I have a Quattro-II 24/5000 and a Cerbo GX with three Fogstar Drift 24/280 batteries on a sailboat. I am having problems when the Quattro-II is operating in Inverter mode: when the AC load goes above ~2kW and the voltage on it’s DC input falls below 25.8V the Inverter trips out and shows its “Low battery” LED. The low battery cut-out is at the factory default 18.6V; obviously nowhere close.
I have enabled the Cerbo’s MQTT server and have a client that subscribes to all messages and logs them to a database, with timestamps. The messages are all as expected up to the point when the inverter drops out, after which there is a couple of seconds of no messages, a stream of empty messages and a bunch of message with the topic “~/logger/0/Buffer/Count”. That looks to me very much like a firmware crash of some kind.
I have raised this with the reseller who sold me the device, but he tells me that Victron does not carry out remote diagnoses and the only way forward so to return the hardware for bench testing. This is frustrating; returning it would be enormously disruptive and in any case the hardware is very likely to pass a bench test. I would have thought that inspecting MQTT logs and remote access via the Cerbo would be far more effective, but I’m told this is not available.
So, any suggestions would be VERY welcome, about how to diagnose the problem, or how to escalate for engineering support.
To be fair, this sounds like a config or installation issue, something not covered by support but the responsibility of the installer. If you are the installer then it is better to hire an experienced professional to assist.
When the system drops and restarts as its designed to do, it’s either an overload or an issue with the battery.
Its a bit difficult to help without configuration data, details and photos of the installation etc.
Your battery bms could be cutting out or you have voltage loss from poor connections. Low battery is a major clue that the issue is on the DC bus or its configuration.
From three batteries there should not be a voltage drop under load. Unless there is a set up issue.
Even at 5kva that is 208A so technically if the they are physically arranged physically right that is 67A per battery.
For this i would imagine there is individual connections to the bus bar from each battery using a minimum of 25mm² cable.
And the inverter itself with a nice feed of 2x 50 or 2x 70mm cable¿ How long are the cable runs are the batteries daisy chained?
Although the likelihood of it being a low cell is also possible.
First step should be a physical inspection and torque checking things. Also a battery health test. Which should be visible on the vrm of the fogstars are sending decent information
Alot can be diagnosed by an installer (who is supposed to be doing this) with the vrm logs.
Only if the equipment has a fault does it logged for a warranty claim. You do need to physically verify though is the voltage drop actually happening or not?
But even then there is leg work to be done before concluding that.
Ah, sorry… I could have given much more background info. I attached an oscilloscope probe to the Quattro’s DC input (ie. not to the batteries, eliminating cable drop); there is a few 10s of mV or ripple but absolutely no voltage drop; not even transient for a few msec. Yes, cabling is substantial but I’m confirming voltage at the Quattro, not at the batteries.
Also, the device runs happily for long periods with 1 to 2 kW loads. During this periods the batteries and the Quattro report almost identical voltages.
I’m not currently using the vSense connections, mainly because there’s no significant drop to the batteries. I could try this, but I don’t think there is any documented dependency on vSense.
Hi VX - thanks for the link to the pre-RMA test form, which I had not seen before.
Tests all look sensible, but in my case I’m expecting that the unit would pass a bench test but is very far from its specified performance in this installation.
In answer to your questions about the batteries: they are 24V units so three in parallel. Cable runs are a meter or two, via the big isolator switches that yacht builders install. But, as mentioned, the Quattro itself is reporting 25.8V on its DC input when it disconnects and reports “low battery”.
Post your remote configuration file. Screenshots, charts, photos. Anything that might help.
Bms limits widgets, cell difference widgets.
At the moment it is guesswork.
If you can connect to the BMS, check what it is reporting.
To force a restart due to low battery is either an incorrect setting, or, if voltage does not drop under load, the battery. Charts will help narrow this down.
So connected terminal to terminal? Then take off from either end of the bank?
How beefy are the interconnects and how big is the cable off the bank?
It then points to batteries tripping as the most likely possibility. Why the tripping is happening would be narrowed down by the current sharing over the physical install.
Sometimes this helps sometimes not as some batteries read and report their voltage before their disconnect. So what is actually at the battery terminals is unknown.
And with 60s between logging the lower dips can be missed.
— So connected terminal to terminal? Then take off from either end of the bank?
No: all three connect through individual circuit breakers to a busbar. Busbar goes through big factory-install switch to the factory-installed main DC busbar. Quattro connects to that via ~40cm cable, again via its own circuit breaker.
–it then points to batteries tripping as the most likely possibility.
No, the batteries really are not tripping. House DC stays up. Batteries have three independent BMSs; they have no way of tripping together. And if only one battery stayed up then ~100A drain is less than half its specified max load.
– And with 60s between logging the lower dips can be missed
The Cerbo published DC voltage via MQTT every two seconds. And the oscilloscope on the Quattro’s DC input would capture a voltage drop of a few microseconds.
I’m sorry if I sound unappreciative of informed and well-meaning advice based on substantial experience; that is not my intention and I do REALLY appreciate the ideas and recommendations
No offense taken here. I don’t have the benefit of seeing the installation so shooting from the hip in the dark is what I am doing.
Sometimes my approach can cause offense. That is also not my intention either. I do see alot of common install issues so started there.
Sometimes even the breakers are a high resistance or bad contact area. Depends on quality and use etc.
If it is not a new install and has recently started doing this is suggests a change in physical installation as settings don’t change.
It s question of whether it the battery, connections or inverter.
Hi Nick – I have access to each BMS through the Fogstar app.
I can sent a photo of three Fogstar Drifts, and of the Quattro. The major “House Batteries” switch and associated substantial cabling was part of the boatbuilder’s factory install and isn’t really visible.
But I’m really struggling to understand the focus on cabling and batteries given that the voltage on the Quattro’s DC terminals, as reported by the Quattro and verified with an oscilloscope, is at 25.8V at the very second that the inverter trips out. There are at least five other devices on the house DC bus that are logging their supply voltage; none report dips. And the lights don’t dim
It’s really just for completeness. We always avoided support type questions on the community as you just can’t be accurate, or necessarily helpful, while emulating a mushroom
It ends up being guesswork.
Typically it is the simple things.
If I had a system restarting with low battery, I would first check my cutoff settings in the inverter (you can attach your remote config file here, there is no PI in it), and if there was no obvious drop under load (verified with a meter as VRM can miss it), the most common cause would be a BMS disconnecting/blocking discharge.
Don’t recall you mentioning what the load is. Some loads will have a transient that can’t be measured but is sufficient to knock a BMS onto it’s back for a moment, particularly relevant with off-grid configs as there is no mains to lean on.
It’s probably helpful for me to describe the load that is causing the problem. Before I do so, PLEASE don’t anybody jump in to tell me about start-up surges, inrush loads etc. because that DOESN’T APPLY in this case !!! The load is a Webasto V-series Aircon. It is an inverter chiller; it starts up draining a few hundred watts and gradually ramps up to a configurable maximum power. I have it set to the minimum of ~2kW, with about 500W of fans, pumps, etc. IT DOES NOT SURGE
I suspect that it has a power factor that is quite far from a resistive load. I can’t see an MQTT topic that appears to be power factor and I don’t think I have a straightforward way to measure it. But given that the Quattro’s spec says it will deliver 5kVA or 4kW I would like to think that it could manage 2.5kW, ever with a wierdo power factor.
Writing this I realise that I should try entirely resistive loads of 3 to 4 kW. I’ll make that my next test.
I tried to upload the Quattro settings file but it seems that the extension .vcsf is not accepted. Which leads me to think that’s not the settings file you are requesting.
This is a VC config file. We’re looking for a veconfigure file which needs either a mk3 usb to get locally or to be remotely pulled via the VRM portal.
