2W1ETN

After a day of getting a new box setup Reflector 21 was go and has been working flawlessly all day on the 29th July 2009 (OK the time was out, NTP installed and that’s fixed).

GB7CD was the fat repeater connected to the new reflector, soon followed by the HotSpot based GB3WE.

Having GB3WE connected it seems quite clear that to a user connected via RF to GB7CD that users on GB2WE appear as though they were using any Icom repeater. Their call sign, and user message all get routed through, even in the Dplus log of GB7CD there is no noticeable difference.

I don’t see how any objection could be based on little more than Myth that has prevented GB3WE and other node adapters being connected to the other UK Reflectors.

The MidStar D-Star repeaters are to start using port B on Ref21, where they also intend to make use of node adapter based repeaters, to complement their current Icom based one.   Â

I understand there has been a lot of interest in exactly where Reflector 21 is located and the kind of hardware and network connections used. Here is the low down, its located in London at a Blueconnex Data centre
The hardware is a Quad Core Xeon, 4GB of Ram, Mirrored disks The network connectivity is provided currently by a Duplex 100Mbps port, which has the capacity to be a 1Gbps port, but I’ve turned it down on the switch, as I doubt we need 1Gbps.

Internet Connectivity to most of the UK is via LINX which should provide a decent low hop count to ADSL lines which so many reflectors use. From other systems connected in different Datacentres around London, the average lag seems to be about 3ms, with about 10ms from Cardiff’s GB7CD over ja.net   Â

By all means feel free to connect your node adapters to the reflector. If it doesn’t work disconnect it. We wont ban you!

So I’ve decided I need more than one additional battery to power the computers and ham radio gear in my Range Rover. Question is just what is the best way to connect 3 leisure batteries together and charge them from the main alternator/battery.

Here is a diagram showing the traditional way of interconnecting 4 batteries to form one larger bank. This is a method that we still see in many installations.

 

Method 1

Notice that the connections to the main installation are all taken from one end, i.e. from the end battery.  The interconnecting leads will have some resistance. It will be low, but it still exists, and at the level of charge and discharge currents we see in these installations, the resistance will be significant in that it will have a measurable effect.   Typically the batteries are linked together with 35mm cable in a good installation (often much smaller in a poor installation). 35mm copper cable has a resistance of around 0.0006 Ohms per metre so the 20cm length between each battery will have a resistance of 0.00012 Ohms.   This, admittedly, is close to nothing. But add onto this the 0.0002 Ohms for each connection interface (i.e. cable to crimp, crimp to battery post etc) and we find that the resistance between each battery post is around 0.0015 Ohms.

If we draw 100 amps from this battery bank we will effectively be drawing 25 amps from each battery. Or so we think.

In actual fact what we find is that more current is drawn from the bottom battery, with the current draw getting progressively less as we get towards the top of the diagram.

The effect is greater than would be expected.

Whilst this diagram looks simple, the calculation is incredibly difficult to do completely because the internal resistance of the batteries affects the outcome so much.

However look at where the load would be connected. It is clear that the power coming from the bottom battery only has to travel through the main connection leads. The power from the next battery up has to travel through the same main connection leads but in addition also has to travel through the 2 interconnecting leads to the next battery. The next battery up has to go through 4 sets of interconnecting leads. The top one has to go through 6 sets of interconnecting leads. So the top battery will be providing much less current than the bottom battery.

During charging exactly the same thing happens, the bottom battery gets charged with a higher current than the top battery.

The result is that the bottom battery is worked harder, discharged harder, charged harder. It fails earlier. The batteries are not being treated equally.

Now in all fairness, many people say “but the difference is negligible, the resistances are so small, so the effect will also be small”.

The problem is that in very low resistance circuits (as we have here) huge differences in current can be produced by tiny variations in battery voltage. I’m not going to produce the calculations here because they really are quite horrific. I actually used a PC based simulator to produce these results because it is simply too time consuming to do them by hand.

Battery internal resistance = 0.02 Ohms
Interconnecting lead resistance = 0.0015 Ohms per link
Total load on batteries = 100 amps

The bottom battery provides 35.9 amps of this.
The next battery up provides 26.2 amps.
The next battery up provides 20.4 amps.
The top battery provides 17.8 amps.

So the bottom battery provides over twice the current of the top battery.

This is an enormous imbalance between the batteries. The bottom battery is being worked over twice as hard as the top battery. The effects of this are rather complex and do not mean that the life of the bottom battery will be half that of the top battery, because as the bottom battery loses capacity quicker (due to it being worked harder) the other three batteries will start to take more of the load. But the net effect is that the battery bank, as a whole, ages much quicker than with proper balancing.

 

Method 2

 

All that has changed in this diagram is that the main feeds to the rest of the installation are now taken from diagonally opposite posts.  It is simple to achieve but the difference in the results are truly astounding for such a simple modification.   The connecting leads, in fact, everything else in the installation remains identical.   Also, it doesn’t matter which lead (positive or negative) is moved, Whichever is easiest is the correct one to move.   The results of this modification, when compared to the original diagram are shown below. Only that one single connection has been moved.

 

After this simple modification, with the same 100 amp load….

The bottom battery provides 26.7 amps of this.
The next battery up provides 23.2 amps.
The next battery up provides 23.2 amps.
The top battery provides 26.7 amps.

This is quite clearly a massive improvement over the first method. The batteries are much closer to being correctly balanced. However they are still not perfectly balanced.

How far is it necessary to go to get the matching equal?

Well, the better the quality of the batteries, the more important it becomes. The lower the internal resistance of the batteries, the more important it is to get them properly balanced.

So that now leaves the question of whether or not there is a wiring method to perfectly balance the batteries.

Before getting to that, it should be pointed out that doing the calculation is not actually required in order to arrive at the ultimate interconnection method. I simply did them to show the magnitude of the problem.

In order to get a better balancing it is simply necessary to get the number of interconnecting links as close as equal between each battery and the final loads.

In the first example the power from the bottom battery passed through no interconnecting links. The top battery passed through 6 links.

In the 2nd example (the much improved one), the power from the top and bottom battery both passed through a total of 3 links. That from the middle 2 batteries also both passed through 3 links which begs the question “why were they not therefore perfectly balanced?”. The answer is that some of the links have to pass more total current and this therefore increases the voltage drop along their length.

And now we get to the correctly wired version where all the batteries are perfectly balanced.

 

Method 3

This looks more complicated.  It is actually quite simple to achieve but requires two extra interconnecting links and two terminal posts.   Note that it is important that all 4 links on each side are the same length otherwise one of the main benefits (that of equal resistance between each battery and the loads) is lost.   The difference in results between this and the 2nd example are much smaller than the differences between the 1st and 2nd (which are enormous) but with expensive batteries it might be worth the additional work. Most people (myself included) don’t consider the expense and time to be worthwhile unless expensive batteries are being fitted or if the number of batteries exceeds 8.

 

This method isn’t always so easy to install because of the required terminal posts. In some installations there is simply no room to fit these. So, thanks to a colleague, we can also present another wiring method that achieves perfect battery balancing.

 

Conclusion

I’ll be opting for method two, as its simple to install, and with the three batteries I’m opting for the difference in amperage between batteries will be somewhat lower than as above. Once installed I’ll take a few photos

I’ve been conducting some tests today to make my own Node Adapter software in place of the VB6 application currently available for use with Satoshi board.
Id have to say my efforts have been so so, considering what I perceive to be the “unopenness” of a number of applications key to the system. I’m not deterred by hopefully by the end of the week I will be able to connect as a DV-Dongle user and save the audio streams.

 

I think I will just stick to packet sniffing port 20001 which is how DV-Dongle users connect. The protocol looks simple enough, but it would be considerable easier with decent documentation.

Last week when testing the Range Rover with everything running I discovered the single 105a/h battery I had placed in the boot wasn’t quite up to the job.

Ok it has 500 potential recharge cycles at 70% usage, but running it flat I’m going to really deplete that potential and thus I’ve come to the conclusion that I need another two.

With three batteries in the boot, with about 300amp/h between them, I shouldn’t need to have to start the engine to apply a charge. The main car battery is disconnected by means of a relay; thus when the engine is off the batteries now in the boot are isolated from the rest of the car.

 

You might wonder, why not just run the engine? Well despite all my efforts of finding an ex police Range Rover, applying additional ferrite filters to a lot of the loom at some expense, there is still on 20 meters a lot of noise which no doubt is coming from the ignition.

I’ve recently purchased from high quality Magnecor HT leads which I’m hoping will reduce the problem. They are Silicone competition leads which promise “excellent RFI and EMI suppression”. Fingers crossed that coupled with some new spark plugs (well why not) the issue of RFI will be gone.

 

I also need to get a move on and find a way to mount the solar panel to the roof. I’m currently seeking a Expedition Roof Rack to enable me to mount it on. Being a 40watt job, it should address the issue of battery discharge when doing very little and keep everything primed without the need to remove the batteries and trickle charge them.

Basically where the vehicle is parked, its half sticks out of the underground car park. So it seems a shame to not make use of the light and charge it up not only when static at home, but also when on a beach or on top of some hill. It will hopefully ensure that whilst receiving at least, the amount of powering being used won’t exceed what the panel is generating (I wish).