solar powered water boiler

I am working on a simple solar system for heating water. The idea is to use sun power and if not enough, use the power grid at night. I have a solar panel and a MPTT boost converter which collects the power from the panel and sends it to the boiler. I am using a timer controlled mechanical SPDT relay to switch between the solar and the power grid - grid at night, solar at day. My problem is that my controller sometimes burns out during the switching of the SPDT relay. What I suspect is that there is a shoot through during the switching time, which places 220V AC from the grid on the MPTT controller output. The MPTT controller is designed for 100V DC max. I noticed also that this does not happen always. It probably depends on the point at which the relay cuts the 220V sine wave, because the relay sometimes sparks, sometimes not. The shoot through is always accompanied with a spark. Is it possible, that the spark (when it happens) crosses the air gap of the relay contact? Is there a simple way to solve the issue? (The overkill for me would be to build my own time delayed spdt relay)

  • \$\begingroup\$ You could use two SPST relays with a RC= time delay for delayed turn on for each of say 100ms using a diode for fast turn off. Using NC in the AC position and power from battery for NO terminals. But is 18V enough to keep the boiler warm? I think you need a battery for storage. \$\endgroup\$ – Sunnyskyguy EE75 Jul 8 '18 at 19:58
  • \$\begingroup\$ Well, currently I have a 220V timer that controlls the relay. It only can turn on and off. With two SPST relays I will need more sophisticated controller, power supply, etc. I was hoping for a simpler solution... \$\endgroup\$ – ivanich Jul 8 '18 at 19:58
  • \$\begingroup\$ What is the PV spec for W, Voc, Isc? and heater spec for W? \$\endgroup\$ – Sunnyskyguy EE75 Jul 8 '18 at 20:02
  • \$\begingroup\$ Directly heating the water would be much more efficient than using a solar panel \$\endgroup\$ – PlasmaHH Jul 8 '18 at 20:02
  • \$\begingroup\$ And about the 18V: I am using boost converter that increases the voltage up to 60V. The more power the panel produces, the higher the voltage. The panel can make 200W, which in a sunny day is enough to warm up \$\endgroup\$ – ivanich Jul 8 '18 at 20:02

Your application seems unsafe for a couple reasons:

  • There is no guaranteed isolation between the contacts even if they configured to break-before-make. In other words, from safety point of view, there is no guarantee that the MPPT controller will be isolated from the AC network. From a product/earth safety perspective this solution is dangerous.
  • Even if the AC network is actually disconnected from the network, the boiler and connected wires have their own internal inductance, and once you open the relay, it will cause a spark (depending when the current crosses zero).
  • If solar grid is wired in a SELV network, you are causing an unwanted risk by sharing AC with SELV on the same contacts of the relay.

The solution would be either:

  1. Use an isolating transformer from AC to SELV to power the boiler
  2. Use an inverter next to the MPPT controller
  3. Use a bolier that is designed for the purpose

As already pointed out, since the power to the boiler resistance is V^2/R, you will get an heating effect that is much lower (and maybe negligible).

  • \$\begingroup\$ Thanks for the answer. Well, I admit, that currently the system is not safe. I was hoping that the SPDT relay will do the isolation. According to your suggestions: An isolating transformer would have power losses that I want to avoid. Same with inverter. The inverter would also pump energy to my neighbours. My boiler is actually designed for both purposes - to be powered by solar and by grid. What I need is, as You noticed, a way to automatically guarantee isolation between AC and MPTT \$\endgroup\$ – ivanich Jul 8 '18 at 20:22
  • \$\begingroup\$ About the heating effect: I forgot to mention that I am using MPTT boost converter. The output voltage varies between 18V and 60V. According to the formula above, 60V ^ 2 / 16 ohm = 225W, which, on a sunny day can warm my boiler for about 7 hours \$\endgroup\$ – ivanich Jul 8 '18 at 20:29
  • \$\begingroup\$ Two logic signals with same trigger but complementary and 1 second Turn on delay = RC and instant turn off (diode*C) to drive DC relays?? \$\endgroup\$ – Sunnyskyguy EE75 Jul 8 '18 at 20:33
  • \$\begingroup\$ Tony EE rocketscientist ... Yes, I am getting convinced that I will have to follow Your suggestion. Seems that there is no simpler way which guarantees isolation. Thanks a lot! \$\endgroup\$ – ivanich Jul 8 '18 at 20:44

There is less chance of a "shoot through" with the properly selected "break before make" relay that has a sufficient voltage standoff rating. Since most relays are designed this way I suggest that there is a different cause for the controller getting fried. The boiler load probably has a large inductive component to it. Whenever the load is actively drawing current and then the relay suddenly opens this inductive part of the load can cause a huge spike of voltage that can easily fry the semiconductor conductor components in the controller. This spike of voltage can also be a contributor to the sparking you see across the relay contacts.

For DC switching the solution for this voltage spike is to add a reversed biased diode (a.k.a flyback clamping diode) across the inductive load to clamp the over voltage and provide a path to dump the current that wants to flow due to the collapsing field in the inductive load.

The same diode scheme does not work for AC switching of inductive loads because of the ever changing polarity of the voltage. In this case there are a class of circuits called snubber circuits that can be installed across the load. In their simplest form the snubber can consist of a series resistor and capacitor. The component values are selected based upon the AC frequency and the inductive characteristics of the load.

Since you have this somewhat non-conventional idea to switch between an AC source and a DC source it will be more difficult to find a voltage spike clamping solution that works well for both cases. So the recommendation that I suggest is to first solve the safety problem as pointed out by others and then figure out the inductive load problem.

Some may suggest putting a DC to AC converter in line with the output of your MPTT controller which provides an isolated AC output to deal with the safety problem. Since such converters are not 100% efficient a certain part of the energy produced by your solar panel will be lost as heat in the DC to AC unit. For this reason I suggest that instead you get a suitable AC to DC switching power supply that has an isolated output and use that on the AC mains side of the circuit. From the mains side there is less concern regarding the conversion losses and with modern component technology it is possible get AC to DC switching power supplies with better than 90% efficiency. With isolated DC now being supplied from the line side the relay is switching between two DC sources and any inductive voltage spike clamping can use the simple and effective diode method.

One has to guess that your boiler is already able to operate on a DC voltage due to that is how you have shown in your diagram. If by chance the boiler is using any motors that are designed to run on an AC voltage source only then you would have to change those out for motors designed to run on DC voltage.

  • \$\begingroup\$ Thanks for the comprehensive reply. My boiler does not have any inductive components inside. It has a simple resistive heater. The it has some inductance, but it is negligible I think. My SPDT relay is a cheap one from aliexpress. I am starting to suspect that it is not really a break-before-make one. Maybe I will have to buy a better one. And to guarantee isolation, I will probably have to use 2 SPST time delayed relays with a dedicated controller... \$\endgroup\$ – ivanich Jul 8 '18 at 20:39
  • \$\begingroup\$ Yes, if your boiler is simply a cold water heater then I understand. Two relays would make it more safe but not as safe as it should be. Relays are always subject to the possibility that the contacts can weld together. And then there is the possibility of an invalid state getting into your controller logic that leads to both relays potentially being on at the same time. It would still be better to isolate the mains for safety sake. \$\endgroup\$ – Michael Karas Jul 8 '18 at 20:50
  • \$\begingroup\$ Maybe I can additionally protect the MPTT controller with a fuse and 2 TVS diodes which will turn on when the voltage exceeds 60V or goes below zero. When one of the diodes turns on, the fuse should blow. \$\endgroup\$ – ivanich Jul 8 '18 at 21:01
  • \$\begingroup\$ The other solution that I am thinking of is to find an SPDT relay with a large air gap that guarantees isolation? \$\endgroup\$ – ivanich Jul 8 '18 at 21:02
  • 1
    \$\begingroup\$ @ivanich You're playing with significant solar power and mains, and you're "hoping" (from another comment) to isolate that using crap off AliExpress? Please reconsider the current path you're on, it's just too easy to electrocute someone or burn your house down. \$\endgroup\$ – marcelm Jul 12 at 18:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.