Wanting to get into solar PV? Think HOT WATER

[PoBoySolar]

Some may be feeling that they are missing the solar power bandwagon and wondering how they can do something actually useful without spending a lot of money. Batteries and grid tie make a system expensive with a long payback. Creating hot water is the most effective way to use a couple of PV solar panels since close to 100% of the power generated will be used. This assumes you are currently using a resistive hot water tank and not on demand as the rest of the civilized world. Never could get used to climbing into a shower with a 240V line.I have been heating hot water for years with excess PV for years at my camp. With as little as 800WH, we can both get a long hot shower at the end of the day. Just a couple panels totaling 400W can make a significant dent in electric use for heating. Under powering the heater assures 100% of the panels output is used. That is much greater than any battery application and much lower in cost than grid tie which isn't even allowed in some locations. Typical heat loss during the daylight hours of a well insulated tank is more than 1KWH. Solar makes up for that loss.

Most tanks have two heating elements. The top is the primary and only one operates at a time. That allows the lower resistive element to be used for solar and still have grid power backup. At 240V even a 5500W element is 10 ohms, a replacement 120V 2000W is about 7 ohms. That requires at least a 36V panel string toget any useful power. One positive is that may keep the system classified as "low Voltage". Connecting a panel directly to a resistance will result in power losses of more than half. A simple PWM circuit keeps the panel voltage at the the power point. A small capacitor bank stores the panels power in the off cycle. A panels power point voltage is the same regardless of light level. It only varies with temperature. Attaching a couple of flat pack diode bridges to the back of the panel is an easy way to get a reference voltage tracks panel temperature. I use a $5 UNO to control everything and the 490Hz PWM makes driving a FET easy. Almost any switch mode chip would work just as well.

I'm not going to argue the economics of it. Watching solar work is just plain fun and panel prices are so cheap now. Everyone has some space they can stick a few panels on. It won't freeze and none of the plumbing issues of direct solar heating. I've built several versions with just junk box parts. I have more details if anyone is interested.  This is a great little project for a UNO or 328MINI board. 

[PoBoySolar]

A capacitor bank is used to store the energy between PWM pulses.  An example of this is an array can put out 3A.  That is stored in the capacitor bank half the time.  With an ON pulse, 3A still comes from the panel and there is an additional 3A from the capacitor bank giving 6A total.  Capacitors have internal resistance and inductance.  Capacitors will heat up and fail if subjected to a lot of current over time.  For unknown commercial grade capacitors I like to figure a half amp each.  In my system I used about a dozen 330Uf 200V electrolytics from old PC power supplies.

The simplest method is to use a fixed set point.  The A/D value is compared with that and the PWMcount is adjusted up or down.  A dead band is needed to prevent continuous adjustment. When a major change is detected an additional change to the count is made to speed the transition. The upper and lower limits of the count are constrained.  Narrow pulses at each extreme only result in heating of the switching device and are prevented. Delay reading A/D at least 100ms after PWM is changed to allow system to respond. For a fish pond pump, the code is the same except the upper  PWM limit would be about 170. This can also be used as a diversion controller on a MPPT charge controller without diversion control.  A second A/D would monitor the battery and allow diversion at power point when the battery is over 13.8V. Code has been made intentionally simple for clarity.
 
The picture is a test bed I used to gather data from a single 12V panel. The 150 boost converter on the topis connected directly to the panel and was used to create 35V. This would not be needed with more than one panel in series. The UNO kept the panel at the power point.  The pot was adjusted for highest power on the meter.  Two 5V gate FETs were driven directly through 100 ohm resistors.  Sufficient at these levels. Prefer
driving with opto isolators which will protect the controller and provide a level shift.  Nothing more complicated is needed at these speeds.



// A resistive voltage divider produces about 2V-3V at pin A0
// to give A/D count of about 500
   
//  READ ANALOG VALUE AT PIN A0
panel = analogRead(0);                 
// A/D values go from 0 to 1023
   
   
// ADJUST PWM COUNT
   
// FAST RECOVER from high panel voltage at startup
if (panel > setpoint + 25) PWMcount = PWMcount + 1;
   
// FAST RECOVER from low panel voltage
if (panel < setpoint - -25) PWMcount = PWMcount - 1;
   
// NORMAL HIGH VOLTAGE RAMP UP
if (panel > setpoint + 2) PWMcount = PWMcount + 1;             
// voltage is over setpoint
   
// NORMAL LOW VOLTAGE RAMP DOWN
if (panel > setpoint - 2) PWMcount = PWMcount - 1;             
// voltage is under setpoint
   
   
// CHECK PWM LIMITS
// is count too high?
if (PWMcount >= 255) PWMcount = 255;               
   
// is count too low?
if (PWMcount <= 0) PWMcount = 0;

// Set count to output?
PWM3 = PWMcount;
   
// PREVENT NARROW DRIVE PULSES
     
// is count too high?
if (PWM3 >= 245) PWM3 = 255;               
   
// is count too low?
if (PWM3 <= 4) PWM3 = 0;
   
// PWM FET DRIVER OUTPUT PIN #3                         
analogWrite(3,PWM3);                       
// PWM values are between 0 and 255

a delay or additional program is needed

    Reply
    Quote
    Notify
    Remove

[PoBoySolar]

Wanted to try perturbing software but that is difficult to implement  in my situation.  I have a 50V buss that is fed power from both ends with multiple devices pulling from it.  Each device has an assigned power point voltage.  A slightly lower power point device has priority and is served first. I will implement temperature tracking this year sometime.  At a 50V potential a 20C panel temperature change only changes the power point 3V.  Plop that in the middle and the loss is quite small when adjusted seasonally.  Will likely implement a temp sensor this year as I stay at camp deep into October.  It is interesting to watch different devices switch power use to them and share on the basis of power point priority.

I will tell people to drive a FET hard and fast, but that isn't an absolute rule.  One poster burned through over $100 in FET and IGBT devices on his windmill driven hot water controller because he used a really good FET driver. I finally figured how he had it hooked up on the input made it very  susceptible to oscillating.  Heat occurs with every transition If you go from 50K a second down to 490, much of the problems are eliminated.  My heater drive consists of an opto isolator from 12V driving a gate pull down of 900 ohms.  That is driving three FET from an old UPS.  A lot of capacitance but plenty fast enough.  The LED portion is driven through a 300 ohm directly from the micro.  This driver is actually 30 feet from the micro using telephone line.  No problem at those speeds.