Axpert 5KVA Max PV Open Circuit Voltage question

[shumbaman]

The Axpert 5KVA model has a stated Maximum Open Circuit Voltage value of 145 Vdc. 

What would happen if the PV array supplied 150 Volts?

The reason I ask the question is I would like to arrange my panels in 3 parallel sets of 4 panels arranged in series thereby delivering a higher voltage with lower current than 4 parallel sets of 3 in series. This would allow me to used lighter gauge cabling.

Also please confirm that in working out the voltage drops on PV system you use the MP as opposed to the OC spec values.

 

[___]

16 minutes ago, shumbaman said:

What would happen if the PV array supplied 150 Volts?

Depends on the design of the MPPT controller. If I recall correctly (I cannot for the life of me find it now), some have the ability to shut their eyes and ears and ignore the high voltage condition, and reconnect when it comes down. These units will safely ignore the occasional high input voltage on cold mornings with bright sun, and work correctly on the other hot days when the voltage isn't as high.

But I doubt the Axpert has one of those.

For the rest, you have to understand something about the two semiconductors that (along with a large old inductor) makes up the heart of an MPPT. Depending on design, you will have either two MOSFETs, or one MOSFET and one Diode. The diode is usually of the Schottky variety, and this is important: It has a lower forward voltage drop, but also a much lower reverse breakdown voltage. The same goes for MOSFETs, they have fairly low reverse breakdown voltages. What this means is that if you put a reverse voltage on them that is too high, they go into conduction mode when they shouldn't, and they do so in a destructive manner.

They blow up in other words.

Now, most designs will have a bit of margin, so 150V MIGHT not blow it up. I'm not sure I would try it though... because on that once in a decade icy cold winters morning with bright sunshine... you might do even more than 150V, and you've already "used" the margin in that case.

[___]

26 minutes ago, shumbaman said:

Also please confirm that in working out the voltage drops on PV system you use the MP as opposed to the OC spec values.

I think that makes perfect sense. At OC values, the current is necessarily zero (that is what OC means), and if the current is zero, the voltage drop is zero.

[shumbaman]

8 minutes ago, plonkster said:

I think that makes perfect sense. At OC values, the current is necessarily zero (that is what OC means), and if the current is zero, the voltage drop is zero.

Thanks plonkster ...I was hoping that the over voltage would simply be ignored and that the unit would simply function as if it were at max voltage. I am not prepared to test it though!

[viper_za]

44 minutes ago, shumbaman said:

Thanks plonkster ...I was hoping that the over voltage would simply be ignored and that the unit would simply function as if it were at max voltage. I am not prepared to test it though!

See this post from the Australian forum that Coulomb made

http://forums.aeva.asn.au/pip4048ms-inverter_topic4332_page35.html

I'm also considering 2S verses 3S. The specifications in the manual are pretty basic:

Max. PV Array Open Circuit Voltage: 145 Vdc
PV Array MPPT Voltage Range: 60~115 Vdc

So what happens above 115 VDC? Does it still PWM, but not track the maximum power point perfectly? If so, how imperfectly?

I decided to try to find out by looking at the binary code. Unfortunately, it's in HCS08 assembler (not that they write it in assembler, of course, but that's what I see). That means I don't get symbols, except for a few from the shared communications, such as the QGS command (SCC to DSP) and (GS command (DSP to SCC). So it's hard going.

SCC = Solar Charge Controller, the part of the PIP inverter that charges the battery from the solar panels. DSP = Digital Signals Processor, the main processor for the PIP inverter.

The code is rather confusing; for example, there is subtraction of two quantities that as far as I can see are never different. So I can't pretend to know all of what is going on. However, the PV voltage (PhotoVoltaic, solar panel input voltage) is one of the shared variables that I am confident is right. So I can see what parts of the code do with the PV voltage.

I see many comparisons with 14500, which is the internal representation for 145.00 V (the SCC deals with hundredths of a volt and amp). When the voltage exceeds this value, I see alarms being set and so on. There are also comparisons with 14000; this is the representation for 140.0 V, where the alarms go off if the voltage stays under this value for a few seconds. All good.

But there is no comparison with 11500, representing the supposed upper end of of the MPPT tracker voltage range. I've done a binary search through the code image to be sure. What I do see are comparisons with 130.0 V. There is a linear ramp from 130 V to 145 V (a range of 15 V), allowing -4 A of maximum charge current per V above 130 V. So at 130 V solar input, the maximum charge current is 60 A; at 135 V, this falls to 40 A (5 V x -4A/V = -20 A), 20 A at 140 V, and zero current at 145 V. There other things that affect the maximum current, for example there is another ramp from 80°C to 90°C. The lower of the two ramps takes effect, so at 85°C (half current) and 140 V (one third current), the larger derating is from voltage, so you get one third current (20 A). At the same temperature and 135 V, the current would be 30 A, since temperature is now more important.

I should point out that all PV currents mentioned are at the battery end, i.e. after the buck converter. The actual PV current will be somewhat lower, due to the "DC transformer" effect of a buck converter.

So what's the 115 V figure about? It's possible that it is buried in there in the form of PWM values sent to the timer registers. But I suspect that 115 V is significant because it's a little under 120 V, the Extra Low Voltage limit. Somehow, they may get to dodge some certification requirement if it's ELV. That's pure speculation on my part.

It seems that operation just above 130 V will be fine, however there will be a reduction of maximum power due to the reduced maximum charge current allowed. But it seems prudent to design any system such that voltages above 130 V will rarely be seen. Again, I point out that this is based solely on references to the PV voltage; there could be other factors that I'm not talking into account.

To me, that makes 3S look more promising. 130 V / 3 = 43.3 V, which is above the MPPT voltage of most 72 cell panels (around 200 W, as mine are). 115 V / 3 = 38.33 V, below the MPPT voltage of at least some of my panels.

With 2S, there is no problem at the upper voltage end, but the problems show up at lower voltages. The MPPT needs a few volts above the battery voltage to operate; I see battery voltage plus 1.0 V and plus 1.5 V in various places. The specifications say 60 V minimum, and there seems no reason to fudge this figure. 60 V for 2S means 30 V per panel; for 3S it means 20 V. Even the poorest usable light would provide 20 V of output per panel, but maybe not 30 V.

So it's still not obvious to me which is best, but at least there are some more clues now. Perhaps those with PIP inverters installed could comment on whether my theory seems to fit their measurements.

[___]

27 minutes ago, viper_za said:

PV Array MPPT Voltage Range: 60~115 Vdc

Probably has to do with the design of the buck converter. Below 60, you obviously have a problem because there isn't enough space between that and the 58.6V (or whatever) that you want to charge the batteries to. It is after all a BUCK converter, it can only step down. On the other end, if you go too high, efficiency suffers. It still works, it might just cause a bit more ripple on the output.

[viper_za]

2 minutes ago, plonkster said:

Probably has to do with the design of the buck converter. Below 60, you obviously have a problem because there isn't enough space between that and the 58.6V (or whatever) that you want to charge the batteries to. It is after all a BUCK converter, it can only step down. On the other end, if you go too high, efficiency suffers. It still works, it might just cause a bit more ripple on the output.

@shumbaman

Yes so rather stick with 3 panels in a string and get the right cables, why take the risk

[KLEVA]

Just to clarify though, the Axpert will accept up to 145V, but only USE 115V max (as per specs). At over 145V the Axpert will SHUT OFF or SHUT OFF PV supply (apparently this is firmware/model specific as to which action it takes). Remember this is the MPPT max voltage it can use, so the current will just be larger than if at 145V, so you will need cables to handle the current at 58-115V anyway (ie: thicker cable than you may have calculated).

Source: Coulomb & Others info on Aus forum & specs.

So, if your string voltage is regularly going to go above 115V, then you are wasting and who knows how the Axpert will react to a constant over voltage. For most 250-300W panels 3 in a string is sufficient, and the Axpert copes pretty well adjusting the MPPT. As Coulombs quoted post above mentions, the info is based on the internal data, and was untested.

[shumbaman]

Thanks for all the feedback. I, like most others on the forum went with Axpert primarily for cost reasons. I firmly believe that Victron offers he best solution but I cannot justify to myself the more than double price.I am in Zim and we pay in US dollars and pay higher prices than you do in SA so being price conscious is essential, especially when you are retired. Many theorists claim that the charging algorithms in the Axperts are not up to scratch and that the transformer-less technology is suspect etc, etc. but after a significant research effort I decided that they offered the most cost effective solution to our problems of 18 hour daily power cuts at the time. Since December though our situation has improved dramatically and like SA we have had little if any load shedding as we are purchasing from Escom at the moment. In fact based on our current regime I dont think that solar (with the exception of water heating) is cost justified as we pay US 10 cents per KWhour. Now I have lived in Africa for over 60 years so I fully understand the concept of what was or what is, is not necessarily the predictor of what will be. In addition as I always remind my SA friends, Africa does not stop at the Limpopo so things could change dramatically in the future so be prepared!

I purchased all that I needed to be effectively "off-grid' back in September last year. Initially I installed what was necessary to give me essential backup for the long outages but I am now looking to complete the installation of 2*5KVA Axperts, 3KW of PV (Suntech) and 12*200 AHours of batteries (Narada Acme). The second Axpert will also have 1200 Watts of Suntech Panels attached (these were left over from a previous project). I am also intending to use Jacos software. My Problem at the moment is that the most ideal location for my panels is 50 meters from the most ideal location for the Inverter/Batteries and I have spent weeks, if not months (I am retired so do have the time!) playing with different alternative solutions.

I want to have fused circuits for each of the 4 strings and combine them into one line somewhere before reaching the MPPT controller. I have major issues with the routing of the cables as I have a flat roofed house and lots of paved areas between the two ideal locations. This all adds up to a cable bill of a minimum of US700 so I have been spending a lot of time investigating ways of reducing this additional cost.

I am now considering a non ideal location for the panels that will reduce the efficiency because of shading. This location will be ideal between the hours of sunrise and 15:00 whilst the sun is northerly but the 15:00 will probably become 14:00 when the sun is further south. Based on this shading regime I am also considering orientating the panels eastward so as to maximize the use of the morning sun and factoring in that that after 14:00/15:00 I wont get much power generation anyway. Having a flat roof house gives me flexibility in the PV installation in that I can make movable frames in terms of the orientation and tilt. I am considering a 30 degree tilt when the sun is in the north and 0 degree tilt when it moves south (our latitude is 18 degrees). Initially I thought of an automatic tracking system as I am into electronics and Arduino as a retirement hobby but all the advice that I get is that its not worth the effort and most of the problems are mechanical.

I would be interested in feedback from those interested in my solar life story" above.

 

Regards to all

[Chris Hobson]

14 minutes ago, shumbaman said:

Initially I thought of an automatic tracking system as I am into electronics and Arduino as a retirement hobby but all the advice that I get is that its not worth the effort and most of the problems are mechanical.

Hi Shumbaman

My wife's family were in Zim and I have spent several happy holidays around Gweru, Mhangura, Marondera and Kariba. A girl I was at varsity with was commodore of Andorra harbour in 2005 the last time we visited Zim.

Being north of the Tropic of Capricorn a tracker may be a good option. Have you thought of an electronic linear actuator? Then your system would be electronic rather than relying on gears etc.

[shumbaman]

6 minutes ago, Chris Hobson said:

Hi Shumbaman

My wife's family were in Zim and I have spent several happy holidays around Gweru, Mhangura, Marondera and Kariba. A girl I was at varsity with was commodore of Andorra harbour in 2005 the last time we visited Zim.

Being north of the Tropic of Capricorn a tracker may be a good option. Have you thought of an electronic linear actuator? Then you system would be electronic rather than relying on gears etc.

Hi Chris...Yes I have thought of this option and looked up pricing the other day. Based on what I have said earlier I just dont know if the expense is worth it. 

Glad you have had good times here. We still have good times  but the last 15 years have tested us and the current cost of living is frightening.

 

Edited April 29, 2016 by shumbaman

[Chris Hobson]

The other option is a remote gate motor and track.

[Chris Hobson]

You can build your own linear actuator. I might even tackle something like this (not that I am looking for extra projects).

 

[shumbaman]

I was also wondering if you could use these affordable stepper motors used in CNC systems.

One thing to also consider though is that the physical size of a 4*3 array swinging more than 30 degrees from the horizontal raises the issue of wind so it would be better to have 4 separate structures of 4 moveable units with each unit having 3 panels. This complicates the design though but not an insurmountable problem.

I might look at this a "phase 3" and go ahead with my non ideal solution of static movable frames and see how things work out.

[Guest]

48 minutes ago, shumbaman said:

US 10 cents per kWh

 

48 minutes ago, shumbaman said:

as we are purchasing from Escom at the moment

Now I want to know from Zuma Inc why am I paying R2.18 incl VAT per kWh, increasing again now by wot, 8%, compared to if I was in Zim I could get same Eskom power at wot, 50%(?) cheaper?

[shumbaman]

When looking at a tracking system you can consider one based on light sensors feeding into a micro controller or simply hard code the coordinates based on time of day and year.

[shumbaman]

5 minutes ago, The Terrible Triplett said:

 

Now I want to know from Zuma Inc why am I paying R2.18 incl VAT per kWh, increasing again now by wot, 8%, compared to if I was in Zim I could get same Eskom power at wot, 50%(?) cheaper?

We are getting an increase soon....not sure how much...but I also think that our rate is subsidized to "placate the electorate"

[Guest]

25 minutes ago, shumbaman said:

"placate the electorate"

Amen! Well said, give that man a Bells!

Ps. I had a very political incorrect sentence typed here when I remembered the post are picked up by Google ... so refer my profile pick, it says it more politely.

[Chris Hobson]

37 minutes ago, shumbaman said:

When looking at a tracking system you can consider one based on light sensors feeding into a micro controller or simply hard code the coordinates based on time of day and year.

Light sensors with Arduino should be easy enough. An anemometer can be used and if the wind speed is above a critical level the panels lie down.

 

Addition: I have a manual (altered four times a year) vertical tilt array facing due North and achieve about a 75% efficiency a solar tracker would take me to 100%. In your case the sun is south of you for a small part of the year but is probably nearly overhead.

 

[___]

1 hour ago, shumbaman said:

the transformer-less technology is suspect

On a previous thread we hashed that out a bit and I think Coulomb confirmed that the Axpert is not in fact "transformerless", depending on what you mean by that. It uses a high frequency design (all the top end ones do now) so it can use a smaller transformer, but the Axpert still has a transformer that it drives with a full bridge arrangement (aka 4 MOSFETs). In this respect it isn't significantly different from the Victron, which does the same thing (though it has a larger Toroidal transformer).

What some people mean by transformerless is that it doesn't have that big chunk of iron in it. That is actually a good thing. What others mean by transformerless, is that it uses a different kind of boost topology to gain higher voltages. The transformerless boost topologies still use an inductor of some sort, and it sacrifices isolation to save costs. Transformers are some of the most efficient devices known to man... we should stop trying to drop them :-)

 

1 hour ago, shumbaman said:

I firmly believe that Victron offers the best solution

The Victron MPPTs also tops out at 150V, though it can actually charge at those voltages :-P. For really high voltage stuff I believe you need to look at Morningstar or Outback. The Axpert does not fare too badly.

[Chris Hobson]

You could look at this http://solarelectricityhandbook.com/solar-angle-calculator.html has Zimbabwe and major cities for data.

[shumbaman]

 

11 minutes ago, Chris Hobson said:

You could look at this http://solarelectricityhandbook.com/solar-angle-calculator.html has Zimbabwe and major cities for data.

I actually have a hard copy of that book and did quite a bit of excel work using their data and combining it with other sources in trying to fully appreciate the effect of orientation and tilt on solar yield.

[Guest]

It has been mentioned by more than one group of hard core solar users the last few years that it is more cost effective to get a panel or two more to make the losses of having the array follow the sun ... BUT that does not take into account if one builds one for yourself for fun. 

Has also been mentioned to have panels face sunrise side, sunset side and the rest topside for the most effective harvesting of sunlight if space is and shade is a issue.

[KLEVA]

Shumbaman - The cable issue... Even though my distances aren't nearly that length, I feel the key would be reducing the number of cables, and using combiner/connecting boxes at the panel end rather than running each string back 50m. The power drop on something like 16mm cable (+- 100A) over that distance should be low enough with 16mm. I am not sure what it costs now, but when I purchased 40m x 2 (one black, one red so 80m total), I paid just over R1300. I am fairly sure that 16mm would be more than sufficient without killing the power, and at least you can connect that into the Axpert if you are careful (don't force it, 10mm is it's comfortable max, but if you are gentle and patient the 16mm fits, or you can safely cut off one core right at the end without affecting the overall cable drop). You don't have to use Solar spec cable from the combiner/connecting boxes, in fact it is more likely to get damaged if there is traffic (including animals) near the cables.

I don't know if that answer helps you (maybe you have already considered that), but US$700 for cable just seems a bit excessive, even for that distance.

PS: Not all Saffers are Rich/Powerful/Money types - some of us have to work for a living and to pay off our solar experiments  The fact that you work in US$ already gives you a head start, especially when you don't lose half your income/savings when the president opens his mouth

 

[KLEVA]

Oh, another sneaky move - if you have to have 2 separate string powers coming back (Note: Many will freak at this suggestion, but hear it out and make up your own mind)... Common out the negative lead. As long as the negative cable has sufficient capacity for the total current, you can quite sneakily share the common negative and only have 2 positives (one to each invertor)... Just a thought... (Now KLEVA grabs a coffee and watches the bunfight  )