Hot Water by Solar Electric direct via MPPT ?

[PNPmacnab]

Many electronic wall warts will work on as little as 50V DC. This is at very reduced current, but can easily produce 100ma to run small devices like a micro. Everyone has a few wall warts collecting dust. I buy these 12V 1.2A supplies shipped from China for only $1.50 total and they even come in an aluminum case. One resistor change insures they will start on as little as 40V. I run full DC heater current thru the heaters mechanical thermostat. A circuit keeps the voltage potential across the contacts at no more than 20V ever. Did a demo with a single strand of copper wire imitating the mechanical contact. I tried to make and break the connection as slow as I could a couple dozen times. Not a hint of an arc. Try that normally and you better be wearing a welding helmet to protect your eyes.

[Tgriff]

OK ... thanks for the catch. My calculator must have had another number in the works. Still . . . 14.5 watts per SSR is significant - even if both not on at same time! It makes the need for forced cooling in a closet with a water heater. I'm thinking of using a 5V computer fan for the SSR heat syncs.

BTW: I watched a video of the recommended installation of conventional solar water heaters, and it required cutting holes in the roof for the plumbing. This alone is a deal-breaker for me - especially with the declining price of solar panels.

A few years ago, the thought of using solar panels to heat water was considered a huge waste - but not any more. I paid $1 / watt for my panels and it is now less than half that price, making the solar panels a cheaper part of the equation. What other unusual applications can be solar powered? Swimming pools, sub-floor heating, geothermal HVAC (now my passion)?

[bcroe]

I checked the SSR specs and it claims < 1.5V "on voltage". At 9.5 amps, this would be 9.5 * 1.5 = 135 watts
power dissipation when on. No wonder they burned up on me!

Better check your slide rule, mine says 9.5A times 1.5V equals 14.25 watts. Bruce Roe

[Tgriff]

on/off time 10 ms - should be fine for PWM control of DC at 50 khz per MPPT specs (50 khz is approx frequency of MPPT controllers)

I don't understand that statement at all. I heat water with PV similar to the German module. I use a fixed power point voltage as it only tends to change seasonally and it doesn't change that much thru out the day to matter that much. That allows me to run the system in parallel with the charge controller to harvest any excess power. The techluk uses an IGBT which wastes a lot of power since saturation is close to 2V. I avoid those SSY modules because of suspicious specs and heating. FET are cheap enough and with several in parallel my heat sink temp rise is barely noticeable.

Noted. Thanks for sharing.

I checked the SSR specs and it claims < 1.5V "on voltage". At 9.5 amps, this would be 9.5 * 1.5 = 135 watts power dissipation when on. No wonder they burned up on me!

For the low voltage power for the thermostats and cooling fan, I plan to use a 120VAC USB power adapter to convert the incoming 96VDC from the solar panels to 5VDC. I looked inside one and it is "front-ended" with a bridge rectifier (so the rest must be a DC-DC power supply) - some testing showed that when supplied with 96VDC from the solar panels, it produced 5VDC just fine.

I already have the SSRs, I just need to figure out how to cool them or replace them.

[Tgriff]

PWM controllers operate with periods of up to several seconds - more than enough to destroy the contacts of a thermostat.

You would need a controller that has arc-fault detection. Most don't. (BTW the German controller you posted a link to replaces the thermostat contacts with its own DC switch, which would work well.)

Arc-fault checking isn't necessary if the supply power is chopped to near-zero several times / second. An additional "controller" isn't necessary either - the thermostat is the controller. We are just simulating AC supply by chopping the solar DC power. No controllers, batteries or inverters, just power to the heating elements in the water heater.

My first design worked, but depended on the AC thermostats to switch the DC on and off. The contacts were quickly fused.

Some Internet research yielded a design which re-wired the water heater to supply low DC voltage to the thermostats and send the thermostat outputs to SSRs, which switched the high DC voltage from the panels. At the rated 9.5 amps, the SSRs had to dissipate something less than 6 watts each. The heat syncs for the SSRs weren't force-cooled, so the SSRs overheated and failed. I purchased new thermostats and SSRs, but left it as originally wired with grid AC power.

I'm still thinking about it though . . .

[PNPmacnab]

on/off time 10 ms - should be fine for PWM control of DC at 50 khz per MPPT specs (50 khz is approx frequency of MPPT controllers)

I don't understand that statement at all. I heat water with PV similar to the German module. I use a fixed power point voltage as it only tends to change seasonally and it doesn't change that much thru out the day to matter that much. That allows me to run the system in parallel with the charge controller to harvest any excess power. The techluk uses an IGBT which wastes a lot of power since saturation is close to 2V. I avoid those SSY modules because of suspicious specs and heating. FET are cheap enough and with several in parallel my heat sink temp rise is barely noticeable.

[Mike90250]

Here's a well respected industry paper on SSR's

https://industrial.omron.us/en/media/090416_OMRON_SSR_TechGuide_tcm849-113231.pdf

after reading it, you may find you need to reduce your switching speed. Repeated high speed switching allows heat to rise rapidly in the SSR, and they burn up. Long cables cause inductive :kick" which frys them.

https://www.machinedesign.com/mechanical/prevent-solid-state-relay-failure-thermal-protection

And this blog/Q&A has a lot of good links and reading :

https://www.element14.com/community/thread/13647/l/have-a-question-about-solid-state-relays-and-their-applications-ask-me?displayFullThread=true

[jflorey2]

Correct - but an AC thermostat will work with "chopped" DC power, as with a PWM controller. When the voltage approaches zero, the arc stops, preserving the contacts.

PWM controllers operate with periods of up to several seconds - more than enough to destroy the contacts of a thermostat.

The damage occurs when the contacts heat up, then fuse after a sustained arc of DC current. Chopping the DC prevents this - and the TechLuck and German controllers do this.

You would need a controller that has arc-fault detection. Most don't. (BTW the German controller you posted a link to replaces the thermostat contacts with its own DC switch, which would work well.)

[Tgriff]

The SSR I was using was something like this:

https://www.ebay.com/itm/SSR-60A-DD-DC-DC-Solid-State-Relay-input-3-32VDC-output-5-220VDC-Control/162952519959?hash=item25f0ba2117:rk:16:pf:0

Specs are:
input 3-32VDC
output 5~220VDC
on/off time 10 ms - should be fine for PWM control of DC at 50 khz per MPPT specs (50 khz is approx frequency of MPPT controllers)

[Mike90250]

SSR's have a max switching limit, and I suspect it's below what's needed for "average joe" to safely use an AC thermostat. 30Hz is about the upper limit for ones I'm familiar with.

[Tgriff]

The contacts in an AC thermostat are not designed for, or generally capable of, controlling DC power.
You have proved that. Bruce Roe

Correct - but an AC thermostat will work with "chopped" DC power, as with a PWM controller. When the voltage approaches zero, the arc stops, preserving the contacts.

The damage occurs when the contacts heat up, then fuse after a sustained arc of DC current. Chopping the DC prevents this - and the TechLuck and German controllers do this.

The approach that I was taking with the SSRs worked as well, until the SSRs overheated. It is a valid approach, but requires re-wiring of the water heater. The above approaches don't require any changes to the water heater - only to the incoming DC voltage. This is a simpler setup.

[bcroe]

I connected 4 - 230 Watt PV panels @ 24V each for 96V @ 9.5A to a standard 230VAC water heater and it heated the water well - a bit too well, perhaps. I found the thermostats had been fused by trying to disconnect the 96V from the heating elements after they had reached their shutoff temperature.

The contacts in an AC thermostat are not designed for, or generally capable of, controlling DC power.
You have proved that. Bruce Roe

[Tgriff]

The TechLuck controller could be valid, though I am less than impressed with it as they use the electrical box as a heat sync and I've seen their claim that supersizing the electrical box will dissipate the heat better.

I connected 4 - 230 Watt PV panels @ 24V each for 96V @ 9.5A to a standard 230VAC water heater and it heated the water well - a bit too well, perhaps. I found the thermostats had been fused by trying to disconnect the 96V from the heating elements after they had reached their shutoff temperature.

I puzzled over this for some time and re-wired the thermostats to control two SSRs (Solid State Relays) to shut off the DC to the elements, but didn't account for the power dissipation and the SSRs overheated and were ruined.

I saw a YouTube video about the TechLuck video, claiming that it switches the DC on and off to implement MPPT control of the power (I assume this it uses PWM, as moderating the voltage would result in too much power dissipation). I've also seen a YouTube video by Engineer 775 confirming that the controller works as expected. I have seen other of 775's videos and he is usually "spot on".

I have seen another such controller, made in Germany: MPPT PWM Heater Charger (https://www.ebay.com/itm/263523458861) that sells for somewhat less than the TechLuck controller.

The bottom line is that using PV panels to heat water in a conventional water heater will work if the voltage regularly goes to zero, as does AC and these controllers claim to do this. The only side-effect is the power dissipation from the PN junction of the electronics. It's something like 0.6V * PV current. In my case, this was 0.6 * 9.5A = 5.7 Watts. This doesn't sound like much, but it must be dissipated, else the controlling electronics will be ruined.

[bcroe]

I have been heating water with a 24 volt 900watt element in my cylinder. Once the battery bank is charged the charge controller reaches 100% SOC and it's auxiliary relay switches a solid state relay which connects the bank to the element

Efficiency doesn't really come into it once your batteries are charged. We did had to light our fire at
least once a week to give the hot water an extra boost and kill any germs.

I would check just how far from the panel MPPT voltage this setup runs. There might yet be room
for much improvement and less lighting fires. For example if the voltage is way low, you could install
another selected element which is a better match, might be in series with the panel connection
to the existing element. Bruce Roe

[J.P.M.]

I have a home made version of this.

https://www.ebay.com/i/272088626625?rt=nc&_trkparms=aid%3D222007%26algo%3DSIM.MBE%26ao%3D2%26asc%3D20160908110712%26meid%3De1a815df671b48789666c5aa0369587a%26pid%3D100677%26rk%3D6%26rkt%3D30%26mehot%3Dpp%26sd%3D222908806209%26itm%3D272088626625

that's all you need. use the panels for something more productive.

On the basis of providing energy to heat water, solar thermal is probably more productive on a per m^2 collector or array area by something like a factor of 2 or so, but the PV option by itself will most likely work out less expensive per unit of energy delivered per unit of money invested and for a lot less maint., even before accounting for required freeze protection in most climates.

In fairly mild climates, PV used in conjunction with a heat pump water heater will most likely be at least as, or more cost effective than simple flat plate solar thermal, and hands down more cost effective than evacuated tubes which are overkill for the task of DHW.