childcarepro avatar image

Cerbo GX Sensor Recommendations

I'm super excited to find that Cerbo GX can monitor my batteries, solar, AND tanks. My stock KIB sensors are nearly useless, reading 2/3 full until gray water starts coming up my drain, signaling 100% full. It's pathetic.

Does anyone know what sensors I can use that will give accurate level readings and work with Cerbo GX / GX Touch? I would like fresh, gray, black, and propane sensors that work with Cerbo GX.

Will this work? Do I really need 4 sensors for just the black tank?!?

cerbo gxtank monitorsensor
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

7 Answers
machined avatar image
machined answered ·

I ordered these two tank sensors and they work great and were super easy to install. Cerbo looks awesome now with everything.

If you end up getting these then get on the beta releases for firmware because there is a lot of options for dialing in the sensors.

WEMA/KUS Boat Fuel/Water level Tank sensor 11.75" (S5-U300) USA 240-33 Ohm Signal

Kus Boat Black Water Level Sensor / Holding Tank Sending 7.5"/ 240-33 ohm

I bought them at and you can pick the right length for your tank. These install from the top of the tank and are basically a bobber around a pole and the waste one has a cage around it. You wont use those nubs for the KIB.

cerbo.jpg (248.6 KiB)
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Thanks, glad to know there are sensors known to work! I didn't want to drill, but this sounds like the simplest working solution. Coming from an Airstream with SeeLevel II meters, my wife HATES the KIB style 0, 1/3, 2/3, full sensors. We've had too many unpleasant surprises with fresh water running out, gray water overflowing, etc.

Where can I find info on how to connect these to my Cerbo GX? Thanks Jim

On your KUS Level Sensor, which wire did you run to the Cerbo? Did you have to go to the KUS sending unit and then to the Cerbo, or just directly to the Cerbo?

Kevin Windrem avatar image
Kevin Windrem answered ·

The link in your first post is to the probes your current KIB sensors use. Yes, it takes 4 probes to give you 1/3, 2/3, full readings.

The item in the link in your second post is the type of sender Cerbo is designed to work with. Keep in mind these need to be mounted through a hole in the top of the tank. You'd most likely need to remove the tanks from the RV to mount them.

I've seen some senders that claim you can mount them through the bottom of the tank but can't remember where I saw them.

I used Garnet Instruments SeeLevel in my trailer. These use sensors that stick to the outside of the tank. But they report to their own panel. Garnet does make a NMEA 2000 sensor system that sort-of works with Victron GX systems. I say sort-of because of a limitation in the current code that displays all tanks in the same tile on the display and the device list. I have a "repeater" application that overcomes this limitation. Note that Garnet does not currently sell the N2K system through normal channels. You'd need to contact them directly. It's also a 3-tank system: Black, Gray, Fresh (no Galley or LPG tanks).

The RV-C versions do NOT work with Victron GX devices.

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I am going to give this a shot. Not a programmer, at least since when I was an 8 year old boy using BASIC on a commodore 64 but I have studied your GitHub write up and think I can pull it off. I ordered the Garnett unit you recommended and the connectors. I will give it a shot this week and write up a quick little summary of my efforts. Thanks Kevin and Ben Brantley for your efforts. Let's hope Victron adds this functionality eventually. Having accurate tank levels is very important to me

And I suspect many others.

You won't need any programming, but will need to run a script from the command line. There's a link in the readme file that discusses how to enable root access on the GX device. You'll need to "ssh" into the GX device from a computer. I'm on a Mac so use it's unix command line interface Terminal. You can run the setup script from there, but if you are on a Windows machine, you'll need to use a tool like putty to transfer the files, then ssh in to run the script from the GX device.

Good luck and let me know if you need any help getting going.

I read your latest GitHub and noticed the tank repeater was not necessary if I updated the OS to the latest beta OS which I did. When I attached my Seelevel display to the cord I made I am not seeing anything on the Cerbo. I have a feeling I may of made a mistake in the cable.

For reference the 3m connector I purchased was labeled 1-4. Using a standard ethernet patch cable I cut one end off and attached the wires as follows:

1 - white/green

2 - orange

3 - orange/white

4 - green

Got any ideas what might of went wrong here?

I think your cable wiring is incorrect.

NOTE: SeeLevel tanks are NOT showing up in v2.70~16 released today!!! The services show up in dubs-spy but not in the Mobile Overview or in the Device List.

There are two "standard" ethernet cable wiring arrangements and both have CAN H and CAN L on Wh-Brn and Brn (RJ-45 pins 7 and 8).

V+ (3m pin 1, RJ-45 pin 6) should NOT be connected

This would be either the Gn or Or wire depending on the cable wiring

V- (3m pin 4, RJ-45 pin 3 SHOULD be connected

This would be either the Wh-Gn or Wh-Or wire depending on the cable wiring

Here's a document that shows the two RJ-45 ethernet color coding schemes. You need to inspect the cable to know which was used.

Here's the pinout diagram I came up with to show the interconnects between the SeeLevel 3m connector and VE.Can RJ45 (and also the NMES-2000 connector).

CanBus pinouts.pdf

canbus-pinouts.pdf (65.7 KiB)

Ok Thanks a lot Kevin. I will double check and hopefully get this fixed. I will also make sure to not download 2.70-16!

I have a standard Amazon bought 568b ethernet cable. I will attach a picture of the 3m seelevel connection I used. I am pretty sure this is right.


When connected I see nothing in the tank display or when in the VE. can Port menu there's no Devices listed. Do I need to maybe change the CAN-BUS profile? Any tips will be greatly appreciated.



connection looks correct.

CANbus setup also looks correct.

I had trouble with the 3m connectors on CAT-5 wire. Pry the blue cap off the connector and make sure all wires are actually making connection.

at a command line prompt run can dump any and see if you are getting messages from SeeLevel. You should see 3 messages in approximately 5 seconds but not evenly spaced (that’s ok). There is other CANbus traffic as well.

Make sure the CANbus cable is terminated at both ends. This can get tricky on the SeeLevel end. At least make Sure there’s a VE.Can termination at the Carbon end.

I checked the wires and even stripped them on one try. I am on about my third connector. I ordered four but they are reusable if you pop the top off it seems. I will take another look.

Forgive my ignorance but I don't know what you mean when you say terminate the Canbus cable. Same goes for "VE.Can termination at the Carbon end".

Dang autocorrect. Last statement should be “Cerbo end.”

CANbus is intended to have 100 ohm resistors at each end of the bus. You should have received two blue RJ-45 terminators with Cerbo. One of these can be used on the Cerbo end of CANbus. The SeeLevel end can’t use the RJ-45 terminators so you need to add a 100 ohm resistor across can H and can L somewhere near the SeeLevel connector. I the cable is short you may be able to get away with only the RJ-45 termination at the Cerbo end.

I did have intermittent connections at the SeeLevel end because the connector they call out is for wires larger than CAT5/6.

3M makes a connector more suited to CAT5/6 wire size. You are more likely to get a better connection than with the one called out by SeeLevel.

You could also splice in larger wires and add the termination resistor at the splice.

Ok copy that. This is what I will try next. I was looking at that Cerbo RJ-45 and wondering what that was for.

Again thanks for the edification and support.

mcornforth avatar image
mcornforth answered ·

Just ordered some differential pressure sensors from Safiery. They look to be robust, accurate, and require no new holes cut in my tank. I believe I saw someone do a DIY version of this with an available automotive pressure sensor, but I decided to pay a little to have someone I can call if I have issues. They'r nice bloaks and do some really awesome work. These sensors do need the GX Tank 140 as they are voltage based sensors.

Here is a link to all of their tank sensors and info on them:

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

woodbri avatar image
woodbri answered ·

I emailed TankEdge about their Moda sensors and got this response "

The Moda sensors work on voltage. 0 to 5 Volts with full usually being somewhere around 3 volts. Our panel sends 10 volts down to the sensor on the blue wire and that voltage comes back as variable voltage on the red wire between 0 and 5 (like I said with full being usually around 3) That amount will vary greatly with each tank and the foil set-up which is why our panel is calibrated to the tank. I'm not sure if your GX Tank 140 adapter will work for this type of set-up or not. We do not have any adapters for that system. Let me know if you have any other questions.

Thank you,
Tonya Adams
541-610-0401 ∙


So this sounds like it would be compatible with the GX Tank 140. I would be interested in knowing if anyoneelse has had success with these on a GX Tank 140. The Moda sensor cost $35/tank

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

childcarepro avatar image
childcarepro answered ·

OMG people, I think I'm onto something here. I found this document at TankEdge:

It states to test the Moda sensor: "…check the voltage coming out of the sensor across the blue to black wires at the sensor. This also needs to be done while the voltage is being sent down to the sensor just as above. This reading will vary depending on the liquid level in the tank. It should not be over 5 volts but could be very small, even a fraction of a volt with low levels."

So the output of the Moda sensor is 0-5V, with 12VDC input. But I just read the Cerbo GX manual and it states it only supports VE.Can, NMEA2000, and analog resistive sensors. Sadly, there is no mention of voltage sensors.

The extra wire I mentioned in a previous post is simply a black ground wire.

It would be so amazing if this could work, but it appears the TankEdge system bucks the trend of all other tank sensors in being voltage senders rather than resistive.

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

It all depends on the output impedance of the sender. I saw somewhere that the pullup inside Venus devices is 680 ohms to +5 volts. That's 8 mA to reach 0 volts. Plus, it also requires Cerbo's new calibration (part of v 2.60 when it's released) to accommodate the voltage range of this sender.

But it's worth someone experimenting with them. If it's close, Victron could possibly tweak their software to support them.

I emailed Tech-Edge Manufacturing asking whether it sent voltage and resistance, the response:

"I’m sorry. Our sensors are capacitive sensing only sending voltage."

I'm willing to try it if you still think it's worth a try. But at this point I have to admit SeeLevel's NMEA2000 might be the best option. I wanted a direct sensor solution for finer resolution and fewer potential points of failure, plus I didn't want to have to hack my Cerbo GX right out of the box.

There are too many unknowns at this point to be confident these senders will work directly with Cerbo.

What is the voltage range Cerbo can tolerate on it's tank inputs? A 180 ohm sender would reach approximately 1 volt, so these senders would produce approximately 5 times that voltage which might be outside the range of the calculations currently used.

Can the sender drive a 680 ohm load to +5 volts directly?

A buffer circuit would be fairly easy to design that could interface the two.

One of us clearly knows what he's talking about... and it isn't me. Hahaha So I'm educating myself... resistive senders have variable resistance that causes variable voltage outputs; capacitive senders output variable voltages.

1. I'm willing to test, but I'm clearly in over my head as far as electronics know-how;

2. As you note, 0-5V is a much wider range than 0-1V; I'm wondering if the Cerbo GX's circuitry even supports 5V or whether it will clip at 1V? If it can handle up to 5V, then simply calibrating could fix this discrepancy?

3. A bigger problem is I'm under the impression capacitive sensors use alternating current; will Cerbo GX have a problem with that?

Resistive senders are simply a variable resistance. That resistance works against the pull-up resister inside the display device to produce a voltage.

The senders you are looking at have a module that reads capacitance between the two strips stuck to the side of the tank. The module converts that to a voltage that is sent to the display unit. I'm sure this will be a DC voltage. But you are right to assume the capacitive level detection mechanism will use an AC voltage, probably fairly high frequency.

The module at the tank will have a finite output resistance. If that's low enough it could work predictably against the 680 ohm pull-up inside Cerbo. If not, the pull-up will offset the voltage the module outputs. In the limit, the 680 ohm resister could overwhelm the module and voltage would end up at 5 volts regardless of the tank's level.

At this point, we just don't know how these two devices will interact.

If the sender's module can provide a usable voltage swing against the 680 ohm pull-up in Cerbo AND Cerbo can be adjusted to accept this voltage swing for empty-full levels then you are golden. Cerbo's adjustments (part of v2.60) are calibrated in resistance so my best guess is it wouldn't be able to handle the full 5 volts of a full tank. That would equate to an infinite sender resistance.

Some form of buffering between each tank module and the Cerbo input will most likely be needed.

I don't have a Venus device with tank inputs so can't do any testing. I also can't get my CCGX to think it has tank inputs so I don't know what the sender resistance adjustment range is. You could check this out by installing the latest Venus release candidate (v2.60~33 at this time).

Wow, thank you, @Kevin Windrem. That was very helpful.

I have reached out to Tech-Edge Mfg (, and they said if I wanted to experiment, their engineers would be able to answer any questions I had. So I think I'll order my Victron setup including Cerbo GX and one Tech-Edge MODA sensor, and we'll see how that experiment goes!

Tech-Edge also makes a regulator for the MODA sensor that raises it to 10V for an older control panel, so I asked if we could lower it to 1V for the Cerbo GX.

Fingers crossed!

OK correction, the regulator brings DOWN the 12V coming from other control panels TO the 10V the MODA requires. I talked to Tonya at Tech-Edge today who had discussed with engineers. So it takes 10V input and outputs 0-5V, but usually empty is slightly above 0 and full is around 3V. She said we could use a resistor to reduce the voltage to 0-1V, but also was concerned about what voltage Cerbo GX outputs TO the senders, as the MODA expects 10V but will work down to a lower voltage but will not function below a certain voltage. She will check with the engineers.

If they can make a version of the MODA that functions within the Cerbo GX's parameters, or at least come up with a mod for it, this could be an amazing solution for Cerbo/Venus GX users.

A resistor to drop voltage probably won't work since Venus devices are pulling UP to +5 volts on each sender line with 680 ohms.

3 volts is more in the ballpark but haven't heard from anyone at Victron what voltages they can deal with, or put another way, what resistance ranges their new software can be set to. Resistance set in the menus will convert to a voltage when the 680 ohm pull-up is factored in.

Not sure if the pull-up can be turned off which would allow the input to be used as a direct voltage input and more likely to work with the MODA senders.

It sounds like the widget Tech-Edge has to reduce 12 volts to 10 volts would be required since a typical RV would have a 12 volt electrical system. Venus devices would not provide power specific for the senders.

@Kevin WindremDo you have any more information since v2.60 was released? I have decided to order my tester Moda sensor to try with my Cerbo GX now that v2.60 has been released.

I have yet to power on my Cerbo GX to my multimeter to see what its output voltage is. I have LiFePO4 DIY cells (4S and 8S setups) but have not decided on which BMS to use, so no power for the Cerbo unless I use another 12V LiIon battery pack with barrel connectors and strip the wires.

I really love what you've done with connecting the SeeLevel II NMEA version, but I'm trying to keep the complexity to a minimum by directly wiring in senders rather than having a whole other system plus an interface between Venus OS and SeeLevel II's system plus any potential funkiness in the NMEA2000 implementations of the two companies.

Nothing more here on my end.

I installed a Tech Edge system in a prior RV and it worked great, would love to find out if their sensors could be used with the Venus GX for my new RV. Have you found out any more information?

@Green Monster I'm about to order a sample Moda sensor or a pack of sensors and try plugging them in to see if it works.

I am FED UP with the nearly useless, empty-1/3-2/3-full KIB-style sensors. I've already run out of fresh water showing 1/4 full, overflowed my fresh tank when sensor shows 3/4 full, had gray water back up into my shower pan showing 2/3 full... twice...

I am following this conversation as I have a Toyota Sunrader I am rehabbing and want to have a full system monitor in place. I have been looking at Simarine and Cerbo GX. But, I went Victron on all my solar and think the Cerbo would be better fit. The tank and LPG monitoring is a big deal in a motorhome. Please keep posting back any results you get.

@childcarepro did you get and try the Moda Sensors with the CerboGX and/or the new GX Tank 140?

Sadly, it's still sitting in my garage. All of it, my whole solar setup and Moda sensor.

childcarepro avatar image
childcarepro answered ·

Thank you, @Kevin Windrem, that was helpful. Haha I guess now it makes sense, one sensor for each "level" it detects. But I thought those sensors used 4 holes in the side of a tank; I suppose the ones I linked to are more accurate than the side holes?

If you have to remove the tanks, that's not easy. And not an option for the propane tank. :( I would love to use SeeLevel stick-on sensors if they would report resistance numbers that could be read by the Cerbo GX.

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

The SeeLevel sensors won't connect directly to Cerbo. The sensors connect to their control panel then on to Cerbo. The entire system is about $350 for 3 tanks -- no propane on this model.

You need to contact Garnet Instruments and ask if you can purchase a model 709-N2K_NLP. It's an OEM model they have sold to end users on a few occasions.

You then need to install and run my tank repeater software that overcomes the one tank per sensor limit for N2K devices.

You would also need to make up a cable to go from the Cerbo GX to the SeeLevel control panel. More details and the tank repeater software is available here. Look it over to see if it's something you are interested in and can handle. Note that Victron does have plans to update their NMEA2000 tank sensor software in a future release but it probably won't make it into v2.60.

BTW, some permanently mounted propane tanks come with a built-in sender. It is unlikely you'll find senders on portable (20-30 pound) propane tanks. But the good news is that if you have such a tank, the sender is resistive and should interface directly with Cerbo GX.

@Kevin Windrem, you're awesome. Thanks for sharing your work with the community! I've checked out your GitHub project page but haven't gotten any further. I'm no Linux expert, but I've been known to type in some commands from time to time to get stuff done. :)

OTOH, since we're not sure whether Victron will adopt RV-C or NMEA2000, maybe I'm better off living with my crappy senders for now until plans solidify? I've heard of people trying to use the 4-level KIB-style senders with the Simarine Pico, so maybe I can see if Cerbo GX can make sense of these readings until we get better sensors?

The other option is to insert 0-190 ohm senders. More labor, but more precise resolution, as we would get continuous resistance readings rather than a quantum reading every 3/8". Correct?

Regarding propane, my little Winnie does have a sender on it, my One Place monitor reports it in 25% increments.

It appears the Pico system uses resistance or voltage sensors and probably exhibit the same issues trying to use the KIB probes as Cerbo would. That is unless Pico has a converter module.

I'm using SeeLevel JS senders on my tanks because they are 6" and 7" tall. I'm seeing 1-2% changes in levels. According to Garnet's documentation resolution should be between 1/4" and 1/2" which is about 4% on my JS sender at 7".

A resistive sender might provide more resolution but not convinced it would be necessary.

As I mentioned previously, there are senders that can mount into a hole in the bottom of the tank but I'd worry about leakage.

Thanks, @Kevin Windrem. SeeLevel is 99% the way to go... wondering whether to wait to see if Victron chooses RV-C or NMEA2000. Sounds like since NMEA2000 is farther along in development, does that make it more likely to be adopted by Victron?

It's reassuring to hear the resolution is higher than 3/8"!

If I went the 709-N2K_NLP route, I imagine it would likely continue to work even after Victron adds support for NMEA2000 natively, correct?

I think N2K is the way to go. RV-C may or may not be implemented in Venus but right now it is not.

You will need to contact Garnet Instruments to see if they are willing to send you the N2K version. It isn't in their catalog and is an OEM for some RV manufacturer. It took a while to get mine but that all depends on their production cycle.

There are other N2K tank systems.They *should* work with my tank repeater but I have no experience with them.

Venus currently supports N2K tank systems that report only a single tank. If an N2K system reports multiple tanks, Venus combines all the tanks in one display so the values and fluid type keeps changing. My tank repeater sorts that out and provides stable displays for each fluid type.

Again, the link to that software is:

I've made it pretty easy to install (in my humble opinion) and will be here to help if you have issues.

There is no reason a Victron native solution wouldn't work with the SeeLevel N2K system. In fact, they have a SeeLevel system in house for development and testing. You would of course at that time want to disable the repeater. But you may wish to keep some or all of the modifications I've made to the Mobile Overview page. My revised tank tile compresses vertically to support additional tanks. Plus I have a completely new layout for the Mobile Overview page. If you like lots of numbers you may like it.

I found a little information on your One Place sensors: They *may* be capacitive from the testing instructions, but would be sending proprietary information to the One Place panel. They still function like the KIB probe types with empty, 1/3, 2/3, full resolution. The LP sender is 0 ohms empty and 90 ohms full but no indication of the number of steps in between. This range does not conform to either the European 0-180 ohm or the US 240-30 ohm standards (empty-full). So your sender may not interface with Cerbo with v 2.54 software. v2.60 will have adjustments that *may* accommodate the 90 ohms full. I have no analog inputs on my CCGX but you could check this now if you want to update to the latest release candidate. New setup menus permit switching to a "Custom" standard then adjust empty and full resistances.

@Kevin Windrem, have you heard of SCAD or TankEdge external sensor strips? I don't know what these sensors output, resistance, voltage, or digital data. But someone mentioned it has one extra wire than their existing tank sensors.

Do you think the Cerbo GX can connect to these sensors directly? The gold standard would be:

1) high resolution
2) high accuracy
3) externally mounted (non-fouling)
4) direct connecting (no interfacing/bridging)

Thank you.

Here's a rewrite to the SeeLevel tank repeater:

Well this is pretty perfect timing. I hope to be connecting all my See level components and the Victron stuff very soon.

childcarepro avatar image
childcarepro answered ·

Would this work? These say 0-190 ohm sending units, available in 150-600mm, and I just need to know my tank depth?

1 comment
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Be wary of this type of sender - very crude resolution. The one I bought only had 4 levels - and not very linear either. They are just reed switches that switch in a resistor ladder.