PIP-4048MS and PIP-5048MS inverters

[coulomb]

........... And yes, it WORKED !!!! Now I've sniffed all communication so I can try to write a better flasher software.

@Martino.Q, well done!


I just noticed: unless your inverter is different to mine, or your local wiring colour code is different, you may have active and neutral swapped for both AC-in and AC-out. Please check these; it could save many hassles getting it right.

Also: only one earth wire? That seems strange to me, and I can't see how it can be safe.

[teba]

[I originally posted this questions in the wrong thread, sorry!]

Hello!

Is there a command (in the patched firmware) that switches the inverter back to battery before the "back to battery voltage" [variabel 13] is reached?
And if yes, is there a command to switch to grid before "back to AC" voltage [12] or "battery cut off" voltage [29] is reached?

Thanks,
Lars

[weber]

Is there a command (in the patched firmware) that switches the inverter back to battery before the "back to battery voltage" [variabel 13] is reached?
And if yes, is there a command to switch to grid before "back to AC" voltage [12] or "battery cut off" voltage [29] is reached?

No. But in both patched and unpatched firmware you should be able to get the same effect by temporarily lowering the back-to-battery setting [13] using the PBDV command, and temporarily raising the back-to-utility setting [12] using the PBCV command, as described here:
https://forums.aeva.asn.au/uploads/293/ ... pgrade.pdf

[teba]

Thanks weber!

Will the values be written to the EEPROM every time it's changed?

[coulomb]

Will the values be written to the EEPROM every time it's changed?

Yes, but the effect on lifetime is minimal, as per this post.

[Martino.Q]

Hi coulomb, I chose the color of the cables. The input and output ground cables are both connected to the inverter iron box.

I have 2 big problems whit arduino:

I use this simple one sketch. https://community.blynk.cc/t/problem-wi ... ed/31864/5
- Found the sam problem whit 'char' , resolved whit : comSerial.write(QPIGS, sizeof(QPIGS));
- Arduino buffer receive just 64byte like in the example before, and the string from inverter is aboute 100byte. I try to use thifferent modify ARDUINO UNO IDE whit 256byte of buffer but don't work.
- The command in the example QPIGS add 2 letters more QPIGS⸮⸮ (poin in the middle and @) and works. QPIRI⸮⸮ not work! Why?

E.c.
Inv request
QPIGS⸮⸮
Buffer:
(000.0 00.0 229.9 50.0 0758 0471 015 410 51.30 04 061 0031 0014

Inv request
QPIRI⸮⸮
Buffer:
(NAKss

Som one have idea what i have to do? or i have to use rapsberry?

Thanks in advance...

[paulvk]

I am using mega1248p and mega644p but not Arduino code using Bascom
What I have is 150 byte buffer running in interrupt mode and waiting for a character 13 to trigger decoding
I found that I had to wait after sending the command for a full response or I only got part of the return
and have to calculate the crc to make certain it was ok.
I was getting part responses just like you have , very familiar!
So its a mater of timing and some times the inverter responds with (NAKss I think because it is busy
so I capture this and resend a second time.

I did post some code for a Arduino mega32p that sat between a PC and inverter
you send it a command eg QPIGS and it calculated the crc then sent the command to the inverter
and relayed the return to the PC calculating the crc and if wrong sent error.

[woetyke]

Hi, can somebody tell me if i can use the automatic transfer switch in the PIP 5048 230VAC if i have 2X 230 line supply.
I mean L1 and L2 without neutral wire. between wires i have 230V, from L1 to earth 115V, and L2 to earth also 115V

if i can't i will need to use an external switcher that disconnects N and L1 while switching.
all our equipment is 230V connected to L1+ L2, like N + L1.

thanks

[coulomb]

Hi, can somebody tell me if i can use the automatic transfer switch in the PIP 5048 230VAC if i have 2X 230 line supply.
I mean L1 and L2 without neutral wire. between wires i have 230V, from L1 to earth 115V, and L2 to earth also 115V

I don't think you can do it safely, because there is no way of centre-tapping the inverter's output when in battery mode. Parts of the load may expect balanced outputs with respect to earth, and I think you'd need a way to ensure that the outputs never float to very high potentials with respect to earth.

There are a few PIP models designed for working with split phase installations, such as the LV5048. I suspect that it contains two 2.5 kW inverters inside, and extra relays. [ Edit: however, these may require a neutral connection, which you say you don't have. ]

[woetyke]

so if i live off grid i also can't use these inverters?

[MarkusRE]

so if i live off grid i also can't use these inverters?

I think your system is working like this, with a phase shift of 180 between L1 and L2 so you can‘t use it.

[coulomb]

so if i live off grid i also can't use these inverters?

As long as all your loads are 230 V and don't have any 115 V loads or any expectation of balanced voltages with respect to earth (called ground where you are?), it may be possible to use an inverter designed for line and neutral in your situation. You would be converting your wiring from split phase to single phase.

You have to be careful of how the neutral gets earthed, especially if there will sometimes be a generator connected. You'd have to carefully consider how any residual current devices (RCDs, you may know them as Ground Fault Circuit Interrupters, GFCIs) will react, and where it or they are positioned with respect to the inverter.

If you're not familiar with these issues, please seek advice from a local electrician.

[woetyke]

thank you guys for all information,

i will do some more research en ask my electrician. futureproof its better to use a inverter with neutral if i can

[tgrauss]

Hello,

I am planning to build a PV array using an hybrid inverter (I don't have a pip-5048ms inverter, but I think that I will change at a later time for 2 of these inverters).
Do you think that it is possible to modify the firmware of the 5048MS to send via rs232 the remaining available PV power.
I want to be able to harvest as much power as possible and divert it to my 3kW electric boiler before (or after, depending of many different variables) charging the battery. This way I won't need as many batteries as most of the energy used is for the boiler anyway.
If it is not possible to modify the firmware that way, what do you use to divert energy to a dump load like the boiler. I have found many systems like this that measure the current, but they are only for systems where we can send energy to the grid, but our inverters can not send energy back to the grid. I was thinking about sensing the current going to the inverter and the energy going out of the inverter and using a triac to pwm power to the heater.
What do you think of this?

[coulomb]

Welcome to the forum.

Do you think that it is possible to modify the firmware of the 5048MS to send via rs232 the remaining available PV power.

No. The only real way to discover how much power is available is to pull as much power as possible, and note what that value is. If you happen to have a full battery and light loads, there is nowhere for that power to go, so it would over-charge the battery.

What others have done is mount a small solar panel next to the others, and use a small computer to find out how much power you can pull from that into a dummy load (or maybe something useful). It might be that you can get away with a tiny panel from a solar garden light, but maybe bigger than that is more useful. Then you scale the output to the output of your array; some calibration will be required.

If it is not possible to modify the firmware that way, what do you use to divert energy to a dump load like the boiler. I have found many systems like this that measure the current, but they are only for systems where we can send energy to the grid, but our inverters can not send energy back to the grid.

I would just use an accurate estimate of the battery SOC. So that's not the value reported by the inverter. Use a BMS that counts coulombs, or use a BMV battery monitor or similar.

... and using a triac to pwm power to the heater.

Unfortunately, the triac will severely distort the current waveform. Things like hair dryers with diodes, and hot air guns seem to really badly affect the PIP's voltage control, so the lights will flicker badly. It could be that with a 3 kW phase controlled load, the inverter may become unstable enough to cause damage.

I believe that there are special inverters designed for running pumps from solar panels; they might produce a variable voltage sine wave (and probably variable frequency, which won't worry your boiler). Those might be able to run from the 48 V battery. They might be able to work directly from the panels, but then there is the issue of isolation; that might be a terrible idea. It's not clear how to control these solar pump controllers; I have not studied them at all. Perhaps a smallish VFD would be better, as they are designed to be controlled (the solar pump ones are just designed to get maximum power from the available solar). If the solar pump type can work directly from your existing panels, it may be that they will just use what is available after you have used some charging your battery and powering loads. In other words, it may be that whatever power you use for battery and loads may look to the pump inverter like less available solar power, so it will automatically do what you want.

[tgrauss]

Hello,

Thank you for the answer.
My current inverter is more of a test inverter than a production one, so I will experiment with it. And if it break, well...
You are right, PWM is not good. Turning on an off on zero crossing is better and should not cause problem. Also, my heating element has three 1kW phases connected together. So I can separate the phases and turn on how many phases I need. So I have more than 1kW available but less than 2kW, just turn on 1 phase. If I have more than 2kW but less than 3kW, just turn on 2 phases... This is also an option available, a little bit more crude but still an option.
I have taken my inspiration from here: mk2powerdiverter : https://mk2pvrouter.co.uk/

I will still see what can be done with current measurements from the grid, the output of the inverter and maybe from the PV and battery. With all these information, I might be able to do something. If I detect that I am pulling more than a certain threshold from the grid, but still not pulling from the battery, then I increase the power that I send to the dump load. Something like that but I will experiment with the algorithm. I will maybe also put a light sensor outside. If I can measure how much light is available and deduct a correlation with the panel production, I might be good.
I will also see what can be done with your suggestion.

I will report back what I will find, either good or bad. If I can find something useful, it might be nice if someone else can benefit
Anyway it is not for now as I don't have the panels yet, not before this autumn at least.

[Mantas]

Hi All,

Please help me out here with an answer. I think I already know the answer but I'm hoping I'm wrong

Is it possible to parallel an Axpert 5K (145V SCC) with an Axpert MK2 5K (450V SCC)?

I found this on the forum:
"** Firmware version 71.71 is for the PIP-5048MG / Axpert MKS II (450 V SCC, PF1.0); from the paralleling constant, it looks like it might be possible to parallel machines with 145 V max SCCs with others with 450 V max SCCs. However, this seems unlikely."
Is the MKS II and the MK2 the same thing?

For what it's worth, I won't have any PV connected to the Axpert 5K, only to the MK2. Or the other way around if that can work. I see the MK2 has a higher Min. Solar Voltage of 120VDC which would mean I need to reconfigure my stings to increase their voltage. Current config only gives me around 100VDC

If it is at all possible, what version FW do I need to install to make it work? Currently, the MK2 is on 71.80 and the Axpert 5K is on 72.70C, which was what was needed to make it work with an older 5KVA/4KW

Thanks in advance!

[coulomb]

Is it possible to parallel an Axpert 5K (145V SCC) with an Axpert MK2 5K (450V SCC)?

It's a tricky question.

I found this on the forum:
"** Firmware version 71.71...

I've updated that today to include some more recent firmware version numbers. Sorry for not keeping that page up to date.

Is the MKS II and the MK2 the same thing?

Where did you see "MK2"? I've not seen it. I would say it's the same thing.

If it is at all possible, what version FW do I need to install to make it work? Currently, the MK2 is on 71.80 and the Axpert 5K is on 72.70C, which was what was needed to make it work with an older 5KVA/4KW

72.70 and its patched versions have the same paralleling constant as PIP-5048MG/Axpert MKS IIs and their patched versions. I used to think that having the same paralleling constant was all that was needed for compatibility, though we had to revise that with our patched firmware (hence pseudo constants 118fz and the like in the parallel compatibility post). But then we found other issues; it seems that the manufacturer has been lax with changing the paralleling constant when the CAN packet command set changes.

Considering specifically the combination 72.70 and 71.80 (for PIP-5048MG / Axpert MKS II, not PIP-5048MK / Axpert King), I would highly not recommend it. 71.80 has five more CAN packets that it decodes and presumably sends; 72.70 will merely ignore them. There is no patched firmware for 71.80 that has the altered parameter ranges of 72.70c, that's why it would have to be factory firmware. Though if you're not using an LFP battery and don't intend to use dynamic charge control, that would not matter, and you could attempt to pair 72.70c with 71.80a.

There is one experiment that should be pretty safe to perform, and would perhaps put an end to speculation. You could temporarily attempt to parallel 72.70c and 71.80 (no firmware updating required), without connecting the outputs. During a few minutes of running like that, see if you get the problem with NE coming up intermittently (see this post). If it does, then it's definitely not going to parallel, and you can put that idea out of your mind. Perhaps power them up such that master and slave swap over if you don't get the problem, to see if it appears when master and slave swap over. Remember that allocation of master and slave is meant to be random, even though usually 99% of the time, they end up in a particular order.

If the above experiment fails to rule this out, and you don't use an LFP battery or dynamic charge control, you might consider flashing the MK2 to 71.80a and paralleling them for real, hoping that the lack of complete CAN command decoding won't be an issue. Again, I don't recommend this. The main missing command in 72.70c is related to clearing a flag that says we are the "reserve" master. That sounds to me like a subtle bug has been fixed.

[Mantas]

Where did you see "MK2"? I've not seen it. I would say it's the same thing.

On the sticker on the side with the serial number. Model Name: RCT-AXMK2-5K-48V. It could just be that RCT slightly altered their part number on their batch?

Though if you're not using an LFP battery and don't intend to use dynamic charge control, that would not matter, and you could attempt to pair 72.70c with 71.80a.

I am getting 2 x Pylontech US3000 units which is LFP but as far as I can see they have their own BMS and I have to put in a r-pi icc unit that has special values specifically for the pylontech batteries. Will this make a difference?

There is one experiment that should be pretty safe to perform, and would perhaps put an end to speculation. You could temporarily attempt to parallel 72.70c and 71.80 (no firmware updating required), without connecting the outputs. During a few minutes of running like that, see if you get the problem with NE coming up intermittently (see this post). If it does, then it's definitely not going to parallel, and you can put that idea out of your mind. Perhaps power them up such that master and slave swap over if you don't get the problem, to see if it appears when master and slave swap over. Remember that allocation of master and slave is meant to be random, even though usually 99% of the time, they end up in a particular order.

I will do this a bit later this afternoon when I can disconnect outputs on my "production" inverter. I'll provide feedback once I'm done.

Thank you once again for your knowledgeable feedback; truly appreciated!

[coulomb]

Where did you see "MK2"? I've not seen it. I would say it's the same thing.

On the sticker on the side with the serial number. Model Name: RCT-AXMK2-5K-48V. It could just be that RCT slightly altered their part number on their batch?

It's not clear to me. Does your model have the round bottom to the display/button area, like this?


If so, it's probably actually an Axpert VM II. This page shows the MKS II and MK2 to be different:
http://www.maitec.co.za/product/rct-axpert-5k-mk2/

Though if you're not using an LFP battery and don't intend to use dynamic charge control, that would not matter, and you could attempt to pair 72.70c with 71.80a.

I am getting 2 x Pylontech US3000 units which is LFP but as far as I can see they have their own BMS and I have to put in a r-pi icc unit that has special values specifically for the pylontech batteries. Will this make a difference?

Yes, I think it will still be easier to control with the LFP flavour of 72.70c than the non-LFP flavour, and you can't parallel the LFP flavour of 72.70c with 71.80/80a. But I don't know how the software in the Pi will react to patched firmware. I guess it will be designed for non-patched firmware, so you should be able to use the non-LFP flavour of 72.70c (LC1_72.70c), and it will have the best chance of paralleling with 71.80/80a.

[Mantas]

Where did you see "MK2"? I've not seen it. I would say it's the same thing.

On the sticker on the side with the serial number. Model Name: RCT-AXMK2-5K-48V. It could just be that RCT slightly altered their part number on their batch?

It's not clear to me. Does your model have the round bottom to the display/button area, like this?


If so, it's probably actually an Axpert VM II. This page shows the MKS II and MK2 to be different:
http://www.maitec.co.za/product/rct-axpert-5k-mk2/

Though if you're not using an LFP battery and don't intend to use dynamic charge control, that would not matter, and you could attempt to pair 72.70c with 71.80a.

I am getting 2 x Pylontech US3000 units which is LFP but as far as I can see they have their own BMS and I have to put in a r-pi icc unit that has special values specifically for the pylontech batteries. Will this make a difference?

Yes, I think it will still be easier to control with the LFP flavour of 72.70c than the non-LFP flavour, and you can't parallel the LFP flavour of 72.70c with 71.80/80a. But I don't know how the software in the Pi will react to patched firmware. I guess it will be designed for non-patched firmware, so you should be able to use the non-LFP flavour of 72.70c (LC1_72.70c), and it will have the best chance of paralleling with 71.80/80a.

I found some specs for it where it actually says:
Product Code: RCT-Axpert 5k MK2
Model: Axpert MKS II 5k

I guess that clears it up.

It doesn't look like the one in the maitec link. Here is a photo of it...

IMG_20200520_121948.jpg (46.49 KiB) Viewed 1949 times

[teba]

Hi coulomb,

I just found out that I can set decimal PBCV (Battery voltage back to utility) values:

Code: Select all

$ ~/src/axpert/axpert.sh PBCV47.3
PBCV47.3	ACK

The manual says the value should be an integer between 44 and 51.
I'm not sure if the value will really be tested for if it is not an integer.

I use your LFP firmware version LF1_72.20e. Is it save to use decimal voltages?

Thanks
Lars

[coulomb]

I just found out that I can set decimal PBCV (Battery voltage back to utility) values:

Huh, well spotted!

The manual says the value should be an integer between 44 and 51.
I'm not sure if the value will really be tested for if it is not an integer.

It's not actually tested for integer-ness, at least in the three firmwares I just checked. The reason for allowing decimal points is probably because some firmwares run on multiple models, e.g. a 48 V or a 24 V machine, and the 24 V machine requires decimals (e.g. 23.5 V). They wisely wanted to keep the protocol the same across these models, so the code allows for values with one decimal place. I think it's just laziness that they don't check for "legal" values; it would not be terribly hard.

I use your LFP firmware version LF1_72.20e. Is it save to use decimal voltages?

The problem is that these parameters, when sent from master to slave and vice versa, are sent as whole numbers. So there could well be confusion, or at least loss of the decimal part of the setting, when using more than one machine in parallel. The master calls the shots, so even if the master and slave ended up with different effective settings, it might not matter, but there could be problems I haven't foreseen. So I don't recommend using non-integer values when you're running paralleled or 3-phase machines, or if you think you're likely to in the future. I note that I started off with one machine, and no intention of running two in parallel, and now I do.

As far as I can tell, when running a single machine, non-integer values should work as expected and be safe, for both factory and patched firmwares. But please let us know your findings.

[ Edit: Added "when running a single machine". ]

[teba]

I actually plan to use some of these in parallel. Possibly also in a three-phase configuration. A second inverter is already lying around and waiting. There is a legal reason I can connect the second inverter not before autumn this year.
So I'll stick to the integer value.

What about the two battery voltages PBCV and PSDV? Is the battery voltage always tested against both voltages? Or is it important that PBCV >(=) PSDV?

I assume that the KettleKomp™ voltage is tested against PBCV and raw battery voltage against PSDV. Am I right?

Thanks
Lars

[coulomb]

What about the two battery voltages PBCV and PSDV? Is the battery voltage always tested against both voltages?

The battery voltage is always compared to the higher value of the back to utility (PBCV) voltage setting, or the battery low voltage (battery low cutoff setting (PSDV) plus 2.0 V (0.5 V for LFP patched firmware)).

Or is it important that PBCV >(=) PSDV?

I would always expect the back to utility (PBCV) voltage to be well above the battery cutoff (PSDV) voltage, but I don't believe that this is enforced anywhere. You might want the gap to be less than 2.0 V (0.5 V for LFP patched firmware), if you want to force the effective back to battery voltage to be a non-integer value.

I assume that the KettleKomp™ voltage is tested against PBCV and raw battery voltage against PSDV. Am I right?

The compensated battery voltage is used for the decision to switch to line (bypass) mode; this partly involves the low battery cutoff voltage setting.

[ Edit: As Weber states below, the low battery warning decision also depends on compensated battery voltage. ]

I can't seem to find the code right now that sets fault code 04 (battery low fault). I believe that it does not use the compensated battery voltage.

[ Edit: "non-integral" → "a non-integer value". ]