Dear Victron its time to change your controls cabling

Q&A and troubleshooting
, , , , , ,
1

Dear Victron Team,

Please accept the below feedback from me, it is intended to be delivered in a constructive way from a customer that is overall happy with the products and systems.

Firstly, I am a data centre controls specialist with extensive experience in Modbus, Lora, RS485, analogue, PLC, sensors etc, and my complaint today covers the choices that Victron has made when it comes to cabling their controls systems.

A few years ago, a 42foot yacht was considered absolutely huge, today a 38foot (12m) Catamaran that is 6.5m wide is considered small, entry level. Class A RVs are also becoming more and more common.

Along the way battery technoloigy has changed with Lifepo4 and other chemistries, but you know all this,.

The issue is this, the electrical loads have shot up, and with it the cost and size of power cabling.

What has not changed is the tech (protocols) cables and connectors for controlling all of the fantastic devices the we use to charge and consume and monitor our energy.

I am fortunate to own a 38foot catamaran with a 12 volt system. My equipment list is as follows:
1: Ekrano GX
2: 5 x MPPT - 4 with ve.direct and 1 with ve.can
3 1 x Skylla IP65 with ve.can
4: 1 x Smart inverter 3kva
5: 2 x Orion XS 12/50 with ve.direct
6: Lynx smart BMS with ve.can
7: Smart shunt with ve.direct
8: 4 x Smart Lithium NG batteries

The controls are made up of the following types:
1: Simple two wire cables for the ATD to the SKylla as it does not speak the full Victron language over ve.can
2: RJ45 for ve.can connections
3: VE.direct cables both with USB/Ve.direct and plain ve.direct
4: M8 connections for BMS
5: VE.can to NMEA

Compare the last data centre I built, which was relatively small at 8 Megawatts and about 2000 racks and 5000 meters square dof space. We had the following controls:
1: Plain analogue cabling dry contacts
2: Modbus RTU / RS485 terminated in phoenix contact terminal blocks.
3: Modbus TCP over RJ45

Thats it.

My main complaint is that VE.Direct is a 100% pain in the backside. Connectors fall out, ve.direct cables from factory vary in the tightness of fit. 10m max length which is not enough for my catamaran of 38 feet.

The use of USB extenders in a marine envronment is pure madness. My Startech 10m (x2) extension and Startech industrial USB hub (x2) has cost a total of 280 Euros, which is nearly 50% the cost of the ekrano GX they are connected to. This is a workaround to a deeper problem. THe USB to Ve.direct cables are an additional cost above the 280 euros, and I bought 6 of them…

The only reason I needed them is the 10m limit of ve.direct cables is not enough to reach from the chart table to the engine bays where the Orion XS DCDC chargers are, or to the various locations where the MPPT chargers are.

I prioritised shorter power cables lengths over shorter control cables lengths which is industry best practice, both in data centres AND in the marine world.

My suggestion is to move to a simple phoenix contact connector and use RS485 across the board. Keep the RJ45 connector if you really want, but in the marine world, that one can be prone to corrosion and is harder to repair.

Decent belden shielded RS485 cable can be used by customers like me who have longer cable runs passing through electrically noizy places, and for shorter runs RS485 really doesnt care.

I have no idea how people use your products on much larger boats, or maybe thats why I see images of Victron screens in weird places, and why the remote console in VRM or the new Android app is so popular, but these are all workarounds in my opinion.

I should add that the Scheiber monitoring systems popular and ubiquitous on french built boats (which I detest) don’t have this problem and have a decent concept of CAN bus or LIN bus) and have at least standardised on a connector (which is positively latched and NEVER falls out) and cabling which seems to be very tolerant of lengths…

Anyway, my rant is over, thank you for reading and I hope you take the feedback as intended, constructively.

Best regards,

Alex

1 Like
2

VE-Direct can work over much more than 10m…although this is neither supported by or recommended by Victron.
The cables can be carefully extended by cutting them in half, and using cat5 to extend, using 1 pair for power - black and red, and one pair each for data, using the red power line as the signal ground.

However, I do agree that - a massive - simplification could be made by migrating entirely to a CAN standard connection for all products. This would suit most environments. RS485 - used by VE-bus connections - is going to have too slow a data rate c/f CAN, and the CAN protocol would cover the requirements of both the VE-direct protocol and the VE-bus protocol.
This move is possibly indicated by the use of VE-CAN on the latest solar chargers.

3

Instead of the Ekrano you could have used the Cerbo (+ GX Touch) and place that at a central place or where the most components are.
Than you would have needed only a HDMI extension to the GX Touch (and a USB power supply near the display).

1 Like
4

Hello,

Thanks for your comment, and in fact that had been my original specification, I was advised by my dealer to choose the Ekrano… Anyway. It certainly would have kept the space in my dashboard clear of cabling.

I still feel that the HDMI extension is a work around, and will gain some distance but doesnt solve the overall long term issues.

NMEA 2000 Bus and VE.CAN have a single cable that runs around the boat / van / house / whatever. Much cheaper in the long run, no USB to ve.direct. Not having many different ve.direct cables etc.

Strangely the GX devices now have 2 (4 in/out) separate CAN ports on the back but only 3 ve.direct, the ratio seems all wrong to me.

Best,

Alex

5

Hi,

Thanks for your comment.

Yes latest MPPTs over a certain size have ve.can, but the brand spanking new Orion XS and smaller MPPTs dont.

And some devices have VE.can but dont talk to BMS for example the Skylle IP65

Thanks for the correction on the RS485 bus speeds, I should have clarified a cabling standard from a communication protocol / standard.

Alex

6

for CAN devices, you have ‘producers’ and ‘consumers’ of information. Most peripheral devices: BMS, chargers, solar controllers etc., are producers of information. They will only respond to targeted commands. So a BMS won’t produce a targeted command for a charger for example.
The Consumers of information like the Cerbo can take all the information, process it and then send targeted commands to the ‘producers’.
This means that if you change the system by adding or deleting a component, then only the ‘consumer’ has to accommodate the change.
It would complicate the code and communication requirements if any producer (e.g. BMS) was also required to produce all the targeted commands required to control the various different chargers that might be on the bus.
Whilst the cerbo only has 3 VE direct ports (3 different PV plane orientations ought to be sufficient), it is possible to add more via a USB port (numerous posts).

7

Indeed, and thanks for correcting me again on the traffic flow for commands. WHilst the BMS does send commands to the batteries, it does not on the ve.can bus. That is driven by the GX device and in my case using DVCC.

Currently I have 7 ve.direct connections as follows:

4 x mppt on a usb hub
1 x inverter on gx device
2 x orion xs on gx device

Then there is an additional mppt on ve.can and an ac charger on ve.can.

Total Cost of ve.direct cables is 63 Euro
Total Cost of ve.direct usb cables is 120 Euro
Total cost of USB hub and cable (Startech) 150 Euro (could be less with cheaper less rugged components, that was my choice)
Total ve.direct cost for project connecting 7 devices: 333 Euro

Now the total project cost for the ve.can device portion:

GX device to Lynx BMS to Skylla charger to MPPT 90Euro

My system is working fine, as I said at the beginning I am very happy with it, but I am trying to make a case that work needs to be done on the controls cabling, and you can see here that the ve.direct cost was ~50% the cost of the GX device itself and massively more complex to install and more costly than the VE.can system. The costs I state above were planned for and the project can be considered a success, however I say agian to Victron that its time to rationalise the control cabling methodologies that it sells.

That, and the JST connector for ve.direct I and others strongly dislike.

Many thanks for reading and sharing your thoughts.

Alex

P.S. I also have their AC ET112 Grid Power meter, and that uses another USB port for its RS485 connection…