BMV702 Installation with 20m cable.

[superdiy]

The Victron battery monitors BMV700 & BMV702 are supplied with a 10m 24 AWG 3 pair UTP (patch) cable, but my preferred location for the display was almost 20m from the shunt. I decided to experiment with a longer cable, but feared that the additional cable length might have an effect on the measurements, especially when the backlight is on and/or when the relay and/or buzzer draws more power and that proved to be the case. For my experiment I took a 30m piece of CAT5e 24AWG 4 pair cable and initially connected 3 of the 4 pairs and left the 4^th pair unconnected. The difference in the displayed voltage readings with the backlight on and the backlight off were about 0.03V, compared to the readings on my Fluke. The difference in the current readings were about 20mA (backlight on compared to backlight off).

 

I then removed the PC board from the shunt and traced the circuit to determine on which wires the positive and negative battery connections and the load connection of the shunt, were running. Since I had a pair of unused wires in the 4 pair cable I then started, by first adding one wire of the unused pair in parallel with the positive (power) wire and then both wires of the unused pair, but the measurements stayed the same. I then added the unused wires in turns in parallel to each of the 2 wires running from the one side of the shunt to the monitor and did the same with the wire running from the other side of the shunt to the monitor. A notable difference was only seen when one of the unused wires were connected in parallel with the orange-white wire. I decided to make use of the additional wire pair (brown & brown/white) and connect that in parallel to the orange/white wire to effectively lower the resistance of the orange/white wire. The result was much more accurate readings and less difference in readings with the backlight and relay either on or off.

[Chris Hobson]

Beautiful. I went to the other extreme and shortened it to 1.5m. Got the last of the goodies today to control power output via the Victron BMV. (DIN rail Bell transformer as a 12V DC power source - 12 V DC relay fed via Victron relay which then controls a 20A AC mini contactor. This was cheaper than getting a contactor with a DC coil and the components were readily available in Graaff-Reinet. Time to finish off my installation.

[___]

Now I'm curious, what was on that wire? If I had to make a wild guess, the ground wire probably serves as both the reference against which the voltage drop of the shunt is measured AND to power the device. So any voltage drop over the length of that wire will affect not only the voltage reading but also the current reading (aka the voltage drop over the shunt). Adding additional current paths reduce the losses and if you reduce it enough it might become good enough.

The other option would be, and here I assume that there is also a wire taking power over to the device on the other end, to power the BMV (especially the backlight) from a separate source, possibly a small wallwart and a LM7812 linear regulator... but I suspect it also takes the voltage reading from those same wires. To make it work perfectly your voltage and current sense wires needs to be separate from the power wires, which is probably doable, though I personally would not have the cajones to take apart a brand new BMV 702 :-)

[superdiy]

The other option would be, and here I assume that there is also a wire taking power over to the device on the other end, to power the BMV (especially the backlight) from a separate source, possibly a small wallwart and a LM7812 linear regulator... but I suspect it also takes the voltage reading from those same wires. To make it work perfectly your voltage and current sense wires needs to be separate from the power wires, which is probably doable, though I personally would not have the cajones to take apart a brand new BMV 702 :-)

 

Yep, voltage reading is taken on those same wires. I would also not take apart a new meter, and that is why I only loosened the shunt PCB and traced that side of the circuit. 

[SilverNodashi]

I would have expected that they would have used a voltage senors with an IC2 bus so one could connect a much longer cable without voltage drop. 

[ZuluAlpha]

I've been curious about this. Interesting. Would the shorter cable not provide less of a voltage drop, and thus higher readings? Also, i suppose getting a nice thick multicore cabtire (spelling?) and making it off with RJ45's could do the trick? Boils down to: has the BMV been 'calibrated' for the drop across the cable supplied.

[superdiy]

On 9/7/2015 at 9:04 PM, ZuluAlpha said:

I've been curious about this. Interesting. Would the shorter cable not provide less of a voltage drop, and thus higher readings? Also, i suppose getting a nice thick multicore cabtire (spelling?) and making it off with RJ45's could do the trick? Boils down to: has the BMV been 'calibrated' for the drop across the cable supplied.

 

Cabtyre might or might not work. The supplied UPT cable contains twisted pairs and twisted pairs are sometimes critical to cancel out noise introduced on the cable - they could have used 6 core comms cable, but they've used 3 pair UTP cable - maybe there is a reason for it.