I didn’t expect the volt meter on my RB450G to be fantastic but I was shocked to find it reading nearly 3V lower than my Fluke, tested at the RB’s power jack while it was running.
Is there any way to calibrate this thing to a known voltage? It’s on a battery system and I’d at least like the running range 21.5-28.4 to be within a half volt.
I doubt it. I don’t think MK uses an IC to measure voltage, rather a voltage divider.
Why does it need to be in the 21.5 to 28.4 range? Connect the RB to a DC-DC converter to stabilize the voltage?
The 450G is on a low voltage cutoff on the battery side of a tycon 24V charger regulator. I just wanted a way to monitor battery voltage remotely and I have a spare 450G.
It’s kind of pointless to advertise voltage monitoring if the error in the reading is nearly the same order as the voltage reading. They could at least offer a linear offset in software to get a little close.
The RB450G has an older style voltage monitor that could change its reading depending on ambient temperature. Newer routerboard devices are much more precise or at least linear error can be observed.
Also, internal RouterOS reading can be a bit off if compared to reading at DC plug due to protection diodes in power delivery chain.
/dev/random
http://wiki.mikrotik.com/wiki/Monitor_input_voltage_on_RB333/433AH
I’ve been using that script on the 450G I have on solar.
I made a modification that “corrects” the RouterOS reading to match the reading we got with a volt-meter at that time. It may not be accurate all the time, but it got closer. I know approximate state of charge of the batteries. It is close enough to know when to start someone on the 2 hour drive out there to charge the batteries and clean the panels.
# the following to create a voltage correction ratio.
# multimeter measured voltage 12.5
:global mmv 125
# mikrotik health voltage while measuring with the multimeter 10.8
:global mtv 108
...
:global voltage ( ( [/system health get voltage] * $mmv ) / $mtv )
Each RouterBoard may need a different “correction” factor, so I planned on setting the script on each device to use measured values for that device. We have not, yet, put another 450G on batteries.