I know that online UPS'es are terribly inefficient, like 85% from AC one side in to AC other side out.

Here is a titbit to add to the calcs, which are way above my head, give me a pic and I tell you a story.
From solar panel to usable AC solar other side solar systems are averaging +-85% efficiency, taking all losses into account, and there are a lot, all perfectly normal. But that is assuming the inverter is +-92% and the charge controller +-96% efficient with some controllers nowadays even up to 98-99% efficient.

 

 

5 minutes ago, edmundp said:

Indeed but  we are still comparing apples with apples - working with 48V nominal across the calculation. The range of difference between 50.92 and 48.4 cannot explain a discrepancy of 3.8kW and 1.65kW?

that's where you're missing the point. 

If you were to draw 80A for 10hours, it won't exactly be 38,4Kw/hr (80 * 48 * 10). Between 50V and 48V you're loosing 1.6Kw. Between 50.9V (40,720) and 48.4V (38,720) you would loose 2Kw. See how much difference a 0.7V makes, over 10 hours?

These calculations are purely on the battery side. I didn't take any power factor or other losses into account. The above calculations are done by multiplying Volt and Ampere, effectively giving you VA (or KVA) values, and not true Watts. If ALL your devices had a power factor of 1, you could use these values as absolute. But since most appliances are inefficient, you can't.  

Perhaps don't try and convert to Watts, unless you know the power factor of each device connected to the inverter.

27 minutes ago, The Terrible Triplett said:

From solar panel to usable AC solar other side solar systems are averaging +-85% efficiency

I think that is a bit low, for just that one branch. The round trip efficiency (from the solar panels, into the battery, back out, into the inverter, and then to the appliance) is around 70%. Now we know that the overall efficiency is the product of the intermediate efficiencies. The inverter, for example, will be around 94% efficient at sort of half load (much worse at lower loads) so lets use that. A good MPPT Charge Controller is over 95% efficient so we'll use that. So if we use the official greek letter for efficiency (which I googled 5 seconds ago), then:

η is overall efficiency

 

• η_cc is conversion efficiency of the charge controller
• η_charge is how efficiently electricity is converted to chemical change during charging.
• η_discharge is how efficiently chemical change can regenerate electricity on discharge
• η_invert is how efficient the inverter is in making AC out of that.

Then (ignoring self-discharge, which is like 1% a month):

η = ηcc * ηcharge * ηdischarge * ηinvert

We already know the efficiency of the inverter and the charge controller to be 94% and 95%, so:

η = 0.89 * ηcharge * ηdischarge

Now for a lead acid battery between empty and 75% SoC, charge efficiency is actually pretty good, above 95%. Similarly, for a battery that is more than half full, discharge efficiency is also very good, lets assume 95%. So:

η = 0.89 * 0.95 * 0.95 ~= 0.8 = 80%

So this is sort of the maximum you can expect when everything is working perfectly, and when you charge slow enough.

Back to the topic at hand though:

I do believe that in looking at the Ah comsumed and just multiplying by the nominal voltage you will make a large error. The largest part of the charge inefficiency is a voltage inefficiency. For a 12V battery, for example, you discharge at 12.5V or less, so drawing 10A constant for an hour would yield you (as near as dammit) to 12.5*10*1 = 125Wh. To put that BACK into the battery, however, requires taking it up to absorption voltage of 14.4, and then put  back the same 10Ah, so 14.4*10*1 = 144Wh. 125/144 ~= 86%.

So you really have to take voltage and Ah measurements over shorter periods and integrate over that, and then I expect you will find that you're likely getting a good 10% more out of those batteries than the BMV might be letting on right now :-)

Edit: I see Silver addressed this already. Only saw that now.

Suggestion: Since you already have a BMV, how about trying out the VRMlogger stuff I posted about, and letting it log to the vrm site? Then you can get numbers you can plonk into a spreadsheet, and sommer test my instructions too :-)

https://github.com/victronenergy/venus/wiki/Installing-VRMLogger-on-Raspbian-Jessie

I half suspect there is going to be issues with setting up vedirect_interface, and it would actually help me if I could resolve that for someone so I can improve the documentation. I personally don't own a real vedirect cable and so that leaves something a bit lacking.

OK. I understand.

That being said - back on my initial question - in terms of efficiency:

Would it not be more efficient to run multiple inverters @ 30-40% than one big tank at sub 10% levels?

I think my major inefficiency (except maybe the Axpert - and hence the question on transformerless (ala Axpert) vs transformer (ala Victron) efficiency) is that I now have one large inverter hovering at sub 10% load for 90% of a day, thus power dissapaiting into "thin air."

If I for instance go for a smaller inverter (with "grid assist" to cover the hairdryer) would I not gain a noticeable amount of runtime to 20% DOD?

2 minutes ago, plonkster said:

Suggestion: Since you already have a BMV, how about trying out the VRMlogger stuff I posted about, and letting it log to the vrm site? Then you can get numbers you can plonk into a spreadsheet, and sommer test my instructions too :-)

https://github.com/victronenergy/venus/wiki/Installing-VRMLogger-on-Raspbian-Jessie

I half suspect there is going to be issues with setting up vedirect_interface, and it would actually help me if I could resolve that for someone so I can improve the documentation. I personally don't own a real vedirect cable and so that leaves something a bit lacking.

Sure thing - no problem in helping. Keen to see VRM for myself...

Sorry I have to: You know I can do that from SolWebRdr already as well ne?  My most humble apologies (really), but the temptation was too great.

 

 

Yes, to a great extend it will help. But, at what cost and inconvenience?

1 hour ago, edmundp said:

Is this the same for transformerless vs transformer based inverters?

Now to answer an older part of the thread :-)

There is only one truly "transformerless" inverter, and that's the high voltage grid-tied units (those that feed off a 380-450V string). Every other inverter will have a boost stage, and every boost stage still uses some kind of inductor technology, usually a transformer because it is very efficient and it gives isolation. In the present context, none of the inverters we're talking about is "transformerless".

Thinking back to the original planning 3 years ago, I remember that "repayment" was a thing that happened faster for large systems. The issue is simply that the batteries have to be replaced and the more often that happens, the longer it takes for the system to pay for itself. For very small systems, the answer to the ROI question is "never".

As I recall, my small 24V system (which is significantly larger than what we're talking about here) just about pays for its own batteries. So on their own, these small systems will cost you money, they are not going to save it. The question is whether they will save enough on the big bank, or on the re-use of batteries down the line, that it MIGHT actually work out okay. On this I'm sort-of feeling 50/50 at the moment. On my own system, I found that as long as I use good second hand batteries every 2 years or so... I save money. With new  batteries, I'd break even (that is, I'd lose money, because that money could have earned interest elsewhere). So in theory the use of those batteries in their "second-hand" incarnation could make or break it. I don't know... :-)

7 minutes ago, edmundp said:

Sorry I have to: You know I can do that from SolWebRdr already as well ne?  My most humble apologies (really), but the temptation was too great.

Well, I know you are sort-of lead developer on that, but what I had in mind was the charts drawn by the vrm site.

18 minutes ago, plonkster said:

Now to answer an older part of the thread :-)

There is only one truly "transformerless" inverter, and that's the high voltage grid-tied units (those that feed off a 380-450V string). Every other inverter will have a boost stage, and every boost stage still uses some kind of inductor technology, usually a transformer because it is very efficient and it gives isolation. In the present context, none of the inverters we're talking about is "transformerless".

Thinking back to the original planning 3 years ago, I remember that "repayment" was a thing that happened faster for large systems. The issue is simply that the batteries have to be replaced and the more often that happens, the longer it takes for the system to pay for itself. For very small systems, the answer to the ROI question is "never".

As I recall, my small 24V system (which is significantly larger than what we're talking about here) just about pays for its own batteries. So on their own, these small systems will cost you money, they are not going to save it. The question is whether they will save enough on the big bank, or on the re-use of batteries down the line, that it MIGHT actually work out okay. On this I'm sort-of feeling 50/50 at the moment. On my own system, I found that as long as I use good second hand batteries every 2 years or so... I save money. With new  batteries, I'd break even (that is, I'd lose money, because that money could have earned interest elsewhere). So in theory the use of those batteries in their "second-hand" incarnation could make or break it. I don't know... :-)

I digress - there are too many IF's involved to make a purely scientific decision.

The thing is - I am already looking down the line (which would still be many years down the road) and seeing doom and gloom and debt and financial ruin and upset SWAMBO to replace my battery bank. And that day will come.

The longer I can postpone that with small spread capital investments the better. Sell off old inverter. Replace with high efficiency smaller unit. Smaller sub systems with smaller I-replace-it-with-bread-and-milk-money. Add panels instead of batteries. And such and such. 

In my case (inclusive of how I use power, when I use power and on what I use power) this makes a lot of sense and could potentially quite radically increase the ROI and longevity of what I have.

I guess the point is mute - there is NO standard solution out there. Not a solution that is a 99% fit. That is only achieved by getting your hands very very dirty, spending LOTS of time on it, and being tenacious as humanly possible. No wonder there are two solar user profiles: Mavoeta RICH, or very technical. Pick your poison.

15 minutes ago, plonkster said:

Well, I know you are sort-of lead developer on that, but what I had in mind was the charts drawn by the vrm site.

Absolutely!

42 minutes ago, edmundp said:

Mavoeta RICH, or very technical. Pick your poison.

Something else in this equation. School fees. And I'm not just talking about the money you throw down a dark pit when you let the magic smoke out of something expensive (though those uncalled for events do provide similar unexpected learning experiences), but I'm talking about spending money on stuff that DON'T save money... and you know full well that it doesn't, but the real value that you get out of it is the knowledge, skill, know-how, and so forth.

How much is the knowledge worth to you? How much is the data worth to you? I've already been told by at least one person that there are very few people who sit on both these chairs: Software developer AND solar/renewable expert.

You want to multiply the learning by installing three small systems? Why the heck not! :-)

42 minutes ago, edmundp said:

Replace with high efficiency smaller unit.

This was pretty much my thinking with the 1.6kva Victron Multiplus. I want to again point people to the Victron Self-consumption white paper, and will link it below. For us city dwellers, you can gain a LOT by using a much smaller inverter. You can spend more on a better quality unit for example (hint hint nudge nudge). What is needed to make it work, is flawless grid fallback (for the hair dryer) and good bypass capability (at just 3.6kva, the Multiplus Compact I have doesn't really qualify as a good bypasser).

I should have bought the 3000VA 24V unit, not the 1600VA. I thought I was going to solve the problem by adding a parallel unit later, but the truth is, I should have bought a slightly larger inverter. That is true. But it is also true that there is no reason on the Good Lord's earth why I would need a 5KVA or 10KVA inverter. In that respect, I fully agree with you. Smaller, more efficient, higher quality.

(Y'all know the 3kva 48V blue inverter need a mere 25w to idle, and that on the smaller units it is as low as 8w?)

https://www.victronenergy.com/upload/documents/Whitepaper-Self-Consumption-and-Grid-independence-with-the-Victron-Energy-Storage-Hub-EN.pdf

Edit: I think I just convinced myself. Time to get the anti-islanding sorted, paper work done, and go hub-1. That's a <10k hit that will produce results almost immediately. Very small results that will likely not give the money back... but... you know. Techie here... not Mavoeta rich guy.

4 minutes ago, plonkster said:

Now to answer an older part of the thread :-)

There is only one truly "transformerless" inverter, and that's the high voltage grid-tied units (those that feed off a 380-450V string). Every other inverter will have a boost stage, and every boost stage still uses some kind of inductor technology, usually a transformer because it is very efficient and it gives isolation. In the present context, none of the inverters we're talking about is "transformerless".

Thinking back to the original planning 3 years ago, I remember that "repayment" was a thing that happened faster for large systems. The issue is simply that the batteries have to be replaced and the more often that happens, the longer it takes for the system to pay for itself. For very small systems, the answer to the ROI question is "never".

As I recall, my small 24V system (which is significantly larger than what we're talking about here) just about pays for its own batteries. So on their own, these small systems will cost you money, they are not going to save it. The question is whether they will save enough on the big bank, or on the re-use of batteries down the line, that it MIGHT actually work out okay. On this I'm sort-of feeling 50/50 at the moment. On my own system, I found that as long as I use good second hand batteries every 2 years or so... I save money. With new  batteries, I'd break even (that is, I'd lose money, because that money could have earned interest elsewhere). So in theory the use of those batteries in their "second-hand" incarnation could make or break it. I don't know... :-)

haha, you had to bring in the negative side of this argument

Yes, a smaller efficient, nut plonkster's response on the batteries says why such a system as a whole won't be. As I said, "at what cost"? 

You need to calculate the cost of everything involved, not just the inverter, or not just the batteries. For example, installing 6x 100W panels is far more expensive than installing 3x 200W panels. 2x 300w gets cheaper, to some degree as well. This is due to the amount of cable, connectors, and panel mounting structures involved. The efficiency of the bigger panels are generally also better. And MPPT charge controllers are also more efficient at higher voltages. 

 

So, here's a curve ball: 

 

To charge a "larger system" (Axpert 5KVA + 3KW PV + 48V/400Ah battery bank), and then charge a couple smaller batteries (100Ah?) from that system, to drive 300W inverters at night would be rather expensive - to the point where the cost/efficient watt doesn't make sense anymore. 

Remember, if you want to beat Eskom cost you don't just calculate your battery R/Kwh. Initially you need to calculate the whole system's R/Kwh. Then do the calculation again after 5/10 year - depending on what needs to be replaced when. 

 

 

This was pretty much my thinking with the 1.6kva Victron Multiplus. I want to again point people to the Victron Self-consumption white paper, and will link it below. For us city dwellers, you can gain a LOT by using a much smaller inverter. You can spend more on a better quality unit for example (hint hint nudge nudge). What is needed to make it work, is flawless grid fallback (for the hair dryer) and good bypass capability (at just 3.6kva, the Multiplus Compact I have doesn't really qualify as a good bypasser).

I should have bought the 3000VA 24V unit, not the 1600VA. I thought I was going to solve the problem by adding a parallel unit later, but the truth is, I should have bought a slightly larger inverter. That is true. But it is also true that there is no reason on the Good Lord's earth why I would need a 5KVA or 10KVA inverter. In that respect, I fully agree with you. Smaller, more efficient, higher quality.

(Y'all know the 3kva 48V blue inverter need a mere 25w to idle, and that on the smaller units it is as low as 8w?)

https://www.victronenergy.com/upload/documents/Whitepaper-Self-Consumption-and-Grid-independence-with-the-Victron-Energy-Storage-Hub-EN.pdf

Edit: I think I just convinced myself. Time to get the anti-islanding sorted, paper work done, and go hub-1. That's a

That is exactly what I am leaning towards. 24V 3000VA unit. You know.

Sent from my SM-N900 using Tapatalk

And now, just as I sat down with the popcorn the debate stopped? 

My vote is cast in favour of Ed and Plonk on this one ... 

It's hardly a debate if you already made up your mind asto what equipment you want to use, and why. 

20 minutes ago, SilverNodashi said:

It's hardly a debate if you already made up your mind asto what equipment you want to use, and why. 

True. But in many ways, we make up our minds the day we put down the cash. Then for six months after that we suffer from severe confirmation bias (money does that, you cannot deal with the knowledge that you wasted your money, so you convince yourself that you made the best choice). Well, 6 months is for cars, which is why "new car surveys" are for people who's owned them at least a year.

I think the model embodied in that white paper -- the self-consumption model -- could be done with other equipment too. Something like a small infini, if such a thing exists. Doesn't have to be Victron... they were just the first guys on the scene :-)

It's hardly a debate if you already made up your mind asto what equipment you want to use, and why. 

But is that not the way of the world? Pleas do not construe this as an attack on any device. My needs are... a bit unique. As many people's is.

Sent from my SM-N900 using Tapatalk

1 hour ago, SilverNodashi said:

It's hardly a debate if you already made up your mind asto what equipment you want to use, and why. 

Well, give Ed en Plonk some credit. Both are quite knowledgeable, both have solar, both are developers, one has Victron and has looked at Axpert, other has a Axpert and is looking at Victron, both will admit if they think they could have done better for themselves, nor are they walkovers and both are open to be swayed IF the debate brings facts that are compelling enough to listen to.

Me, what do I know, but I am very much open to a change of perspective, if the debate makes sense, and so far the two of them make a lot of sense. 

Strictly speaking, I got it wrong. It's not confirmation bias. It's the "sunk cost fallacy".

https://youarenotsosmart.com/2011/03/25/the-sunk-cost-fallacy/

I used to drive a BMW. An old BMW. I owned a Toyota before that. I kinda know how you convince yourself you didn't do the wrong thing :-)

Strictly speaking, I got it wrong. It's not confirmation bias. It's the "sunk cost fallacy".

https://youarenotsosmart.com/2011/03/25/the-sunk-cost-fallacy/

I used to drive a BMW. An old BMW. I owned a Toyota before that. I kinda know how you convince yourself you didn't do the wrong thing :-)

He he he he. Have the t-shirt for that as well. Peugeot.

Sent from my SM-N900 using Tapatalk

12 minutes ago, The Terrible Triplett said:

Well, give Ed en Plonk some credit. Both are quite knowledgeable, both have solar, both are developers, one has Victron and has looked at Axpert, other has a Axpert and is looking at Victron, both will admit if they think they could have done better for themselves, nor are they walkovers and both are open to be swayed IF the debate brings facts that are compelling enough to listen to.

Me, what do I know, but I am very much open to a change of perspective, if the debate makes sense, and so far the two of them make a lot of sense. 

Here's the hard truth: No inverter will ever suit everybody's need, nor work perfect in every environment. Victron, Axpert, ABB, MLT, Sunnyboy,  Microcare, Shneider, Infinisolar, Imeon, Goodwe, etc, etc. All these inverters came into existence cause the developer probably couldn't find an inverter which suited his exact needs. Or he saw a gap to make money, and hope his products will be used by many people. Just because it's expensive, doesn't make it the best. And just cause it's cheap doesn't make it the worse either. All of these inverters have some problems / bugs / drawbacks. The questions is: which will work best for you?

So Ed, are we going to do the split systems, with all the Plonkster said?

I already have. 1kw 12v pack on 600w pure sine. Saves about a kw on consumption during night time. Total cost - R1200 and a bit of fiddling about. 😃

Don't ask how yet. Still doing a proof of concept....

Sent from my SM-N900 using Tapatalk

Edit: No coc required. When I move I pack it up and away.

Had sparky in today, to give me a CoC for my light connections.

For my solar system, I am in the clear!!!

If your solar system needs Eskom:
- And you connect it to the DB board, as we all know, you must have a CoC.
- If you connect to Eskom using a wall outlet, that does not require a CoC. It is the same as a UPS.

Obviously no-one is going to connect a grid-tie inverter to a wall outlet, that is DB board.