One of the things that irks me about an Axpert is its PV connection. You are limited to 145V DC and  I have a 17m home run between the combiner box and the inverter which is a two core 25mm2 armored cable. The PV terminals are 10mm2. As a compromise I shoved 16 mm2 cables into the 10mm2 terminals.

After some connecting and disconnecting the point of the cable looked like this.

 

A large size bootlace ferrule has made it much neater

The added bonus it the point of the ferrule just just squeaks into the PV terminal. I made a small sample to test on the inverter that has no PV before tackling the main inverter. I am very pleased with the results.

9 hours ago, Chris Hobson said:

You are limited to 145V DC and  I have a 17m home run between the combiner box and the inverter which is a two core 25mm2 armored cable.

Hi Chris, this is purely as info. I looked at your Panel ratings and if i am looking at the correct Yingli data sheet one can expect max 30.4 volt and 8.24 amps from each panel under load. Using an Axpert, I take you have 3 parallel strings of 4 panels in series. Total of 121.6 Volts 24.72 Amps. For this comparison I use a PF of 1 

If I read that into an app called Electrical Calc Pro  (You can double check on this  " https://photovoltaic-software.com/DC_AC_drop_voltage_energy_losses_calculator.php ")

As follows

1.  121.6 Volt DC at 24.72 Amps over 17 meter (Up to Inverter) using 25mm 2core copper cable will result in a 0.784V  volt (0.654%) volt drop 
2. 121.6 Volt DC at 24.72 Amps over 16 meter (Up to Isolator) using 25mm 2core copper cable will result in a 0.738V  volt (0.607%) volt drop 
3.  120.86 Volt DC at 24.72 Amps over 1.5 meter (From Isolator to Inverter) using 10mm 2 core copper cable will result in an iditional 0.069V (0.057%) volt drop

Should you chose to replace the 25mm from the isolator to the inverter with 10 mm 2 core you will loose and additional 

(0.784 - ( 0.738 + 0.069 ) ) = 0.073 V at 24.72Amps = 1.8Watt.

 

Just to show that you can comfortably and safely (and without Compromise) reduce the wire between your isolator and inverter to 10 mm. 

 

Further two calculations worth mentioning. 1: Using 16mm all the way and 2: using 10mm all the way.

• 121.6 volt 24.72 amps over 17 meter all the way on a 16mm cable will result in a 1.22V (1%) volt drop. An additional 0.43 volts x 24.72 amps = 10Watt los compared to 25mm   
• 121.6 volt 24.72 amps over 17 meter all the way on a 10mm cable will result in a 1.92V (1.58%) volt drop. An additional 1.136 volts x 24.72 amps = 28Watt loss compared to 25mm 

If you chose to use a 10mm cable over the full 17 meter you have to live with only a 28watt extra loss. I know we all feel the bigger the better on dc cable , but its not entirely true. When determining wire / cable size we aim to keep the volt drop below 5 percent, and as you can see we are not close to 5 percent in these calculations. Contrary to common believe, there is no danger in using smaller wires , as long as you stay within electrical guidelines and specifications.  

  

Thanks @Jaco de Jongh 

Thank-you for the link above it is better than the calculator I was using http://www.solar-wind.co.uk/cable-sizing-DC-cables.html

I did my calculations slightly differently. A solar panel behaves like is a constant  current device delivering current which is close to I (P Max) early during the day, once it is getting good irradiation.  Any further increase in power is due to an increase in voltage. On a cold day this is easy to see as the panels settle in a 85V thereabouts whereas the normal operate at a lower 70-75V  and on one particularly hot day here in the Karoo when the temperature  was over 46°C it dropped to 68.8V. So lets repeat the process using 72V (since that is the voltage I largely operate at) and 32 Amp (4 strings)

Using 25mm2 my losses are 1%

Using 16mm2  my losses are 1.5%

Using 10mm2 my losses are 2.5% (bordering on unacceptable).

The 25mm two core was part of the 36V DC battery light supply that was made redundant by the installation of solar so I used it as it was free. I could have used 10mm2 for the last metre but there is a part of me that says that's not really acceptable. For my second inverter you will be pleased to know I will again be using 25mm2 up until the DC circuit breaker (because it is free) and then 10mm2 as there will only be 16A flowing through it.

 

47 minutes ago, Chris Hobson said:

I could have used 10mm2 for the last metre but there is a part of me that says that's not really acceptable.

I understand, and it remains your choice. Like I said Common believe makes us think we cant go smaller and bigger is better,  but I assure you, you will not compromise your system at all by changing the last meter and a half. 

47 minutes ago, Chris Hobson said:

32 Amp (4 strings)

I assumed 3 strings of 4.... but the highr amps will affect the calculations quite a bit.  

12 hours ago, Jaco de Jongh said:

keep the volt drop below 5 percent

I thought it was 3%?

I suppose in the end it doesn't matter too much, as long as you're not heating the cable to dangerous limits, and you can live with the loss.

For Victron ESS configurations, we want max 0.4V drop (For a 48V setup, 0.1V per 12V increment) between the battery and the inverter for certain functions to work correctly. That's less than 1%. At least it only affects the battery/inverter link :-)

58 minutes ago, plonkster said:

I thought it was 3%?

As  a matter of fact I think you are correct,  3 % on DC and 5 % on AC, 5% stuck in my head as 90% of my planing / design at work is done on AC. So out of habit I refer to the 5%. 

Edit: @plonkster  you had me confused there for a moment, I had to look through my previous calculations on DC circuits  at work and it was also done at a 5% volt drop limit.  Then i quickly had a look at a copy of the SANS regulation. its and older one, Borrowed the latest to someone... and this is what it says.

 

The 3% might be an amendment, but if memory serves me correct, its more a recommendation in the solar world (To minimize losses I guess).

Come to think of it, for limiting energy losses on solar I would stick to 3%, for other DC applications 5% will do. 

 

42 minutes ago, Jaco de Jongh said:

its more a recommendation in the solar world

Correct, I wasn't talking about what the rules might say, I was thinking about the general advice dished out by solar people :-)