Archived Forum

The two documented methods of diverting surplus power (Mk2 and PLL) operate by cycling a resistive dump-load between its on and off states as required to balance consumption with generation.  By this means, the energy state of the premises remains almost constant, and no charge or penalty is applied by the meter.  These systems are reactive, with a rapid response time of around 30 ms.

Some commercially available alternatives offer proportional control which is predictive in nature.  This method of control is more complex and can lead to difficulties when conditions change, such as when there is a sudden increase in the baseload. 

My first attempts to divert surplus power used predictive logic to drive a proportional Carlo Gavazzi SSR .  The perfomance was OK, but nothing like as good as the later "Mk2" version.  A top quality "basic" Crydom SSR is less than half the cost of a proportional controller.  The performance is more dependent on the sketch that is running on the associated processor than with the output stage.

Thanks for the reply.

I already have an Arduino set up to monitor PV generation and household consumption so I'm half way there. Just need to decide how to send the excess PV power to the immersion without importing any power.

So it is possible to use either type of SSR to achieve proportional diversion ?

If so, are any other components required, apart from the Arduino ?

Using any of my Mk2 PV Router sketches (see Summary Page for details), a 5V Arduino with the standard V and I sensors can divert all surplus energy using nothing more than a basic on/off SSR. 

The "phase-angle control" version operates like a dimmer switch, so is effectively providing proportional diversion.  This method of control has EMI implications.

Thanks again.

I'm very new to Arduino coding and your sketches are way beyond my level but I'm hoping to try and incorporate some of your code into my own setup to get this working, albeit on a more basic level, before progressing to something more sophisticated.

Could you possibly explain (in layman's terms) how an SSR is used to vary the AC power output.

Is this perhaps done using PWM output on the Arduino ?

Hi Scott,

I have a cheap SSR purchased from eBay switched by the PLL sketch to power my immersion heater. It's been running fine for over a year now and when the sun is shining my water gets nice and hot.

These sketches have the capability to switch the relay based on the power consumption in a cycle, so a 3kW immersion heater can be switched fast enough that it doesn't consume the full 3kW. As Robin said, this is much like how a dimmer switch works for reducing the light output of a bulb.

Well that's my take on it, it may now be completely accurate, but it allows me to understand enough to avoid exporting all the electricity I generate. I can genuinely say my gas bill has also dropped, along with my electricity bill.

Hope that helped.

Allen

Thanks for that Allen.

I'm new on this forum and still trying to get my head round it all, there's a lot to take in.

I have a power reducer fitted to our immersion which cuts the power down to 1.5kw which helps but I'd like to get a power diverter set up to make better use of our spare PV output.

So basically, you're using the Arduino to switch the SSR on and off rapidly ?

Do you have a link to the PLL sketch you're using ?

All the details can be found at this link http://openenergymonitor.org/emon/pvdiversion/pll.

The reason I preferred this sketch over Robin's was purely because I use a DS18B20 temperature sensor to cut off the diversion once my cylinder reaches temperature. I do this because the sketch will use the number of cycles diverted to estimate how much power has been transferred, so if you don't turn the immersion off (and not just rely on it's thermal cutout), the sketch thinks it is diverting 3kW, when actually the immersion has shutdown.

I'm also toying with the idea of another CT clamp to measure the diverted power. But I like to monitor how my cylinder temperature changes (when the kids are showering for example!)

This is also an interesting thread http://openenergymonitor.org/emon/node/1535

Thanks again.

Much of it is way over my head but I have an idea for a basic diverter until I can get to grips with the more advanced stuff.

If you want to develop your own diverter, then that's fine and you'll find quite a lot of information here. If you want a proven design that's known to work well, then one of Robin's Mk2 variants or MartinR's PLL is your best bet. You won't need your 'reducer', these designs attempt to exactly balance the average immersion load against the surplus PV so that nothing is exported, only failing when the controlled load cannot accept any more energy (for whatever reason). I take it you've read the article "Choosing an Energy Diverter".

Hopefully, your supplier's energy meter is compatible. Most are, but there are a few about that will only play nicely with a phase-controlled load.

Could you possibly explain (in layman's terms) how an SSR is used to vary the AC power output.

Hello Scott,

Here's an analogy that may help visualize how Robin's diverter works, and the part the SSR plays.

Let's say you have a bucket with a 10mm hole in the bottom. The bucket is getting filled with water from a 10mm (ID) hose.

If you put stopper in the hole, without shutting off the hose, the bucket eventually overflows. If the hole remains open, the bucket eventually goes dry, or the water falls to a very low level. However, if you used your finger to cover the hole till the bucket was half-full, then repeatedly covered and uncovered the hole (at at rate dependent on the bucket fill rate) you could: keep the bucket from overflowing, keep it from running dry, and, maintain the water level in the bucket between an upper and a lower level.

The hose filling the bucket represents your PV system production. The hole in the bucket represents the excess that would otherwise escape to the grid. Keeping the water level in the bucket between the two levels, represents a balance between energy imported from, and energy exported to, the grid. The water level in the bucket rising and falling is analogous to what happens with Robin's diverter. The difference between the  upper and lower "water levels,"  or more correctly, energy levels, represents what often gets referred to as the "sweet zone," which is the upper and lower limits of energy that can pass through your utility meter, without causing the meter to tally, i.e. charge you for, the energy.

That is, essentially, the principle behind Robin's diverter. It uses the concept of an "energy bucket" and your meter's "sweet zone" to do the same thing with your excess PV energy, i.e "burn" some of it with a DHW cylinder immersion heating element (or other load, if so desired) rapidly and repeatedly, so that sufficient energy is consumed to prevent export, but not so much as to cause import.

The diverter uses the SSR as a switch to connect and disconnect the load (immersion element) to the mains, rapidly and repeatedly, such that your excess PV energy is consumed heating your water, very little or none of it slips away to the grid, nor do you import any energy from the grid. That happens only when excess energy is produced. When your energy demand exceeds your PV system output, the diverter stops connecting the immersion element to the mains, i.e. the SSR isn't triggered, and you import the needed energy from the grid. When demand falls below your PV system's production, the diverter senses the change, and the diversion process resumes.

I've generalized things to a large degree, there's quite a bit more to the diverter's internal workings, and hopefully, I've not confused you further.

Regards,

Bill

Cheers guys, really appreciate your help with this.

I've knocked up some code which will divert power at staged intervals, i.e, 50%, 75% and 100% depending on the amount of excess power available.

A bit crude, admittedly, and no where near as efficient as the more advanced setups here but it'll do till I build up enough knowledge to attempt one of the devices on this excellent site.

Scott

If the load is activated on alternate mains cycles, it will consume approximately 50% of its normal rating.  That will be fine for when precisely that amount of surplus power (SP) is available, but there will generally be some degree of mis-match. 

If the SP is slightly more than half the rating of your load, some SP will leak away to the grid.  If it is slightly less, there will be a net import of power.  Because of the way that a reciprocating flow of energy is recorded by the meter, this mis-match could become expensive. 

The Mk2 and PLL diverter systems that are described on this website do not suffer from this effect because the energy state of the premises is actively maintained at a constant level.

There is a detailed description of the workings on the Mk2 Router in the final section of the online description document.

I've been reading a few of the threads on here and the mist is slowly starting to clear.

I'm thinking of going for one of the slightly more expensive proportional SSRs . Think that might be a better route considering my limited knowledge of electronics.

Must say, there's some fantastic work been done here by some talented folk !

I'm slightly worried that on your own admission you have limited knowledge and yet you're shying away from a tried and tested solution (which is how I'd describe both MartinR's PLL and Robin's designs). Between us here, we've had enough experience to be able to solve easily pretty much any problem you might have with Robin's or Martin's; but if you head out on a limb you might come up against something we haven't seen before and it will be that much harder to get you going.

I'll second that as a user of Robin's Mk 2 "product." It works, works well and the support is unbeatable, as Robin is the software developer for said product.

I appreciate your concern and offer of support. I'll no doubt be calling on it a good deal before I'm finished.

My intention is to start off with a basic device similar to ones I've seen mentioned in other threads then upgrade to one of the devices you refer to once my knowledge and understanding of them improves.

Don't get me wrong, I can now see the advantages these more advanced solutions offer.

...well you have been warned!
You seem to be taking the more difficult route, but hey, good luck.

Paul

I have to agree with the comments above. I am no technical wizard, but following the designs already tried an tested on here, was able to produce a working diverter that has served me well for over a year. I myself went out on a bit of a limb because I used an SSR, when most of the designs talk about using a triac.

This lead to some problems, but the guys on here helped me out, and I'm really happy with my solution. I have no idea how a proportional SSR would work to solve this issue, digital sounds so much easier to me.

Hi guys, if anyone can devise a better way of diverting surplus power reliably using standard hardware, I would be most interested to hear about it ...

The concept of "surplus power", as I have written about on many occasions, is not strictly accurate in this application.  With the diverter systems that are described here (Mk2 and PLL), action is taken when there is a surplus of energy. This is done by switching a dump-load on and off as necessary so that the energy state of the premises is kept at a near-constant level.

While the diverter is operating in this way, some power will always flowing at the supply point, either import or export.  The meter continuously monitors the reciprocating flow of energy, as does the diverter, so the two systems remain in step.  Crucially, by keeping the energy fluctuations within a certain range (the meter's sweet zone), penalty charges can be avoided.

Any system that does not measure continuously is likely to give incorrect results in the presence of reciprocating energy flows.  

Hi all,

I am also at the point where I am looking for solutions to increase my self consumption after my 3.12 kWp array was installed on my roof.

I read through all (actually I don't think I managed ALL) the threads related to the power diversion devices but must be honest; I didn't understand more than 50%...;-)

Nevertheless; I need to divert power to my geyser (that's how immersion heaters are called in South Africa) to increase my self consumption.

I have a emontx v3 monitoring the grid power and house power. An emonglcd is happily displaying the data. As far as I understood correctly I just have to add the output board from you, Robin, to the mix or use a SSR only and then I would be up and running using the correct sketch.

As mentioned further above I am in sunny South Africa, so ordering the item from UK is possible but it's not practical.

Therefore I rather would like to get the parts locally and would prefer the SSR route which is hooked up to the emontx.

Is my basic understanding correct and what SSR is to be used. I understand a Crydom D2425 would be sufficient. My immersion element is a 2kW at the moment (originally 4kW which I exchanged to 2kW after installation of the 3.12 kWp array). Is it advisable to increase the element to 3kW to match the max output of the array; I believe so....?!

Your comments are highly welcome, thank you! 

"I read through all (actually I don't think I managed ALL) the threads related to the power diversion devices but must be honest; I didn't understand more than 50%...;-)"

But did you read the introductory articles: Choosing an Energy Diverter?

One possible fly in the ointment is your supplier's meter. Read the Resources article about energy meters. It's possible that yours might have a very small "energy packet", and in that case burst mode won't deliver the savings that you expect. You'll need to do phase control.

You could revert to your 4 kW heater. As it is controlled, it will only draw the exact power necessary to balance the house consumption to match the PV output. But a 3 kW heater will draw less current hence you will have less problems with harmonics or dip. If your house consumption is such that you're not exporting or only exporting a little when the heater is on, it's probably not worth changing the heater for a 3 kW one. If you're exporting a lot, the 3 kW will get the water hot in less time, but it won't accept any more energy. It's for you to make that decision based on how much excess PV you have and for how long (both summer and winter), and your cost/benefit analysis.

Any SSR that can be driven by a little under 3.3 V and can handle the load should be fine. But make sure you get the right sort, i.e. one that recognises and switches on the zero crossing if you're going to be using burst mode, or one that doesn't if you're going to be using phase control.
But make sure you get one from a reputable supplier - we've had horror stories about cheap ones.

Depending on how much it takes to drive the SSR, you might find that your emonTx V3 needs a 5 V power supply, as the internal one wasn't designed with external loads in mind.

Hi Robert,

thank you for all the input and information; that makes things quite a bit more transparent.

yes, I read the article but I realize only now that it answers quite a few of my questions.

I have one of the "disc type" meters (3phase but my Elec reshuffled the main db that all circuits are on one phase incl PV) which doesn't spin backwards. Here in South Africa we don't get anything for electricity fed back to the grid; that is why I want to increase my self consumption as much as possible without battery back up. Having said that I believe I shouldn't have any problems with the "burst mode" and too small packets?

What would be a reputable brand for a proper SSR?

Thanks!

You should have no trouble with that meter, provided that it does go backwards a little way before it hits the ratchet (until they change it!), I think Robin has posted a video showing one of those see-sawing backwards and forwards as the burst-mode system operates.

I'd go to a reputable supplier for your SSR, rather than just look for a reputable brand - there are cheap forgeries of reputable makes around, and a reputable trade supplier will have a lot to lose if they get it wrong.

ajscheid, I'm not sure which sketch you are currently using, but these ones all support diversion of surplus power in an emonTx V3 setting (info from my Summary Page):

Using an emonTx V3.2 as a Mk2 PV Router This describes using the emonTx V3.2 rather than an Arduino to supply the processing power. Posted, on 14/2/2105, at http://openenergymonitor.org/emon/node/10171.

Using an emonTx V3.4 as a Mk2 PV Router This describes using the emonTx V3.4 rather than an Arduino to supply the processing power. Posted, on 20/2/2105, at http://openenergymonitor.org/emon/node/10203.

Using an Emontx V3.4 as a MK2 PV Router with phase angle control This describes using the emonTx V3.4 rather than an Arduino to supply the processing power. Posted, on 09/07/2015, at http://openenergymonitor.org/emon/node/10865

As Robert has mentioned, the Crydom 2425A relay may not work reliably when driven from a 3.3V system.  A transistor stage could perhaps be added to boost this control voltage to ensure reliable operation.  Inverting the sense of the output signal may be helpful in this case (by changing the order of LOAD_OFF and LOAD_ON in the enum definition).

In some parts of South Africa, I understand that there are penalties for exporting, hence there is a financial benefit of consuming any surplus power on-site.

Thanks Robin!

Yes, you are right; generally speaking the feed back of generated power is not yet regulated here. Some say feed back at this stage is illegal. I have a grid-tied system which goes down as soon as the grid goes down. That is why I received a COC for my system.

Presently I am sending quite a big amount of power back to the grid (5.6 kWh during the last 24 hrs) and don't get anything for it (I doubt that NERSA get going anytime soon) . As my disk-type meter is not turning backwards (which obviously would be illegal anyway) I want to keep as much power on the premises as possible.

I am currently using the standard sketch on a emonTx V3.2. I read through the respective post you posted above and decided, in combination with the comments of Robert further up, that your output board with my existing emontx V3.2 is actually the way to go. The only thing I am worried about is the kW limitation on CT4; according to the specs the CT4 input is limited to 4.5 kW max. which is used to monitor the grid supply in the mentioned sketch. But it clearly could be possible that the grid supply is higher during a day without sunshine where you have to override and the maid has to wash and gives a sh** and starts the tumble dryer at the same time. I have seen spikes as high as 6kW during the past 15 days since my PV system is installed.

Wouldn't that kill the input leaving the emontx obsolete for diverting the surplus power? Which in return would mean that your full router would be the better solution for me....

Does that somehow make sence?

Cheers,

Antonius

Hi Antonius,

I'm not sure what you mean by "the standard sketch".  Is it a discrete sketch which uses the emonLib library, or one of my "continuous" versions?

The emonTx V3 has low-valued burden resistors for CT1-CT3 which are intended for whole-house monitoring.  I believe these are 22R.  For CT4, a higher value of burden is fitted which allows greater sensitivity albeit over a reduced range.  The range of this channel could be increased by simply adding an extra resistor in parallel with the one that's already on the board. 

With each of my "Mk2 Router for emonTx V3" sketches, CT4 is used solely for balancing purposes.  The range of this channel only needs to exceed the ratings of the PV system and the dump-load, whichever is the greater.  If whole-house monitoring at the supply point is also needed, that should be done using one of the other CT channels with its appropriately sized burden.

Even with its default resistor, I would not expect the CT4 channel to suffer any damage as a result of being overloaded.  Like a cartoon character, it will spring back into life when the sunshine returns. 

A triac-based output stage could be driven directly from your 3.3V emonTx.  With a low-current trigger, this should be possible even when powering the unit from an AC/AC adapter.  If you would like one of my output boards, just send me a PM; or there's a contact email address on my Shop page.

Hi Robin,

I am using the single phase discrete sketch for the emontx  v3.2.

That all sounds very good and seems to be the way to go. I send you an email to the Shop address.

Cheers,

Antonius