Offgrid solar, limited controller options for both MPPT and dump load?

[richard.cs]

I've been looking through charge controller spec's, and so far as I can tell there seems to be a choice of the following:

• Dumb PWM controllers open circuiting the array to regulate
• PWM controllers supporting dump loads, basically the same but applying load to regulate (dead simple, I've built a few of those)
• MPPT controllers, without a (simple) way of connecting a dump load, reducing power when needed by walking away from the max power point
• MPPT controllers with a load output that can kind of be configured as a very low frequency PWM dump load, coming on at some SOC and off again at some lower SOC, like 90% and 85%. Basically very crude bang-bang control.

What I am struggling to find are controllers that keep the PV at its max power point and then bring a variable load (high frequency PWM or true buck) up gradually to absorb excess power and only move away from the max power operating point if that load doesn't absorb enough power to avoid overcharging. Clearly this approach is the optimum for using all the available power from the array (if you have a useful dump load). As it needs extra components I'd expect it to be a premium feature, but it looks like it barely exists at all. Have I missed something or is this 5th kind of controller very rare? The only one I have seen is an Outback controller than will PWM a load at 200 Hz via an external SSR. I've actually found more forum threads asking for this feature than I have found devices supporting it, which is a bit odd for something so obvious.

It's non-trivial to implement as an add-on to an existing system (with a type 3 MPPT controller) either, controlling a dump load based on reported SOC gives very low frequency control and excessive cycle use (as option 4), doing it by battery voltage seems like it'd inevitably get confused by multi stage charging algorithms. The best I can come up with is watching for the array voltage to rise as the MPPT controller shifts operating point to wind down the power but that then gets confused by array temperature. Still you could kind of see it working, a separate system linearly ramping up a dump load between some range of array voltages that is above the max power point under normal conditions would allow an MPPT controller that knows nothing about dump loads to work in this mode.

It just feels like a feature that should be common, at least on higher end controllers. Maybe it's just me falling into the engineering trap of trying to over-optimise things. 

[tszaboo]

Batteries don't mind microcycling. Toyota tested on the Ni-MH battery 100000 cycles with 20% SoC for their hybrids. Regular lithium will do just as well, if you decrease the SoC change the number of cycles goes up almost exponentially.
And I guess it would be an EMC nightmare to regulate that load.

[Marco]

• Dumb PWM controllers open circuiting the array to regulate

Why is this dumb?

Sure you can't cut off the current from a PV string in kHz range without HAMs coming to lynch you, but you don't need much capacitance to do it at say 100 Hz.

[richard.cs]

Batteries don't mind microcycling. Toyota tested on the Ni-MH battery 100000 cycles with 20% SoC for their hybrids. Regular lithium will do just as well, if you decrease the SoC change the number of cycles goes up almost exponentially.

I'm not sure I'd consider 20% a microcycle, but I suppose 5% is getting that way. I thought most people considered microcycling as <<1% at a few Hz not 5% at milliHertz. I take the point though, if the cycles are small enough not to dominate battery life despite their large number, then a simpler approach with bang-bang load switching has a lot in its favour. "Small enough" is quite chemistry dependent of course.

Why is this dumb?

"Dumb" in the sense of being a simplistic control scheme, I wasn't intending to imply that it is "bad". It's perfectly functional when the array voltage is well matched to the battery and you're not trying to squeeze the last few % out. Today though the economics tends to drive us towards using the mass-produced grid tie panels (cheaper per Watt due to high volume) with active conversion rather than array matched to battery.

You both seem pretty worried about EMC, I get it, but that's a filtering issue. If you get the filtering right, which admittedly has a cost (design and then components), then however you switch you can still have negligible HF on the external wiring. Any controller doing MPPT is already switching internally at tens of kHz, because MPPT is just a buck converter with an additional control loop wrapped around it, and you need high frequencies to make the component values reasonable. I am sure plenty of controllers out there are EMC disasters on both input and output ports, but it doesn't have to be so.

[Marco]

If you are dumping load, efficiency doesn't matter (as long as the magic smoke doesn't escape). If you don't have to dump load, you can bypass the PWM with a relay.