Aux1 waste not thought

[dgd]

To clarify my first post about improving aux1 wastenot which was a bit confused, I was thinking about an Amps value (diversion) that would be relative (added) to the current needed as read from the shunt/whizbang to hold the Abs/Float voltage and that would then start Aux1 wastenot after a delay only when the array has enough power available.
That would improve aux1 wastenot with minimum diversion losses (vs aux1 float diversion)and let the users have the choice to keep the whizbang on aux2 and still use end amps to control the Absorb charge.
Erik
PS: Maybe that it could also solve the flickering issue with Aux2 wastenot pwm through an inverter? Or maybe that it doesn't worth it? I don't know.

Erik,
Sounds like a great idea.
The only problem I see with this is that the classic could never know what current is available from the array when it is limiting current at absorb or float. To discover this it would need to temporarily disable current limiting to allow maximum current to flow.
This will probably not be a good thing for battery charging as during this test time the battery voltage may rise too high.
Then how often would this test need to be done? Every minute or 10?

I can't help thinking that to get this array excess power while absorbing is not so easy without the use of the aux2 pwm control. In any case I never really appreciate the way it had to be implemented in that the PWM start was at a predefined voltage below absorb and as the charging reached the absorb voltage the pwm increased duty cycle until full on.
Where there was a struggle for PV power to raise the battery to absorb this stealing of scare power by introducing a load progressively could prevent absorb being reached.

The only true time of excess PV power could be in float then gradually introducing a load would be the way to go IMHO.
This method would be outside the classics capability as its really only concerned with battery charging and soc.

However there is a relatively easy and cheap solution.
That is to try and bring on float as soon as possible and just use aux1 float high to switch in a load.
The cheap way is to over panel, so that in poor weather you still get good power. In good weather just switch out some panels

Dgd
I

[SolarMusher]

To clarify my first post about improving aux1 wastenot which was a bit confused, I was thinking about an Amps value (diversion) that would be relative (added) to the current needed as read from the shunt/whizbang to hold the Abs/Float voltage and that would then start Aux1 wastenot after a delay only when the array has enough power available.
That would improve aux1 wastenot with minimum diversion losses (vs aux1 float diversion)and let the users have the choice to keep the whizbang on aux2 and still use end amps to control the Absorb charge.
Erik
PS: Maybe that it could also solve the flickering issue with Aux2 wastenot pwm through an inverter? Or maybe that it doesn't worth it? I don't know.

However there is a relatively easy and cheap solution.
That is to try and bring on float as soon as possible and just use aux1 float high to switch in a load.
The cheap way is to over panel, so that in poor weather you still get good power. In good weather just switch out some panels

Dgd
I

David, after a few days using aux1 wastenot, I'm a bit disapointed with how it works. It works fine when sunny but too much problems when cloudly. I've finally set hold time to 0.1sd and delay time to 1.30mn/2mn to give it a chance to float when absorb reach EA after 1mn and stop it to supply power from batteries when on float mppt. I believe that you're right with aux1 float high but I would like to give the classic/tristar pwm a last try before to switch definitively to aux1 float high only (if I only could get it right).
Erik

[SolarMusher]

Finally, was successful today on cloudly weather using aux1 wastenot with Vhigh -0.2/delay 90sd---Vlow -0.4/hold 1sd. This way the classic holds Absorb or Float when enough power from the array, classic checks every 90sd for enough power if not it disconnects quickly in 1sd, checks another time 90sd later and hold aux1 wastenot if there's enough power available to hold the load. Not as good as Aux2 pwm but better than Aux1 Float high for my system when sunny and enough power to supply these 1200W elements on absorb (10h30am to 12h30-13h pm until float usually).
Erik

[boB]

"checks every 90sd for enough power"

sd ?

[zoneblue]

I reckon the ultimate way to do this is with arduino or similar. You get a small reference PV, measure its Isc, ADC in, compare that with what the classic is producing, and use the difference to PWM your HWC SSR. So long as the reference pv is the same mC type with teh same glass, youll should get a very very accurate result.

[SolarMusher]

"checks every 90sd for enough power"

sd ?

Allo boB, 90sd => 1mn30.
Aux1 connects at -0.2V under Vabs (58.8V) after 1mn30 delay time and quickly disconnects at -0.4V under Vabs (58.8V) in 1 second hold time or less. When you watch the classic, it looks like "it was checking every 1mn30 if there's enough power at the array to hold the diversion load and if not it keeps a steady abs or float voltage waiting for a next power check". This 1mn30 delay time (or more) allows the EA/Wbjr on Aux2 to end aborb in the 1mn time under EA without too large voltage variations because of the quick disconnect time of 1 second or less. In fact, from what I saw an independent Vlow setting under -0.4V would be better (something like -0.2V) but the classic sets Vlow relative to Vhigh so it's seems to be impossible to set it both at -0.2V.
Anyway it's not so bad for me, as it can use a good part of the absorb time to heat water at higher voltage which provides more btu to the elements than a float voltage around 53.2/53.5V.
Did you read the arc fault alarm that I've had when using both, classic and tristar pwm?
Any thoughts?
Erik

[SolarMusher]

I reckon the ultimate way to do this is with arduino or similar. You get a small reference PV, measure its Isc, ADC in, compare that with what the classic is producing, and use the difference to PWM your HWC SSR. So long as the reference pv is the same mC type with teh same glass, youll should get a very very accurate result.

Hi Zb,
Interesting! I would like to have more time to learn about arduino. I was pretty sure that it can be done but for now I'm just looking at making an efficient DC diversion with limited means. Before to spend a few more bucks in pricey efficient elements, I would like to be sure that it will work the way I want  .
Erik

[RossW]

I've been playing for some time (couple of years - can't rush into these things!) with my system, working out "opportunity power", and I'm doing it this way.

1. I calculate the suns current position in the sky
2. I calculate from that what the "cosine-corrected" solar radiation should be on the ground
3. I compare that figure to what I'm actually measuring from my pyranometer.

The value calculated at (3) tells me how much actual sun I've got to play with, adjusting for high cloud, dust, smoke, humidity etc.

4. I then calculate, based on (3), and the position of the sun and the known position of my panels, how much power I "could" be getting. There is room for improvement here as I'm not actually measuring the array temperatures, which are substantially higher than ambient.
(edit: this step is necessary, as 4 of my arrays are on trackers and 2 are fixed, so the off-axis angle for them is constantly changing)
5. I subtract from (4), what I'm measuring I'm actually  taking from my solar arrays (to power loads and charge batteries).

The difference between them (the value calculated at 5) is how much "opportunity power" I have to potentially play with.

My longer-term aim is to use a VFD to run a compressor for a heat-pump. In summer this would run a chiller that I can use for cooling both during the day and overnight (very low power requirements overnight - fan and circulating pump is basically all it needs), or to use the heat-pump for heating and domestic hot water in winter. The VFD would allow me to ramp the compressor up or down to use the amount of "spare power" available from moment to moment.