Tonight's power related shenanigans.

[paulca]

It's winter.

The solar panel charge controller cut me off.

Plan B.  I came up with it last year.  I have a 20A bench PSU doing nout in the garage beside it.  I normally use it to give the battery a yearly slap in the face with 10A from 11V too full charge to fizz it up (it's old).  It's only December 6th.

The plan was to use the mains powered 20A bench PSU to support the solar panel load in winter.  Basically a monitor on the load/battery voltage which .. before the cutout... cut's in the PSU at 12V, 3Amp.  That will support the load, but not actually charge the battery (much).  When the voltage next hits 13V the PSU will be cut off.

For now I'm running it manually but will kill it before I head to bed.  I want it on a remote switch at least.  (I have checked the power supply and it's fine with a back current from the battery).

Panel: 50W
Battery 100Ah Sealed LA Marine (Start) battery.  Somewhere between an actual marine battery and a cheap leisure battery I believe.  3.5 years old.  Charged for well over 1000 days in a row.  Flattened completely twice to 9V.
Load:  Intermittent Boost/Buck converter powering 12V stuff, bench stuff, headphones, USB chargers, QC chargers etc.  This will expand.

The unsaid obvious is that the PSU will only supply current such that the battery remains at 12V.  It's above 12V, no current will be supplied.

[fourtytwo42]

Hi Paul,

Won't the bench psu trickle discharge the battery if it's attached yet unpowered ?

I commiserate about winter, I was trialing some new shading software and blew the main inverter, ever since waiting for replacement IRFB4227's that are virtually unobtainable AND we have the postal strike..................

42

[paulca]

Hi Paul,

Won't the bench psu trickle discharge the battery if it's attached yet unpowered ?

I commiserate about winter, I was trialing some new shading software and blew the main inverter, ever since waiting for replacement IRFB4227's that are virtually unobtainable AND we have the postal strike..................

42

To be tested.  The power supply is an RDTech jobby.  It's designed to tolerate LiPo charging by idiots.  So I know it will just measure the battery voltage and supply no current and take no current if the load is higher voltage that's output is set to.

I haven't tested if there is back current flow when the PSU is off.

In the spring I upgraded the load end of the system, running a 6mm cable to the office 10meters away so it's easier to access and use.  The plan was also to upgrade the panel from 50W to 330W and start looking for a bargain 100Ah lithium battery.

Life kinda got in the way with other expenses and I still haven't upgraded the panel.  So I kinda knew I was going to end up here and kinda knew I was going to have ato have a mains backup if I want to continue to use that 12V power rail over December and January.

At the moment, the battery is at 12.3V and charging at 3.3W.  The PSU is showing those 12.3V but 0.00A.  I haven't measured the waste power of the DC brick + PSU though.  It's probably a couple of watts!

[fourtytwo42]

Yup the sun is really low in the sky now, I ended up with my panels on the ground (after moving house) and all sorts of obstructions are throwing shade at them hence the new software trials.
Have you got yours on the garage/shed roof ?
I reckon even a small amount of elevation avoids some of the shading.

[paulca]

It's sitting on a garden table.  Pointed south.  But it has the house in the mornings and the hedge in the evenings.  It only gets about 4 hours direct sun in winter even if it's not cloudy!  Less surprising is, should the sun actually come out on a clear cold day around noon, it will still produce about 38W!

[paulca]

The plan was to buy a roof mount kit from a local solar supplier who will ship to DIY consumers.

For the 330W panel + twin row portrait rails to mount 3 such panels was around £300 delivered.  The fact my current charge controller will probably current limit at 200W even on a sunny winter day... can be taken care of later with a 40A upgrade £150.

[fourtytwo42]

For the 330W panel + twin row portrait rails to mount 3 such panels was around £300 delivered. 

Ouch!
Mounting hardware certainly adds up, fortunately I brought mine with me from the old house and just used 150mmsq fence posts to ground mount it, but should I buy some more panels (that I am considering) then I will also need more expensive mounting hardware. You see people on "off grid" outback style sites using all sorts like scaffold poles, angle iron and even plain timber but none of that would be allowed here even though it's probably perfectly serviceable 

It annoys me we have such draconian limitations otherwise I would put up a wind turbine too!

[paulca]

IIRC the panel was £170.  The mounting hardware, was the full kit with all the little bits and bobs that add up.  Rail joiners, rail end caps, edge and mid panel clamps, slate hooks, 3 inch wood bolts.  All that added up + VAT and delivery to the other 130 odd. 

It's only this year it has been practical to order such things.  Previously the only company that would sell them could only ship rails by custom courier in a van.  So they had to ship them to Waterford and then drive them in a van all the way to me 250 miles north.  So it cost £45 + VAT delivery, JUST for the 3 x 4m rails.  The other stuff came separate at £20 p&p.

Local "solar experts" are just "grant shops" who have barely got enough certification to put up a full system and don't want anything to do with even selling hardware to end users. 

My other plans for solar, at the lower level should start once my latest AliEx order arrives.  I'm going nano, micro and seeing if I can write a self sustaining MCU doing it's own MPPT on a tiny little panel.  So I ordered a range of tiny panels to play with.  From a 1.5V 10mA thing that's barely 3cm across to a 6"x6" 250mA 6V panel.  I figure I can play with buck convertors (or even boost) convertor MPPT circuits on a breadboard and hopefully not toast too many mosfets or MCUs!  Do that with the 50W panel I'm more likely to set the smoke alarm off!

[fourtytwo42]

You certainly have a good plan there Paul, mine is now a 1Kw high voltage (120V+) system and the mosfets are direct driven by the mcu. They blow frequently if I try to breakpoint the code......something that somehow slipped my mind hence the present problem.
The problem with scaling up can be the voltages involved, I have to use a very wide input range buck converter to provide the MCU power, this wasn't too bad till the power stage needed a/some fan/s THEN the wide range buck converter had to produce enough power for the fans as well......
As for the power stage I started off with just a "simple" boost for an immersion heater, then it became buck/boost to widen the operating range, then it became a transformer output full-bridge to enable inverter operation.
So what started of as simple, got very complicated indeed however it is reliable and does work as long as I am ultra careful tinkering with the MCU.
Good luck along your journey and of course have lots of fun

[mag_therm]

I have just finished the first part of a similar 12 UPS for the electronics bench, computers, ham radio gear, and network stuff.
The power is a home brew DC-DC converter 14V 20Amp. Link is 24 ~36 VDC
Battery is a #27 Marine, about 85 A.h. I decided not to use LiFe because they are only rated at about 50% capacity at float duty.
So would have cost about 8 times > than the Lead Acid
So far it is working well. Conducted noise is about -88 dBm @ 2 ~ 20 MHz, it has eliminated about 8 plugged switching power supplies from near the radio receivers. Load is less than I initially audited, about 2 A when the computers are idle

The remaining noise source is the 120V powered computer monitor which is radiating from the top corner right near a HF receiver.
Not sure what to do about that.

[paulca]

50% capacity at float duty?

By "float duty" you mean under a constant trickle?  A Life charger won't trickle.  It will just stop.  It might "ping" on and off once in a while.

If you are substituting in a LiFe in place of a lead acid, then as long as you pick you cell voltage alignment carefully the cells will pretty much just ignore your float voltage.  "Boost" phase will have charged them well above that anyway.  As long as your "float mode" is aware it might be the receive of the current!

I might be wrong, but I think 4 3.2V nominal LiFe cells gives you 12.8V nominal.  The cells I believe go up to a max of 3.8V?  Meaning a peak charge voltage of 15.2V, completely compatible even with an equalizer charge circuit.  The bottom end of 2.8-2.9V per cell also aligns with the general usage pattern of not taking it below 12V on load too.  You would need to set your load cut out to be strict about the 12V minima though.

[mag_therm]

I have read in a few places, and on one of the USA battery manufacturer's site, that LiFe are not intended for float.
Here is another  one I just found:
https://walkingsolar.com/can-you-float-charge-lifepo4/

For "float" I mean the old type of simple lead acid DC UPS where the charger takes the load with the battery floating in parallel.
In that case the LA battery floats at 2.25 to 2.3 Volt per cell, and come up to full charge like that. eg 13.8 Volt
So power failure has the full capacity available eg 85 A.h.  Battery can run like that for years with occasional outages.
The telepower and power generation station  UPS that I worked on in the past had Lead Acid battery life of 20+ years

If trying to do that with LiFe, there are 3 problems ( as I read):
1) The charge voltage to get full charge into battery needs to be about 14.5 V or more.
If floating at 13.8 V, the battery will only charge to about half capacity approx.

2) But Floating a LiFe continuously at 14.5 V +, causes chemical erosion problem as described in some sites.

3) To purchase a reasonable quality retail #27 battery, 85 A.hr cost me ~ $120 inluding core, expect 4 years+
When I compared reasonable quality LiFe, the same 85 A.h considering the reduced capacity would have cost  more than $800.
I looked at options of smaller 12V batteries in parallel or a single large one, and both cost more than $800.
And while the LiFe lifetime is better than Lead Acid for number of discharges,that does not matter much in a UPS.

[tautech]

FYI we downunder had the worst winter for solar in the last few years with well below normal UV levels resulting from all the crap in the stratosphere stemming from the big Tonga eruption a few months back.

This gives some insight to what's going on:



Winter, spring and early summer thus far (Dec 22) have been very wet here but UV levels seem to have recovered if our PV charge controller logging is anything to go by.