Hi folks,
I am off grid and use a generator when there isn't any sun...
I have a couple of 6V batteries, wet, and charge them with an inverter welder that is managed by a small control system.

From the point of view of battery charging this has some issues but that is not what I am asking about in this thread.

My broadband, being in a very long rural line is to say the least somewhat temperamental.
With the router set at its default profile I get about 700kbps on a good day. If I set a custom SNR, I am using a Billion router, I can push this to around 1.24mbps on a good day.
A bad day is when it is very wet, particularly if it's been set for some time, but it doesn't have to be raining. Even heavy most for a prolonged period will drop the bit-rate and even cause a complete dropout on occasions.

My line is at least 8k, phisically, and tests tyo12k or more in bad weather.

So... I am actually fairly impressed that there BB works at all. I do not understand the wet connatation as BT have tested and say that there does not appear to be a wet, shorted, joint so I can only assume that the water has some other affect, perhaps it's dyelectric affects the transmission line characteristics.

My problem though is that the system drops out when the battery charger is running, at least it does some of the time. I can't determine any sort of pattern. From time to time the batteries will charge and the BB is uneffected, even when it is already poor. Other times what looks like a solid 1.2 signal will drop out in seconds when the charger is switched on and return just as quickly when it is turned off.

The charger will be adding ripple to there DC but the router PSU is a boost followed by a buck converter to produce stable 12V and both the DSL line and power have fairly chunky ferrites on then, 3 turns each, as common mode noise filters.

Now Obviously I am missing something but I doing not know what to look for or where/how to test it.

I have several reasonable multyi-meters and a modestly capable scope but that are not going too help if I don't know what I am looking for or what to do about it when I find the problem.

Sorry to be a bit thick but it simply isn't something I know enough about to fiddle with in any meaningful way.
I would very much appreciate any suggestions with respect to tests or modifications I could do/make that may help.

I have no issues when solar charging, although this may change when I add there PWM controller I am working on at the moment.
The charger is approximately 3m from the router, which I could change but that would involve a longer TEL/DSL line which originates within 3m of the charger and can't easily, as in I would need a new building, be moved.

Looking forward to being educated...
Thanks,
Al

 

... Wonder if the welder/generator has a stray magnetic field when it is on. Maybe a compass could detect it.

 

Well I expect the answer is going to be EMF or direct noise on the supply but I am not aware of any mechanisms that compromise electronics in the presence of static fields...
I suppose sufficient flux could conceivably saturate cores in magnetics but I haven't ever encountered it.
That said, I could be wrong, which is why I am asking in the first place.

However since you can't shield magnetically if a magnetic field is my issue, static or otherwise, my only option would be to change the proximity.

Anyone think gausse testing in the immediate area is worth looking at and if so why?

Al

 

You could try beefing up the common mode filtering to see if it makes a difference. At least that might help to home in on the source of your woes.

 

There could be a ripple or pulse super-imposed on the charging line between the generator and the batteries. Take a look with a scope. Maybe an electrolytic capacitor between the charging line and ground would be beneficial.
Another idea is to establish a common ground point between the generator and the battery group.

 

Hello,

Does the charger have a metal housing?
Is the cable of the router shielded?
There might be an EMI-RFI problem.
The following page of the EDUCYPEDIA has some links abou the EMI-RFI:
http://educypedia.karadimov.info/electronics/emc-emi.htm

Bertus

 

Well I expect the answer is going to be EMF or direct noise on the supply but I am not aware of any mechanisms that compromise electronics in the presence of static fields...
I suppose sufficient flux could conceivably saturate cores in magnetics but I haven't ever encountered it.
That said, I could be wrong, which is why I am asking in the first place.

However since you can't shield magnetically if a magnetic field is my issue, static or otherwise, my only option would be to change the proximity.

Anyone think gausse testing in the immediate area is worth looking at and if so why?

Al

Humidity changes and loss of signal indicate oxidation of the bare or tinned copper on your non-soldered connectors. Many Cat4/Cat5 cables are connected to a jack by simply wedging a wire into a fork of a sprint copper alloy. Coaxial cables are crimped and then screwed into its Mating connector with a connection of some spring copper tension. All of which can get oxidized at the interface.

The insulation in your cables can also become compromised and all breaches should be in an enclosure (connectors). Finally, remember that humidity is not 100% every time it rains, it is at 100% when there is fog. Cables can hydrate and cause attenuation. Outer cables are not easily penetrated but once water (humidity) penetrates through a cut or crimp connector it works down 3, 6 or 12 inches from the connector. Inner wires are not insulated as well as the outer jacket of the bundle.

You could try cutting 12 to 18" from each cable end, adding new connectors and assembling. Your high resistance (oxidized) connections will be gone. Humidity compromised cable will be gone (assuming there are no issues along the main length of the cable).

 

Useful stuff folks, thanks.

Charger in grounded metal enclosure - Yes
DSL / 12V power shielded - No

Poor / oxidised connections - Probably but they are outside my control on BT's line. typical rural installation IE chucked in the hedge at the roadside.

I will give the scope a to but wouldn't know how much noise is going to represent a problem for either the power of BB lines.
Anyone know what a BB signal should look like?

It did occur to me that I could try running the router from a separate battery whilst charging the main one, if it's still drops out then it has to be noise on the DSL line given that the power will be totally isolated.

Thanks,
Al

 

Hello,

How are the charger, battery and router connected at the moment?
Could you draw a picture of the set-up?
Perhaps a filter between battery and router might help.

Bertus

 

The charger is directly connected to the batteries and the charger current is managed based on the battery voltage.

The router, as with several other things, has its power supplied by a voltage stabiliser...
I have a boost converter trust lifts true battery voltage to 36 on the intermediate DC bus.
Connected to this are a couple, more to add as I need them, of buck converters that drop the 36 to whatever I need for the device in question... For the router this is 12v

0V is common, IE. none of the converters are isolated.

As far as I know the charger is also none isolated which means that battery negative would be at the same potential as my nutral-earth bond.
I have a power inverter connected, two actually, but neither have anything to do with the problem and do not affect its manifestation.

I can do a picture but it is so simple I thought that the text would be sufficient... happy to do one though if you think it will help.

Thanks,
Al