Axpert Firmware 72.70

[Don]

I have taken the plunge and "upgraded" my firmware to version 72.70. As there is no manual available to guide us as to what the new available setting actually mean, I have just left them at the default values for now. At least nothing tripped or blew up when I restarted the inverters. 

[Coulomb]

I only came across the 72.70 update file yesterday. I've started a comparison with firmware revision 72.60. I've come across tiny changes, like the first two fields of the Q1 command response are different, but it doesn't affect later fields.

But the big one is the function that controls charging (SCC and mains charging). It's about twice the size it was in revision 72.60. I'd say that they have realised the essential problem with charging that the patch was addressing, but the new software is so complex that it's hard to know for sure. I'm running 72.70 myself as of late today, so I'll watch it closely in its actual behaviour as well as the theoretical analysis I get from reading the firmware.

The bulk and absorb phases, which were essentially merged in firmware revision 72.60 and earlier, are a little separated now. There is even a timer now for transitioning from CC (bulk) to CV (absorb) phases. I haven't found any significant effect of this distinction so far.

With respect to the new parameters, there was a very brief post about this in the AEVA forum:

"32 param is bulk charging time. Options: Automatic (default) or time (5-900 minutes)

38 param is enabled or disabled to change dry connect purpose. If disabled (default) you can use dry connect to manage generator support or something like that with battery voltage conditions, but if 38 is enabled you can dry connect in another way: for example trigger the grounding box to connect neutral and grounding of AC output together. "

In summary: there is much new code for handling charging, but I'm still deciding whether they have gotten it right yet.

Edited October 11, 2023 by Coulomb

[Don]

Thank you very much for the feedback @Coulomb. I have been running the firmware for about 2 days now also watching it closely. Looking at the charging graphs, it looks to be nice and smooth. 

[ebrsa]

Thanks for trying out the 72 70 firmware @Coulomb and @Don. I look forward to your feedback before I install 72 70. At present I am very happy with @Coulomb ' patched firmware. The 72 70 version is still incapable of equalising my Trojan T105 batteries which require a voltage of 64.8V. However the bulk charge voltage is 59.4V so maybe there will be some benefit. I will just try and exercise some patience until you post more information and appreciate the generosity of all the time spent on the matter.

[SilverNodashi]

On 2/4/2017 at 0:23 PM, Coulomb said:

I only came across the 72.70 update file yesterday. I've started a comparison with firmware revision 72.60. I've come across tiny changes, like the first two fields of the Q1 command response are different, but it doesn't affect later fields.

But the big one is the sInvChgControl() function that controls charging (SCC and mains charging). It's about twice the size it was in revision 72.60. I'd say that they have realised the essential problem with charging that the patch was addressing, but the new software is so complex that it's hard to know for sure. I'm runnng 72.70 myself as of late today, so I'll watch it closely in its actual behaviour as well as the theoretical analysis I get from reading the firmware.

The bulk and absorb phases, which were essentially merged in firmware revision 72.60 and earlier, are a little separated now. There is even a timer now for transitioning from CC (bulk) to CV (absorb) phases. I haven't found any significant effect of this distinction so far.

With respect to the new parameters, there was a very brief post about this in the AEVA forum:

"32 param is bulk charging time. Options: Automatic (default) or time (5-900 minutes)

38 param is enabled or disabled to change dry connect purpose. If disabled (default) you can use dry connect to manage generator support or something like that with battery voltage conditions, but if 38 is enabled you can dry connect in another way: for example trigger the grounding box to connect neutral and grounding of AC output together. "

In summary: there is much new code for handling charging, but I'm still deciding whether they have gotten it right yet.

some nice changes there!

[SilverNodashi]

On 2/5/2017 at 11:11 AM, ebrsa said:

Thanks for trying out the 72 70 firmware @Coulomb and @Don. I look forward to your feedback before I install 72 70. At present I am very happy with @Coulomb ' patched firmware. The 72 70 version is still incapable of equalising my Trojan T105 batteries which require a voltage of 64.8V. However the bulk charge voltage is 59.4V so maybe there will be some benefit. I will just try and exercise some patience until you post more information and appreciate the generosity of all the time spent on the matter.

I doubt if the new firmware will allow the equalization charge of 64.8V, due to internal hardware limits. i.e. the capacitors in this unit are 60V Caps. There were some talks on the Australian forum about replacing them. If you have the know-how and components, it would be worth-while to replace those components. 

[ebrsa]

@SilverNodashi I am hesitant to replace the caps without a circuit diagram. Then there is the issue of guarantee which will most certainly be void if I change the circuitry. It would not solve the problem of equalising T105s either as the inverter seems capable of a maximum voltage of 61V which fall short by 3.8V. Elsewhere on the forum it was said that the caps are 63V. Perhaps the HA02s that I installed today will solve most of the problems. I will just be patient and wait to see what @Coulomb finds out regarding the 72.70 firmware.

[SilverNodashi]

2 hours ago, ebrsa said:

@SilverNodashi I am hesitant to replace the caps without a circuit diagram. Then there is the issue of guarantee which will most certainly be void if I change the circuitry. It would not solve the problem of equalising T105s either as the inverter seems capable of a maximum voltage of 61V which fall short by 3.8V. Elsewhere on the forum it was said that the caps are 63V. Perhaps the HA02s that I installed today will solve most of the problems. I will just be patient and wait to see what @Coulomb finds out regarding the 72.70 firmware.

Well, your warranty is only a year or two, then you can replace the parts  But, as you say, with the current firmware it wouldn't make much difference. Perhaps, in future, if they release another firmware that can allow higher voltages (obviously for appropriate hardware), it might help? Here's the EVA post about replacing the caps and MOSFETs: http://forums.aeva.asn.au/forums/pip4048ms-inverter_topic4332_post56103.html#56103

[SilverNodashi]

@PaulF007 you mentioned some time ago that your Axpert floats at about 50W, without any load. Here's an explanation (and somewhat solution) as to why: http://forums.aeva.asn.au/pip4048ms-inverter_topic4332_post63663.html#63663 Scroll up/down to the post on "13 January 2017 at 8:42pm" by "Weber". 

[PaulF007]

Hey @SilverNodashi since I "upgraded" my reader things were much better and the losses are at a minimal , I had a look at my previous readers source and it seemed (and I do use it loosely as I don't know halve the stuff I see) like the interpretation of the inverter values were not that accurate , hence the apparent losses. It also does help to have gain some experience since I started. The one thing I never checked was how the values compared on the inverter directly but things are running smoothly , for now ,  and I am very happy with the info I get nowadays.

The losses now measure about 4% to 8% of load during night time and that I would attribute to the power factor not being too good on transformers ectr,

I also are looking forward to the current tests they guys are doing with the new firmware..

[Chris Hobson]

54 minutes ago, PaulF007 said:

 I had a look at my previous readers source and it seemed (and I do use it loosely as I don't know halve the stuff I see) like the interpretation of the inverter values were not that accurate.....

Surely the reader can only report the data the inverter sends?

[ebrsa]

@SilverNodashi thanks for the link. I have about 10 months to go on the Axperts' guarantee so I am hoping for an easier solution than hacking the Axperts. To my mind it may just be easier to build a seperate power supply of the eaqualisation voltage and equalise the batteries infrequently. I am going to measure battery voltages later in the week after the HA02s had time to do their magic. Perhaps I should make Voltronics aware of the issue and see if they have a solution.

[Chris Hobson]

1 hour ago, ebrsa said:

@SilverNodashi thanks for the link. I have about 10 months to go on the Axperts' guarantee so I am hoping for an easier solution than hacking the Axperts. To my mind it may just be easier to build a seperate power supply of the eaqualisation voltage and equalise the batteries infrequently. I am going to measure battery voltages later in the week after the HA02s had time to do their magic. Perhaps I should make Voltronics aware of the issue and see if they have a solution.

Hi Ed

Voltronic do have a high voltage version, the Axpert V series (not to be confused with the Infinisolar V series which is an Axpert with grid tie capabilities - I just about confused myself but never mind) which can go to 61V and can have a scheduled equalise. It still however cannot equalise Trojans which are the best vented batteries out there so unfortunately they have missed the mark.

[ebrsa]

Hi Chris,

It seems that the 72.70 firmware posted by Dave Morley also works on existing Axperts. Both Don and Coulomb have posted loading it and I doubt that they have the high voltage version of Axperts. It will be interesting to hear about their experiences.

[Chris Hobson]

HI Ed

Jip the 72.70 firmware is for the MKS series of Axperts. The 72.20 adds two program functions which @Coulomb has described  There has been some confusion with folk saying 75.10 was the latest firmware when in fact is is for an entirely different range and extracts of Axpert manuals have been quoted which are the Axpert V series manuals which has greater functionality than the MKS series which is the range common in South Africa.

[ebrsa]

Hi Chris,

This is all very interesting and the various versions are bound to give rise to much confusion. Perhaps if we all write to Voltronic and ask that they make all firmware available on their website with proper documentation we may just be persuasive. It sure will help to keep the inverter firmware current with all the benefits we hope it may bring. Will also get rid of the present confusion. After all they seem to sell enough off the products to be able to afford and justify a few hours of their webmaster's time. Ubiquity, whose communications equipment we use, have all the latest software on their website and there is even a monitoring program which allows me to flash the units over the air whenever a new version becomes available. Never had a problem with that or bricked a unit  even though the monitoring program is a beta version. But they are a US company, not a Taiwanese one.

[Chris Hobson]

11 minutes ago, ebrsa said:

But they are a US company, not a Taiwanese one

That I did not know. In which case one should contact the US office and sh!t on their heads - Manuals in Chinglish - confusing model and series names - poor web support. They are good inverters (TTT is about to correct me) and brilliantly priced. 

 

Sorry you were obviously referring to Ubiquity rather than Voltronic. (facepalm).

[ebrsa]

Sorry Chris, there seems to be a misunderstanding. Ubiquity is the US company and Voltronic is Taiwanese to the best of my knowledge.

[Chris Hobson]

14 minutes ago, ebrsa said:

Sorry Chris, there seems to be a misunderstanding. Ubiquity is the US company and Voltronic is Taiwanese to the best of my knowledge.

No I got it spent to much time in the sun this morning.

[ebrsa]

I just sent Voltronic an email from their website to ask how I may obtain a copy of the latest firmware. Also suggested that they should make new firmware and documentation available on their website. Perhaps if we all do it something positive may happen. So please everyone, give them a gentle push. They sure need it.

[PaulF007]

22 hours ago, Chris Hobson said:

Surely the reader can only report the data the inverter sends?

Hey Chris

Man it seems like the interpretation of the data was tricky in the beginning. All I know is that since I am running the new reader the "losses" is much lower than in the beginning. I also now tend to look at the BMV watts as a measure and not the inverter watts just for incase . 

[Guest]

Been thinking about this a lot lately.

Having worked with Morningstar and Victron, harvesting data from both makes with data pre-calculated direct from the device to give you all the salient information plus the fact that you get very infrequent frimware updates, that you can program these makes exactly to your batteries needs, for all makes of batteries, equalising feature as standard, what is in the back of my mind is that you also get cheap and dirt cheap MPPT controllers from Taiwan and China.

Comparing what MPPT controllers cost, from the main brands that has been around for years, tried and tested, keeping in mind that when we tend to say you get what you pay for when that need suites us, would you connect your expensive battery bank to one of those cheap MPPT's?

I have, long time ago, used cheap controllers, best I could afford at the time, for the rest where unafordable then. Today not a chance. That battery bank sitting there costed a shiite lot of money. I will provide the controller to make sure it gets what it needs. 

 

Ps FWIW. Because I had data from the like of Victron / Morningstar, I compared that to data from other makes that are quite favoured on some forums. No comparison at all bar the very bare basics like volts and amps.

[Coulomb]

On 2/4/2017 at 9:02 PM, Don said:

 I have been running the firmware for about 2 days now also watching it closely. Looking at the charging graphs, it looks to be nice and smooth. 

I've found that the same old bug (whereby the Axpert can go to float too early) is still there. I also demonstrated it on my own system. There is a difference in that with firmware revision 72.70, it will take ten minutes (was 30 seconds for 72.60 and earlier) to go to float, if the current happens to be below the threshold. Most of the time, the threshold will be the total charge current (setting 02) divided by 5. For example, if this setting is 40 A, the threshold current will be 8 A, so if the charge current is 7 A or less for 10 minutes, it will go to float.

I believe, but have not had the chance to verify, that this will happen if you set a fixed absorb time > 10 minutes as well. Also, this will also happen if the charger is internally in the bulk stage. However, I think you can only tell bulk from absorb from one of the fields of the undocumented Q1 command).

There is also no path from the absorb stage back to the bulk stage. This should happen any time that the battery voltage falls significantly below the CV battery voltage setting (setting 26).

Anyone can verify the bug. Set the output source priority (setting 01) to Uti (Utility powers load as first charge priority, but this also makes utility charging easier). Change setting 11 (maximum utility charging current) to 2 A. Set the maximum total charge current (setting 02) to 20 A more more. Turn off your PV breakers or isolators. It should start utility charging at a pretty constant 2 or 1 A (1.9 A displays as 1 A). After 10 minutes, which would be nothing like a full charge, it will go to float (CHG LED on solid).

To show that it will charge properly under these conditions, change the maximum utility charge current (setting 11) to more than a fifth of the total charge current (setting 02). For example, setting 11 = 30 A, setting 02 = 50 A. Now start a new charge cycle, 

Weber and I are testing a patch to 72.70 to fix this bug. 72.70 does seem to control the battery voltage a little better (there seems to be less overshoot), so it seems worthwhile to produce another patch.

[Don]

Tank you for the information @Coulomb. We are anxiously awaiting your 72.70 patch. Just wondering why Voltronic don't get you to write their firmware for them? Those Chinese obviously seem to be be confused and struggling to get it right. 

[Chris Hobson]

2 hours ago, Coulomb said:

Weber and I are testing a patch to 72.70 to fix this bug. 72.70 does seem to control the battery voltage a little better (there seems to be less overshoot), so it seems worthwhile to produce another patch.

Thank-you for all your hard work.

The overshoot is one of my biggest gripes with the PIP/Axpert.