Archived Forum

Tinbum has experience of a diverter and battery charging. If he doesn't notice this thread and respond, try a PM.

cheers, already pm'ed him hence the post.

You didn't even hint at that! PM me if he doesn't respond and I'll nudge him. Failing that, the obvious answer is you'll need to speak to the charger manufacturer to get their help. The ideal situation would be that the charger can accept a control input, so you leave the charger permanently powered and drive the control signal instead of switching mains. It needs more thought than it's had so far because you'd need to extract another output and condition it to suit the charger.

Hi,

I use Eltek chargers of 3 differing types.

Initially I was using these 1kW chargers;

http://www.ebay.co.uk/itm/351177203790?ru=http%3A%2F%2Fwww.ebay.co.uk%2F...

They are switched on the AC using AF mode but I did have to change a small relay within them as the contacts didn't last. (Same relay but with different contact material, also used a socket to aid later replacement if needed- haven't had to replace since ).

I still use the above but now I've also added 5 of these 175W chargers as one of my meters won't work with the diverter without some import being registered so I try and do the switching with small loads;

http://www.ppauctions.com/archived/archived_lot.php?arc_id=13837

With these I leave the AC on and switch the DC output, once 5 of these are on a 1kW turns on and the sequence starts again. (175W output but input is about 210W)

I then have some Eltek flatpack2 HE chargers

Hope that helps.

Nick

hadnt thought about switching the dc side. i wonder if better control couldnt be achived as Robert suggests using an output from the mk2 that can vary the current from the charger.

Roberts idea would be ideal and probably is feasible but I don't have the skill at the moment or time to learn it. It would certainly help me with my meter from hell!!

You probably want to consider the life of the battery too.  I wonder what all that micro-cycling does to the battery after a year or two.  I don't know enough about battery chemistry to know if's likely to cause an issue.  Any chemistry gurus out there?

http://batteryuniversity.com/ seems to be a good source of general information.

They do sell pulse chargers for batteries that are supposed to help the battery- slightly different I know to what I'm doing though.

Even with direct DC MPPT chargers you are still at the whim of what the sun is doing. I work on the philosophy that you know the battery will be damaged if you don't charge it and that its only the initial 'bulk charge' and some of the 'absorption charge' that will be subjected to the varying voltage and current. I can't see a better way to do it without importing from the grid or only using a part of the pv.

When there's some excess PV, but not enough to fully satisfy the charger without importing, what's the maximum frequency you'll switch it at?  Do you bang away at it as if it were an immersion heater, or do you adopt a much gentler algorithm?

Just came across this paper on pulse chargers which is fairly interesting:

http://www.ipenz.org.nz/ipenz/publications/transactions/Transactions98/e...

although it's 16 years old, so may not represent the state of the art.  But it certainly has a few cautionary statements such as:

A low cost charger may consist
solely of a transformer and full-bridge rectifier, and in the cheapest of these the output is often
determined by the mains voltage and the turns ratio of the transformer. Usually such a circuit has no
output filter so that during float charge operation the 100Hz AC ripple can produce micro-cycling of
the battery consisting of small charge/discharge cycles which are increasingly detrimental over time.

and in its conclusions:

A simple ‘unintelligent’ invariant pulse charging approach was shown to dramatically decrease
charging time but is also capable of gassing the battery and must be applied with care. In this
technique specific pulse widths and/or magnitudes must be chosen carefully to fit each lead-acid battery
and may need modification with battery age.

No much gentler.

I use them in AF mode and now its even more gentle since I added the five 175W chargers.

With just the 1kW chargers they were set for being on a minimum of 2 seconds though they do ramp their output up- hence switching the ac.

Now with the 175W they are on for about 10 seconds minimum unless the sun were to go behind a cloud when they are set for 0.5 seconds minimum. The 1kW's come on once five 175W are on.

All the chargers are switch mode.

Do you have multiple chargers attempting to charge the same battery simultaneously, or is each charger responsible for its own sub-set of batteries?

They all do one big battery bank 48v and about 1200Ah.

Do the chargers co-operate with each other, or are they all oblivious of their competition?

The 5no 175W all have a single control unit where the voltage is set so they all work at the same voltage.

The 4 no 1kW all have separate voltage adjustment  though I have set all these to the same voltage. I have just bought a proper Eltek  housing for these (as link above) so will be putting them in that which may interlink the voltage settings.

The Flatpack 2 HE chargers are all software controlled and do communicate with each other and also do staged charging.

The batteries are also charged directly from some PV using a Moringstar MPPT 60 charger. This in effect actually does the main control of the charging due to the setting voltages I put into the Elteks, which varies between summer and winter. The elteks basically put the bulk charge in and keep the batteries near/in the absorption stage.

No much gentler.  I use them in AF mode and now its even more gentle since I added the five 175W chargers.

Can you remind us what AF setting you are using?  My recollection is that your supply meters have very small "sweet zones" so I would expect that your AF setting would have to be very restrictive.

Depending on the sketch that you are using, the AF parameter will be in one these forms:

float offsetOfEnergyThresholdsInAFmode = 0.1; // <-- must not exceed 0.5
const int postMidPointCrossingDelayForAF_cycles = 25; // in 20 ms counts

My meters have changed (a few times). They used to be 1100J but  I now have one meter with a zero sweet zone and the other with 3600J

My sketch is old so I have;

float offsetOfEnergyThresholdsInAFmode = 0.4; // <-- must not exceed 0.5

that's not where I used to change it though,

I preferred to change what you used to call

#define JOULES_PER_WATT_HOUR 3600 // 0.001 kWh = 3600 Joules

Changing the value of JOULES_PER_WATT_HOUR to anything other than 3600 has never made sense to me.  My later sketches include a SWEETZONE_IN_JOULES declaration so that this aspect can be tidily changed.

An offset value of 0.4 means that your two energy thresholds are at the 10% and 90% points of the energy bucket.  Given that there will always be some overshoot, that's a very relaxed setting.  It's interesting to know that this setting is acceptable to your 3600J meter.  I generally set this parameter to 0.1 so that only 20% of the available range is in use.  This should be OK with most types of meters that are to be found in these parts. 

With hindsight, it was probably unwise of me to mention the limiting value of 0.5.  In practice, no system will work when set to this value because there will always be some overshoot which will result in penalties being correctly applied by the meter.

Given that there will always be some overshoot, that's a very relaxed setting. It's interesting to know that this setting is acceptable to your 3600J meter.

I had to use as much as I possibly could of the sweet zone with the first meters because it was so small and with trial and error I found this worked fine with no import at all.

That sounds a good approach to me.  Use as much of the penalty-free zone as you can get away with. 

If there's any drift between the two measurement systems (meter and Router), there is bound to be an occasional glitch/pulse.  The more of the 'sweet-zone' that is used, the more frequent such chargeable events will be.  If your new 3600J meter is working nicely alongside your various dumploads, it sounds as if you've got the AF setting just right.

It does work really well, so much so that when the meter was first put in, I thought it was faulty. It took 3 days to clock up 1kWh (that's its resolution). Even now its only reading 112kWh and it was put in on 23rd July, half of that has been in this the last month, which has been dull.

I just wish I could resolve what to do about the other meter.

I'm wondering how things are going to be with the new smart meters and if they are going to be rolled out for 3 phase supplies as well. I know BG are rolling out their smart meters now but has the final spec actually been finalized? 

Hi Tinbum

I think i'm following the setup you have. Why are you switching the DC output from the rectifiers though? Do the rectifiers not like having the AC switched? A surge issue? If the DC output is turned off does the rectifier use any juice?

Do you then have an additional inverter to get back to 240vac or are you using the 48v directly? (ie am i missing another part of the puzzle?) If you have an inverter is it just one or do you have a clever way to step up the output capability as well?

Thanks 

Hi,

The 175W are dc switched.

The 1kW and the 2Kw are ac switched.

The only issue I had was, as I described above, with a small relay in the 1kw. Since swapping that relay I've not had any problem.

When I had that relay problem I purchased some really cheap DC SSR's on ebay to try incase I hadn't solved it. As I had these I decided to use them when I added the 175W chargers. These 5 chargers use a total of 25w when all the DC SSR's are off, as shown by one of those plug in meters, and 1200W when all are on.

I had initially though about using DC switching but quality SSR's (Crydom) for DC were a lot more expensive. As the 175W chargers only output about 3A max I thought it worth trying the cheap ones. I'm not sure that I would risk them for 1kW or 2kW though.

I have one 4.5kW inverter and have this set on a 'sell mode'. This is controlled by an external photocell so that it comes on below a certain light level. I have this set a 4A (1kW) when we are up and at night I turn it down to 3A when I go to bed. This covers our base load. We will still import if we turn a kettle on when its dark or during the day if the PV isn't enough. It generally works very well though. Its also slightly inefficient as for a small part of the day the pv will be supplying when the inverter is also supplying. The chargers will put this back in the batteries though, with inefficiencies. I haven't made that part any better as I have an SMA system ready to be installed that will work with totally automatic use of the batteries day or night.

Hi,

Cool gotcha now. 

Your inverter allows you to say how much it will produce? Or are you just restricting it to 3A by some cleverness? & hence knowing it will only make 1kw or 4.5kw.

How do you use the AC? Does it just connect into your consumer unit or are there parts of the system that gets the free power but not all? 

How do you handle disconnecting from the grid when the grid power goes/comes back?

Are you moving to an SMA Island setup then? Or are you going to use the inverter with the battery built in?

Are you not happy with your setup or is it just easier to get SMA's? (very interested)

If you're dumping yours.... errr let me know ;-)

Thanks

You can set the inverter in 0.5A steps with a press of the button.

Everything get the power and I'm always connected to the grid though I could connect it so I could run off grid- which I have done for a month while we were having our DNO supply upgraded. It ran well off grid, but the diverter introduced a lot of flicker during the day and iI had to limit the max PV output. Switching 3kw loads on an inverter with a diverter makes it very hard work for the inverter to keep everything within range!!

No, I have a 3 phase sunny backup set and the auto transfer box to install. That's going to take some setting up using a diverter as well!!

Sorry, the system I'm using will continue to be used as we actually have 2 supplies and the SMA will just be on the one.

The main disadvantage with my system is that it wont automatically supply power from the batteries if you turn a kettle or oven on during the day when the PV isn't enough or on a night.

You can set the inverter in 0.5A steps with a press of the button.

Isn't the inverter a voltage source?  Is that setting a trip current?  What's the advantage of winding it back when you go to bed?  What happens if you don't?

[EDIT]

Ahh... never mind... I was thinking you were off grid.  When you're on-grid it's more a current source, and I guess you're trying to reduce your battery sourced exports overnight?

What inverter have you got that allows you to change its output?

How does it sync its waveform with that of the grid when its connected (back) to the grid?

Just checking I've got it, you've got the PV making 240 connect to the grid. spare PV gets rectified to 48v battery, which then gets inverted back to 240 which is then connected to the same AC as the consumer unit, grid & PV. ie you've got "input" from the grid, PV and inverter, whilst dumping spare into the batteries?

I didn't think you were allowed to connect a Sunny Backup - its missing something can't remember - but its not for UK connections? Or are you using a Sunny Island and a backup ?

I take it you've got the G59 raley built into the system as your inverter is connected to the grid?

(I'm debating about what I want to do - as long as I'm not frying anything!)