Emulating solar battery function
Emulating solar battery function
This is basically an electronics question about matching solar panels to a battery (and/or direct to a load) in such a way as to maximise overall power from the panel.
If you hook up a solar panel to a battery (such as a lead acid battery) the panel output voltage will tend to sit at the current battery voltage (during charging phase). Unfortunately this may mean that the panel is not outputting it's full Watts potential.
Here is why:
The watts (power) output of a panel is given by Volts x Amps.
For a typical panel (hooked to lead acid battery) during a charging phase this may be something like 11.9v x 4.8A = 57.12Watts
A solar panel will usually have an output current that is fairly constant throughout it's voltage range - so short circuit current may be 5A, current at 12v may be 4.8A and current at 18v may also be 4.8A, but currents will drop at voltages that are ABOVE the peak output voltage.
If we calculate the wattages being produced by the panel at some different voltages we see the following:
Ov (short circuit) x 5A = 0W
1v x 5A = 5W
5v x 5A = 25W
12v x 4.8A = 48W
18v x 4.8A = 86W
20v x 3.9A = 78W
Most "12v" solar panels have a "peak power" output at around 18volts - so if I buy a 100W (12v) panel that has it's peak output at 18v it will put out the following current:
100W / 18V = 5.55A
However, if the battery is at 11.9volts then the ACTUAL power coming out of the panel is 5.55A x 11.9v = 66W which is noticeably lower than the potential 100W this panel can produce when allowed to run at it's maximum power point.
The problem here is that a lead acid battery has internal characteristics that end up wasting a significant amount of the possible power from the panel.
And if we add a PWM solar controller in series with the batery we get even less power out of the panel because the controller will "taper" the charge rate to safeguard the battery.
Those of you who know about solar charging will by now be screaming "use an MPPT controller dummy!!"
Yes its true - an MPPT charge controller will allow the panel to remain at 18V (or whatever the controller measures as the peak output voltage of the panel or array) - HOWEVER - the total power output is still limited by the battery voltage and it's charge state (sometimes a battery cannot accept faster charege without being damaged - and the controller understands this).
So - how can i "replace" a standard 12v lead acid battery with something that emulates an 18v battery (or whatever my particular solar panel lists as it's preferred maximum power point)?
I could make up a new battery from qty 9 individual "2v" cells - the same way that a "12v" battery is made up of qty 6 "2v" cells (ok in reality it is something like 2.2 volts per cell...). If I did this and my solar controller was designed for corrrect charging of an 18v battery it would allow me to better maximise the total power out of my panels.
However - I am not likely to find a controller that is happy with an 18v battery, and I am equally unlikely to find an 18v radio, TV, LED light, fan, etc etc.
So - back to my question: How can I emulate a storage battery (of a specific voltage) without being tied to the somewhat limiting chemistry of a 12v lead acid battery?
It would be nice to be able to hook a solar panel (with nominal Max power poit of 18v or so) direct to a storage unit (battery emulator) that sat happily at 18v
An ultracapacitor (supercapacitor) can absorb the 18v without dragging it down to 12v - but supercapacitors have the opposite problem to what we see with lead acid batteries - the capacitor voltage will float as high as the panel can go - which might be 20volts for a "12v" crystalline panel or even 30volts for a "12v" amorphous panel. Which destroys electronics of course.
So is there any way I can hold the output voltage of a panel to it's max power voltage?
Is there any way to emulate an MPPT controller in a "manual" way - so that instead of the MPPT controller intermittently switching off and going throught its normal calibration and sampling routine i could set a predetermined voltage point which i know matches the label on the panel (or array of panels)?
One of the uses I have for my panel array is to power a small domestic air conditioner in the height of summer - but for this I need to ensure that my panels are running at peak wattage (ie panel voltage floats to 18v). I need an 18v "battery" in the system to allow the panels to stay at this voltage, but also a "12v" battery (without the usual lead acid chemical inertia) to interface the 12v inverter that will be producing the mains 240v.
Here is my schematic:
(PLACEHOLDER for image)
(Dammit the forum won't accept multiple space charcters so I can't draw my ascii art schematic...) (see below)
NOTES:
1) The DC-DC converter must be smart enough to never drain the emulated battery below 17.5v
2) It would be great if the 12v battery on the front end of the mains inverter could be replaced by an emulated battery or eliminated altogether. Maybe there is a controller somewhere that is capable of combining several of the functions shown in the schematic?
If you hook up a solar panel to a battery (such as a lead acid battery) the panel output voltage will tend to sit at the current battery voltage (during charging phase). Unfortunately this may mean that the panel is not outputting it's full Watts potential.
Here is why:
The watts (power) output of a panel is given by Volts x Amps.
For a typical panel (hooked to lead acid battery) during a charging phase this may be something like 11.9v x 4.8A = 57.12Watts
A solar panel will usually have an output current that is fairly constant throughout it's voltage range - so short circuit current may be 5A, current at 12v may be 4.8A and current at 18v may also be 4.8A, but currents will drop at voltages that are ABOVE the peak output voltage.
If we calculate the wattages being produced by the panel at some different voltages we see the following:
Ov (short circuit) x 5A = 0W
1v x 5A = 5W
5v x 5A = 25W
12v x 4.8A = 48W
18v x 4.8A = 86W
20v x 3.9A = 78W
Most "12v" solar panels have a "peak power" output at around 18volts - so if I buy a 100W (12v) panel that has it's peak output at 18v it will put out the following current:
100W / 18V = 5.55A
However, if the battery is at 11.9volts then the ACTUAL power coming out of the panel is 5.55A x 11.9v = 66W which is noticeably lower than the potential 100W this panel can produce when allowed to run at it's maximum power point.
The problem here is that a lead acid battery has internal characteristics that end up wasting a significant amount of the possible power from the panel.
And if we add a PWM solar controller in series with the batery we get even less power out of the panel because the controller will "taper" the charge rate to safeguard the battery.
Those of you who know about solar charging will by now be screaming "use an MPPT controller dummy!!"
Yes its true - an MPPT charge controller will allow the panel to remain at 18V (or whatever the controller measures as the peak output voltage of the panel or array) - HOWEVER - the total power output is still limited by the battery voltage and it's charge state (sometimes a battery cannot accept faster charege without being damaged - and the controller understands this).
So - how can i "replace" a standard 12v lead acid battery with something that emulates an 18v battery (or whatever my particular solar panel lists as it's preferred maximum power point)?
I could make up a new battery from qty 9 individual "2v" cells - the same way that a "12v" battery is made up of qty 6 "2v" cells (ok in reality it is something like 2.2 volts per cell...). If I did this and my solar controller was designed for corrrect charging of an 18v battery it would allow me to better maximise the total power out of my panels.
However - I am not likely to find a controller that is happy with an 18v battery, and I am equally unlikely to find an 18v radio, TV, LED light, fan, etc etc.
So - back to my question: How can I emulate a storage battery (of a specific voltage) without being tied to the somewhat limiting chemistry of a 12v lead acid battery?
It would be nice to be able to hook a solar panel (with nominal Max power poit of 18v or so) direct to a storage unit (battery emulator) that sat happily at 18v
An ultracapacitor (supercapacitor) can absorb the 18v without dragging it down to 12v - but supercapacitors have the opposite problem to what we see with lead acid batteries - the capacitor voltage will float as high as the panel can go - which might be 20volts for a "12v" crystalline panel or even 30volts for a "12v" amorphous panel. Which destroys electronics of course.
So is there any way I can hold the output voltage of a panel to it's max power voltage?
Is there any way to emulate an MPPT controller in a "manual" way - so that instead of the MPPT controller intermittently switching off and going throught its normal calibration and sampling routine i could set a predetermined voltage point which i know matches the label on the panel (or array of panels)?
One of the uses I have for my panel array is to power a small domestic air conditioner in the height of summer - but for this I need to ensure that my panels are running at peak wattage (ie panel voltage floats to 18v). I need an 18v "battery" in the system to allow the panels to stay at this voltage, but also a "12v" battery (without the usual lead acid chemical inertia) to interface the 12v inverter that will be producing the mains 240v.
Here is my schematic:
(PLACEHOLDER for image)
(Dammit the forum won't accept multiple space charcters so I can't draw my ascii art schematic...) (see below)
NOTES:
1) The DC-DC converter must be smart enough to never drain the emulated battery below 17.5v
2) It would be great if the 12v battery on the front end of the mains inverter could be replaced by an emulated battery or eliminated altogether. Maybe there is a controller somewhere that is capable of combining several of the functions shown in the schematic?
- Attachments
-
- BatteryEmulationRequirement.jpg
- (16.98 KiB) Downloaded 161 times
Last edited by greengeek on Sat 11 Apr 2020, 20:53, edited 4 times in total.
Re: Emulating solar battery function
Put your ascii art (with space characters) inside code.../code delimiters (with square brackets round code and /code) to preserve space markers.greengeek wrote:(PLACEHOLDER for image)
(Dammit the forum won't accept multiple space charcters so I can't draw my ascii art schematic...)
wiak
WeeDogLinux forum: https://weedoglinux.rockedge.org/viewforum.php?f=4
Tiny Linux Blog: https://www.tinylinux.info/
Check Firmware: http://murga-linux.com/puppy/viewtopic.php?p=1022797
Tiny Linux Blog: https://www.tinylinux.info/
Check Firmware: http://murga-linux.com/puppy/viewtopic.php?p=1022797
Re: Emulating solar battery function
wiak wrote:Put your ascii art (with space characters) inside code.../code delimiters (with square brackets round code and /code) to preserve space markers.
wiak
Code: Select all
----------
| ----------------
| | Mains |
| + - ------------------ | Inverter | /---------\
"12v" | -------- | | |------- | | |
solar |_________| 18v |__________| Highpower |_________| | |___| Aircon |
panel | 18v out | store | 18v out | DC-DC | 12v out | 12v | | | |
| |________| | converter | | Batt? | | \_________/
| (emulated) | | | | |
---------- (battery) ------------------ ----------------
-
- Posts: 721
- Joined: Sat 31 Mar 2018, 08:01
- Location: Rakaia
- Contact:
Re: Emulating solar battery function
Code: Select all
----------
| ----------------
| | Mains |
| + - ------------------ | Inverter | /----------\
"12v" | ----------- | | |--------- | | |
solar |_________| 18v |__________| Highpower |_________| | |______| Aircon |
panel | 18v out | store | 18v out | DC-DC | 12v out | 12v | | | |
| |_______| | converter | | Batt? | | \________/
| (emulated) | | | | |
---------- (battery) ------------------ ----------------
Just a small correction.
!2v lead Acid batteries actually produce 13.2 volts (6 x 2.2v) when in good condition and require a charger to supply over 14 volts to charge.
A quick charge will supply a higher voltage.
Most modern battery chargers work on the current through the battery rather than the voltage .
These voltage I gave were those shown by a digital charger, not a stand-alone volt meter which may show slightly lower figures, though not by much.
!2v lead Acid batteries actually produce 13.2 volts (6 x 2.2v) when in good condition and require a charger to supply over 14 volts to charge.
A quick charge will supply a higher voltage.
Most modern battery chargers work on the current through the battery rather than the voltage .
These voltage I gave were those shown by a digital charger, not a stand-alone volt meter which may show slightly lower figures, though not by much.
"Just think of it as leaving early to avoid the rush" - T Pratchett
Between the Solar battery and the 18V storage you need a reverse polarity high power diode to prevent "back charging" the solar cells. The "|" is positive.
---->|-----
---->|-----
Linux user #498913 "Some people need to reimagine their thinking."
"Zuckerberg: a large city inhabited by mentally challenged people."
"Zuckerberg: a large city inhabited by mentally challenged people."
- BarryK
- Puppy Master
- Posts: 9392
- Joined: Mon 09 May 2005, 09:23
- Location: Perth, Western Australia
- Contact:
greengeek,
I read your first post, but cannot see what the problem is.
If you have a MPPT controller, it will translate the panel output, let's say 18V at 5.5A, to whatever it charges the battery at, say 13.5V, at 7.4A (slightly less than that, due to losses in the controller).
The full power from the panel is being extracted. There is no reason for the MPPT controller to throttle the charging current back, if the deep-cycle battery is of adequate capacity.
Note, I have used a deep-cycle 45AH AGM battery for years, have just started a project to go over to lithium:
https://bkhome.org/news/
Is there something about your post that I have not understood?
I read your first post, but cannot see what the problem is.
If you have a MPPT controller, it will translate the panel output, let's say 18V at 5.5A, to whatever it charges the battery at, say 13.5V, at 7.4A (slightly less than that, due to losses in the controller).
The full power from the panel is being extracted. There is no reason for the MPPT controller to throttle the charging current back, if the deep-cycle battery is of adequate capacity.
Note, I have used a deep-cycle 45AH AGM battery for years, have just started a project to go over to lithium:
https://bkhome.org/news/
Is there something about your post that I have not understood?
[url]https://bkhome.org/news/[/url]
Yes, sorry it may not be very well explained.BarryK wrote:I read your first post, but cannot see what the problem is.
That is where the problem lies - the MPPT will ALWAYS throttle back the charge rate UNLESS the battery is so big that it can swallow everything the panel puts out.If you have a MPPT controller, it will translate the panel output, let's say 18V at 5.5A, to whatever it charges the battery at, say 13.5V, at 7.4A (slightly less than that, due to losses in the controller).
The full power from the panel is being extracted. There is no reason for the MPPT controller to throttle the charging current back, if the deep-cycle battery is of adequate capacity.
Trouble is - with lead acid batteries - having a large enough battery capacity in order to swallow maximum charge on a sunny day means that there is significant battery degradation due to undercharge on a cloudy day.
This is what I call the "inertia" of the lead acid battery system - the battery chemistry never exactly matches what the user needs from it. You end up having to waste money and energy building the system around the "health" of the battery.
With Lead aicd batteries you just can't win.
What I need to do is build an alternative storage medium that does not waste effort tapering charge and measuring battery charge state.
One of the options I thought about is a bank of supercapacitors surrounded by some form of electronics that will limit the discharge rate of the supercapacitors. I hoped that this might be available (or in development) somewhere since EV cars are now so common and they have high efficiency requirements - obviously not suitable for lead acid and all it's problems. Other alternatives are solutions such as heavy mechanical flywheel, molten salt storage, compressed air etc etc but they are probably beyond my means.
But how efficient is the MPPT/leadacid combo in reality I wonder?Note, I have used a deep-cycle 45AH AGM battery for years, have just started a project to go over to lithium
A 45AH battery should really only release about 30AH in one cycle or else life cycle is reduced. So here are my rough calculations:
100W panel in full sun for 8 hours = 800WH. If the MPPT controller converts this to 12v (nominal) then the available output is 800/12 AH
(= 67AH)
However, if this is hooked up to the 45AH hour battery (depleted to 15AH from previous day's use) then only 30AH of the available 67AH gets used.
However if the battery is completely discharged then we end up using 45AH of the 67AH available - but this is still wasting a lot of available power and you may well be replacing that battery every year or two (as has been my experience) unless you are lucky enough to find extremely high quality batteries that can tolerate tha level of discharge.
Of all my batteries I have only found Yuasa to be reliable long term but i bought them years ago and can't find their equal now.
Your suggestion of using Lithium instead of Lead acid is a logical step - although I am aiming at a panel array producing enough energy for a 1000W portable airconditioner and I freak out at the thought of using Lithium for that sort of loading. (Finding a suitable Battery Management System would be a pre-requisite).
Which leaves me with the idea of directly hooking the panels to electronics that avoid the losses inherent in the charge cycle. Problem is - the electronics need an energy storage system that smooths out variations in both load and supply - which is the need I want to solve.
I have qty12 150W panels so I should in theory have more than enough power to run the load - but I just need to ensure the "storage medium" keeps the panels running at their 18v maximum power point. I don't see an MPPT doing that unless I hook it to Lead acid batteries - unless there is a lithium MPPT available?
I feel that the Nissan Leaf drive system has something to offer in terms of its power handling capacity - except that the Leaf system voltage is much higher than I am prepared to risk.
Last edited by greengeek on Tue 14 Apr 2020, 20:42, edited 1 time in total.
And thanks for the link to your lithium project. Lots of good info there.
Especially the comments re lead acid limitations:
https://bkhome.org/news/202004/portable ... mping.html
Especially the comments re lead acid limitations:
https://bkhome.org/news/202004/portable ... mping.html
Yes that will work to some extent - although as BurnIT pointed out the actual battery voltage is higher than nominal so 3x "6v" in series would probably settle at something more like 19v. Definitely useable but not quite ideal.dancytron wrote:Three 6 volt tractor/golf cart batteries hooked up in series?
The big problem with lead acid batts is the amount of wasted energy in trying to match the panels to the required charge/discharge requirements of the battery chemistry.
Worrying about battery health destroys the efficiency of the system and the potential power output of the panels
You mean two panels in series? Yes, that would give me a system voltage of 24v which is a good idea because there are less transmission losses in higher voltages.Smithy wrote:How about two 12v solar panels and a step down converter...
However - this will still give me the same issues of needing some form of storage medium that holds the panels at their max power point - which would mean a storage medium voltage of 36v instead of 18.
Same issue - just a different voltage.
If I can find a step don converter (DC=DC converter) that is capable of handling the currents I need then I assume it may have large capacitors inside as well as the robust regulation circuitry that may be adequate for my needs.
The problem seems to be that inverters and control electronics seem to be designed around the voltage ranges specific to lead acid charging.
Now that Lithium is on stream maybe there is a wider choice of voltage converters and inverters - but do they have the ability to be set for voltages like 18v?
That is my issue.
Not 10000w, just 1000w. I don't have it running yet - I have been buying new panels over the last year to replace my older, much smaller setups. I looked at some portable air con units this summer and the smallest I saw was 750w but most were around 1000w so I reckon it should be doable with a direct connection to the panels on a hot sunny day (next summer!). If I can sort the right electronics.Smithy wrote:So that’s like a 1 bar electric fire in old money. 1kw.
10000w right? That’s a big output. Does it run on car/jet/boat batteries ok? Besides the loss that you mention with lead acid.
Periodic use or 24/7?
I would be setting it up to start the air con automatically when the panels are at full output and the air temp is upwards of about 25 degrees C.
For my other loads, like lighting and PC etc I have a couple of 150AH lead acid batts in parallel and I will keep those as a separate system. I also have two 100AH batts in series (24v system) for power tools and heavier intermittent loads. I did have an Engel fridge on there at one point but it was poorly insulated and ran constantly, draining the batteries too low on cloudy days.
Really hoping I can get my monster array up in time for next summer.
Right (oops added a nought).
Voltage regulators/ step down converters can be built fairly easily or you can get them cheap on the web, I would think the simplest path to aircon would yield the most efficiency, heatsinks spring to mind,
I only do little fountains and fairy lights lol, but I did manage a lake district type cascade waterful and stream into a pond for someone, but it was mains powered. Cool project greengeek.
Voltage regulators/ step down converters can be built fairly easily or you can get them cheap on the web, I would think the simplest path to aircon would yield the most efficiency, heatsinks spring to mind,
I only do little fountains and fairy lights lol, but I did manage a lake district type cascade waterful and stream into a pond for someone, but it was mains powered. Cool project greengeek.
- Moose On The Loose
- Posts: 965
- Joined: Thu 24 Feb 2011, 14:54
Just a few comments about optimizing charging in a home brew system.
You want to have either a panel that makes just under the batteries voltage or just over. Having to do one where the source may be greater or less is always going to lead to a less efficient charger. (I will side bar to explain why) I will assume that the battery voltage is higher than the panel for the description of the method but the same idea translates to other cases.
Begin side bar
Generally non-isolated DC-DC converter designs relay on the fact that the average voltage across an inductor is zero in the long term. This part is often called the "working inductor". One end of the inductor is rapidly switched between the higher of the two voltages and ground with a duty cycle that makes the average on that end equal to the DC voltage on the other end. If which of the sides has the higher voltage is not known, additional switching parts are needed to switch on either end of the inductor. These parts add losses.
Yes, I ignore CUK and SEPIC designs here
End side bar
Assuming the battery is not currently charged and can take all the power the panel can give, you want to maximize the charging current into the battery.
Ignoring current limiting for the moment: The switching parts of the booster are driven by a PWM signal or signals. To boost the voltage more, the switch to ground is left on for a greater percentage of the time.
You need some way of sensing the current in the battery. They make ones that work from the magnetic field the current produces. They are idea for this sort of case. Accuracy is less needed than repeatability.
The design does this:
At about a 1 Hz rate wobble the PWM percentage up and down a bit while watching the current. Which direction gives the greater charging current is the direction you want to move the average PWM slightly.
The difference in the two gives a measure of how far to change the PWM average but given how slowly the situation changes, you can just increment and decrement if you want to avoid servo design topics.
Where Puppy Linux comes in:
1) You can run LTSpice and then ExpressPCB under wine to make your own hardware.
2) It is fairly easy to make your circuits interface via RS232/USB to a PC running Puppy Linux. This PC can do much of the dithering and average setting etc in a bash script running in the back ground. The user interface could be provided via a tiny web server or gtkdialog.
You want to have either a panel that makes just under the batteries voltage or just over. Having to do one where the source may be greater or less is always going to lead to a less efficient charger. (I will side bar to explain why) I will assume that the battery voltage is higher than the panel for the description of the method but the same idea translates to other cases.
Begin side bar
Generally non-isolated DC-DC converter designs relay on the fact that the average voltage across an inductor is zero in the long term. This part is often called the "working inductor". One end of the inductor is rapidly switched between the higher of the two voltages and ground with a duty cycle that makes the average on that end equal to the DC voltage on the other end. If which of the sides has the higher voltage is not known, additional switching parts are needed to switch on either end of the inductor. These parts add losses.
Yes, I ignore CUK and SEPIC designs here
End side bar
Assuming the battery is not currently charged and can take all the power the panel can give, you want to maximize the charging current into the battery.
Ignoring current limiting for the moment: The switching parts of the booster are driven by a PWM signal or signals. To boost the voltage more, the switch to ground is left on for a greater percentage of the time.
You need some way of sensing the current in the battery. They make ones that work from the magnetic field the current produces. They are idea for this sort of case. Accuracy is less needed than repeatability.
The design does this:
At about a 1 Hz rate wobble the PWM percentage up and down a bit while watching the current. Which direction gives the greater charging current is the direction you want to move the average PWM slightly.
The difference in the two gives a measure of how far to change the PWM average but given how slowly the situation changes, you can just increment and decrement if you want to avoid servo design topics.
Where Puppy Linux comes in:
1) You can run LTSpice and then ExpressPCB under wine to make your own hardware.
2) It is fairly easy to make your circuits interface via RS232/USB to a PC running Puppy Linux. This PC can do much of the dithering and average setting etc in a bash script running in the back ground. The user interface could be provided via a tiny web server or gtkdialog.