Jump to content

Using BMV 702 Relay


incagarcilaso
 Share

Recommended Posts

Hi. I have a heat pump water heater (Ariston) and it has the option to connect a relay from a solar installation that will start the heat pump in a certain mode when the relay is activated. I have difficulty in making out whether this is compatible with the BMV 702 relay. In the heat pump instruction manual it says: "If you have a PV system to be connected you can connect a bipolar cable from the inverter to the electrical box on the right hand side of the product. Connect said cable to the SIG2 connector. Caution: 230V signal". So the instructions are sparse and the difficulty I'm having is understanding why they say "Caution 230V signal" as I thought the relay was potential free and, what's more, this receives the signal rather than sending it. The BMV specs say that the relay is "potential free" but also gives it a rating of "60V/1 amp max". But if it is potential free why does it have a rating. Is this 60V DC or AC? Does this mean that I cannot connect the BMV relay to the heat pump, which says "Caution 230V signal"? - I don't want to melt the BMV.

I wonder how this cable connection can be made? Once connected, my aim is to set the relay on the BMV to activate when the batteries reach 100% charge and deactivate again when they drop below 97%.

Link to comment
Share on other sites

Well.... the input on the heat pump is going to take a voltage, where the presence of that voltage will indicate that the pump should/may turn on, and the absence indicates it should remain off? If yes, then it cannot be really potential free, and I'd put the "bipolar cable" (who knew cables could have personality disorders?) and other strange things down to a Chinglish translation.

So the question is then, what kind of voltage does it want as a signal? It would seem, from the warning label, that it takes a mains signal, 230V. So the question then comes down to one we have often discussed, whether you can switch 230VAC with the BMV relay. The consensus has been that since the BMV relay is rated for 60V DC, and because an AC arc extinguishes much easier than a DC arc, the relay should be good for 230VAC, provided you remain within the current restriction of 1 amp. If you are concerned, then rather use the BMV to switch another relay (rated for 230VAC) using a lower voltage (say 12V).

I know TTT has been using his BMV to switch 230V relays forever, but if it was my own system, I'd want a low voltage back there, for the simple reason that I don't much like low voltage and high voltage wiring inside the same enclosure.

Edit: Perhaps what they mean by potential free, is that the SIG2 connector goes directly to the coil of a relay/contactor and has no other connection to the mains, as opposed to the alternative where there is a live supply there and connecting a switch to it will close the circuit.

Link to comment
Share on other sites

22 hours ago, plonkster said:

Well.... the input on the heat pump is going to take a voltage, where the presence of that voltage will indicate that the pump should/may turn on, and the absence indicates it should remain off?

Edit: Perhaps what they mean by potential free, is that the SIG2 connector goes directly to the coil of a relay/contactor and has no other connection to the mains, as opposed to the alternative where there is a live supply there and connecting a switch to it will close the circuit.

Thanks Plonkster, that helps a little. In answer to the question, the relay on the heat pump doesn't switch it on or off. The heat pump is always on in a certain mode and runs when needed (twice a day for 45 minutes normally) to keep the household water temperature at 55 degrees. What this solar relay does is change the operating mode of the heat pump while the relay is energized to make the most of the solar production at peak times. So what it will do is when batteries are charged to 100% it will change the mode of the heat pump to boost and heat the water to a higher temperature while there is extra solar production. Then when the relay is de-energized the heat pump switches back to its standard operating mode. That way it will probably be sufficient for it to come on once a day instead of twice, and this will be during the period that it is supplied by solar production.

Can anybody help with specifics on how to couple up the BMV relay in this case? I see there is an additional problem that the BMV is 60 VDC but the machine it is connected to is 230VAC. What do I need in the middle to get this to work? Is there a relay that will accept a 60VCD input to energize it and the pass that to energize a 230VAC relay?

I'll see if I can scan the circuit diagrams on the heat pump so that you can see if it really is potential free. I imagine everyone already has access to the BMV specs.

Link to comment
Share on other sites

BMV is also potential free. So one could use battery to pass current through the BMV through the heatpump relay and energise a AC contactor with a 48VDC coil. TTT contacted Victron who said that the BMV relay could switch 230V AC. There was a looong discussion about this on the forum a year ago. I'll try and find it.

Link to comment
Share on other sites

31 minutes ago, Chris Hobson said:

BMV is also potential free. So one could use battery to pass current through the BMV through the heatpump relay and energise a AC contactor with a 48VDC coil. TTT contacted Victron who said that the BMV relay could switch 230V AC. There was a looong discussion about this on the forum a year ago. I'll try and find it.

That's promising. I felt sure this would have been discussed on the forum and had a look but could not find anything so posted this.

Link to comment
Share on other sites

The simplest option is to pass 230V through the BMV relay and push that into the SIG2 connector. I would assume that whatever is inside the heat pump will be a 230V relay or contactor drawing less than 1 Ampere, so that should be all you need to do. You would of course have to test it first and make sure it really uses less than 1 amp.

Second option would be to use a lower DC voltage, get a contactor or relay of the same voltage, and wire the contactor/relay to the BMV in the same manner. Then use that secondary relay/contactor to switch 230V into the SIG2 connector.

What I'd probably do is use a 24V signal. Standard irrigation stuff, and 24V relays are easy to find. Small 24V transformer, switching through the BMV, then run that 24V wire all the way to the heat pump. Place the 24V relay at the heat pump, and connect 230V via that relay to SIG2. This way you have a low voltage signal going to the heat pump and no high voltage AC on the back of your BMV.

Edit: The 24V scenario is how I switch my well-pump. In this scenario I use a RainBird irrigation computer that already HAS a 24VAC output for "master valve", I pass that through a rectifier (half-wave was good enough) and switch a small 24V relay (broken garage door openers and daylight switches are good sources for such 24V relays) which in turn powers the pump. This means I need only a thin 24V signal wire running to the pump, so I avoid a whole lot of compliance and safety issues that way.

Link to comment
Share on other sites

P.S. I also have the CCGX on my system and thought about using its relay because it is 220VAC but its configuration seems more based around starting a genset and so wouldn't work for my application, since what I want to do is start an appliance when the battery is fully charged rather than when it is discharged. Perhaps I'm wrong and the config parameters on the CCGX are flexible enough to activate a deactivate a relay for this type of application?

Link to comment
Share on other sites

In normal operation the dry contact on Axpert closes the relay to start the gennie. You could have it powering a normal closed contactor. So the signal from Axpert to start the gennie in fact cuts the power to the heatpump.

Please tell me how you use CCGX in your system. I might end up merging a Victron and Axpert system at some point in the future.

Link to comment
Share on other sites

1 hour ago, Chris Hobson said:

In normal operation the dry contact on Axpert closes the relay to start the gennie. You could have it powering a normal closed contactor. So the signal from Axpert to start the gennie in fact cuts the power to the heatpump.

Please tell me how you use CCGX in your system. I might end up merging a Victron and Axpert system at some point in the future.

Ahh, good idea, that way I could reverse the NO / NC states. The only problem is that the Axpert works the relay according to setting 12 (point back to utility), which in principle I never reach, except in unusual circumstances. For this reason I had to rule out using the Axpert's own dry contact. It would work for energizing the relay because that is based on setting 13, which I do reach on a daily basis but at the other end of the relay scale setting 12 doesn't work for this application. The relay changes the mode of operation on the appliance rather than just switching it on and off so I do need it to switch the relay both ways every day - first when the batteries reach 100% (this will change the appliance mode to immediately heat the water to a higher temperature, making use of excess production) and second it needs to de-energize the relay a few hours later when battery charge state reaches, say, 98% (even thought the heat pump probably won't still be running at that point because it will already have heated the water to the set temperature).

I use the CCGX because it is very reliable for reading system state, locally and remotely - it is the most reliable system I have used for this. Of course, it does not read the Axpert values directly so while I cannot see exact net PV production and total household consumption in the CCGX I can see the difference between them and so know more or less what is going on. Because I have the BMV in the system this does give me all the essential information in the CCGX for the battery state (ultimately the most important data). With this I find that the CCGX works well enough with the Axpert. Furthermore, all the additional features and configuration possibilities are still there with the CCGX and can be applied with an Axpert, like logging directly into the CCGX screen on the local network (or even remotely if you want to risk it) and using all of the alarm types and features. Loggin in to the CCGX live screen remotely means that you can operate the CCGX from there as well as see all it's values. Here's an image of one of the screens of the CCGX attached to an Axpert, showing general system state and one of the BMV readings.

 

ccgxrem.PNG.b6491d8d23297a0c7587d8a425983f6b.PNG

Link to comment
Share on other sites

On the CCGX under Settings/Relay you can configure the relay. I'm not sure how the generator start and pump functions work, I assume the relay is then controlled by another subsystem, but the other two settings make more sense. The default is to activate the relay when there is an alarm condition, but you can also activate it manually.

If you set it to manual, you can control it in software using dbus, see attached images.

Doesn't look like an SoC based setting is possible out of the box, but it would be easy enough to make your own, at least if you have some python experience. You can look at my bms-bridge code to get an idea how to query and write a setting. That project has everything to create your own little service on the device. The constant querying of the value is a little inefficient, it is possible to subscribe to changes, but it's probably good enough.

Selection_062.png

Selection_063.png

Link to comment
Share on other sites

23 hours ago, plonkster said:

 

Quote

The default is to activate the relay when there is an alarm condition

This could be the solution except that I have several alarm rules set up in the CCGX so that I receive timely information on system state by e-mail and I don't want the relay to activate for all of these alarm rules, of course, just one. It's odd that the CCGX doesn't allow to selectively apply the alarm rules that you set up to the relay so that not all alarm conditions activate the relay but only the one(s) you want. In this application I just need the relay to activate when my batteries reach full charge and then deactivate again when they have started to discharge.

Given this shortfall in the options for the alarm relay, it looks like the only way to use the relay on the CCGX would be as you describe by setting it to manual and including a software layer in the middle.

 

 

 

Link to comment
Share on other sites

1 minute ago, incagarcilaso said:

including a software layer in the middle.

Hehehe make no mistake, that's how all these things work. A feature request comes in, and someone implements the code. It's not an intentional shortcoming, just not enough people asking for it, and because there is a relay on the BMV that already does the job there's probably not much demand. The CCGX did however gain software SoC estimation in 2.00, so perhaps it now makes more sense to add such a thing.

Link to comment
Share on other sites

1 minute ago, plonkster said:

Hehehe make no mistake, that's how all these things work. A feature request comes in, and someone implements the code. It's not an intentional shortcoming, just not enough people asking for it, and because there is a relay on the BMV that already does the job there's probably not much demand. The CCGX did however gain software SoC estimation in 2.00, so perhaps it now makes more sense to add such a thing.

Sure, and getting these implemented is not easy - but some things just seem so obvious that you wonder why it needs a request. I think most users will have more than one alarm configured ...

Yes, and I will end up using the BMV but wouldn't it be good to have two separate relays at your disposal on a system with CCGX installed witht the BMV doing SOC (pretty common hardware stack).

Link to comment
Share on other sites

14 hours ago, incagarcilaso said:

... - but some things just seem so obvious that you wonder why it needs a request.

Makes me think of when the earth was flat.

True story, I wanted to switch circuits based on SOC, for volts just did not work any which way we tried, way back in 2010/11. Contacted all the local and international solar device manufacturers asking for a solution. No-one even responded. Then I asked the right person, whom knew of BMV's, problem solved. Ps. I swear my emails was read and someone said: Interesting, this makes sense. Lets look into it.

Like today on PF it is the only way to do it, switch on a BMV's SOC. But I think we have gone and became a wee bit complicated. There is a simpler mechanical way to do it with relays and a BMV, along the same lines as I have done with my box, with just one more relay added, as I did in the end.

 

Any case, I have found that what is obvious to one person, is a WOW idea for the next person. 

Also find that us humans tend to get stuck in a rut, don't want to change, to complicated or too simple an idea. 

Or we all tell each other an idea is good as it gets and that's that. That is until someones comes along with a new idea, but then again we tend to chase them away for upsetting the apple cart. Or sometimes that new idea grows into something even better.

Or one day another person brings along the same new idea as the 1st person, repackaged and better sounding maybe, now 2 people says it is good, then the next person tries it and the next and suddenly that becomes the new norm.

OR as in case of if enough people asked for something, makes me wonder maybe they didn't think of it / knew it was possible where if it was there, it would have been the new norm. Then again if you add the new idea that is not asked for, sometimes you open a whole new can of worms. (EDITED)

Bottom line, nothing stays the same, we don't like change, sometimes we get very complicated, but then again new ideas cannot be kept silent indefinitely.

Never a dull moment I tell you.

Link to comment
Share on other sites

10 hours ago, The Terrible Triplett said:

Makes me think of when the earth was flat.

That was a really really long time ago. Perhaps a little OT, but seriously, people don't realise how long we've known that the earth is spherical. Often in some kinds of debate you'll hear someone trying to pry that back into the time of Galileo (and Galileo then gets backported into the "dark ages", he was a contemporary of Descartes for crying out loud)... but at that point the question was Heliocentrism (whether the earth or the sun is in the middle) and even at that point it was already sort of known that the sun goes in the middle because people like Copernicus has been working on it for decades, but delayed publication because of the backlash he feared... and the backlash, contrary to what people will tell you, was actually from academics who were married to the ideas of Aristotle (who lived like a 400 years BC) and the mathematics of Ptolemy. There... that's some stuff you didn't know but now you do :-)

10 hours ago, The Terrible Triplett said:

Any case, I have found that what is obvious to one person, is a WOW idea for the next person. 

I tend to agree. It's actually not that obvious, especially in the moment. I can imagine that faced with the prospect of having one relay and a hundred things to do with it, the decision was made to do one obvious thing with it (it's a monitoring device, let's allow people to do monitoring with it, such as switching on a big red light when there is an alarm condition) and then provide a Manual option for everyone else. YAGNI, it's a software thing.

Since Venus 2.00 (December 2016) there is a kind of software SoC support, basically the CCGX tells the Multi what the MPPTs are doing and it comes up with a better SoC. In the past the Multi would just default to 85% because it didn't know what the MPPTs were doing. It might make sense to add such things now, along with the other feature that's come up here before: Can I switch a relay when there is a high current draw (which almost seemed weird when it turned out that the BMV doesn't support that kind of thing).

Edit: If you really want it, I would suggest heading over to the venus issue tracker and logging an issue. It will probably be labeled as an enhancement, but then at least others who want it can comment on the issue and it might gain some momentum.

Link to comment
Share on other sites

I like your history lessons Plonkster. I like.

25 minutes ago, plonkster said:

... especially in the moment.

Today the new norm is, Min SOC, Max SOC, panel watts possible, between these times and if there is a huge load ... do something.

I think it could be done with one relay.

Link to comment
Share on other sites

2 minutes ago, The Terrible Triplett said:

panel watts possible

That one is still somewhat of a problem. Long example following...

In a recent version of Venus, the feedback mechanism was changed to incorporate solar production into the maths. So if you set the "Maximum Discharge Power" setting to 1000W, and there was also 1000W of PV coming in, then the Multi would be adjusted to allow as much as 2000W to go to the loads (the rest comes from the grid). Once the batteries reach the minimum configured SoC, it would only allow the 1000W of PV through. It was an awesome feature and though I had to patch it a little, I used it for a long time. But there was a problem with it...

The problem is that you have to very carefully calculate how much power you really have available, taking into account losses and conversion efficiency. This is almost impossible to do, and the developer did what most people would do (and what I would argue is the right thing to do): make a simple approximation that is good enough, that is, assume that 92% of the input power makes it to the output. The trouble is that these numbers aren't constants, they vary according to load and temperature, and as such you will always be off slightly. In this case, it was off towards the discharge side: When all solar power was passed to the loads with the batteries at MinSoC, it would slightly overestimate the available power and end up slowly discharging the batteries past the minimum level. Not a big problem if you have enough PV to correct this later, but during prolonged cloudy periods it is a problem.

The alternative is to slightly underestimate the available PV, but that has a problem too. The remainder of the current (the bit you underestimated) will end up charging the battery, and if that is then enough to push the battery voltage up to the point where the MPPT pulls back, you have an interesting domino effect where you continually underestimate the power and set the feedback too low, causing more current to go to the battery, causing the MPPT to pull back, causing the entire thing to wind down until nothing is pushed back.

In short, you cannot just ask the MPPT what the current charging power is, that is not an indication of how much is really available.

The patch I used was a simple one: It looked at the current SoC, if it was below the minimum, it would perturb the feedback current and make it pessimistic, causing the battery to slowly charge. Once the Soc reaches the minimum, it would perturb it the other way, make it slightly optimistic, causing it to discharge slowly. The effect was that the battery would hover around the minimim SoC within about 0.5%. Downside: not sure this is good for the battery over long periods.

My new idea is this: Use something similar to the congestion protocol used by TCP/IP. That works like this: Ramp up the download speed until packets go missing (the congestion point), then halve it. Ramp up again from the halfway point until packets go missing, then pick a halfway point between the previous congestion point and the current one. The effect is a binary search for the congestion point until you are just under it.

Similarly here: Ramp up the feedback power until the charger cannot maintain the set voltage (whether that is float of absorb), then halve it. Ramp up from there until it happens again. And so on and so forth, binary search for the solar power level, and then you stay just under it.

When I have time one day.

Link to comment
Share on other sites

Jip, after reading all you have said over the months, that is why we have not added panel watts yet into the equation.

I got a raw idea from my controllers. Inverter runs lekker at max load whilst the controllers says it is in float. At the time the panels are running at near max watts.

So I though, mmmm, if I set a variable to say that if the panel watts coming in charging the batts are = or > said variable and SOC is Y, start using solar. If the batts takes the whole day to charge, that is fine.

But, if by say variable of a pm value, the batts are not yet at Y SOC, then switch back to Eskom to allow controller to charge to set SOC that I need for the evenings.

It is a balancing act, and it will change all the time based on weather and time of the year.

So ja, it is not yet ready to go into development phase.

Link to comment
Share on other sites

Something else I considered is for the software to learn what the curve is like. It is relatively simple to know when the MPPT is at max power, just look at the voltage and if it is below float, then you know whatever is coming from the MPPT right now is the max it can do. Over time the software can learn what the curve looks like, and then (this is the important sort of trick), we can use some characteristics from PV panels. They are constant current devices, and the current is more or less proportional to the amount of insolation, whereas Vmp comes down at the same time. See attached image. So if you have a current and a voltage point, armed with the known curve of the array, you could in theory predict what Pmax will be.

 

Selection_065.png.68a2d226649b6b1bf2e564156c9c23c4.png

 

Trouble is, if you're over by 20 watts and the batteries are close to full... you could end up in the death-spiral I mentioned earlier. So you'd still need to fine-tune it which kinda tells me, just go with the simpler option :-)

Link to comment
Share on other sites

On 5/10/2017 at 11:36 AM, plonkster said:

... just go with the simpler option :-)

And that is the truth, nothing but the truth. 

Jip, we have the prediction curve. Proof of concept works lekker till variables come in like clouds and all that.

So idea I had was to measure the irradiation at the panels and add that to the calcs, till I saw the costs of good ones.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...