frater secessus

frater secessus

retired snowbird

backchannel - comments on solar advice

A reddit poster gave some solid advice to a beginner. I have a few thoughts that are not appropriate for that beginner’s consumption, but may be of use to people who have already gotten their feet wet.

solar panel connectors

1) the connectors on solar panels and controllers are really standardized, you almost can’t go wrong.

While most panels do have MC41 or MC4-compatible connectors, smaller panels and panels from “solar generator” companies often have different connectors. A cynic might suspect this is to lock customers into their own product line.

There is also an H4 connector2 in the wild; it is technically different from but interconnectable with the MC4. The majority of MC4 connectors out there are knockoffs and not actual

on the general requirement for MPPT controllers

2) get an mppt solar controller not a pwm one.

PWM is a reasonable default. Unless one has a technical reason to use MPPT (see previous link) an MPPT will average something like 15% more harvest for 100%-200% higher price, which isn’t a great cost/benefit ratio. Claims of “30% moar power” are true, it just don’t apply during most of a normal day so the average daily increase is less than 30%.

If one is short on power the rule of thumb is “it’s usually cheaper to add a panel than to buy an MPPT”.

sizing an MPPT controller

3) your solar controller will need to be rated for 10a for every 100w of solar, for example if you are connecting 200w of solar you will need 20a controller.

That is roughly true for PWM controllers.3

But MPPT controllers are sized on average harvest rather than panel rating. 10% overpaneling is a common minimum overpanel recommendation, and a Victron has demonstrated that overpaneling as high as 30% can be optimal, covering 99% of the power available. 15% might be decent target, as it’s halfway between 0%/parity and the 30% max suggested overpanel.4

The primary reason for overpaneling is to limit cost: upsizing an MPPT controller is expensive, sometimes prohibitively so. The trick is to buy just enough MPPT capacity to meet actual needs.5

factors that may influence overpaneling ratios

  • factors suggesting more aggressive overpaneling (nearer 30%)
    • in areas of low insolation (Pacific Northwet, Canada, northern Europe)
    • flat-mounted arrays
  • factors suggesting less aggressive overpaneling (nearer 10%)
    • when insolation is abundant (American southwest, lower latitudes)
    • seasonally/monthly-tilted arrays

how overpaneling is calculated

Working backwards from the PV power ratio given by Victron’s MPPT Calculator they are using ~14.5v as a constant for deriving the overpaneling ratio.6 So to find the overpaneling ratio of a particular array and controller

  1. panel wattage / 14.5v = panel amps
  2. (panel amps - controller rated amps) / controller rated amps

so…

  1. 300w / 14.5v = 20.689655172
  2. (20.689655172 - 20 = 0.689655172) / 20A = 3.4% overpaneled.

3.4% isn’t aggressive enough IMO; we have (slighly) overpaid for too big a controller.

small systems

It’s hard to match up small systems for value:

  • 200w on a 10A MPPT (38% overpaneled, poor controller will be hammered and we will be missing out on harvest)
  • 200w on a 15A MPPT (8% underpaneled, wasting some money on a slightly-oversized controller). This is as close as we will get to optimum with 200w.
  • 200w on a 20A MPPT (31% underpaneled, big waste of money on the controller rating)

In this case I’d try to add another 100w panel on the 20A, for about 3% overpanel.

bigger systems

some examples

  • 350w on a 20A MPPT (21% overpaneled)
  • 500w on a 30A MPPT (17% overpaneled)
  • 600w on a 40A MPPT (15% overpaneled)
  • 900w on a 50A MPPT (24% overpaneled)

other fun constants

If we want to find the amount of panel for a given overpanel percentage (again, using 20A controller):

  • 0% overpaneling7 (parity): 14.5v x 20A = 290w
  • 10% overpaneling: 15.95v8 x 20A = 319w
  • 15% overpaneling: 16.675 x 20A = 334w <– sweet spot for average conditions?
  • 20% overpaneling: 17.4v x 20A = 348w
  • 30% overpaneling: 18.85v x 20A = 377w

Bonus examples from my own lunatic experiments:

  • 200w on a 10A MPPT (38% overpaneled - controller is too small, but it was what i had on hand). The poor epever ran maxxed out half the time but never even got warm. Still have it but gave away the 2x 100w panels that fed it.
  • my previous setup 570w on a 40A MPPT (2% underpaneled)
  • my present setup 750 on 45A MPPT (15% overpaneled). I follow good weather in the desert southwest and might bounce off the 45A a few seconds a day near solar noon around summer soltice. The rest of the time the daily max is <45A.

  1. the connector was designed by the Multi-Contact company. The company was later bought by Stäubli Electrical Connectors but the MC designation remained. ↩

  2. Amphenol Helios 4. ↩

  3. the actual limit is incoming current. We usually leave 20% headroom to account for Perfect Storm conditions, so a 30A PWM could accept an array of ~24A. Most 12v panels have an Isc of <6A per 100w so the watts/10 formula roughly works: 200w = ≤12A. 12A x 1.2 = ≤14.4A. Since PWM can’t “round down” like MPPT we have to go up a size to 20A. 300w is ≤18A x 1.2 = ≤21.6A so we need a 30A, or at least a 25A if we can find one of those. The rule breaks down around 400w, since we only need ≤28.8A (18A x 1.2) so a 30A works. But it is very cheap to upsize a PWM controller and by the time we get to 400w lots of folks will be running MPPT for other reasons. BN: use the actual Isc from the panel under consideration rather than the 6A max discussed above. Three of this renogy 100w panel would only be 18.756A (3x 5.21A x 1.2) and so would run safely on a 20A MPPT. ↩

  4. not coincidentally, MPPT-controlled systems are typically ~85% efficient after various losses. So +15% might be a good bet to offset those losses. ↩

  5. Q. if we size for average harvest what happens when we get unusually great harvest? A. If the sun is especially powerful that day the overpaneled controller will clamp output at rated current (20A or whaterver); no harm, no foul. We trimmed a few watts and saved the cost of moving up to the next-bigger controller. ↩

  6. I think they are basing this on a typical Vabs value, since most systems will be hitting that setpoint around local solar noon when they will be making max power. There is no point in using something like 12.0v since the charging voltage will usually only be that low when in the morning when the sun is weak. Reminder: max power output is rated current x charging voltage. In the morning when the bank is 12.0v a 20A controller can make 240w at that 20A limit. Later when the bank voltage is 14.6v it can make 292w at the 20A limit. ↩

  7. Victron would call this 100% overpaneling, and 30% 130% overpaneling. I think this is a translation error, and they mean “percent of rated output” rather than “overpaneling”. ↩

  8. ie, 14.5v x 1.1. Then x 1.2, and x 1.3, et cetera ↩

Updated:

You may also enjoy