Isolate solar charger PWM pulses

[solvejskovlund]

TL;DR Summary

How could we stop house wires to act as antennas for the pwm charging pulses from the solar charge controller?

After we got the solar array working, we made all lamps, laptop chargers, most kitchen appliances... run on 48vDC (so that we can have the inverter switched off unless we're running any 220v appliances). Then everyone in the house got a headache. We traced this down to being indoors at daytime. Further we traced this to be caused by constantly changing electric fields being stronger than if we lived under the most powerful power line around here. We have measured up to 1400V/m in the living room. This all seems to be caused by the mppt (pwm with some extra logic) solar charge controller making high frequency pulses to charge the battery. All wires in the house seems to be acting like antennas for this high frequency pulses, causing our headaches.

How can we stop these pulses from spreading out to all cabels in the house?

The current workaround is that we manually disconnect all house wires when the solar array is connected. That is - either we have the battery connected to solar for charging, or connected to the house wires for consumption. Never both connections at the same time.
(It turns out that the 220V wires from the inverter does not have this problem. Problem is only related to the cables that are connected directly to the battery.)

I'd like to know if anyone has some ideas of how to leave both the charger and the house wires connected without getting these pulses around the house.
We've tried:
* using shielded wires (grounding the shield) - did help some, but not enough to be worth the effort to replace all house wires.
* put a constant voltage (45v) circuit between the battery and the house wires - did nothing.

We've thought about putting a huge capacitor in parallel to the battery bank, but we're worried this will confuse the mppt logic.

What could work (when we get more batteries) is to pull one battery out of the bank for consumption, leave the others for solar charging. When this battery is drained, pull another out of the bank, put the first one into recharging mode (from the bank) and once charged, return it into the bank. Circulating the batteries this way would probably solve the problem, but it would require an electrical engineer to switch on a light bulb, or some arduino setup controlling 30ish relays (planning for 6 batteries requiring least 5 relays for each battery (positive to bank, positive to recharge, positive to house wires, negative to bank/recharge, negative to house wires). Besides I'm not sure how to do the recharge in order to return the battery to the bank.

Just to avoid any confusion. The headache is felt at first. Then traced to be related to house wires being connected/disconnected. Then this all has been confirmed using a EMF meter "GQ EMF-390". The meter sets off an alarm in the living room while solar is charging. Once solar array (or the house wires) is disconnected the meter shows 0.

 

[Baluncore]

We've thought about putting a huge capacitor in parallel to the battery bank, but we're worried this will confuse the mppt logic.

You need a low-pass filter between the battery and the house. That should block the electrical noise from the MPP charger from reaching the house. The LPF would be made from a big capacitor as you suggested, but there would be an inductor between the battery and the capacitor. That inductor will hide the capacitor from the MPP controller. It may require a couple of LP stages to get good attenuation of the MPP noise.

Is the problem worse when the battery is low and being charged, or when the battery is fully charged ?

 

[solvejskovlund]

It may require a couple of LP stages to get good attenuation of the MPP noise.

Is the problem worse when the battery is low and being charged, or when the battery is fully charged ?

I haven't noticed if the battery SOC matters. I'll do some tests aiming to see if that may be the case. We've set the system to keep battery level in the interval 25%-80%.

Any ideas of what size of capacitor and inductor would be suitable?

 

[Baluncore]

Any ideas of what size of capacitor and inductor would be suitable?

Computation of the inductor and capacitor values will depend on the total current drawn by the house circuit, and the switching frequency of the MPP.

 

[solvejskovlund]

I tried the only 50V+ capacitor I found in my box, and the largest (by size) inductor I found there. They were 100uF and 47mH. The inductor packaging doesn't say how much current it can handle, but I doubt it will like more than 50mA. I connected them as inductor pin1 to battery+, capacitor pin1 to battery-, capacitor pin2 to inductor pin2. Then I connected the load wires to capacitor pin1 and 2.
According to the calculator at electronicbase.net/low-pass-filter-calculator they should give a cut off frequency of 73Hz.

As for the load I used my multimeter set to display vdc and a 3k3 resistor.
Checking with the EMF390 meter, there were no change at all.

One may state that the filter will not work unless there is a higher load pulling current through. Well, the problem in the house is independent of load. In daytime there is usually 0 load. We need a solution that works with 0 load.

As I said, the problem is that the house wires act as antennas. It does not matter if positive or negative wire is disconnected. BOTH has the be disconnected in order to make problem go away. Based on this and the LC-filter test, I don't think a passive filter is the way to got. There is a need to somehow isolate both wires from the battery.

Now as the battery has managed to be fully charged (within our set limit), I can also state that there is no difference if the battery voltage is 46,8V (this morning) or 53,2V (now).

 

[anorlunda]

Summary:: How could we stop house wires to act as antennas for the pwm charging pulses from the solar charge controller?

We have measured up to 1400V/m in the living room

What? Are you sure you are reading that Meter right?
Maybe microvolts/m

 

[Baluncore]

As I said, the problem is that the house wires act as antennas. It does not matter if positive or negative wire is disconnected. BOTH has the be disconnected in order to make problem go away. Based on this and the LC-filter test, I don't think a passive filter is the way to got. There is a need to somehow isolate both wires from the battery.

Your claim that house wires act as antennas is just an excuse and a distraction because it is unmeasurable. It suggests you are on a witch hunt.

If there is noise on both terminals of the battery then I would suspect you have no ground on your battery circuit and it is floating. Not only would it explain the noise problems, but it would also be dangerous as insulation will breakdown and there is no path for static discharge or nearby lightning strikes to reach ground.

The components you have on hand will not do. That is no way to rule out an engineering solution. You must find out and specify what maximum current your battery will be required to supply.

If both sides of the battery did need to be isolated, then you need two inductors with one capacitor, but first you MUST sort out your grounding problems and document all the circuits.

 

[solvejskovlund]

Each battery have a 40A fuse. I'm planning for 6 batteries i parallel. That is a total of 240A. That is for powering the inverter. Not relevant for this problem.
The DC distribution to the house wires (which is the relevant for this thread) passes through a 15A fuse. However, in daytime, when this is a problem, the current drawn is 0.

We're using a hybrid charge controller. It does combine solar charge controller, generator charge controller and inverter into one box. This box has terminals for:
Solar +
Solar -
Battery +
Battery -
Generator N
Generator L
Out (inverter) N
Out (inverter) L
Ground

We don't have a generator connected. All the other terminals are connected. We were wondering if Battery negative should be connected to ground. Also if solar negative should be connected to ground. As we could not find any connection schemes were this was done, we chose not to do so. When I tried shielded cable the shield was connected to ground. I'd guess that if the negatives were supposed to be connected to ground, they would not be separate terminals. Is this thinking wrong?

 

[Baluncore]

Is this thinking wrong?

Yes. You cannot advance by making assumptions because you will then never know what is real. You must get the documentation and read the specifications.
Take care.
You may need to check the equipment with a meter, or remove the covers to identify the grounding of the separate circuits.

 

[Rive]

We have measured up to 1400V/m in the living room.

That would be comparable to the EMP of a high altitude nuclear blast. Something is just a bit fishy there with that measurement You won't have just headache by that, but damaged electronic equipment too.

Each battery have a 40A fuse. I'm planning for 6 batteries i parallel.

Please stop. That thing is not something you are supposed to tweak around with. Get a professional.

Then everyone in the house got a headache.

Usually headache is not caused by EM, but by ultrasound. I suggest to change the direction of your investigation.

 

[berkeman]

Please stop. That thing is not something you are supposed to tweak around with. Get a professional.

Usually headache is not caused by EM, but by ultrasound. I suggest to change the direction of your investigation.

Both are good points.

@solvejskovlund -- Please get help with your wiring from a professional electrician, and please talk to your family doctor about your headaches. We are of little help on both issues.

Thread is closed.