off grid solar e-scooter charging facility

bobc

10 kW
Joined
Jan 20, 2011
Messages
993
Location
Knutsford England
Hi, thought I'd share "what I've done" and what I'll do next.
A couple of years ago I bought 600W worth of cheap PV panels and bolted them to my south wall. I never intended to grid-tie & go for the feed-in-tariffs, this was always a standalone supply for my workshop. I was also given a load of old deep cycle SLAs. So I charge the SLAs direct from the solar and regulate by dumping spare electricity into resistors squeezed against my hot water tank. Until recently I only used the SLAs to supply LED lighting & a laptop in my workshop.

Last summer I switched off my gas supply, the water was hot enough to shower all summer.

This year I bought an emax scooter which I use for commuting. So I have to charge it overnight. The SLAs bridge this gap perfectly. But there's not any hot water any more..... And now winter is approaching, I don't get enough during the day to charge the scooter. BTW I just bulk charge the scooter using an ebay boost converter - this works great.

But - back to the plot - I need more juice from the PVs. So I'm going to make an MPPT converter. These are available for a few hundred quid (at the 600W level) but don't do quite what I want. This will net me 30 to 40% more power from the arrays. OK it won't run me through the depths of winter, but I won't want to use the scooter if it's so bad outside, and should get me hot water again in the summer!

I need to make a 50A, 25V buck converter, & I'll control it with one of the £1.50 arduino nano's
PVtankup.jpg
I'll document the converter on here in case anyone wants to re-create :)
 
Looking at the angle of the roof below the PV panels , IF that roofing material was a whitish (?) color, or best yet ( impractical) a mirror..... You would get increased output. Especially during low sun angle winter months.
 
bobc said:
Hi, thought I'd share "what I've done" and what I'll do next.
A couple of years ago I bought 600W worth of cheap PV panels and bolted them to my south wall. I never intended to grid-tie & go for the feed-in-tariffs, this was always a standalone supply for my workshop. I was also given a load of old deep cycle SLAs. So I charge the SLAs direct from the solar and regulate by dumping spare electricity into resistors squeezed against my hot water tank. Until recently I only used the SLAs to supply LED lighting & a laptop in my workshop.

Last summer I switched off my gas supply, the water was hot enough to shower all summer.

This year I bought an emax scooter which I use for commuting. So I have to charge it overnight. The SLAs bridge this gap perfectly. But there's not any hot water any more..... And now winter is approaching, I don't get enough during the day to charge the scooter. BTW I just bulk charge the scooter using an ebay boost converter - this works great.

But - back to the plot - I need more juice from the PVs. So I'm going to make an MPPT converter. These are available for a few hundred quid (at the 600W level) but don't do quite what I want. This will net me 30 to 40% more power from the arrays. OK it won't run me through the depths of winter, but I won't want to use the scooter if it's so bad outside, and should get me hot water again in the summer!

I need to make a 50A, 25V buck converter, & I'll control it with one of the £1.50 arduino nano's

I'll document the converter on here in case anyone wants to re-create :)

Constructive comments for your setup:

Are you currently running a charge controller at all? If not you are destroying your battery bank. You can get an MPPT Charge Controller on Amazon for like $89 on Amazon see link below (cheaper than you, me or anyone outside of China could ever build one).

https://www.amazon.com/ECO-WORTHY-Charge-Controller-Regulator-15-30/dp/B00FF1KGT4/ref=sr_1_29?s=lawn-garden&ie=UTF8&qid=1477448968&sr=1-29&keywords=mppt+charge+controller

Also you see that lower roof angle in your picture with the "3"on it? Your panels need to be facing south (which it looks like they may be) and mimicking the angle of that #3 roofline. Id be surprised if you ever got 200 watts out of that 600 watt panel setup with those things standing almost vertical as they are.

Lastly, the way you are heating your water is incredibly inefficient, get rid of that fire hazard. For the price of 1 decent panel you can have one of http://www.ebay.com/itm/Solar-Water-Heater-Collecter-Panel-Kit-Easy-hookup-free-shipping-Built-In-USA-a-/381559686147?hash=item58d6bac403:g:rQAAAOSw9r1V8aN4 which heats water better than that and frees up your power to charge your scooter or run appliances.
 
Also I never saw a increase in efficiency going from a PWM charger to an MPPT until I reconfigured my battery bank from 12 to 24v.


Heres my setup, it'll run about anything in my house except the electric dryer.

https://www.youtube.com/watch?v=TRnkC41b-Hc
[youtube]https://www.youtube.com/watch?v=TRnkC41b-Hc[/youtube]


build thread: https://endless-sphere.com/forums/viewtopic.php?f=41&t=82860&p=1235439#p1235439
 
Constructive comments for your setup:

Are you currently running a charge controller at all? If not you are destroying your battery bank. You can get an MPPT Charge Controller on Amazon for like $89 on Amazon see link below (cheaper than you, me or anyone outside of China could ever build one).

https://www.amazon.com/ECO-WORTHY-Charg ... controller

Also you see that lower roof angle in your picture with the "3"on it? Your panels need to be facing south (which it looks like they may be) and mimicking the angle of that #3 roofline. Id be surprised if you ever got 200 watts out of that 600 watt panel setup with those things standing almost vertical as they are.

Lastly, the way you are heating your water is incredibly inefficient, get rid of that fire hazard. For the price of 1 decent panel you can have one of http://www.ebay.com/itm/Solar-Water-Hea ... Sw9r1V8aN4 which heats water better than that and frees up your power to charge your scooter or run appliances.

1) I'm not going to 24V - I'd have to lose a panel to fit in the space & end up with an even number(so I can series them), or go boost charger with associated extra losses.
2) of course there's a controller on there - it's what controls the water heating! - when SLA voltage exceeds 13.5V it starts to dump amps into the water heating (pulsewidth modulated of course)
3) the wall mount was because my roof faces the wrong way. The angle gets better in winter, which is sort of nice, but shadows from neighbour's houses then start to be an issue. The wall mount is easy to install and best of a bad job considering my available space... BTW 200W - you're having a laugh, you must live in the tropics! North UK we're at 54degrees latitude, near vertical is not too bad........(sin 54 is 81%)
4) fire hazard :) thanks for your concern, but I see max temperatures around 60degrees - I have 30 metal clad power resistors clamped in parallel to 100 gallons of water in a copper tank, I'm actually pretty relaxed about it :) The water heating is a handy by-product of how the battery charging is controlled. Yes if I'd only wanted solar water heating I would have done it directly!!!
5) MPPT efficiency increase; at the moment the PV voltage is limited by the SLAs to 13.5V. With the MPPT in the PV voltage will be in the 18 to 20V zone with about the same amps. I presently max out at 30A into 13.5V - that's over 400W from a "600W" system. The MPPT should get that nearer 530W, - still less than 600 due to the suboptimal angle. I can't understand what you're trying to say regarding the battery voltage.
Overall it's nothing like an optimum system - but it's taken advantage of a load of cheap and free stuff & given me some fun at the same time - I'm pretty happy with it. The solar cells went in 2 years ago, the SLA's & charging control/waterheating a few months after that, & they've been working ever since. I've clearly not been too hard on the batteries because they still run the emax charger overnight OK
 
Wish i could get those minutes of my life back I spent discussing your setup. Ive got 400watts with another 4 panels going up this week. I live in San Diego.

Even panels used in Alaska don't have that steep of an angle. Gross.

Why not post up some pics of your setup (controller, batteries, wiring etc)?
 
Don't take the hump st - I still don't think you appreciate how far north the UK is:- San diego is at just 32N lat. A vertical PV there would effectively be at ~1/2 output. I'm at 54N lat, putting me at ~81% output (more 'cos I'm not actually vertical). FWIW anchorage alaska is at 61N lat, just about 1 degree north of the top end of UK mainland. There are lots of informative websites on the internet that will explain how the vertical angle of a PV might be chosen advantageously depending on geographical latitude. Further consider that I have an excess of power available in summer but need better output in winter.
panels used in Alaska don't have that steep of an angle. Gross.
Absolutely a slope would be better; but from a start at 81% there's not a lot to shoot at....
BTW I expect craneplaneguy is right about reflections from the roof below helping a bit :)
 
This is a good solar checker.
http://www.sunearthtools.com/dp/tools/pos_sun.php

In the UK we need a panel that can be tilted in winter and summer to keep the system operating efficiently around 65-70 degrees on December 21st and around 35-40 degrees on June 21st and always face true south not magnetic south (northern hemisphere) so they catch the noon peak energy point.
UK is not the best for solar panels, closer to the equator is best then theres less Adjustment needed to keep max efficiency through the season's.
UK is best for rain and wind a little wind turbine on the eave of the roof above the panels to back up the cells and a few forklift deep cycle cells for storage may be a good idea.

Some good basic info here.
http://solarpaneltilt.com
 
Years ago when I was off grid, I happened to notice the amps displayed on my charge controller was higher then I had ever seen before...... doesn't matter what it was (I sure don't remember, this was over 30 years ago) but higher then ever before is the point. A bright clear sky winter day, with lots of snow on the ground it was. My little PV array (300 watts, it seemed huge at the time) was mounted on a home made rack with a friction device to hold it at the correct angle, due to strong winds and or poor design, the rack had slipped, it was now vertical, not my preferred 60 degrees I liked in the winter months. Since the bottom of the array was only 24" above the snow covered ground, it was obvious to me that what was happening was a double refraction:the sun hitting the panels directly in the usual fashion, and also bouncing off the snow, in effect. Whatever, it worked! For sure, I'd think something whitish in color.....bright anyway, would give you a measurable amount of increased power.

Just for grins, try a large piece of white cardboard, something free anyway, simply laid on the roof (do it with no wind) and see what happens. Nowadays, with my large grid tied system and the lowered price of PV, my racks, still home made, are welded permanent at 45 degrees, close enough! I currently (pun intended) have over 13,000 KWH in my utilities system credited to my account, so the incentive to wring every last watt out of my array is less.
 
Hello bobc;

Interesting project you have going on, always fun to play around with solar and batteries.

From my experience working on solar systems, here are some tips for you, many solutions to improve your system, it is really up to what you want and your budget.

Regarding the mppt, I think you are overestimating the benefits it will provide you.

"MPPT efficiency increase; at the moment the PV voltage is limited by the SLAs to 13.5V. With the MPPT in the PV voltage will be in the 18 to 20V zone with about the same amps. I presently max out at 30A into 13.5V - that's over 400W from a "600W" system. The MPPT should get that nearer 530W, - still less than 600 due to the suboptimal angle. I can't understand what you're trying to say regarding the battery voltage."

The batteries are what define the voltage. Therefore, your mppt will not run at 18 or 20v, it will also run at 13.5V as well. You will also be limited to the same 400W power, because your system is limited by the current through the charge controller, 30A max I believe ?. If you switch to a 30A mppt, you will have the exact same issue. You need at least a 50A setup to get the full 600W to your battery in 12V.

What the mppt will do for you is provide a bit more energy in low light conditions and hot days. Usually, 10 to 15% increase, not much more except specific cases (near shadows or very unoptimized system in the first place, hard to tell without an accurate modeling of your system).

If you consider switching to 24V battery bank, with 30A you are more than fine, and you will be able to get the full charging power (600w) when needed. For about 120 pounds you can order a 30A mppt. Considering the price of charge controllers, that s usually a better financial option to get to 24V and these mppt can handle up to 100 or 150V therefore since you have 5 x 12V panels i think, no worries to get them all in series with such a converter, no need to loose a panel.

By the way, a buck converter is nothing like an mppt. A mppt is able to track a maximum powerpoint. Buck converter can work only as a current limiter, and can be setup on one particular max power point, but that is highly unefficient as the sun intensity always vary. With an arduino though, you could get it to become an mppt, but I would not recommend much this path except if you like playing with electronics for hours; in that case go for it, there s many info online about it.

Regarding making hot water from PV, typical solar heaters provide a much better efficiency per m2 of 70% vs 15-20% for the PV panels. You can always do it with your PV for sure, but in term of cost it is far from optimal. Now in your case, you have already the panels and the DC heater, so why not, but that really not the best way to use this excess: I would feed it to the house grid ( not resale it at all because that is way to complex legally speaking, but a simple grid tie inverter plugged to a wall plug), and it would lower your electricity bill by providing daytime running appliances with solar energy. I suppose you also have a electrical water heater, so I would tune this one to run when the electricity is fed in.

Next time if you can share your exact setup it would be more easy to comment, and also what are you trying to achieve. I m still wrapping my head around to know if you take showers in the winter ^^
 
Regarding the mppt, I think you are overestimating the benefits it will provide you.
This all depends on the particular PV setup. I know it's all temperature dependent, but if we say my array has open circuit voltage of 22V and max power point at 19V, then when i run it at 14V output I am only getting about14/19 of the available power. From a 600W panel that would be 442W. A bit more in reality as the current at 14V would be slightly higher.
An mppt is just a carefully controlled power converter of some sort, could be buck, boost, sepic, cuk, flyback, forawrd, whatever you want, the intelligence in the control is what tracks the MPP.
In my case the buck converter will take in the (up to) 31 amps at 19V from the PVs & convert it to 42A at 14V for the batteries. The PV array voltage is 19V - (it's maximum power point) - and the batteries are at 14V - (suitable charging voltage). No a simple buck converter is not a maximum power point tracking system. But it can be the 'muscle' in one. In the (made up) numbers above that is a 35% increase in charging current. If the buck converter is 95% efficient that would come down to a 29% increase.
I get that increase at medium and high power levels, less at low levels when the array open circuit voltage droops.
FWIW I've been developing these things sporadically since 1984 - that first one was a BP experimental PV station of about 100kW IIRC at Marchwood in surrey & I did the MPPT control for it when I was a postgrad at UMIST. Most recently I implemented MPPT in a (commercially available) industrial AC drive product specifically for remote off gird pumping applications in the developing world.
Yes, if I'd wanted to do the solar thing for monetary gain, I'd have got the full 5kW of solar arrays & bought ready made solutions from approved suppliers to qualify for the government's feed in tariffs. I'm not, this is for fun. This post was intended to be about how it works and how to design, not "look.. I've bought a box & plugged it in...." Am I posting in the wrong section?
 
Hello bobc,

Glad to see another person with experience in solar then !

Sorry to have been a bit harsh, but I have to say your system is quite inefficient overall from my perspective as a designer of an off grid system, with no interest in feed in tariff whatsoever.

You provided very little information that justify your particular technical solutions, so for me it is really hard to understand what and how.

I am sure you have your reasons, so it would be nice to understand them. What was the costs of the various components for instance ? Or why would you rather use the solar energy to make hot water, why you could divert it to lower your electricity bill, and create hot water with a very simple and cheap solar heater that would be paid by your electricity savings ?

It would also be quite interesting to have some efficiency numbers : for instance I would be really intrigued to know the efficiency of charge/discharge of your old sla ? Or Wh produced per day ?

Anyway, I just add to come in and say: please people don't think this is the optimal option for you! (especially the hot water part).
It can be technically done though, and you seem to have succeeded on that part as it works as it should.
 
Tried to find the old marchwood station on the internet - yes, looks like it was a very early prototype "solar farm" and operated successfully for 5 years before being dismantled in 1989. It is variously described as 80kW and 30kW so I assume either they removed some of the panels at some stage or are confusing kW with kWh.
solarBP.GIF
No mention of me or UMIST in my trawl, but I think we were brought in late on to 'make it work'.
It seems to be widely described as the first big grid tied solar installation in UK - anyone know of anything earlier
 
Started looking at the power converter design. Whatever I do will need an inductor that can take up to 40Adc without saturating.
And it has to work at several tens of kilohertz. It will be the biggest and most expensive thing in there.
Ebay and the catalogs have nothing so I thought I'd look at what I can make from a PM39 ferrite core. RS do a gapped core set (#2125016) with Al = 250nH in N27 material (good to over 100kHz) and Ae = 370mm^2 and costs £8.12
If I put 14 turns on that I get (250n x 14 x 14) 49uH. Magnetic flux density is B=iL/NA which at 40A is 378mT. That will do I reckon - the PM39 is a 39mm/50mm length/diameter cylinder, and the winding window should be OK for 14 turns at 40A.
If I want current ripple to be less than 10A pk-pk then from E=Ldi/dt, "off time" = 50u * 10/14 = 35.7us which would give pwm period of 35.7 * 19/(19-14) = 125us. So 8kHz PWM and this inductor would give 10A ripple
I'll aim for 20kHz PWM and (therefore) 4A ripple I guess

To control the "power point" of the PV array, the buck converter needs to control its input voltage (the array voltage), not the more customary output voltage control. Which is sort of why you can't get an "off the shelf" thing to do this. The loop is pretty similar but the control sense needs to be inverted. I still need the customary output control loop to control battery charging as well - fortunately the old unitrode 3843 switchmode chip lets you do this by adding an external opamp.
 
Here's the 50uH 40A inductor. I turned the bobbin out of a piece of scrap nylon I had laying about. The winding uses 6 strands of 1mm diameter copper, This will probably get warm with 40A, but hopefully not too warm
Ind50uH40A.jpg
 
My first time with kiCAD, I'll get this board routed on my the CNC. Plan is to use 2 sided board but just rout one side, the other (top) side will be the GND plane so GND nets will be soldered on top, the others underneath in the conventional way. kiCAD took a bit of getting used to, I was getting very annoyed with it at one stage, but when I got my head around their "philosophy" (DRC avoidance rather than checking/fixing) it sort of all made sense & progress was much more rapid.
mppt40.JPG
The schematic is included below. It's nice & simple but there is a thing to watch out for, the battery + is wired to the PV+ so the negatives are at different potentials. bottom line - if you use the SLA batteries to power the laptop & plug in the USB to the arduino on the board, some expensive smoke will come out of your laptop.......
View attachment mppt40.pdf
The charger SLA float voltage can be trimmed via USB, as can all the MPPT parameters (typically on P&O system you'd want to adjust voltage step size and step rate)
Regarding the use of old PTH tech, the plan is to make a descendant of this project into an educational tool at some stage, & leaded components are easier for the hamfisted to solder up....
 
Checked the saturation current level of the main inductor. I do this by discharging a chunky capacitor through it. The scope trace below shows the discharge current with a very obvious non-linearity at 42A - that's where the inductor saturates (the theory above said 40A). The LC circuit continues to ring afterwards but the current never goes back over the saturation level.
Current was measured using an old pearson probe (I think it's a calibrated CT + load resistor)
IMG_0065.JPG
 
Not had time to debug this yet, but here's most of the hardware with small bits checked (e.g. 5V supply output from 3843)
DSC01689.JPG
The arduino will go on as phase 2, the thing will basically work without it
 
I love the prototype it's awesome, Good informative thread wish I could add input to help but you have it all covered well done no flys on you.
 
I started writing the arduino code. It's simple enough for now and the PWM analog outputs are working correctly at 50kHz prf
It samples array volts and amps continuously during the hill climber step and uses an average power for step direction control. I hope this adaptive oversampling will give good performance, it's nice & easy to do anyway.
Once again the standard arduino PWM control for analog output is a bit pants so I used an extended PWM library somebody has done for free download. This gives me PWM at 50kHz so it's easy to filter & retain good dynamics. (at 50kHz there are 160 steps between 0% and 100%)
Note it is likely this code WILL become more complex to cope gracefully with sunrise & sunset (there's potential for these things to lock up trying to control to more array volts than the array's open circuit voltage) and to put USB data logging in there.
Code:
// solarmppt - maximum power point tracking controller for my PV array
// I have a 600W array, charging a set of 12V SLAs
// inputs to the arduino are
// A0 Battery voltage   1023 = 16V BVscal = 16/1024 =
// A1 PV array voltage  1023 = 35V PVscal = 35/1024 = 
// A2 PV array current  1023 = 54A PIscal = 54/1024 =
// A3 array V loop control
// outputs are
// D10 Battery volts trim 255 = 15V, 0 = 12V
// D9  Array volts trim   255 = 21V, 0 = 15V
// D13 LED red during bulk charge, off for float
// stepsize = array trim step size scaled like D11
// steptime = array step duration scaled in millisec

#include <PWM.h>

#define BVscal  0.0048        // multiplier for battery voltage ADC
#define PVscal  0.0006        // multiplier for array voltage ADC
#define PIscal  0.0006        // multiplier for array current ADC
#define steptime 1000         // hill climber step time in milliseconds
#define stepsize 4            // hill climber step size (down, 1 more up)
#define trimlim 160           // 160 is 100% duty cycle on D9 and D10 at 50kHz 
#define batterytrim 80        // battery float voltage trim
#define bulkthresh 1000       // when A3 > this, it's floating

byte  arraytrim;
float PVvolts;
float PVpowsum;
int   Nsamples;
float avepower;
float avepowprev;
byte  PVtrim;
boolean  steppingup;
int   timetarget;

void stepup() { //array voltage step up
  if ((PVtrim + stepsize) > trimlim) {
    PVtrim = trimlim;
  }
  else {
    PVtrim += stepsize;
  }
}

void stepdown() { // array voltage step down
  if (PVtrim < stepsize) {
    PVtrim = 0;
  }
  else {
    PVtrim -= (stepsize - 1);
  }
}

void setup() {
  // put your setup code here, to run once:
  // initialize serial:
  Serial.begin(9600); // serial control via the usb uart
  Serial.setTimeout(10000); // 10s serial timeout
  // set up ports
  pinMode(9, OUTPUT);   // array trim
  pinMode(10, INPUT);   // bsttery trim
  pinMode(13, OUTPUT);  // bulk charge LED
  //initiallise
  digitalWrite(13, LOW); // charger off
  //measure cells
  analogReference(INTERNAL);
  PVvolts = analogRead(A0) * PVscal; // dummy read
  PVtrim = 80;  // full width at 160 when freq = 50kHz
  InitTimersSafe(); //use PWM library
  SetPinFrequencySafe(9, 50000); // do hi F PWM for better dynamics
  analogWrite(10, batterytrim);
  analogWrite(9, PVtrim);
  Nsamples = 0;
  PVpowsum = 0.0;
  steppingup = true;
  avepower = 0.0;
  timetarget = millis() + steptime;
}

void loop() {
  do {
    PVpowsum += PVscal * analogRead(A1) * PIscal * analogRead(A2);
    Nsamples += 1;
  }
  while(millis() < timetarget);
  timetarget += steptime;
  avepowprev = avepower;
  avepower = PVpowsum / Nsamples;
  Nsamples = 0;
  PVpowsum = 0.0;
  if (avepowprev > avepower) {
    steppingup = !steppingup;
  }
  if (steppingup == true) {
    stepup();
  }
  else {
    stepdown();
  }
  analogWrite(9, PVtrim);
}
 
After spending lots of time tracking down bugs with the routed circuit above, I've lost patience with it and sent off £25 to china for some proper PCBs
With luck things will go a bit smoother when they arrive :)
 
Got to get my mpcnc working then I could have routed you a better circuit board as a one off. did some soldering today, am using little cat5e/6 sockets and breakout boards to send the stepper signals to the motors. They are all done now. Am off to berlin later to partake in a stag doo. wont be back till Wednesday. :shock: hopefully catch up soon.
 
Thanks Andy, but I'm done with routed PCBs now; flash whisker shorts & badly tinned joints. It has got a load worse with the higher temperature lead free solder. I'm not going to live for ever......
Have a great weekend! Best wishes to pips tom & ailsa
 
I like your DIY approach to even commonly available parts :)

I too would love to do an off-grid system for charging the scooter. Hard to justify the spend when it costs me 70c a day to run, but we all love a good project :) Plus we rent, so nothing ever seems worthwhile.

We have an instant electric hot water system - 20 amps per phase - 14 kW!
 
Back
Top