Surplus Control of a Battery Charger Load Bank

Getting Started
1

Hello all, this is my first post.

Currently using diverted surplus solar power to charge a 100Ah 16s LifePO4 battery however its largely a manual process that need automating. Battery power is then used for peak clipping and to supply the houses static load. It works very well and has been in use since Sep 2025. The diverter exports about 100w to grid for one or two seconds as its decision logic works out which way to go having said that a PJ1103C dual sensor says that a continuous 100W is being exported back to grid except at night, not sure why.

Thinking of using a binary up/down counter/integrator to control a charger load bank. The chargers are rated as 100W, 200W, 400W, 800W. A diverter needs to have two control outputs and an operating dead band. With =>100W of grid consumption the counter counts down removing load, if solar surplus is >=100W counter counts up increasing load. When in between the two the counter stops counting.

When the load bank cannot remove all surplus power it is dumped to an air heater. The PV array is 1.5kWp so little or no surplus power is wasted, essentially all of the surplus power is put to use.

So what diverter is best suited to my needs? :slightly_smiling_face:

2

Welcome, John, to the OEM forum.

In answer to the question: almost certainly a home-grown one.

But I’m a bit confused buy some of your terminology. You mention charging a battery, then a “charger load bank”, then an air heater. The battery and charger - or is it chargers? - and the air heater are fairly obvious, but what is the “charger load bank”? Is it for heating domestic hot water, storage heaters?

Based on Robin Emley’s “Mk2PVRouter” (see PV Diversion in ‘Learn’), I put together a 3-channel control for 3 immersion heaters (for a B&B), one tank fed the kitchen hot water & owner’s bathroom, and the other two tanks were linked top & bottom and supplied the guests’ bedrooms. The order of priority was the kitchen first, then the other two tanks equally in turn, and finally export to the grid.

It sounds to me as if you want something very roughly like that, but including the battery charger(s) as well as the heating loads.

3

If I’m reading it correctly, his charger load bank together with the air heater are “dump loads” to absorb excess PV production. (although it’s not really correct to call the charging of batteries a “dump” function)

Which adds up, as after the batteries are fully charged, if he still has excess PV energy to dump, it goes to the air heater.

Then again, I may be off in left field with this one.

4

Bill, that’s correct. The kit being developed is for Thailand or other such places where overhead supplies feed pole transformers and the consumer is charged or not allowed to export back to the grid. The charger load bank does indeed absorb excess solar power in a controlled way; in this case its the diverter’s auxiliary control relay that says when excess power available. The diverter is essentially being used as a ‘zero’ export controller.

5

Thanks Robert for taking the time to reply. The French Mk2PVRouter is of interest as it appears to have the capability for two control relays that may or may not be programmable in the way desired.

The present diverter’s control relay is being used to flag excess power, in this case its not to export to immersion heaters rather the desire is to redirect the excess to a battery charger load bank. At the moment a single variable current charger is manually set so that the diverters display reads zero or 0.1kW and needs to be tweaked to ensure it stays that way, tiresome but it meets proof of concept. The other way charging is done is turn on a 1kW charger, depending on the level of surplus the diverter display may very from zero to say 500W which is purchased from the utility. In this case the purchase price per kWh is essentially less than full price. The auto load bank is to reduce the surplus and keep it at around zero kWh consumption but in any regard even if one purchases kWh its not wasted it gets reused later on.

6

OK, I was thinking of Robin’s original English software in particular, running (in the application I mentioned) on an emonTx V2 or generally, on an Arduino Uno, which it will with minor changes.
Note, I am not thinking of using the “burst mode” with a cycle time measured in a few mains cycles, I suspect you were already assuming much longer times, in the order of tens of seconds or minutes, between adjusting the value of the load.

How do you do this, is it switches giving fixed steps or a continuously variable knob you turn?

The way I am looking at the problem is, it all hinges on the details of what amount of control you have and how you exercise this control over the amount of energy diverted into the battery, i.e, in effect, the charge rate. The more precisely you can do this, the better. It’s only when the battery can’t accept more energy that the air heater comes into play.

If you were to go down the homebrew (Arduino) route, you could have a lot of relays or equivalent switches, limited by the number of output pins available, to control either separate chargers or dump loads, or even to drive a motorised potentiometer up and down – replacing a front panel knob – and so giving very fine control, if that’s what one of your battery chargers can do.

You would get to decide, by setting the level of the “energy bucket” (in Robin’s terminology) where each output switched on or off, so by arranging these in the right order, and maybe with some interaction which is easy if you’re writing the software, you get to control exactly what is used to consume the solar energy at any given level.

To do this with something commercially available, I think you’d need to be looking at an Industrial grade programmable controller.

7

Robert, you raised some very good points.

Whilst the design is intended for the likes of Thailand it would create IMHO a renaissance for Solar Diverter technology. Sometimes everybody’s thinking gets stuck in the rut of established practice and pushing power back to the grid is one such case. Think of the tradeoffs DNO and regulators have to make to make our very old electrical distribution system safe and reliable.

Of course people with large oversized solar arrays are obliged to export to grid and earn a SEG 4p/kWh but what if the culture changed. No need to spend thousands on a large arrays, tailor size to meet ones home consumption needs and like me dump the small excess to air heater or indeed an immersion heater the rest going to battery.

As for the variable current 58.4v charger, there is indeed a knob on the front that adjusts current from 2A-25A the problem I have is that I cannot find a beast with APFC that is similarly adjustable. The consequence is that its PF ranges from 0.55 (2A) to 0.6 (20A) and whilst I am not charged for the reactive load component it will be creating THD components in the mains sine wave that I am not at all happy about.

Another issue arises in that the proposed fixed output 100, 200, 400 800 chargers with APFC, they need to be high spec medical type PSU as these feature galvanic isolation yet another challenge to be met. Here the charge rate is set by the 4 bit binary counter dynamically, no motorised knob but the idea of one was considered and dismissed as too complex with long term reliability issues.

Again if the culture changed, NOT a chance, the only power going back to the grid would be the intermittent 100w or so needed for the diverter to make its mind up as what to do. I believe the Thai regulations do allow a small amount of export power for 30s and only for control purposes.

8

Why do you need medical-grade galvanic isolation for a domestic battery? Or are you saying the only controllable battery charger with a respectable power factor is the medical sort?

9

It depends on the type of GTI.

Most Chinese models are dc coupled that means that a protective Type A RCD device’s coil may see a dc current that saturates it rendering it inoperable so a Type B must be fitted. If a battery is used to power the inverter it may see a dc path through to the mains. On the other hand German SMA inverters are ac coupled that is an interposing transformer is placed between the dc input and the battery. Well that’s my slant on things, to err on the side of caution.

In a failure mode no ac mains should find its way to the battery and by the way my Chinese 25A adjustable charger is ac coupled and disconnects itself from the battery when fully charged, similarly my Chinese IP67 1kW fixed current charger with PF of 0.98 is also ac coupled and it too disconnects from the battery when fully charged. The intended four PSU for the auto role must have CC capability as well as having ac (high impedance) coupling.

In the UK the design complexity of battery inverter interfaces is why the balcony-solar regs only allow solar panels at the moment but amendments will made next year 2027 to allow balcony-solar to dump excess power to battery and thereafter to grid or so I believe.

10

For the sake of fully explaining the problem, let’s say you can obtain (say) four a.c. coupled chargers with a decent power factor and with outputs in a convenient binary sequence in the ratio 1, 2, 4, 8 as you mention, giving you 15 charging rates/dump loads, and you have designed a controller which can measure the exported power and enable one or more chargers. Is it feasible to parallel the outputs on the assumption that each charger will either be operating at its current limit, or it will reduce its output if its target battery voltage is reached (even if another charger is still charging having a slightly higher voltage setting). I can see the potential for instability here.

One alternative thought: the variable current 58.4 V charger with the knob on the front that adjusts current – what is behind the knob on the other side of the panel? How does it interact with the electronics, and can this be changed so as to be done by something external (e.g. opto-couplers or reed relays)? Solving this would allow you to use the charger, then the problem transfers to: can you externally in parallel with setting the current, apply external and adjustable power factor correction, and if needs be harmonic filters, to clean up the circuit supplying the charger? I appreciate it’s a complication you’d prefer to avoid.

11

I have the 100W PSU/charger, a Mean Well NSP 100, linked here is the NSP 200 but its part of the same family. The 60V NSP 100 is CC with the clamp voltage set manually via a pot to around 57.5v at which point it starts folding back. Its DC output is connected directly to the battery. When the diverter threshold set at 100W is exceeded it brings on the control relay and both the NSP 100 and the adjustable charger. Both are quite happy coexisting without conflict.

The NSP 100 is first to fold back and turns off as the threshold voltage is exceeded. Then as the next threshold 58.4v is approached the adjustable charger starts to fold back and only this time it maintains a very small top up charge keeping the battery fully charged just as long as the control relay allows it.

So I have the 100w and prospectively the 200w and each can have their clamp voltages set which is fiddly but near enough so each other contributes to charging without conflict. All controlled charging has mains inrush current limiting as a precaution. So in theory the missing 400 and 800w can be similarly setup. It just happens that this family has a logic control that turns them off but at the moment its cycling the mains to the chargers that control them.

As for the adjustable charger, finding the Chinese OEM that makes it is to say the least challenging, so asking them to provide a one off custom job is really a non-starter simply because their commercial supply lines are so well protected.

Still missing though the two MK2 PV Router control relays and have messaged the French to see if its possible. I might add that I would be buying a fully built and tested device rather than self assemble as I just don’t have the time at the moment.

12

The French do not appear to be making or selling Mk2 for the moment, reasons unknown.

The prospective design has moved on to five chargers. On the working prototype the adjustable charger is set to its lowest value 2A and is always powered as long as the system says their is a qualified surplus . It starts to fold back at around 58.2v. The other four NSP series PSU have their outputs clamped or adjusted by pot to 57v. As all 5 are connected to the battery each device senses the same voltage and each pushes current into the battery without conflict at their own CC rates. As the battery voltage rises all four will begin to fold back however the 5th is still pushing current into the battery and by 58v is the only one doing so. It ends up topping off the battery at 58.4v. As the battery voltage begins to fall slightly the 5th charger keeps it topped up cutting off each time 58.4v is reached. By contract the 1kW power charger when used disconnects and stays off when 58.4v is reached.

13

I wasn’t aware of that, nor I suspect was Robin as we remain in contact and he hasn’t mentioned anything to me.

As you want to use commercially produced and readily available parts with no or minimal modifications, then what you’re suggesting does indeed look to be the only viable solution.