PWM and SSRs - a bit of advice please

Thread Starter

BigBlueCar

Joined Sep 23, 2013
14
I'm trying to use up as much as possible of my 10kWp of grid connected solar PV without exporting any to the provider. I have a thermal store with 3 x 3kW heaters (230V) but of course, they should only consume whatever power is available, which means at least one of them being on PWM or something similar when Mr Sunshine isn't quite so bright.

There's at least half a dozen well tried devices on the UK market that will do this for 3kW max. but none for 3 x 3kW (t.b.h. there's maybe one, but I'm fed of asking their one and only technical man/owner to ring back).

So I found this controller in Europe, which looks like the cat's whiskers. https://solarcontrols.cz/wattrouter_mx.html

100A maximum measurement - tick, Windows management interface - tick, DIN rail mounting - tick, multiple temperature probe inputs - tick, timer schedules - tick, LAN, RS485 & USB connection - tick, 6 SSR outputs for proportional or PWM control - tick tick tick !!! .
Plus another two relay outputs that can control some 20A contactors for me if I want to block switch two of the three immersion heaters. It really is a nice bit of kit and the price is pretty good too.

However, how do I lash the outputs up to the heaters, what SSR specification, what heat dissipation is needed and so on?

This is going to be a step beyond my base skill level so I want to get it absolutely right. I'm not even sure if I should be using PWM or "proportional" control since the 80+ page manual doesn't give any guidance.

The SSR outputs appear to have various options, including 0-12V 200Hz to 10kHz and Sallen-Key filter, whatever that is????

I can post some screen shots of the specification and of the management software if this helps, otherwise they can be found here:
https://solarcontrols.cz/archives/eshop/WATTrouterMx_EN.pdf
https://solarcontrols.cz/live/wrmx/?lng=en
https://solarcontrols.cz/en/wattrouter_models.html

I'd be really grateful for some solid advice on how to control the heaters and possibly then expand the six outputs to other, power-hungry devices for when the water is fully warmed through.

Cheers,
Ian
 
SSRs all have some losses and your best bet might be an electromechanical relay. You don't need the PWM to be very fast because you're running a heater. Especially with the large thermal masses involved, you could probably set your PWM period to a full minute.
 

Thread Starter

BigBlueCar

Joined Sep 23, 2013
14
Many thanks for that. I don't think I explained myself very well so please forgive me if I go around the houses a bit.

An electro-mechanical relay is going to be no use to me because it is either full on (at 3kW) or full off. My Solar PV is anything between zero watts and 10kW, with frequent short term (< 1minute) spikes both up and down. Thermal mass of the water is irrelevant to the problem, it just affects the heat up time for a mean level of Solar PV output.

EM relays would mean that I would be intermittently drawing power from the grid for the heaters, regardless of sample period, when the option to demand heat from the gas boiler instead is much cheaper. Commercial devices typically have a sample interval of 1-2 seconds which prevents the system getting caught out by a sudden cloud across the sun. Not only that, but I don't want three 20A contactors bouncing about all the time and sounding like an old fashioned telephone exchange.

More practically, all we do is turn off the gas boiler and allow the immersions to operate only up to the total solar PV power output during daylight hours. As pure resistance heating devices they will happily run at less than full power if we offer up less than the nominal 230V.

I need something that will chop up the AC on every cycle and only let, say, 2250W through to the heater if that is as much as the solar PV is kicking out. Or even better, 3kW + 2250W if the solar PV is doing 5250W total. If a cloud comes along and power briefly drops to 4100W total, then the second heater only runs at 1100W until it brightens up again.

So..........
  • Cold tank, 5250W of solar PV being produced - one immersion at full 3kW (until it gets to temperature), second at 2250W, third one off.
  • Top of tank gets to temperature, 5250W of solar PV - top immersion switches off on thermostat, other two run at 3kW and 2250W.
  • Ditto but with only 4100 of solar PV - one immersion off on thermostat, second one runs 3kW, third one runs 1100W
  • Big grey clouds cut solar to 950W - the first immersion heater that's not at temp runs at 950W
  • and so on.
The attached image is the output yesterday from the smaller of my two inverters, so multiply by 2.08 for my total output.. As you can see, electromechanical relays are a non-starter.

This is the most popular device doing said magic on the UK market https://www.marlec.co.uk/wp-content/uploads/2016/01/Solar-iBoost-FAQ-010815.pdf?v=79cba1185463 but as I said, it can only handle 1 x 3kW. By chopping up the AC on every cycle it effectively drives the heater at less than full voltage.

The Czech device I was asking about is much more sophisticated but being more flexible it needs external SSRs with better than 50Hz switching rates to control the power.I've not looked in detail but I'm guessing it will be 200Hz, i.e. two on and off events on every half cycle of AC.

I don't care about SSR losses because this is free energy from the sky now that I've paid a lot of money for the solar panels and inverters. So long as those losses are low enough for me to manage any cooling requirement without too much hassle.

I'm amazed that this isn't a more common problem but then again, most solar PV installations are more modest than what I installed. I'd have put in even more if the local grid provider would have allowed it.

Can someone give me guidance along the lines of the original question please?

Ian
 

Attachments

Last edited:
SSRs all have some losses and your best bet might be an electromechanical relay. You don't need the PWM to be very fast because you're running a heater. Especially with the large thermal masses involved, you could probably set your PWM period to a full minute.
Seems to me, you've suggested the ideal solution right here: "As pure resistance heating devices they will happily run at less than full power if we offer up less than the nominal 230V"

Why don't you just tie the heaters directly to the DC output of the PV array? Why do you need an inverter at all?

Your make-up power is gas, right? Have a high level thermostat to turn off the PV input if all the water is hot. Have a low level thermostat run the gas boiler if the water is too cold (in spite of the PV input).
 

Thread Starter

BigBlueCar

Joined Sep 23, 2013
14
Seems to me, you've suggested the ideal solution right here: "As pure resistance heating devices they will happily run at less than full power if we offer up less than the nominal 230V"

Why don't you just tie the heaters directly to the DC output of the PV array? Why do you need an inverter at all?

Your make-up power is gas, right? Have a high level thermostat to turn off the PV input if all the water is hot. Have a low level thermostat run the gas boiler if the water is too cold (in spite of the PV input).
That's a whole different can of worms.

Aside from re-engineering a 12 month old PV system and having to lay in a different cable to bring in DC from an outbuilding where the Solar PV is, I currently get paid something for every measured kWh I generate, plus another fee for 50% of it, which is what the government scheme assumes I'll be exporting. Sadly, it's not the 45p per kWh that folks got in the early days, but it's enough to give a 9-11 year payback on the investment. Disconnect it from grid AC and I kiss it goodbye. That's before my 10 year guarantee on individual panels is invalidated for mucking about with the installation.

Plus, I have to pay for the other 50% of electricity currently used for dishwasher, washing machine, dryer, etc. that are "free" to run during daylight.
 

Thread Starter

BigBlueCar

Joined Sep 23, 2013
14
If you don't want to use burst firing, then use a Triac for phase angle control.
Max.
It's beyond my comfort zone and/or talent, but I thought high power SSRs were effectively triacs anyway?

But please tell me more. If the unit is going to do the control, what in terms of components to I put between it and the immersion heater(s)?


.
 

MaxHeadRoom

Joined Jul 18, 2013
19,977
SSR's are generally ON/OFF (switch devices), Normally SSR can be used for on/off for relatively long periods at a time, but a Triac, which is a basic component of an SSR, can also be used to control individual phase angle of each half cycle of AC current, from 0v to maximum.
A typical lamp or fan 'dimmer' operates this way.
Max.
 

Thread Starter

BigBlueCar

Joined Sep 23, 2013
14
SSR's are generally ON/OFF (switch devices), Normally SSR can be used for on/off for relatively long periods at a time, but a Triac, which is a basic component of an SSR, can also be used to control individual phase angle of each half cycle of AC current, from 0v to maximum.
A typical lamp or fan 'dimmer' operates this way.
Max.
Cheers Max. Are there any example circuits or any component info you could point me at?
 
With a 12A load it's going to be all about losses. I would propose MOSFETs over a Triac or SCR. At 12 amps, a typical triac (Vf = 1.4V) will dissipate 16.8 watts. Even a modest 500V MOSFET will do better. ON Semi's FCH023N65S3 has an on-resistance of 23 mO. Two of these in a back-to-back configuration will only dissipate 7W and that's distributed over a pair of packages. If you put 3 in parallel, you would hardly need a heatsink. MOSFETs are easier to drive (at least, at these low frequencies) and easier to parallel than thyristors.

https://www.onsemi.com/pub/Collateral/FCH023N65S3-D.PDF

If cost isn't too critical, a pair of these bad boys would be essentially bulletproof and would bring your total dissipation under 4W, 2W each in two packages that can handle a KW

https://www.littelfuse.com/~/media/electronics/datasheets/discrete_mosfets/littelfuse_discrete_mosfets_n-channel_ultra_junction_ixfn170n65x2_datasheet.pdf.pdf

In either case, tie source-to-source and gate-to-gate. The two drains are the two ends of the switch thus formed. A mere 10 volts at practically zero static current will drive the gates. I frequently use a photovoltaic opto-coupler to drive this kind of arrangement. That might be a little slow for your needs, though. I have used a lot of Toshiba TLP191 but TLP3905 would be faster.
 
Top