Author Topic: Solar panel control, a different approach?  (Read 5613 times)

0 Members and 1 Guest are viewing this topic.

dyslexicbloke

  • Newbie
  • *
  • Posts: 24
Solar panel control, a different approach?
« on: July 14, 2010, 06:01:00 AM »
Hi folks . Would appreciate your thoughts in respect of my latest musings.

I currently have 120w rated (I wish) array of solar panels on a 12 volt system.
Each individual unit is 15w and they are currently all paralleled and connected to my batteries via an off the shelf controller (Simple voltage regulator).

I am not short of space and the climate in the UK is more often than not cloudy or at least hazy so I use amorphous units as they appear to give better output in less than ideal conditions and handle partial shade much better than crystalline units. They are available for less than 3.00 / what if  you look around.

I am about to double my array size and will need a new regulator which set me thinking.

1. To avoid any power control losses perhaps I could simply disconnect some panels if the DC bus voltage were to go too high. Relays, at least magnetically latched relays with no holding current required, would do nicely. I know I would need a controller but a simple sequencer with a little histolysis is easy enough to build from opamps.

2. If each panel had its own relay they could be tiny . Thus cheep.

3. Is there a solid state alternative to this, just the power stage, which has similar negligible losses in a steady state.

4. Is there any benefit to switching pares of panels into a series configuration when solar radiation is low, could I possibly get usable power when conditions are bad. Anyone tried this?

Look forward to your comments,
Al

tsveti

  • Newbie
  • *
  • Posts: 3
Re: Solar panel control, a different approach?
« Reply #1 on: July 15, 2010, 02:20:32 PM »
 Hi dyslexicbloke,
   I am currently building my own charge controller. Your solar panels are designed probably for 17V nominal voltage. They probably have about 21V without load.
  In order to obtain max efficiency from the solar panels, you have to make them work on the maximum power point on the volt-amper curve. The max power point depends on the type of the panel and on the ambient temperature. It is not easy to design a controller which maintais all the factors. You could buy such a controller, but if I am right, you (like me) want to build this controller on your own.
   I decided to make a compromise and to buid my own controller which will not change the point according to the temperature and and type of the panel. The controller will change the current consumption from the panels in order to keep the voltage on the panel terminals =17V.
   In fact it is Switching PWM controlled STEP-DOWN converter with 2 inputs for PWM control. Through the first input I will control the output voltage - it will be stabilized at 13.8V for charging the accumulators. The second input will change the current through the panels so the  voltage on the panels remains = 17V.
   I use the obsolete PWM controller TDA1060, which have 2 additional inputs for PWM control, power PMOS transistors, inductor, shotky diode, low-inductive electrolitic capacitiors and some other components.
   I hope such a controller will be suitable for the initial experiments. If you are interested ,I could draw and publish the schematic when I finish and test the assembly.
   Regards
   
   

dyslexicbloke

  • Newbie
  • *
  • Posts: 24
Re: Solar panel control, a different approach?
« Reply #2 on: July 15, 2010, 04:49:00 PM »
I would be very interested ... thanks.
The only thing I will say at this point is that if you are tracking for panel voltage and letting the current float, which makes sense, then you will have to vary the charging voltage, hence current, to match the available wattage.
if an overcharge situation were to occur you will need to dump some load, which seems a little pointless, or reduce the panel current which would increase the voltage.

My point is .... you can only fix one side of the equation, if you track for max power from the array you must put it somewhere.

Sounds like a good starting point though.

Since you mention an inductor I assume you are aiming for a buck circuit, is that right?
I don't know much about this sort of system but I cant help wondering if PWM into a passive integrator wouldn't work just as well.
mmmm .... As I typed that I think I realised the problem, please tell me if I am correct.
an integrator would dissipate power from the charging resistor, whilst an inductor, presumably used in flyback mode wouldn't .. is that right?

Feel free to point out where I am going wrong if I am
Al

JeffD

  • Jr. Member
  • **
  • Posts: 50
Re: Solar panel control, a different approach?
« Reply #3 on: July 16, 2010, 12:10:38 AM »
Check out this system here http://www.timnolan.com/index.php?page=arduino-ppt-solar-charger.  I built a solar MPPT and battery charge controller in 2008 and then in 2009 discovered Tim Nolan's site who had done something very similar to what I had done.  The only difference was my buck converter wasn't synchronous and was less efficient.  I quickly upgraded my buck converter to his design which improved efficiency by about 12% over my old buck converter.  Using N-mosfets with much lower Rdson than the P-mosfet I was using made the biggest difference.

This is the algorithm I use when charging.  In the bulk stage (stage 1) of charging, the controller performs mppt on the solar input in order to get the most from the panels.  When the battery is in the absorption stage (stage 2) the controller throttles back on the current (decreases PWM duty cycle to the buck converter) from the solar panels in order to maintain absorption voltage based on battery temperature.  Once the absorption stage is completed then the current is throttled back more to maintain float voltage based on battery temperature.

Opera House

  • Sr. Member
  • ****
  • Posts: 264
Re: Solar panel control, a different approach?
« Reply #4 on: July 19, 2010, 02:36:51 PM »
If you operate the panels at 17V there will be absolutely no gain in power.   It will be a nice regulator.  You need to operate in the 35V range to get past the nominal voltage losses of the switching regulator.

ghurd

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 8070
Re: Solar panel control, a different approach?
« Reply #5 on: July 19, 2010, 03:30:32 PM »
First, forget relays.
There is no reason to use a relay.

Second, no offense intended, but if you are asking this question, then you should probably not get into DIY MPPT controllers.

Concerning question #3, yes there is a solid state alternative.  mosfets.
It is easy to use some type of controller to trigger a mosfet.
The mosfet can short the solar panel before the power reaches the battery.  It can be much faster and more accurate than anything with a relay.
By tiny and cheap, how does $1 and TO-220 size sound?  To me it sounds smaller and cheaper than a relay.  Plus it will basically never wear out.
This is my kit's version of it.
It shorts the solar power between the blocking diode and the panel.
It is an old school version of solar, but it works and seems to fit the requirements you are asking for.
Since mosfets use basically no power at all, then a single controller can switch many many (practically an unlimited number of) mosfets.
The only issue is keeping the amps in a mosfet low enough so the mosfet does not get too hot.
If you have Two 4-packs of the ebay Taiwan panels then you might consider something similar.



Concerning question #4, if the solar radiation is so low a single panel can not charge a battery, the placing them in series is not going to help much.
Certainly not worth the effort, cost and complexity.  Might be able to measure a difference, but it will be very tiny.

G-
www.ghurd.info<<<-----Information on my Controller

snake21

  • Full Member
  • ***
  • Posts: 143
Re: Solar panel control, a different approach?
« Reply #6 on: July 20, 2010, 12:29:14 AM »
wow........ghurd,you r simply the boss of controllers.in fact,i was looking something like that because i will be adding some solar panels to my system very shortly.
what i understood here is that the kit will cancel the power generated by the panels when the battery is fully charged.am i right.?

will this small system have negative effect on the panels?

i have one question,does solar panels need dump load?i believe  the answer is no

ghurd,please can you provide some additional imformation about how this system works,i mean,the solar one you just mentioned.

thanks

ghurd

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 8070
Re: Solar panel control, a different approach?
« Reply #7 on: July 20, 2010, 09:24:17 AM »
Hi Snake,

Correct.  The panel from the powers is 'canceled'.
It never makes it to the battery, therefore it does not need dumped.

It will not harm the panels.

The good thing about your system Snake, is you can use the same ghurd dump controller you are using now.
Simply add more fets and blocking diodes for the solar panels.  Connect the new fet gates to the existing fet gates.
No modifications to the controller or dump loads are required.

G-
www.ghurd.info<<<-----Information on my Controller

tsveti

  • Newbie
  • *
  • Posts: 3
Re: Solar panel control, a different approach?
« Reply #8 on: July 20, 2010, 03:08:53 PM »
Hello dyslexicbloke,
   I drew a simplified block diagram of my assembly. In fact it is not a MPP tracker. It allows connecting the solar panel to an accumulator which capacity must be chosen so the max PV current on the accu side does not exceed 10-20% of accu capacity. I do not use current limitation for lower power dissipation.
   There are some considerations:
 1. Both inputs of PWM controller have the same action. When the voltage on them increases, the PWM increases;
 2. Each of them limits the PWM of the other. For instance if the input1 sets a PWM 50%, the input 2 can only decrese the PWM below 50%. It can not increase the PWM above 50%. The same for the input 2.
   Now let's imagine that there is enough irradiation and the accu is discharged. The Accu voltage will be below 13.6V and the op amplifier2 will compare this voltage with Uref2 and will increase the voltage on ist out. This will increase the PWM and the current draw from the panels. In meanwhile the incresed current is going to decrease the voltage on the panel. The opamp1 will compare this decrease with Uref1 and result will be decreasing the voltage on opamp1 out(Uref1 is on the (-) input). The PWM and current goes down and PV voltage returns to its set value about 17V.

   Whet the accu is charged the voltahge on it will reach about 13.6V and opamp2 will compare it with Uref2 . Its out will decrease the PWM and current and this way the accu will not be overcharged. This opamp will keep the voltage on the battery about 13.6V.
   You are right, when the battery is charged, the current will be low and the PV voltage will be highgher. At this situation you have nowhere to put this power. I foresee overvoltage protection of PWM controller and the gate of MOSFET.
   I also will disconnect the accu during the night due to leakage current through the PV.
   Soon I will prepare the whole scheme to publish.
   This is not a MPPT, but it charges the battery, prevents overcharging, has acceptable efficiency(it is switched mode buck converter) and keeps the PV work around the nominal voltage. The input and output voltages are adjustable.
   I look forward for your comments.
Regards


   
i resized and made your photo display please keep photos on this site under 640x480 and read the posting photos sticky down in site news
Kurt
« Last Edit: July 21, 2010, 04:27:32 PM by kurt »

ghurd

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 8070
Re: Solar panel control, a different approach?
« Reply #9 on: July 20, 2010, 04:43:01 PM »

   I look forward for your comments.


I may be missing something, but it looks kind of complex for the potential gain.
Pots R1 and R2 seem to lock it into a basically fixed input and Battery voltages?

I applaud the effort even if I do not understand the circuit.
G-
www.ghurd.info<<<-----Information on my Controller

Opera House

  • Sr. Member
  • ****
  • Posts: 264
Re: Solar panel control, a different approach?
« Reply #10 on: July 21, 2010, 12:57:33 PM »
So what are the op amps for?  I think you are a long way from designing anything.

dyslexicbloke

  • Newbie
  • *
  • Posts: 24
Re: Solar panel control, a different approach?
« Reply #11 on: July 22, 2010, 05:14:25 AM »
Interesting stuff folks .... Thanks.

We don't have any mains power and my batteries are 6X 110A/h with a bench charge current of around 8A/cell ... more than the panels will put out.
I have 8x 15W panels and 2X 40w Panels, all amorphous, which I have found perform better in poor conditions than the more efficient mono ones.
Space / Size isn't an issue for me and there was little difference in cost / Wat.

I also have a generator, with a home grown auto start and stop circuit, that charges the pack if the voltage drops too far.

My intention was to have the two 40w panels, which i recently bought, connected all the time, I don't expect this will ever significantly overcharge the pack.
I just bought latching relays, no holding current, for the 15W panels that only cost 1 each, I think of them as remotely controlled switches.

I here what you are saying RE buck switching losses, it was one of the things I was looking for an opinion on and I am not surprised.
Obviously I will not be attempting to do PWM with relays .... They will however provide an easy method of switching panels on and off with effectively no losses.
I am going to try the series connection idea, it cant do any harm afterall, and I will post the results, good or bad.

Thanks for the advice thus far
Al

dyslexicbloke

  • Newbie
  • *
  • Posts: 24
Re: Solar panel control, a different approach?
« Reply #12 on: July 22, 2010, 05:37:41 AM »
G,
Thanks for your time so far ........
A couple of specifics if you would.

I have never used MosFets but was under the impression that they were tricky to drive.
I get that the gate is field effect and therefore has negligible currant requirements but I also thought that there would be significant power dissipation across them, hence the requirement for a heat sink.
Are these assumptions incorrect?

Why not relays as switches, are these not going to offer the least losses for anything that doesn't require rapid switching?
especially if there is no requirement to provide a coil current to maintain their state.

I fully understand the concept of MPPT, but have little idea when it comes to designing switch mode supplies, certainly not enough to make it worth playing with one.
Do you know of any resource that I could use to educate myself.
I am quite capable if handling the basic building requirements, analogue and digital, we all have to start somewhere you know.

Thanks in advance
Al

DamonHD

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3054
  • Country: gb
    • Earth Notes
Re: Solar panel control, a different approach?
« Reply #13 on: July 22, 2010, 07:54:25 AM »
FETs, if driven hard and of a high enough rating (and low enough on resistance), should generate relatively little heat.

You probably will want some kind of heatsinking for serious work, but it shouldn't be too horrific.

However, I'm speaking in theory here!

Rgds

Damon

ghurd

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 8070
Re: Solar panel control, a different approach?
« Reply #14 on: July 22, 2010, 08:54:28 AM »
Hi Al,

I think your assumptions are based on high voltage and high power systems.
A 12V system with a handful of amps is a different animal.

If you want the battery maintained at a fairly stable voltage, even within say 0.25V, then the relays will be switching pretty fast... and wear out.

High voltage, high amps, and high frequency make mosfet switching tricky.
It is pretty simple to switch in a 12VDC system.

Not sure what the contact rating (voltage drop or resistance) of your relay is,
but a common IRFZ44N is rated at 0.0175 ohms Max, typically under 0.014 ohms.

If that's not low enough for you, the $2 fet on my website, FDP7045L, is rated 0.0039 ohms typical, and that's with 50A going through it.
And it is even logic level so it is compatible with logic level ICs.
Sadly, that one is not suitable for anything but 12V systems.

The IRFZ44N is pretty much bullet proof in a 12V system, as long as
- Schottky flyback/freewheel diodes are included on inductive loads
- The Gate is driven to at least about ~12V
- The Gate is not allowed to float
- The (simplified) power dissipation is reasonable
(max ohms x I x I = ? Watts, which I like to keep it under about 1W, so a $1 IRFZ44N with a small heat sink should deal with as much as 8.5A without any problems)

I go overboard on keeping the heat in anything far below what most people do, but it works for me.
I learned if something is hot, it Will fail.  Eventually.
I also learned that relays Will fail.  Eventually.

Heat Losses?
Who cares?
In a (non-MPPT) reasonable controller design, the only time the fets are wasting heat is when the battery is at full voltage, and the controller is trying to stop the battery voltage from climbing by using or stopping the extra power anyway.  So the 1W of loss or heat is no issue.

"Do you know of any resource that I could use to educate myself."
Some general stuff about driving fets?  Maybe my original post about the ghurd controller.
It shows and explains how a 0 or 5V signal switches a logic level N-fet.
That N-fet switches a P-fet, which switches off-board power N-fets.
And it explains why it starts and ends with N-fet(s).
http://fieldlines.com/board/index.php/topic,129060.0.html

If you mean about MPPT, then probably a Masters degree in EE.   :-\
Things that seem totally unrelated can have a major effect on the circuit.  Which direction a cap sits on the circuit board, how close it is to other components, even the length and layout of the traces on the circuit board, can create all kinds of confusing issues.
There were a couple postings here on FL about modifying computer PSUs to MPPT or switching supplies.

MPPT falls in a dead zone of what I play with, and is something I would rather pay a couple $hundred for a factory made unit that works properly instead of several $hundred and countless hours making one that does not work as well.

BTW-  I expect the series panels, in decent sun, will put out about 10% more than a single panel.  45% loss compared to standard parallel connection.
In poor sun, series may make a couple dozen ma in conditions where the parallel panels would have just stopped making any ma, but those conditions will not last long (measured in minutes?).
I expect the control and monitoring circuit will use more power than switching the configuration gains per day.
G-
www.ghurd.info<<<-----Information on my Controller

dyslexicbloke

  • Newbie
  • *
  • Posts: 24
Re: Solar panel control, a different approach?
« Reply #15 on: August 14, 2010, 09:01:23 AM »
Series panels ......
I did some test and confirmed the general concensus here ... not worth the effort for the gain.
That said the latching relays did what I needed turning individual panels on and of as needed.
RE High frequency switching with relays .... I never thought this was an option, I know it would fail.

I have played with FET's though, my genny field control module failed so I had to build a new one ... all in all not as tricky as I thought, only two blue smoke events and my genny is working again.

mmmm .... now I am off topic. Since I now like FET's I am going to try and build a PWM charge controller but I guess that is a new thread.
Thanks for all the input folks, it was a big help

Yyrkoon

  • Jr. Member
  • **
  • Posts: 84
Re: Solar panel control, a different approach?
« Reply #16 on: September 26, 2010, 06:41:59 PM »
So what are the op amps for?  I think you are a long way from designing anything.

I am not an electronics genius, but I am assuming the op amps are used as comparators. Or more correctly, used to sense voltage on the input, and output stages of the circuit. After Which ,  the op amps tells the PWM what duty cycle, and frequency the FET should operate at.  All pretty much basic Buck circuit knowledge. That is, up until the use of the op amps, and the use of a PWM to directly drive a MOSFET   ???  But who knows. Maybe the author kept it simple for one, or more various reasons.

All of this is really not a bad idea, assuming it does work. And, really is not that far off from a MPPT. Just need a MCU, a bit of code to monitor / control various things, and Viola!

OperaHouse

  • Hero Member
  • *****
  • Posts: 555
Re: Solar panel control, a different approach?
« Reply #17 on: September 28, 2010, 11:43:55 AM »
The switching chip has these op amps built in so they serve no purpose.  As I said, this design is long on the leaf and short in the can.