Any tips on my setup?

[Wrybread]

Thanks for the great information Sensij, as usual.

And knowing the "opportunity load" term led me to this thread, where someone almost goes down the same road I just did (by reading the ModBus output from their controller and controlling the load externally):

https://www.solarpaneltalk.com/forum...pload-question

And there's the usual distracting bickering with TheTrollThatShantBeNamed about it in this thread:

https://www.solarpaneltalk.com/forum...0-excess-power

And there's some interesting discussion on the topic here. Amazing how well people can collaborate when they're civil:

http://outbackpower.com/forum/viewto...=2865&start=25

but it would be interesting to see what carefully collected data on a system like yours would show.

I just added logging for PV current, PV voltage, battery current, battery voltage, and battery SOC (as reported by the controller). I'm averaging each of those over 15 minute periods, let me know if you think I should do that differently. It's easy enough to change it to 5 minute intervals or whatever.

I would just suggest that you think through the possible failure modes so you don't end up sending power at the wrong time.

Avoiding turning on the fridge before my batteries have had time to fully charge is of course the big riddle. Currently I'm turning on the fridge when my charge controller reports an SOC of 85% or greater and a battery wattage of more than 300 watts. Another idea is to turn on the fridge X minutes after going into float with a sustained wattage. Let me know if you have any recommendation.

Avoiding over discharging my batteries is relatively easy with my setup. Since my fridge is only 2 way (AC or propane) and not 3 way (AC, DC or propane) I have to power my fridge from my inverter, which will cut off before my batteries get too low. I don't expect that will ever happen, and would mean the Pi died or my program crashed while the fridge was turned on, but it's a nice safety to have.

That said, I sure wish my fridge let me power it with DC directly, since I'd rather not leave my inverter on all the time. But I don't care enough to upgrade my fridge.

[sensij]

From there it was easy to make a little app so I can monitor my panel remotely. Here's a screenshot of what it looks like in the morning with the sun coming up. The value in the center and on the graph is wattage. I plan to add the rest of the relevant data from the charge controller (pv voltage, pv amperage, battery SOC - to whatever extent that's useful) and the temperature inside the fridge and its power source at some point.

Nice programming! A couple years ago, another user was planning a 48 V system with some questions similar to yours (and a whole lot more, if you search the OP's name). Then, as in this thread, I was advocating for increasing the PV capacity beyond what the conventional "panel STC rating / battery voltage = charge current" calculation would suggest. I posted a model of what the actual charge cycle might look like in terms of volts and amps on an ideal day, but it would be interesting to see what carefully collected data on a system like yours would show.

https://www.solarpaneltalk.com/forum/off-grid-solar/off-grid-solar-panel-systems/22226-help-me-finalize-my-off-grid-system-please

It looks like you might have the ability to generate those curves, especially if you experiment with installing the 3rd panel.

The ability to use excess charging power to an "opportunity load" once the battery achieves the desired charge level is definitely a feature in some charge controllers. One implementation is an "auxiliary output" that can provide a control signal to an SSR based on the battery SoC. Using a Pi to manage that with a less expensive controller is clever, I would just suggest that you think through the possible failure modes so you don't end up sending power at the wrong time.

[Wrybread]

> https://www.rv.net/forum/index.cfm/f...print/true.cfm

Cool Project!

But sadly it looks like the range of use for those diversion loads is limited. From the Morningstar FAQ:

What are acceptable dump/diversion loads?

Diversion loads acceptable for use are water heating elements and air heating elements (i.e. purely resistive loads). DO NOT use light bulbs, motors, inverters, refrigerators, or other electrical devices for diversion loads. These loads will sustain damage or cause the controller to disconnect the load (leaving the system unprotected from overcharge). Only heating elements should be used. Diversion load sizing is addressed in your diversion controller

Here's an interesting thread where someone recommends a voltage sensitive relay such as this one:

http://www.ebay.com/itm/Voltage-Sens...-/301360777164

For the Tracers, its a shame that it's not built-in, since it's so close. The external load can be turned on when there's light (PV input), which I know is a common feature on charge controllers, but currently there's no way to set a more specific threshhold.

[ewarnerusa]

The Tristar does this. That is what the diversion load support feature is on it. I followed a thread on rv.net about a guy who used it to divert the current to a 12V water heater element.
"Diversion Charge Control
In the diversion mode, the TriStar will use PWM charging regulation to divert
excess current to an external load. As the battery becomes fully charged, the
FET switches are closed for longer periods of time to direct more current to
the diversion load.
As the battery charges, the diversion duty cycle will increase. When fully
charged, all the source energy will flow into the diversion load if there are
no other loads. The generating source is typically a wind or hydro generator.
Some solar systems also use diversion to heat water rather than open the
solar array and lose the energy."

EDIT: I found the old thread.

https://www.rv.net/forum/index.cfm/fuseaction/thread/tid/25824208/print/true.cfm

[Wrybread]

I am also surprised a solar charger doesn't come with control to direct the power to a "dump load" but then again you can still use the power directly from the panels if your batteries are fully charged so you really aren't losing anything if you set up your system to take advantage of the power generated by already having the load connected. If you don't have this load then you are the one to blame.

Do you know of a charge controller that will turn on that secondary load only after the batteries are fully charged? If the system isn't smart enough to prioritize the battery bank first, then I can't see a way to use the power productively without (often) getting in the way of charging the batteries.

[SunEagle]

I'm surprised being able to switch the power over to another application (a fridge for example) isn't more common on charge controllers. It seems like such an obvious need. In the language of some people earlier in this thread, you're "losing money" or "screwing yourself" once those panels stop outputting power. Personally I don't see it that way since I have yet to have to pay for any sun, but it's still nice to be able to do something with the power once my batteries are charged up. Or maybe some charge controllers can already do this?

If you have a wind turbine the charger usually has the option of redirecting the energy to a "dump load". That is built in to keep the turbine from either over spinning or burning something up if you don't have a way of stopping it by initiating a "break" or tethering the blades so it doesn't turn as fast.

I am also surprised a solar charger doesn't come with control to direct the power to a "dump load" but then again you can still use the power directly from the panels if your batteries are fully charged so you really aren't losing anything if you set up your system to take advantage of the power generated by already having the load connected. If you don't have this load connected then you are the one to blame and are losing the energy generated.

[Wrybread]

I'm surprised being able to switch the power over to another application (a fridge for example) isn't more common on charge controllers. It seems like such an obvious need. In the language of some people earlier in this thread, you're "losing money" or "screwing yourself" once those panels stop outputting power. Personally I don't see it that way since I have yet to have to pay for any sun, but it's still nice to be able to do something with the power once my batteries are charged up. Or maybe some charge controllers can already do this?

[Wrybread]

> Yeah, my Tristar PWM 45 interfaces with a RS232 cable. I need to use a USB converter with it!

Ha. But actually, from some perspectives, that's a good thing, since that data is very easy to read by other gizmos. But yeah, in the modern era, I think being able to read data over the network (on your phone for example) is crucial or, at least, extremely handy.

[Wrybread]

And looking at more product pages, it looks like they might have added the webserver feature to the MPPT-45a? It's certainly commonly mentioned on the features list. The product listing that Sunking posted says it has "Ethernet: fully web-enabled interface to a local network or internet; view from a web browser orsend email/text messages" for example. But people in the Amazon reviews specifically say it doesn't have that feature. Dunno, maybe newer models do?

[ewarnerusa]

Yeah, my Tristar PWM 45 interfaces with a RS232 cable. I need to use a USB converter with it!

[Wrybread]

It's too bad Morningstar hasn't put their remote montoring features on their lower end charge controllers yet. On their MPPT-60 you can plug the controller directly into a network and monitor the settings from any gizmo attached to your network, or (if desired) over the internet. Note the method this guy is using for example:

https://www.youtube.com/watch?v=X2XBguHAK_w

As far as I know that feature isn't available on any other models though. There's conflicting info about that though, since this Amazon listing for the MPPT-45a says it has the ethernet port, but according to the reviews it doesn't. I don't know, and it wouldn't solve my needs anyway, since I want to use the power from my panels for something else once the batteries are fully charged.

[ewarnerusa]

Agreed, the laptop consumes a lot of power to have plugged in all day. I have tried to work around that by using a 12V adapter to cut out any inverter use, having the laptop display off unless needed, and also put a 12V outlet timer on with a schedule built around solar harvest. So I can keep the laptop running all night and day if the sun cooperates, but it doesn't address the issue of using so much harvest just to keep the laptop running. It frustrates me that I have this Microsoft tablet from a few years ago that won't run the MSVIEW software! The tablet would use less energy than the laptop. But this only becomes an issue when I'm in one of my tinkering moods, usually the setup is left alone to do its thing.

[Wrybread]

The main problem with using a laptop for 24/7 monitoring in an off-grid application is that the power draw is pretty massive compared to gizmos like the Raspberry Pi. There's also space issues, and longterm reliability issues since there's moving parts (fans and harddrives). Also from a quick Google search it looks like MSVIEW is Windows only, so it's not ideal if longterm reliability is critical.

I should also mention that the method I'm using would work with a Morningstar too, since they output their data the same way as the Tracers (via ModBus).

[ewarnerusa]

I have seen that monitor referenced on RV.net and follow a thread about it. I do have an Amazon-purchased digital voltmeter that I installed in the wall with an on/off switch. It isn't reading directly at the batteries so I'm seeing all kinds of bias but it serves its purpose. At least it has a potentiometer so I can calibrate it.

In hindsight, I suppose MSVIEW wouldn't offer you a way to trigger relays based on performance. But that is way out of my league anyway. It does offer .csv file logging of parameters, so maybe that can be used to trigger something.

[Wrybread]

Sorry I should have said, a Raspberry Pi is a little computer about the size of a deck of cards. It draws very little power (2 to 4 watts) so I can leave it running 24/7 without much impact, and it runs the Linux operating system so there's almost never a need to reboot it. Super reliable and configurable. I like to say its like a blender or any other appliance, it just does its thing super reliably. I'm running it "headless", meaning no monitor or keyboard attached (I configure it over the network), but I could connect a monitor to it to display data. But viewing the data on my phone is more convenient for me, and I leave an old phone stuck to the fridge permanently displaying the data.

The Tracer has some stock software to monitor and configure it from a laptop as well, and I've found it useful for seeing longterm data, but I don't use it much for seeing current conditions since it's a bit unwieldy to connect a laptop every time. Besides it doesn't give me the flexibility of doing things with the data, like automatically triggering a relay when X conditions are met. And of course laptops take too much power to leave running 24/7 just for this type of application.

As far as monitoring the voltage of your batteries, if you ever want something more permanent Amazon sells a battery monitor that's pretty well reviewed for under $20. I've always wanted to try it:

https://www.amazon.com/bayite-6-5-10.../dp/B013PKYILS

I have the Trimetric so I don't really have a need, but it's so well reviewed that I'm tempted to get one just to play with.

> Who says solar doesn't work for off grid camping, right?

Hopefully no one? Or people who haven't actually tried it? It works fantastically well!