Jump to content

Solar System Review Please


andrew9484

Recommended Posts

Hi All,

I am pretty much done with my installation.

1. Please review my current installation / wiring and let me know if you see anything missing / unsafe / could be improved.

Final step will be to add another Earth Leakage unit to my DB board and then use changeover switches for some selected circuits. I have marked the new section with red dotted line.These are circuits that could run from battery / solar but unless there is extended power outage I would like to leave them on municipal supply (eg washing machine).

2. Please review my proposed additions(inside red dotted line) and let me know if any issues.
3. Any comments appreciated!

 

Thanks,

Andrew

solar.pdf

solar.jpg

20170408_152736.jpg

Link to comment
Share on other sites

Looks like an axcellent job and I will be stealing elements of your layout in my own upcoming installation.

From what I have read on this forum and other areas I would suggest adding 1 PV panel (if you have the space) for 3x3 strings rather than 4x2 strings, this will give a peak string voltage of 110v rather than the current 73.6v which is close to the drop out voltage of 60v, this arrangement should give a better result in less than optimal conditions like early morning, late afternoon, hazy day etc.

 

Link to comment
Share on other sites

Hi Andrew and Pilotfish

Those 320W panels are probably 72 cell panels and depending on Andrew's location I would be hesitant to have strings of three due to temperature compensation of the Axpert's Voc max and MPPT range. Two 72 cell panels in series should not be anywhere near the dropout voltage. As an example a single Kyocera 320W panel's Voc is 49.5V

 

Link to comment
Share on other sites

Both of you guys are right. Axperts are made for 3 panels in series setup, 315/320 watt max size, but a 2 panel setup would also work. The 320W might just be a "little" to big for the Axperts with 3 in series. I have 315 watt panels. Yesterday morning my Axpert took a swan dive again on the solar production side down to 0 watts for about  30 seconds and then taking off again like a rocket. With the "cold weather" about 10 deg. I could not believe the open circuit voltage. I think it went up to 138 volts or something. I normally only see 126-128 volts, just before I let them loose out of the starting blocks. I have 110 volt LED panel lights on the PV side. I thought to myself, they must be gone, but I see they are still up and running. :)

Link to comment
Share on other sites

Hi Guys,

I understand your concerns regarding Voc of 3 high power panels in a 3 string, and believe that they are valid for freezing temps on a crystal clear day on 21 December with a panel aligned perfectly with the sun (so about 4' at that date in Joburg where I live) - but in any other instance it is extremely unlikely that 3x CS6X-320 will exceed 145v (48.3v per panel).

If you believe that the conditions described above are a likely scenario and that you will EVER see 1000W/m2 under any conditions at your location then you should stick with the 2 panel strings. If you use the far more realistic NOCT rating of 800W/m2 and allow all other conditions to be absolutely ideal as described then you will still be below the max rating. 

I am not pressing the above to get anyone else to change their mind, but because I intend using 3x Yingli 310W strings in my own installation based on the above with very similar spec - if I am making a mistake I would like to learn that now.

The Kyocera 320 panels mentioned by Chris appear to have a very high Voc and I wouldn't like to push my luck with those, and of course once you add the 50% safety margin required by the engineering fraternity then my theories are blown out of the water.

Link to comment
Share on other sites

29 minutes ago, pilotfish said:

I am not pressing the above to get anyone else to change their mind, but because I intend using 3x Yingli 310W strings in my own installation based on the above with very similar spec - if I am making a mistake I would like to learn that now.

They will work just fine.

30 minutes ago, pilotfish said:

If you believe that the conditions described above are a likely scenario and that you will EVER see 1000W/m2 under any conditions at your location then you should stick with the 2 panel strings. If you use the far more realistic NOCT rating of 800W/m2 and allow all other conditions to be absolutely ideal as described then you will still be below the max rating

The readings from my setup correlate is a lot closer to the NOCT rating of 800W/m2, so I tend to agree with you. 

Link to comment
Share on other sites

I achieve very close to my theoretical maximum early mornings and evenings in winter. A panel reaches operational voltage very quickly when the sun starts shining on it. It does not need 1000W/m2 to get there. It needs that to achieve its Pmax. The Axpert also has the tendency during the day to shut down the MPPT briefly to recalculate when things get busy like a variable load like an iron on a day with cloud edge effect. I would be very cautious of exceeding the SCC Voc max even momentarily. You might just get a "been there done that" badge in the post from TTT.

My temp compensated Voc for a string of 3 panels is 115V. Today is not even really cold and have a peep at PV Voltages before the MPPT really kicks in.

58f5ba3c8c2c8_Screenshot(97).thumb.png.31d584584a8210e8e0e9336175508168.png

Link to comment
Share on other sites

23 minutes ago, Chris Hobson said:

The Axpert also has the tendency during the day to shut down the MPPT briefly to recalculate when things get busy like a variable load like an iron on a day with cloud edge effect. I would be very cautious of exceeding the SCC Voc max even momentarily. You might just get a "been there done that" badge in the post from TTT.

Mine did the swan dive thing yesterday. I saw the voltage go up to 138V on my system. The temperature was 10 degrees outside. What makes it worse for me is the tracker, as theoretically, the panels should be very close to 100% aligned with the sun.  It did it again this morning, but the voltage only went up to 132V.

Link to comment
Share on other sites

1 hour ago, Chris Hobson said:

You might just get a "been there done that" badge in the post from TTT.

:D:D:D:D

As a matter of fact, been there done that ... on purpose, for I wanted to see if the marketing material was true. My controller did switch off.

To get high voltages are easier than not. As per Chris, early morning cool panels and during a cool day with cloud effect can push my small system quickly to the brink. I frequently get panels exceeding their 930w (max 1043) and the 400w array (max 439) respectively. If I put in thicker cables, I would get that even more I guess. Not worth the cost though.

I would not push a controller to hard for too long. Even if it switches off (wastage), the components are pushed.

Link to comment
Share on other sites

I believe the BMV shunt should be on the negative side. The monitor itself is powered by the RJ-style cable that runs from the shunt, and if you put it that side there will be no potential difference to power it with.

Link to comment
Share on other sites

Since the buck converter inside an Axpert SCC is likely to be synchronous, a major factor in determining losses is the switching frequency. A lower switching frequency reduces losses in the inductor core, rectifier and MOSFETs. The lower end of permissible switching frequency is determined by the physical size of the caps and inductor (You want to lower the frequency further you need to increase the size of these components). Switching frequency is determined by the Vin /Vout  ratio. The larger the ratio the greater the switching frequency. Based on this the Axpert' s MPPT is probably most efficient somewhere in the midrange between 60 and 115VDC. I must admit I am at the edge of my understanding of how a buck converter works (more reading Mr Hobson!).

Plonky/SuperDIY and anyone else who tinkers is my reasoning sound?

 

Link to comment
Share on other sites

11 minutes ago, Chris Hobson said:

Plonky/SuperDIY and anyone else who tinkers is my reasoning sound?

As I understand it, with buck converters, higher is better when it comes to switching frequency... except it has an ugly side-effect: Skin effect!

I'm also somewhat at the edge of understanding here, but I don't think the Vin/Vout ratio depends too much on the frequency. Vin/Vout is directly proportional to the mark/space ratio of the PWM signal you drive it with and independent of the switching frequency. The greatest trouble with Vin/Vout, far as I know, is that the MCU used in these chargers only have a certain number of steps available.

For example, I looked into using the ATMega 328 (aka arduino) for this. My frequency of choice was 100Khz, and this isn't doable with the arduino's "user friendly" api, which limits you to around 32khz. You have to dabble directly with the registers. To get 100Mhz from the clock speed of 16Mhz, I had to set the prescaler to 80, which left me with only 80 discrete steps, or 1.25% steps. You have to multiply this with the Vin/Vout ratio to get an idea of how precisely you might hit the MPPT, for example, if I had a ration of 3:1, then I can only hit the MPPT with around an accuracy of 3.75%, which is probably okay... I don't know. I kinda lost interest after proving that I can generate 100khz using a small atmel chip :-)

The advantages to higher frequency is smaller components. Think of it as a storage tank (up on poles in the air) and another storage dam (down on the ground), the tank being the inductor and the dam below the capacitor, with a big old valve that fills up the tank. You open up the valve and fill up the overhead tank fairly quickly (high pressure aka voltage), then you close the valve and let the tank drain into the dam below. The objective is to keep the dam below at a steady level (output voltage). The higher the flow out of the dam below, the longer you have to open the valve. The faster the valve can operate, the smaller the two tanks have to be.

But as I said, the curse is that at a certain point, skin effect mandates that the wire in the inductor has to be thicker, so this only works up to a point.

In addition, there are isolation constraints. Usually the MOSFET switches in the positive line. N-channel FETs have lower on-resistance than P-channel FETs, so you want to use an N-channel FET. In the Positive line. The FET needs 20V ABOVE that to switch on properly, so you need a supply of 20V ABOVE your highest battery voltage to switch that thing on. Which means you need an isolated supply (or a bootstrap/charge-pump circuit). So many charge controllers need to isolate that circuitry from the low-voltage MCU stuff (running at 5V usually), and the cheapest way to do this is using an opto-isolator. But optos cannot switch that fast and they get expensive as they get faster. So that also limits the speed.

Of course, you also want that thing to not be audible, so you want to be above 30khz. Optos generally limit you to <=50khz.

Synchronous vs Asynchronous. Doesn't make much of a difference. Async uses a diode -- usually a Schottky -- for the flyback half of the cycle. Sync uses a second FET. FETs have lower on-resistance than Schottkys. But it makes driving more complicated.

So, not an expert, but from lots of reading about this, it seems you will generally have a switching frequency of between 40khz and 100khz (which determines the size of the cap/inductor), and depending on the chosen frequency that will then place certain limits on the discrete steps possible from the MCU, and that will then determine the optimum vin/vout ratio.

Final bit: Saturation. You can saturate the inductor with magnetic energy. Back to the tank analogy, it is possible to overflow the overhead tank. If the pressure/flow is very high (aka high Vin/Vout), and the smallest opening time of your valve is too long, that tank overflows (aka inductor saturates). But far as I could see, your switching FET's max voltage is usually the limiting factor rather than saturating the inductor.

I wish I was an electronic engineer and I could tell you that is absolutely true, but as it stands that is merely as I understand it.

The one MPPT I actually did tests on, you should all know by now since I mentioned it so often, is the Microcare units. That runs at 40khz.

Link to comment
Share on other sites

6 hours ago, andrew9484 said:

Is there a problem?

First of all, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

The way the batteries are connected in the diagram, the top string will work the hardest, followed by the string second from the top etc.etc.

If you really have no other option than to use more than 1 string, best way to ensure that all strings are utilised equally is to ensure that all the cabling in each string is of the same size (gauge) and length and to connect the ends of all the strings to the same connection points e.g. busbars.  Then never link the midpoints of the different strings directly - you can do it via fuses, but not directly as per the diagram.

Then for safety reasons you have to fuse each string separately before it is connected to the busbar.

All that said, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

58f857e193efa_EqualResistanceperString2_wm.jpg.8ebe042afffce08a54d80458c1a7cf01.jpg

Furthermore a battery balancer is recommended - for parallel strings of up to 4 batteries you can get away with something like one HA-02.  The only advantage of using more than one HA-02 balancers for parallel strings, is an increased balancing current, but once the batteries are balanced after the balancer was connected for the first time, the balancer will not work that hard any more and one balancer should be able to keep the bank balanced. If more than one balancer is connected in parallel (more than one pair of wires per balancer in parallel) the balancers might fight each other and actually not balance effectively - discussion for another day...

HA-02_Multiple_strings_wm.png.9b710012d92fcad410e05c3844c8ee0e.png

In my experience all failed cells I've ever came across went "short-circuit" and not "open circuit". I know @Chris Hobson said that he has seen cells which have failed and went "open circuit" - that will not be such a big deal, except that the remainder of the string containing the faulty cell will effectively be disconnected from the bank and of no use at all, until it is discovered maybe years later. 

 

The problem comes in with a cell failing and causing an internal short-circuit. To simplify the following explanation I will use a standard lead-acid cell voltage of 2.25V.  If you now have 24 cells in series, you have a total string voltage of 54V.  Since all the strings are connected in parallel, the voltage of every string will be 54V.  Now if one cell in one of the strings fails and becomes shorted (0V) the 54V across the string (maintained by the other parallel strings) needs to be divided by the number of healthy cells in the string - 54V / 23 cells => 2.35V.  To get these cells to 2.35V each they need to be charged and they will be charged by the other parallel strings.  This will result in high (charging) currents flowing into the faulty string which will cause heat.  When the battery bank is charged by the charge controller / inverter the voltage per cell will rise even more, because the voltage of the battery bank will go higher during charging - again this will cause more heat in the faulty string.  Furthermore, when one cell in a battery fails, another and another usually fails soon after that; now when 2 cells in the same string fail, the voltage in the remainder of the cells in the string will rise to 2.45V each, if another cell fails, the voltage in the remainder of the cells will rise to 2.57V each, you get the picture. Higher and higher currents will flow into the faulty string, causing more and more heat and eventually you might have a thermal runaway, exploding batteries and some pyrotechnics.  Now imaging all of this happens when you are away for the weekend and there is nobody at home to react to the BMV's mid-point alarm (which might not even detect a problem in the first place, especially not if the midpoints of all the strings are tied together)...

So bottom line, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

Link to comment
Share on other sites

1 hour ago, superdiy said:

First of all, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

The way the batteries are connected in the diagram, the top string will work the hardest, followed by the string second from the top etc.etc.

If you really have no other option than to use more than 1 string, best way to ensure that all strings are utilised equally is to ensure that all the cabling in each string is of the same size (gauge) and length and to connect the ends of all the strings to the same connection points e.g. busbars.  Then never link the midpoints of the different strings directly - you can do it via fuses, but not directly as per the diagram.

Then for safety reasons you have to fuse each string separately before it is connected to the busbar.

All that said, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

58f857e193efa_EqualResistanceperString2_wm.jpg.8ebe042afffce08a54d80458c1a7cf01.jpg

Furthermore a battery balancer is recommended - for parallel strings of up to 4 batteries you can get away with something like one HA-02.  The only advantage of using more than one HA-02 balancers for parallel strings, is an increased balancing current, but once the batteries are balanced after the balancer was connected for the first time, the balancer will not work that hard any more and one balancer should be able to keep the bank balanced. If more than one balancer is connected in parallel (more than one pair of wires per balancer in parallel) the balancers might fight each other and actually not balance effectively - discussion for another day...

HA-02_Multiple_strings_wm.png.9463a5d5d948844ae03e0202ff5e9865.png

In my experience all failed cells I've ever came across went "short-circuit" and not "open circuit". I know @Chris Hobson said that he has seen cells which have failed and went "open circuit" - that will not be such a big deal, except that the remainder of the string containing the faulty cell will effectively be disconnected from the bank and of no use at all, until it is discovered maybe years later. 

The problem comes in with a cell failing and causing an internal short-circuit. To simplify the following explanation I will use a standard lead-acid cell voltage of 2.25V.  If you now have 24 cells in series, you have a total string voltage of 54V.  Since all the strings are connected in parallel, the voltage of every string will be 54V.  Now if one cell in one of the strings fails and becomes shorted (0V) the 54V across the string (maintained by the other parallel strings) needs to be divided by the number of healthy cells in the string - 54V / 23 cells => 2.35V.  To get these cells to 2.35V each they need to be charged and they will be charged by the other parallel strings.  This will result in high (charging) currents flowing into the faulty string which will cause heat.  When the battery bank is charged by the charge controller / inverter the voltage per cell will rise even more, because the voltage of the battery bank will go higher during charging - again this will cause more heat in the faulty string.  Furthermore, when one cell in a battery fails, another and another usually fails soon after that; now when 2 cells in the same string fail, the voltage in the remainder of the cells in the string will rise to 2.45V each, if another cell fails, the voltage in the remainder of the cells will rise to 2.57V each, you get the picture. Higher and higher currents will flow into the faulty string, causing more and more heat and eventually you might have a thermal runaway, exploding batteries and some pyrotechnics.  Now imaging all of this happens when you are away for the weekend and there is nobody at home to react to the BMV's mid-point alarm (which might not even detect a problem in the first place, especially not if the midpoints of all the strings are tied together)...

So bottom line, parallel battery strings are not advised.  Rather invest in a single string of higher Ah rating.

Hi. Thank you for the detailed feedback! I will review in greater detail later and see what i can do. I realise what i have is not the best solution but was result of trial and error and 'upgrades' to battery storage. I will add the midpoint fuses. I am also planning to connect the BMV voltage free contact to my house alarm.

Link to comment
Share on other sites

  • 2 months later...

I correct the post. Stupid question. I received the HA-02 battery balancers purchased by e-bay. The conecctions for an battery configuration  of 8 Trojan 6 volts in series are elemental. I also thank you for your attention @superdiy

 

Edited by Elmichi
Correction
Link to comment
Share on other sites

  • 1 month later...
28 minutes ago, maxomill said:

why link he midpoints

Far as I know, it's  a cheap way to monitor the midpoint of all strings (assuming you have two or more), because the BMV only has one input that you can use for either a starter battery, midpoint monitoring, or a temperature probe. The fuses is there so that if anything bad happens in another string -- say a cell goes short circuit-- it takes the fuse out.

Normally you don't want to link midpoints, so this is a compromise, acceptable because it is fused :-)

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...