Advice on high power (25A) SPDT SSRs

Forums › Gadgets, robotics, home automation, electronics (including wearables) › Advice on high power (25A) SPDT SSRs

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#153861 10-Oct-2014 09:36

Hi,

I have recently had some solar PV panels installed and am trying to come up with clever ways to use the excess PV being generated on these sunny days, since Meridian only pay 25c for the first 150kWh exported each month. I have come across the very clever chaps at OpenEnergyMonitor (http://openenergymonitor.org/) and in particular one bloke named Robin Emery who develops and sells what he calls the MK2PVRouter (http://www.mk2pvrouter.co.uk/).

The idea is this thing will monitor your feed to the grid using a CT sensor and voltage sensor. As soon as you begin exporting to the grid it will start 'flicking on/off' a specified load, typically a hot water immersion element, just enough to bring your export back to zero. As your solar generation fluctuates this thing will adjust the speed of the HWC element cycling in order to keep your import/export at zero. Very clever stuff!

I have the added complication that I already have a solar hot water system via evacuated tubes. The controller for this system is able to manually boost my HWC at specified times, via a big juicy relay. So mine is configured to boost the HWC at 4am if the cylinder temp is below 50 degrees.

My HWC has two elements, one in the middle (hooked up to the solar H2O controller) and one at the bottom - currently disconnected but which I want my PV router to be controlling, since I want to heat the entire cylinder with this excess solar energy when it is available.

Unfortunately my HWC is not in the same place as my distribution board, and I only have a single 240VAC circuit to the HWC. So I have to be very careful that I don't have both elements on at the same time, since the wiring is not rated for that kind of load.

So attached is my current setup and the proposed setup (a picture is worth a 1000 words and all that!).

The question is this - can anyone see any problems with this design? And secondly does anything know of any SPDT SSRs that would be suitable for this type of application? The switching rate will be very quick - multiple times per second when it is active so it will need to be very robust and fast acting.

Robin is using these SSRs (http://uk.farnell.com/jsp/search/productdetail.jsp?SKU=1200241) in his latest iteration but they are only SPST and as you can see from my design I really need SPDT relays.

I am hoping someone one here will be able to point me in the right direction!

Cheers,
Ben

EXISTING HWC WIRING DIAGRAM

PROPOSED HWC WIRING DIAGRAM

1 | 2

5705 posts

Uber Geek
+1 received by user: 3666

Lifetime subscriber

#1151217 10-Oct-2014 11:38

You could always manufacture a STDP relay from two STSP relays, with a a transistor inverting the control signal to one of them.

Probably smarter would be to change the control signalling rather than the power. Have one STSP relay connected to each element... connect the inputs of these together. And have a STSP relay connected to each power source (240V mains, solar panel). Connect the outputs of these together, to the inputs of the others. Now you can connect any input to any output. I guess you would need to be careful to not turn on both the mains and the solar panel inputs at the same time. And you need to be careful not switch both outputs on.

But... How much solar power are you expecting to generate at 4am? Or indeed how much solar water heating?

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#1151232 10-Oct-2014 11:55

Hi Frank - thanks for your thoughts. Yeah, in normal operation the PV router will never be active and switching the lower element at 4am, which is the only time the middle element is switched on. So in theory it should never happen, but the system has to be safe and prevent it from ever happening. It is possible for example to hit the 'Boost' button on the solar controller which will turn on the middle element, and if the PV router is active then you could easily end up with both on - which equals bad times for my wiring!

I have been doing some reading and 2xSPST SSRs with a NOT gate on one of the control signals would work, as you suggested.

Or I could just have a single SPST SSR on the bottom element feed, controlled by my PV Router. The middle element would be connected directly to the mains, via the existing solar relay. Then have a NC mechanical SPDT relay sitting before the SSR, which gets its control signal from the solar controller signal. So when the solar controller signal goes HIGH the mechanical relay goes OFF and the SSR becomes isolated. This would ensure that if the middle element was on, the bottom element couldn't be. And as long as the solar controller signal is OFF the PV router is free to do its thing.

Sheesh - all getting a bit complicated...

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#1151239 10-Oct-2014 12:06

And a circuit diagram...

Aredwood

3885 posts

Uber Geek
+1 received by user: 1749

#1151616 10-Oct-2014 22:40

The problem with what you are trying to do. Is that this system will only ever draw 3kW or nothing. Even if you switch the element quickly it will simply appear as a pulsed load to the electricity meter. Meaning it will record imported and exported power at the same time. Also those mains solid state relays are based on TRIACs. With an internal circuit that switches the TRIAC on at the 0 crossing point of the mains cycle. And they only switch the TRIAC off again at the 0 current point of the mains cycle. This means the absolute max switching frequency you can get from 1 of these is 100Hz.

A far better way of achieving what you want to do would be to connect the bottom element to the DC wires between your panels and your solar inverter. Then you will need to use the Arduino to monitor current at the switchboard. And then control the element either via a big MOSFET. Or a "buck" inverter. When the element is on, the solar inverter will simply think that the panels are not producing as much power, so will not send as much power into your mains wiring.

I am assuming that the designer of the system you have linked to is on a "nett metered" billing plan for his grid connected solar. Which means he can "get away" with a system that simply switches the element on in bursts. In NZ you have to have a meter that separately records imported and exported power. Which means that a system that can only control the element in bursts could actually cause higher power bills.

Aredwood

3885 posts

Uber Geek
+1 received by user: 1749

#1151618 10-Oct-2014 22:46

Also does his Arduino sketch compensate for varying power factors?

KGM

5 posts

Wannabe Geek

#1151703 11-Oct-2014 09:20

Heatsinking is important with SSR's. I know of several failures that have led to fires with SSR's (inefective heat sinking being the cause in a couple of cases), so would advise caution.

A relay with high current contacts or small contactor may be more appropriate for the switching device. The only downside is that an audible thump or click will be heard when it operates.

Mighty Ape

Shop now at Mighty Ape (affiliate link).

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#1151719 11-Oct-2014 10:00

@Aredwood - very interesting to read what you said. I am not an expert of things power related, my background is in software. I would be happy to share the sketch if you are interested? This system is in use in quite a few places apparently, both in the UK and all over the world, including NZ. A couple of Geekzoner's have them installed from what I understand.

I wasn't aware that our import/export meters operated in a different manner to the UK. From what I understand in the UK they just have a secondary meter installed if they have solar generation. This meter is used to report back to the authorities how much solar has been generated, and they are paid a 'Feed in Tariff' of around 50% of whatever they generate, regardless of what they use or export. So I guess this goes along with what you are saying, i.e. that the main meter is a 'nett' meter. The system developers talk about the meter 'sweet spot' which is typically 3600 Joules. As long as the PV router keeps the nett import/export variation within this range then the meter won't register a 'pulse' and thus cost you any money.

But as you say, if our import/export meters are separate registers, then we won't have this 'sweet spot' to take advantage of - if my understanding is correct.

Unfortunately my HWC is in a different location to my distribution board and inverter, so a direct connection from the DC side of the inverter to the HWC element is not possible, and is far from my area of expertise!

@KGM - yep there would be a suitably large heat sink attached to the SSR. Because the SSR has to switch so quickly (in theory) a mechanical relay is not really suitable. It will be fine for the other relays as they are only switched once or twice a day.

nickb800

2735 posts

Uber Geek
+1 received by user: 829

Trusted

#1151721 11-Oct-2014 10:02

Just checking, how is the solar hot water controller powered? If you switch the mains supply to the controller off when turning on the lower element then it may do funny things when it reboots. I'd want to be giving the controller a constant supply and putting an SSR on the heating element output of the solar controller. You might be planning to do this, I just couldn't tell from your diagram

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#1151722 11-Oct-2014 10:04

Yeah sorry, the solar hot water controller is powered via a standard wall socket, which is on a different circuit to the HWC element circuit. So it will have power always.

Handle9

11925 posts

Uber Geek
+1 received by user: 9675

Trusted

Lifetime subscriber

#1151730 11-Oct-2014 10:09

Aredwood: Also does his Arduino sketch compensate for varying power factors?

It'd be a unity power factor as its a purely resistive load wouldn't it?

Aredwood

3885 posts

Uber Geek
+1 received by user: 1749

#1151875 11-Oct-2014 16:21

Handle9:
Aredwood: Also does his Arduino sketch compensate for varying power factors?

It'd be a unity power factor as its a purely resistive load wouldn't it?

I was referring to the combined power factor of all loads in your house. Because you are measuring the current in your incoming mains cable. And the aim is to try and keep the power drawn from the network as close to 0 as possible. But since the power company doesn't charge for reactive power (KVAR) There is no point in trying to reduce the amount of reactive power drawn. Only the real power.

Geekzone support

Support Geekzone with one-off or recurring donations Donate via PressPatron.

Aredwood

3885 posts

Uber Geek
+1 received by user: 1749

#1151898 11-Oct-2014 17:07

SumnerBoy: @Aredwood - very interesting to read what you said. I am not an expert of things power related, my background is in software. I would be happy to share the sketch if you are interested? This system is in use in quite a few places apparently, both in the UK and all over the world, including NZ. A couple of Geekzoner's have them installed from what I understand.

I wasn't aware that our import/export meters operated in a different manner to the UK. From what I understand in the UK they just have a secondary meter installed if they have solar generation. This meter is used to report back to the authorities how much solar has been generated, and they are paid a 'Feed in Tariff' of around 50% of whatever they generate, regardless of what they use or export. So I guess this goes along with what you are saying, i.e. that the main meter is a 'nett' meter. The system developers talk about the meter 'sweet spot' which is typically 3600 Joules. As long as the PV router keeps the nett import/export variation within this range then the meter won't register a 'pulse' and thus cost you any money.

But as you say, if our import/export meters are separate registers, then we won't have this 'sweet spot' to take advantage of - if my understanding is correct.

Unfortunately my HWC is in a different location to my distribution board and inverter, so a direct connection from the DC side of the inverter to the HWC element is not possible, and is far from my area of expertise!

@KGM - yep there would be a suitably large heat sink attached to the SSR. Because the SSR has to switch so quickly (in theory) a mechanical relay is not really suitable. It will be fine for the other relays as they are only switched once or twice a day.

NZ meters for solar setups definitely record the imported and exported power separately. To stop people from using the grid as a "battery". If the power companies didn't do this. We would have the silly situation of people exporting lots at midday, competing against renewable generation. And then drawing lots from the grid in morning / evening, using fossil fuelled generation. Yet not paying for it.

Im assuming that the UK has mostly fossil fuelled and nuclear power stations. Therefore they won't have this problem. And more of their cooking / heating / hot water would be via gas. So the morning / evening peaks won't be as bad there.

The system you have described will still work. It just won't be the best. Also don't bother sending me the sketch. As my software skills are not that great. Im good at the hardware side of electronics.

Another way of doing it - Wire this between the mains and your bottom element. http://www.aliexpress.com/item/Free-shipping-New-4000W-220v-Adjust-SCR-Voltage-Regulator-Motor-Speed-control-Dimmer-Thermostat/694392288.html It is basicly a really high current leading edge light dimmer. Then remove it's variable resistor, install an opto isolator IC, and a few small fixed resistors. And modify the Arduino sketch to output a varying voltage depending on how much current you want to "dump" into the cylinder.

The only unknown will be how the meter (and the arduino sketch) will interpret the bad power factor of a partial draw from the dimmer controlled element, other loads in your house, and the unity power factor outputted by the solar inverter. Assuming it is a good brand.

SumnerBoy

2079 posts

Uber Geek
+1 received by user: 306

ID Verified

Lifetime subscriber

#1152091 12-Oct-2014 10:01

The guys over at the OpenEnergyMonitor forum have confirmed they have a 'nett metering' system in the UK. One of them developed a 'phase angle control' variant of the PV router which they claim should work with our 'import/export' meters but this is all getting a bit too complicated for me, and outside my field of expertise. I wouldn't be confident altering that dimmer module to do the load controlling either, not that keen to be playing around with 240VAC/3kW loads!

Appreciate your feedback on all this but I am beginning to think this might be something to shelve for a while. The fact is I am on day/night rates, so am only paying 13c/kWh at night, which is when I am currently heating my HWC (when required). This only happens in winter when the solar evacuated tubes don't produce enough hot water to heat the cylinder - in summer I use close to zero electricity for hot water heating.

Therefore with the buyback rate at 25c/kWh for the first 150kWh (from Meridian) I am better off exporting excess PV during the day. And even after I drop down to 10c/kWh (after the first 150kWh) I am only losing out 3c/kWh compared to if I was diverting PV during the day (and therefore not exporting) and not heating the HWC at night. Some calcs; assuming the cylinder is on for an average of 1 hour each night to boost, running at 3kW, that is only costing me 9c a day. Assuming this is only happening during winter, so for 6 months say, that is only costing me $16 a year extra. That is a very conservative estimate as it doesn't take into account low PV days where I wouldn't be able to divert anyway, and would have to heat at night.

Those SSRs are over $60-$70 plus the extra bits and pieces like enclosures and heat sinks. I would be looking at $100 easily. So it would take over 6 years to pay back the investment. This isn't too bad I guess, but only saving $16 a year makes me wonder if it is all worth it.

Good fun working it all out tho ;).

1eStar

1604 posts

Uber Geek
+1 received by user: 375

#1152139 12-Oct-2014 12:22

Power the lower element rather than the mid element.

Aredwood

3885 posts

Uber Geek
+1 received by user: 1749

#1154044 15-Oct-2014 00:09

SumnerBoy: The guys over at the OpenEnergyMonitor forum have confirmed they have a 'nett metering' system in the UK. One of them developed a 'phase angle control' variant of the PV router which they claim should work with our 'import/export' meters but this is all getting a bit too complicated for me, and outside my field of expertise. I wouldn't be confident altering that dimmer module to do the load controlling either, not that keen to be playing around with 240VAC/3kW loads!

Good fun working it all out tho ;).

Phase angle control is the technical term for how light dimmers work. So it sounds like they have already developed the same system as what I described. If you have a few chest freezers. Another option might be to add some smart control to them. Dual stage thermostats to try and only run the compressors when there is lots of solar production. You would be storing "negative heat" if you know what I mean. The compressor will probably draw about 500W But will have a bad power factor. Approx 1000KVA.

I will prob buy 1 of those phase angle controllers to do some experiments with. As I think that my smart meter doesn't measure correctly some types of current waveform. Which would explain why im getting high powerbills with only 2 people in my house. Last bill $161 Yet no electric hot water, no electric heating, no electric cooking. Power is mostly used only by motors and electronics. Only things in the house with heating elements is the kettle, dishwasher, and my soldering iron. And bills were virtually the same last year when there was 5 people in the house.

1 | 2