem with multiple attachments and the recently updated forum software.

Ironically, the forum software needs to be patched to host the patched firmware 

 

[ebrsa]

@Coulomb my 2 Axperts (Proline branded) have now been running flawlessly for a week with your and @weber's version 73.00a firmware. They seem to run cooler too, I guess because the fans are running continuously although quite slowly. I also installed an 8 inch mains fan in a box at the top between the inverters which blows air into the tops of both inverters.

Is it likely that further modifications to the code may be made before the final version is released or will the Beta just be declared final.

 

 

[Coulomb]

On 18/03/2018 at 4:34 PM, ebrsa said:

@Coulomb my 2 Axperts (Proline branded) have now been running flawlessly for a week with your and @weber's version 73.00a firmware.

Great to hear that.

Quote

They seem to run cooler too, I guess because the fans are running continuously although quite slowly.

The fans also run faster by a factor of 5/3 when charging, solar or AC. I think that the idea is to keep the electronics cooler during the day (most charging is via solar, presumably), so it's still fairly quiet at night, but gets extra cooling when the sun shines.

Quote

Is it likely that further modifications to the code may be made before the final version is released or will the Beta just be declared final.

Yes, there will be no further modifications. We were going to make the present version final a few weeks ago, but various things cropped up. There will eventually be a version that acts exactly the same, but displays the firmware version page as LC1 or LF1 instead of BC1 or BF1.

[ebrsa]

Thank you for the information @Coulomb. Must say I find the new display options most useful. Sure hope you and @weber will continue your mods for as long as you find ways to improve the Axperts' performance. Thanks again for it all thus far.

[Coulomb]

Weber and I have released patched firmware 73.00b, to address the error 90 issue.

http://forums.aeva.asn.au/viewtopic.php?p=67159#p67159

[___]

@Coulomb, I'm not a member of the other forum but I read through that quickly. Some ideas.

First, 30 seconds seems a bit short. Just thinking about the Victron BMV's full detection (tail current drops to below a certain configurable percentage for a configurable time) defaults to at least several minutes, and the general advice is 10 to 15 minutes.

I think the reason for the 2A figure might be for practical reasons. I read somewhere (might have been your forum) that the Voltronic doesn't have a current sensor on the battery side, instead it infers the current from the AC measurements (The Multi does that too). This means battery measurements under about 1A is inaccurate (the same is true on the Multi), so quite likely the lowest whole number it can actually measure is 2 ampere. You guys know the guts of that thing better than I do, so it is possible that I am way off base here... but the number rang a bell for me. On planet blue it is very similar, the lowest sort of accurate number you can set is 2 amp per parallel unit.

[bravla]

Hi,

I have a 5 kva inverter like axpert with 80 A MPPT controller ( no name brand) it had 72.90 version and today it stops running and show error 90. I tried to flash but the reflash tool stops with connect fail see attached .pictures.  I test the communication with monitoring tool and it works. I changed the cables etc Please let me know if you have any ideea why I cannot conect to inverter. 

P.S: the inverter is made for Europe.

 

Thank you

 

[ibiza]

If you have ICC connected to the inverter you must remove that cable.

[bravla]

hi, 

what is ICC cable

[ibiza]

Your question tells me that you don't have ICC which is a monitoring software. Go and check on http://iccsoftware.co.za/ for downloads. Very helpful tool that monitors lot of thinks like: batteries, pv input, ac out and much more.

[Allen Hobson]

Hi There,

TO be clear about patching the firmware, do we need to install  the base first of 73? before installing 73.00b?

Or is 73.00b all inclusive?

(Now to hunt down the cable  ) I am still on 52...

[weber]

73.00b is self-contained. You do not need to load 73.00 first.

However I note that the change we made in going from 73.00a to 73.00b is only intended to allow our patched firmware to run on older models of genuine Voltronic inverters, not clones. Coulomb and I feel it is perfectly legitimate for Voltronic to protect their R&D investment by disabling their firmware on clones, so please don't ask us to defeat this protection. And it's quite possible we couldn't, even if we wanted to.

We feel sorry for people who bought clones without being aware of it. It's not your fault. But unfortunately, we can't help you without effectively helping the clone maker, and we consider that to be unethical. So you need to take this Error 90 problem up with your supplier.

[Coulomb]

12 hours ago, bravla said:

Please let me know if you have any idea why I cannot connect to inverter. 

Error 90 is undocumented, so we're all speculating here. Your reflash problem could be independent of error 90, caused by one of the several niggling reasons why the reflash tool won't connect. For example, maybe COM1 isn't your USB to RS232 adapter, but some other (virtual) serial device. Maybe the USB to serial adapter doesn't work in this application; some seem to work with more PC variants than others.

But assuming that you have not made one of those errors, one theory is that Voltronic Power have done something to trap clone manufacturers that improperly copy their firmware. The version number 72.90 is new to me; it could be that this is a clone manufacturer's minimally changed copy of the official firmware 73.00.

If this is true, the clone manufacturers are doing something illegal and unethical. To fix it, they'll have to write their own firmware from scratch, including the flash boot loader that is presumably bricking their clients' machines. (Again, this is if the cloner trap theory is valid). As an additional penalty, they'll have to either provide some special tool to flash program the new firmware, or every customer will have to send their CPU daughter board or perhaps the whole machine back to the factory (or have a new one sent out). Unfortunately, this greatly inconveniences the customers as well, who have done nothing wrong, merely wanting a good deal (and who doesn't?). I guess that Voltronic hope that this will push more customers in the future away from the no-name clones to the branded products. Of course, it might also backfire, turning customers away from a manufacturer who would cause them so much pain. I guess we'll see which way it goes, once more assuming that the cloner trap theory is correct.

Another possibility is that error 90 has some other meaning that I don't know of, but it seems relevant that the machine stopped with error 90 before you had flash updating problems.

[bravla]

thank you for yours answer I will try to change the CPU board with other with old firmware to see if error wil disapear.

 

[Allen Hobson]

1 hour ago, weber said:

73.00b is self-contained. You do not need to load 73.00 first.

...genuine Voltronic inverters, not clones....

How does one find out of it is a clone orr true Voltronic? Model Numbers?

[Gabriël]

hi @Coulomb, maybe you can help me here ?

[weber]

21 hours ago, Allen Hobson said:

How does one find out of it is a clone orr true Voltronic? Model Numbers?

I'm sorry Allen, but I don't know a general rule. The number of brands and model numbers that genuine Voltronics are sold under is huge. You can see a partial list at the start of this page. The only clone brand I'm aware of so far is "Must". But there are undoubtedly other brands and model numbers in both categories.

[Coulomb]

The beta version of patched firmware 73.00c is available now:

http://forums.aeva.asn.au/viewtopic.php?p=67551#p67551

It has minor tweaks to improve battery current accuracy (though minor, they took a lot of effort and time, mainly by Weber), and fixes two more bugs that were present in the original factory firmware, regarding settings changes from the LC display and buttons.

Edited May 8, 2018 by Coulomb
Link wasn't live; it's a beta version

[Gabriël]

On 2018/04/01 at 6:36 PM, ibiza said:

If you have ICC connected to the inverter you must remove that cable.

that was the reason for my firmware update showing the exact image of the reflash tool in  bravla's post

 

On 2018/04/01 at 6:15 PM, bravla said:

I tried to flash but the reflash tool stops with connect fail see attached .pictures

disconnecting sorted the update issue, i do not think i had the error messages, my lcd light just went dead - THAT WAS A SCARY MOMENT!

[Coulomb]

7 hours ago, gabriel said:

disconnecting sorted the update issue, i do not think i had the error messages, my lcd light just went dead - THAT WAS A SCARY MOMENT!

So you have performed an update successfully now?

Re the scary moment: you have to remember that during reflashing, the main software is completely disabled; you are talking only to the boot loader. With a successful download, all that usually happens is that the LCD and LEDs merely freeze at whatever they were displaying last; you might not even notice that anything is different if you weren't looking for it. But when a reflash has comms errors or doesn't start properly, it can be in a sort of limbo between normal operation and normal firmware updating, so that neither the Reflash tool nor the LCD seems to respond. That's when you need to keep a clear head and figure out how to rescue the operation.

The other thing to keep in mind is that there is only one serial port that talks to the outside world. So if you have ICC talking away on the USB port while you are trying to update via the RS232 port, the two sets of serial data are going to get badly intermingled; they might even wire-OR and mingle at the bit level, I've never sorted that out.

[___]

7 hours ago, Coulomb said:

you are talking only to the boot loader

They almost all work this way. I worked for a company who made software for fitness watches. The watch has a bootloader and when you update firmware over a USB cable, the main code is not running, just the bootloader. I wrote code for a chinese card payment terminal once. That had a bootloader as well, we even implemented enough HTTP in it to do remote downloads. Turns out the device also has its own bootloader, holding the right keys kicks it into a kind of passthru so that you can communicate directly to the Motorolla GSM modem in it to update THAT firmware! This recipe is so old you almost cannot imagine anyone doing it differently.

Even the Arduino has a bootloader (in fact, that's the entire magic of it!). For a second or so right after power on it checks the condition on its input pins, and if right, it accepts a new program on the serial port and writes it to flash. That means you don't need an expensive hardware programmer, the chip is its own programmer. You can literally make your own arduinos by flashing the free bootloader to new ATMega 328 chips (which is the point of it being open source).

The usual trick is that the main firmware has a checksum and if it doesn't match the bootloader doesn't load it. Instead it just remains in bootloader mode. That means you can restart the firmware upload without bricking it. Alternatively, the other way is to always check if someone wants to upload firmware for half a second right after power on (I believe the dreambox works that way). As long as you don't trash the bootloader, you can almost always recover.

[Gabriël]

1 hour ago, plonkster said:

As long as you don't trash the bootloader, you can almost always recover.

for me, the electro-technically challenged guy, the best sentence in your comment

[JT1]

On 2018/05/08 at 5:12 AM, Coulomb said:

The beta version of patched firmware 73.00c is available now:

http://forums.aeva.asn.au/viewtopic.php?p=67551#p67551

It has minor tweaks to improve battery current accuracy (though minor, they took a lot of effort and time, mainly by Weber), and fixes two more bugs that were present in the original factory firmware, regarding settings changes from the LC display and buttons.

Hi @Coulomb I have installed the 73.00c with 4 inverters in parallel but the text is not displaying well. See attached for picture of master (Should be HS), the Slave (SL text) displays fine. Other settings text also not readable on some settings can upload a few if needed. The inverter automatically sets Parameter 1 back to UTL after a few hours. I then force it to SBU / SOL but it will set it back to UTL again. Any ideas ?

[Chris Hobson]

6 hours ago, JT1 said:

.... the text is not displaying well. See attached for picture of master (Should be HS), the Slave (SL text) displays fine.

@Coulomb and @weber changed the text a bit to make it more readable. Some of the 7 segment  script is difficult to implement (the letter M was one of the ones they had trouble with).

[Coulomb]

14 hours ago, JT1 said:

Hi @Coulomb I have installed the 73.00c with 4 inverters in parallel but the text is not displaying well. See attached for picture of master (Should be HS), the Slave (SL text) displays fine.

That's not a bug, that's a feature!    You''re supposed to look at it like this:   The "MS" stands for MaSter. As far as we can tell, "HS" is their attempt at making two seven-segment characters look like "MS", using their rather poor "font".

It's impossible to please everyone, but the changes are all meant to be improvements. They should all be mentioned in the AussieView™ manual.

Quote

Other settings text also not readable on some settings can upload a few if needed.

If after reading the manual you think that something is unreadable or missing etc, please post specifics. In particular, the keys don't work the same in information display mode.

Quote

The inverter automatically sets Parameter 1 back to UTL after a few hours. I then force it to SBU / SOL but it will set it back to UTL again. Any ideas ?

All I can think of is that perhaps some monitoring software is running in the background (e.g. you might have tried WatchPower and thought that closing the window would stop it from running. You need to use the system tray to exit, as explained in the middle of this post (NOTE: the software in that post is not the latest).

Edited July 5, 2018 by Coulomb