Yup, If I do such a thing, it will be with a very big FET. The issue with the FET is the transition period. During that time, it dissipates power, but after that it's just a small resistance. The resistance isn't a good thing though, higher resistance on that path. MOSFET and then latch it with a real relay? Perhaps just a really big powerful contactor?

Probably leave all this foolishness and go home to your Moma is more like it :-)

On 7/2/2016 at 9:32 AM, plonkster said:

Coulomb makes a good point. I thought of this PSU I removed from a Phillips Blue Ray player (the mechanical parts failed). This is a very decently put-together little board with proper isolation between the mains and the low end, and good air gaps as well.

But note the input stages. In the green, you see a capacitor. That's a capacitive dropper. From there it goes into the (rather small!) transformer in the red, and then it is turned into DC by the four diodes also in the red. Big electrolytic tank to make it smooth. Then it goes into the SMPS, which is in the blue block. That is, two stage PSU, first drop the current with a cap, then drop the voltage somewhat with a transformer, then do the rest with an SMPS.

So... how would this thing fare on a MSW inverter? I have no idea really. My gut feeling is to say that that little transformer has far too little iron for me to try it. In addition, capacitive droppers also have far worse power dissipation on MSW. So did the manufacturer take that into account when he specced that cap in green? Do I want to risk it?

:-)

A bit late to the conversation, but just trace that circuit again.  The cap in green is usually not a capacitive dropper in this instance, but rather a filter cap connected between live and neutral on the incoming AC - that is the cap Coulomb mentioned which might fail because of the sharp rise and fall times.

The "transformer" in the red block is usually a dual inductor / coil filter, not a transformer - effectively an inductor in both the live and neutral lines wound around a single iron core.

Not according to the symbols at the bottom of the board... :-)

Sent from my GT-I9195 using Tapatalk

2 minutes ago, plonkster said:

Not according to the symbols at the bottom of the board... :-)

Sent from my GT-I9195 using Tapatalk

The symbol of the dual inductor looks similar to a symbol of a transformer, but it is wired differently - just trace the tracks on the board.  It also has far to little windings to handle mains if used as a transformer.  

That would actually make a LOT of sense. Just by the physical orientation of it, the coil is likely a dual inductor. The cap is in series though, I don't think I traced that wrong.

Sent from my GT-I9195 using Tapatalk

24 minutes ago, plonkster said:

That would actually make a LOT of sense. Just by the physical orientation of it, the coil is likely a dual inductor. The cap is in series though, I don't think I traced that wrong.

Sent from my GT-I9195 using Tapatalk

Just check the cap as well, that cap looks a bit small to be used as a dropper in such a circuit.

If I judge that cap by its size it is probably a 10nF or 22nF. Using the formula for capacitive reactance ¹/_2*Pi*F*C  for 10nF, it gives you a capacitive reactance of 318K ohm and using ohms law (^V/_R) that will give a current of roughly 0.75mA at 240VAC. Say the secondary side of the SMPS gives 12V only (usually 5V, 12V and sometimes a few other voltages as well) and ignoring all losses etc. that would result in a current of about 15mA @ 12V. That might not even cover the load resistance on the SMPS output stage.  

Fine, I'll check it again when I get home :-P It's going to turn out that you're right... I just know it :-)

Sent from my GT-I9195 using Tapatalk

14 minutes ago, plonkster said:

Fine, I'll check it again when I get home :-P It's going to turn out that you're right... I just know it :-)

Sent from my GT-I9195 using Tapatalk

No worries, just my 2c. 

Tell me, 2 separate battery banks, both 12v, one connected and powering load.

So bank 1 gets a wee bit low so BMV gives me a audible warning just before it is flat.
So I connect bank 2 in parallel, then I disconnect bank 1.

What is the problem if done manually?

Just now, The Terrible Triplett said:

Tell me, 2 separate battery banks, both 12v, one connected and powering load.

So bank 1 gets a wee bit low so BMV gives me a audible warning just before it is flat.
So I connect bank 2 in parallel, then I disconnect bank 1.

What is the problem if done manually?

Nothing. You need to switch off / disconnect the load before you do it, so it's a manual process. This is fine if you can be there every time todo it. 

Something else that triggered my thoughts on this, I wonder if a dual battery regulator, as used in 4x4's would work in this case? The "national luna" regulator in my 4x4, as far as I understand, will charge both batteries when the engine is on, when the engine is off it blocks the 2 batteries. i.e. if there's a fridge connected and the 2nd battery runs flat, it doesn't drain the 1st battery. 

SO, what if it was connected in reverse, so that the two batteries could be charge separately, but then when there's load it engages both batteries. It has a 400A peak current so it would be fine for smaller installations.  

 

This whole idea is plausible and could help in using older & newer batteries together. BUT, batteries also drain at different rates, so I guess if you want to go this route you would need to monitor each battery and cut it off at a safe level as well. 

Off cause, if the research yield good results you could mix batteries, and in turn save some cash on batteries in the long run

 

Silver, you are now on the right track with where Ed is going. 

But he will share once he is ready. 

Why can I not connect the bank 2 to same busbar as bank 1 that is powering, and then disconnect bank 1 for the same busbar?

16 minutes ago, SilverNodashi said:

Something else that triggered my thoughts on this, I wonder if a dual battery regulator, as used in 4x4's would work in this case? The "national luna" regulator in my 4x4, as far as I understand, will charge both batteries when the engine is on, when the engine is off it blocks the 2 batteries. i.e. if there's a fridge connected and the 2nd battery runs flat, it doesn't drain the 1st battery. 

Yup, that's precisely where I got the idea too. It's a very simple device, it literally latches the batteries together when the voltage goes above a certain threshold. The idea here would be the exact reverse.

Those 4x4 things also suffer certain drawbacks. The second battery is never floated at quite the right voltage, due to the long cable run to the back. When you drive long distances, and where you have a solar panel on top to help, it's usually not a problem. For the most part, people don't go camping every weekend either so the average guy probably doesn't even notice. But it is there, and there is a solution as well: A boost converter that charges the second battery from the first one :-)

20 minutes ago, SilverNodashi said:

batteries also drain at different rates

That was the surprising bit, at least initially. When they are the same chemistry, then in terms of DoD or percentage capacity they don't drain at different rates. Each battery contributes pro-rata according to its capacity. The only caveat, as I mentioned earlier, is that at very low levels of discharge this isn't 100% true. For the first bit, the big guy does almost all the work.

In real life this means that when you parallel a new string with an old one, the new string's life is shortened. The result from that paper, which I will link again below, is that when you discharge to deeper levels, your new bank will "almost" reach its design lifetime, and that what kills it is the sulphation caused by "insufficient overcharge". Basically, to make it work you have to slightly overcharge the bank (because the older string "fills up" first and has to suffer some overcharge while you attempt to get the other one there as well), and if you don't get this right then the new bank dies of suphation.

But the objective here is precisely to avoid the sulphation issue via separate charging.

Of course the proof of the pudding is in the eating.

Here it is again, for those who haven't seen it yet.

http://neuralfibre.com/paul/wp-content/uploads/2007/05/can-we-now-sin.pdf

Assuming lead acid, yes. 

Ed is way ahead of us here ... 

4 minutes ago, The Terrible Triplett said:

Silver, you are now on the right track with where Ed is going. 

But he will share once he is ready. 

Why can I not connect the new bank 2 to same busbar as bank 1 that is powering, and then disconnect the bank 1 for the same busbar?

While it's under the load? I don't know but I guess it might work? Or send your inverter in for repairs again it something pops

 

Jip, never do that.

We are not talking solar panels. Just inverter and batteries.

Just now, The Terrible Triplett said:

Jip, never do that.

We are not talking solar panels. Just inverter and batteries.

You're still running DC volts and Amps - the source doesn't matter too much. 

BUT, and try this at your own risk, if the load is connected already (i.e. inverter + light and fishtank on), and the load isn't too high, it may probably not be an issue to add another battery. Once connected the current will be divided between the batteries. Could it cause damage? Probably. I don't have any unwanted inverters lying around to test with, so do this on your own. But, please share a video

1 minute ago, SilverNodashi said:

But, please share a video

No, I am not going to try it ... had enough of smoke lately. 

But I still cannot fathom why it will be a problem.

11 minutes ago, The Terrible Triplett said:

No, I am not going to try it ... had enough of smoke lately. 

But I still cannot fathom why it will be a problem.

hehe, it would be fun!

DC volts don't charge phase (not 100% correct, but I'm trying to bring across the message) between positive and negative. Thinking back to science class you learned that electrons flow between negative and positive, right? Those electrons don't ever stop flowing, unless the circuit is broken. Anything in between (resistor / LED light / motor / buzzer / etc) will keep the electrons flowing. 

When there's more volt & amps, more electrons flow. Think about the municipal water coming out of the tap at about 3 or 4 bar pressure. That's similar to Volts in this case. The amount of liters you get per minute is similar to Ampere. You can make less water running, but the pressure will be the same. So, at 12V you can have 1A or 100A. Current is "drawn" and not pushed. i.e. when you connect a light 10W bulb it only will draw 0.83A current. A 100W bubl will draw 8.3A. And a 1000W bulb will draw 83.3A. Simple math. Let's forget, for a second, about inefficiencies, volt drop, etc. 

Now, if you draw 83.3A from your 12V battery, and connect another 12V battery, that 83.3A gets split between the two batteries - google " current divider law". 

Let's use water as an example again: Imagine you want to water your garden and open the tap, you get say 100L water/minute at 3 bar. You can reduce the amount of water/minute by adjusting the tap, but the pressure stay the same. At 4 bar, for example, when the tap is open and you connect a sprayer to the hose, you get a lot of splatter. Now, imagine doing the same with a hose connected to a pool pump. MUCH more splatter. In fact, you'll probably battle to fit the sprayer while the pump is running. 

This is what happens when you connect / disconnect a DC power source to a load. The bigger the power source, and the higher the current being draw, will determine how easily it is to connect / disconnect it safely. Chances are you will just see a big arc and burn the cable / terminal / bus bar a bit. BUT, on disconnect, you might actually ^arc^ weld the cable and bus bar, or the terminal and cable together, causing it not to disconnect. 

 

 

An AC circuit on the other hand changes phase between positive, zero and negative. It does it very very quickly though - 50Hz (50 cycles per second) so the current drops (albeit very quickly and for very short periods of time) and doesn't cause this phenomena to happen. 

Check this out:

 

 

 

 

P.S. I'm trying to explain this in laymen's terms, so please forgive me if you're an electrical engineer - the right stuff is sometimes too difficult to understand or explain. 

This guy, every time I watch his videos ... I see myself. 

Ok, I think I understand better. Will let it sink in a bit.

29 minutes ago, SilverNodashi said:

Now, if you draw 83.3A from your 12V battery, and connect another 12V battery, that 83.3A gets split between the two batteries - google " current divider law". 

That is a good point, in fact, if one bank is fairly empty and you connect a well-charged one while under load, you might just have to handle all of the 80+ amps.

You definitely don't want to do this close to vented batteries either.

Definitely go with a solid state thing I would say. Or solid state with relay in parallel, close solid state first, then the relay. No spark that way, and you get the low impedance latch you want.

This is beginning to feel more and more cowboy to me. Perhaps better left well alone :-)

Now if only the small Victrons could be paralleled, like the 350W and 800W phoenix models... :-) Unlike the Axpert, as far as I know you can parallel inverters running off different battery banks with the Victron.

Still... might make sense with two 1600VA or 1200VA compacts... which will set you back a good 30k or more... :-)

10 minutes ago, plonkster said:

That is a good point, in fact, if one bank is fairly empty and you connect a well-charged one while under load, you might just have to handle all of the 80+ amps.

You definitely don't want to do this close to vented batteries either.

Definitely go with a solid state thing I would say. Or solid state with relay in parallel, close solid state first, then the relay. No spark that way, and you get the low impedance latch you want.

This is beginning to feel more and more cowboy to me. Perhaps better left well alone :-)

Now if only the small Victrons could be paralleled, like the 350W and 800W phoenix models... :-) Unlike the Axpert, as far as I know you can parallel inverters running off different battery banks with the Victron.

Still... might make sense with two 1600VA or 1200VA compacts... which will set you back a good 30k or more... :-)

There's a reason why it's easier / safer to go with a bigger system if the need requires it. 

OR, split the loads in the DB board. i.e. run all your lights (will probably be less than 150W, depending on your house) on one inverter + battery + charge source, the TV + PC / etc on another, and anything else on another. You could probably run the fridge, garage motor and alarm / electric fence together with the lights as well - see this as "critical load #1". TV, PC, cellphone chargers, etc, as "critical load #2", and pool pump / electrical tools / etc a "no critical load"

This way you could also salvage and mix batteries. Your lights could probably run 3 times longer (in terms of cycles) than the rest of the same model battery. 

The other option is hardware specifically designed to take more than one battery bank. Take the SolarEdge GTI. In it's original habitat, you have solar panels with boost converters on the roof, feeding a single 450VDC line that comes down to the inverter, and then you have the final stage on the ground (man, Afrikaans and German is so much more accurate here, Wisselrigter, Wechselrichter, that just says exactly what the last stage does).

So what the Tesla battery does is use a boost converter to feed the SolarEdge inverter. You could feed it with two boost converters running from separate batteries, as long as the boost converters are designed to handle such paralleling (the "optimizers" that goes on the roof are precisely such devices), you could theoretically chain as many batteries as you want (up to some physical maximum, no doubt).

Of course that is similar to paralleling two inverters. And I am not aware of any such hardware that already exists, other than the mentioned solaredge inverter, which is only half the solution.

There's a reason why it's easier / safer to go with a bigger system if the need requires it. 

OR, split the loads in the DB board. i.e. run all your lights (will probably be less than 150W, depending on your house) on one inverter + battery + charge source, the TV + PC / etc on another, and anything else on another. You could probably run the fridge, garage motor and alarm / electric fence together with the lights as well - see this as "critical load #1". TV, PC, cellphone chargers, etc, as "critical load #2", and pool pump / electrical tools / etc a "no critical load"

This way you could also salvage and mix batteries. Your lights could probably run 3 times longer (in terms of cycles) than the rest of the same model battery. 

Now you are seeing the light...

Sent from my SM-N900 using Tapatalk

8 minutes ago, SilverNodashi said:

OR, split the loads in the DB board. i.e. run all your lights (will probably be less than 150W, depending on your house) on one inverter + battery + charge source, the TV + PC / etc on another, and anything else on another. You could probably run the fridge, garage motor and alarm / electric fence together with the lights as well - see this as "critical load #1". TV, PC, cellphone chargers, etc, as "critical load #2", and pool pump / electrical tools / etc a "no critical load"

I like.

Have a feeling that more and more people are considering alternatives, coming over to the light side. (pun intended) 

To split it more succinctly, you have a distinct difference between devices that are:
(1) only used at night (lights);
(2) small devices that are on 24/7 like alarm systems;
(3) Then the rest - garage and gate motors tend to have a built in battery backup.
(1) and (2) tend to me small loads that are perfectly suited for 12v batteries and inverters, huge savings versus 48v batteries. And you may have a few spare from a replaced 48v bank.

Have a 'spesmaas' it will become more prominent into the future due to no Eskom failures, with the big banks being monitored making more people aware of their battery use resulting in them starting to eye them batteries thinking: I have to replace these one day. What is the ROI now really?

The penny drops real hard the first time you realise: Wait a minute, I do not have the AH I use to have in the bank.

Unless you are or on your way to go off-grid, batteries for city use does not make sense if you make an effort to save power at night. Even less so if you can feed back to Eskom, legally. Batteries in cities only makes sense if there are regular power failures.

8 minutes ago, edmundp said:

Now you are seeing the light...

Sent from my SM-N900 using Tapatalk

I've "seen the light" a long, long time ago already. Speaking my mind here is different. My dad's house' load used to be split like that. The cost of the extra cable was probably not worth the savings on the inverters. But now he has a couple small battery banks which don't all need to be replaced at once.