EG4 12K/18K/FlexBoss Firmware FAAB-2122 (Now replaced by FAAB-2525)

[Rezin777]

There is nothing you can do to mitigate this failure unfortunately.
More cooling is always an advantage on any electrical device, but it is not the driver behind this failure.

Thank you for that.

 

[Rezin777]

Well, it wouldn't surprise me if they kept the attachment points the same, but no way to know unless you take one apart. It would lack certification either way I imagine unless it was some official upgrade kit. There is also the possibility that the issue isn't with the attachment to the bus bar itself. It could even be purely a paperwork issue and there might not be any physical change at all, but this seems less likely to me since this isn't the kind of industry where you make customers keep coming back to buy more. If we were talking about an aircraft or something like that I could see a manufacturer doing that (and selling "software" upgrades that didn't actually change anything but the certification).

*This is copypasta from AI (Grok) so take it as you will.*

Based on available data, the Eaton BR and BSB B3S circuit breakers have bus stabs that are mechanically equivalent in terms of design and compatibility for Eaton BR busbars. Here’s a detailed breakdown strictly focusing on the stab specifications:

Bus Stab Comparison

• Eaton BR:
  • Stab Design: Curved clip, approximately 0.75 inches in length, designed to grip the busbar in Eaton BR, Cutler-Hammer CH, and compatible Westinghouse/Challenger panels. The clip has a specific curvature and tension for secure contact with the busbar's geometry (typically tin-plated aluminum, rectangular or slightly notched).
  • Busbar Compatibility: Eaton BR load centers and specific legacy panels. Some models include a rejection tab for CTL (Circuit Total Limiting) panels to prevent improper installation.
  • Dimensions: The stab aligns with a 1-inch wide breaker footprint, ensuring proper seating in the panel’s busbar slots.
• BSB B3S:
  • Stab Design: Explicitly designed to match Eaton BR’s curved clip geometry, with the same ~0.75-inch length and contact profile. BSB documentation (e.g., for B3S series used in solar/inverter applications like EG4) states the breakers “install in the same port locations as Eaton BR,” confirming identical stab shape and dimensions.
  • Busbar Compatibility: Compatible with Eaton BR load centers and BR-compatible systems (e.g., GridBOSS, EG4 inverters). No evidence of a rejection tab, suggesting non-CTL design for broader compatibility.
  • Dimensions: Matches Eaton BR’s 1-inch wide breaker footprint, with stabs engineered to align with BR busbar slots.

Key Findings

• Stab Equivalence: The BSB B3S stab is functionally identical to the Eaton BR stab in terms of clip shape, length, and contact area. This allows BSB B3S breakers to plug directly into Eaton BR panels without mechanical issues.
• Differences: No mechanical differences are noted in stab design based on available data. However, BSB breakers use hydraulic-magnetic trip mechanisms (vs. Eaton BR’s thermal-magnetic), which could affect performance but not physical stab compatibility. BSB’s UL-listed variants ensure regulatory compliance for BR panels, but non-UL models may vary in quality.
• Caveat: While stabs are mechanically the same, always verify panel compatibility via the load center’s UL listing (per NEC requirements). Mixing brands risks poor contact or code violations if not explicitly approved.

If you have a specific panel model or need further confirmation (e.g., via part numbers), I can refine the analysis.

 

[stephen.m]

There is nothing you can do to mitigate this failure unfortunately.
More cooling is always an advantage on any electrical device, but it is not the driver behind this failure.

Thank you for that, even though that’s really bad news.

I haven’t seen this particular question posted yet. How does this failure mode behave if you have 2-3 fb21 in parallel? Does having multiple unit give any sort of redundancy? And do having multiple units reduce the risk of one unit failing?

 

[Quattrohead]

It is not a stress related failure, so multiple inverters will not make a difference. It is a failure of the board that controls the current flow to and from the battery, the inverter will still work fine as far as putting out AC current all the way up to its limit and converting PV into usable power.
There are many tens of thousands of these units out there that are not impacted by this problem, obviously a batch issue.
I imagine that tech support and RMA were swamped to start with but systems are now in place to handle this and I believe resolution times are now within reason.

 

[stephen.m]

It is not a stress related failure, so multiple inverters will not make a difference. It is a failure of the board that controls the current flow to and from the battery, the inverter will still work fine as far as putting out AC current all the way up to its limit and converting PV into usable power.
There are many tens of thousands of these units out there that are not impacted by this problem, obviously a batch issue.
I imagine that tech support and RMA were swamped to start with but systems are now in place to handle this and I believe resolution times are now within reason.

Thank you for that.
If you guys are able to identify risk based on batch and serial numbers, I’m sure the majority of us will appreciate knowing it their units are at risk or not, even if a recall scenario is unfeasible.

Thanks again

 

[geordi]

*This is copypasta from AI (Grok) so take it as you will.*

I believe I have one of the V2 (grey box) GridBoss units, but the selection of breakers does not bother me with my usage. I DO however have a selection of BR series breakers from my old load center (I changed that out to a HOM style), so I will attempt to attach one of those BR into the GridBoss and see whether it locks to the rail or not for the people who want to use lower rated amperages in the smart slots

 

[Zerog2k]

man the https://monitor.eg4electronics.com website seems super flakey this morning... I can only imagine it's related to this rollout...

 

[Rezin777]

I believe I have one of the V2 (grey box) GridBoss units, but the selection of breakers does not bother me with my usage. I DO however have a selection of BR series breakers from my old load center (I changed that out to a HOM style), so I will attempt to attach one of those BR into the GridBoss and see whether it locks to the rail or not for the people who want to use lower rated amperages in the smart slots

Eaton BR fit quite well, even in a V1 Gridboss. Considering they are acceptable in later versions, and what I've gleaned from the internet regarding Eaton BR vs BSB B3S, I would personally be very comfortable using them if I needed to do so. That's just me though, and I am not an electrician, engineer, lawyer, or insurance investigator. YMMV.

 

[blueforce]

It is not a stress related failure, so multiple inverters will not make a difference. It is a failure of the board that controls the current flow to and from the battery, the inverter will still work fine as far as putting out AC current all the way up to its limit and converting PV into usable power.
There are many tens of thousands of these units out there that are not impacted by this problem, obviously a batch issue.
I imagine that tech support and RMA were swamped to start with but systems are now in place to handle this and I believe resolution times are now within reason.

Does this affect the FB18s as well or is it specific to the FB21?

You say this is a batch issue but batches of what? The FB21s or the specific boards? My friend in the area bought his 1st FB21 in November and by August his 4th FB21 died. All of them succumbed to the same DC board issue.

 

[geordi]

He's had FOUR of them all fail? Yikes. I bought my FB21 in March and it was dead in July, but there were signs of impending failure for several weeks prior to the last shutdown. Only my own unfamiliarity with the issue (and mistakenly thinking it was the battery BMS instead) caused me to not recognize the problem until just before the final battery-open error.

What I had been told was that they had several suppliers (my assumption - all near each other in China) and that the boards from a specific supplier had a specific defect (that was not disclosed to me) and this wasn't something that could be resolved in software, it is a hardware component failure. It is also not a field serviceable issue - not that I'd want to try that when it is behind a warranty void sticker - and the company DID KNOW which serials were affected.

Probably to try and avoid a massive flood of tech support, they have not reached out proactively to the customers to alert them that they might have an affected serial number... Well THAT seems to have backfired from other reports here. So maybe they should just suck it up and issue that recall / contact the owners who they can SEE have affected serials connected to their monitoring server, and go from there. THAT would be a customer-first proactive approach instead of waiting for each unit to fail if they are certain that each affected unit WILL definitely fail.

It certainly seems like they are pretty sure that the hardware failure is a matter of WHEN rather than IF.

 

[Quattrohead]

12/18kPV LCD Version 23 Firmware is now live.

 

[RARankin]

12/18kPV LCD Version 23 Firmware is now live.

Is that valid for the add-on screen for the Flexboss series as well, or is that forthcoming?

 

[blueforce]

He's had FOUR of them all fail? Yikes. I bought my FB21 in March and it was dead in July, but there were signs of impending failure for several weeks prior to the last shutdown. Only my own unfamiliarity with the issue (and mistakenly thinking it was the battery BMS instead) caused me to not recognize the problem until just before the final battery-open error.

What I had been told was that they had several suppliers (my assumption - all near each other in China) and that the boards from a specific supplier had a specific defect (that was not disclosed to me) and this wasn't something that could be resolved in software, it is a hardware component failure. It is also not a field serviceable issue - not that I'd want to try that when it is behind a warranty void sticker - and the company DID KNOW which serials were affected.

Probably to try and avoid a massive flood of tech support, they have not reached out proactively to the customers to alert them that they might have an affected serial number... Well THAT seems to have backfired from other reports here. So maybe they should just suck it up and issue that recall / contact the owners who they can SEE have affected serials connected to their monitoring server, and go from there. THAT would be a customer-first proactive approach instead of waiting for each unit to fail if they are certain that each affected unit WILL definitely fail.

It certainly seems like they are pretty sure that the hardware failure is a matter of WHEN rather than IF.

Yes, 4. He, like me, run at the upper limits of units abilities for battery DC>AC

 

[Redpineapple]

@RARankin

Since we're sharing... here is what I got from EG4.

FB21 Firmware / LCD Screen

 

[rich0]

Probably to try and avoid a massive flood of tech support, they have not reached out proactively to the customers to alert them that they might have an affected serial number... Well THAT seems to have backfired from other reports here. So maybe they should just suck it up and issue that recall / contact the owners who they can SEE have affected serials connected to their monitoring server, and go from there. THAT would be a customer-first proactive approach instead of waiting for each unit to fail if they are certain that each affected unit WILL definitely fail.

Well, it is reasonable enough to do it in an orderly way, but I do tend to agree with you. If their supply is limited they have three competing priorities:

1. Replace units that are already breaking in the field
2. Replace units that are likely to break in the future.
3. Sell units to new customers.

I don't think anybody would disagree with prioritizing #1, which might mean delaying #2 if you're struggling just to keep up with those who are already down. Somebody who is without power should take priority over somebody who is merely at risk in ending up in that situation.

I think most would say they ought to prioritize #2 over #3, and I agree in principle, but there is also the practical reality that this is a fairly small company as far as I can tell and they do need cash flow to pay for everything. Every unit they sell to a new customer brings in cash they can use to pay to fix the existing units. I really do hope they intend to do just that, but none of us are served if they just go bankrupt. Even a bankruptcy court will allow for this sort of thing as usually creditors would prefer getting all their money over time vs just getting a small fraction right away. It just sounds sleezy, and I can't really say anybody is wrong to feel that way. Without seeing where their money is going it is hard for me to be too critical. Certainly the dividends shouldn't be getting first priority in a situation like this.

 

[geordi]

They might very well be operating with your three priorities, but in order 1, 3, 2 right now. That's OK and probably the most logical as it keeps some cashflow happening while resolving the units that definitely have had issues that resulted in customers calling / emailing in.

As it should be, happy customers who aren't having issues are probably just fine living their lives for a while, even if they are members of group #2.

But as soon as reasonably practical and group #1 is resolved / shipping has caught up, then there probably should be an active plan (maybe first bought / first swapped) to reach out to the affected #2 group members and say "Hey, you might not be having this issue just yet, but we'd like to schedule a swap out at a time of YOUR convenience, so that your system doesn't decide for you to leave you in the dark" or something to that effect.

This doesn't seem like a safety hazard, I don't think this will fail spectacularly bad and cause any actual losses, so there's probably no reason for a full-on press release like the auto industry does when there's a flaw that could cause the tires to fall off your car or the steering to disconnect. That's obviously a little more important. So they can stretch it out, provided anyone that joins group #1 is able to get information and get resolved in a timely fashion. My hope is THAT is what is happening now.

 

[wpns]

12/18kPV LCD Version 23 Firmware is now live.

Was that supposed to be a link to that firmware?

 

[fnnwizard]

Hmmm...
Hardware issues? Aren't the "Bosses" assembled there in TX?
Would that have anything to do with it?

 

[robby]

It would be critical/nice to have full transparency since they (LuxP and Eg4) depend on the user base for a lot of the QA/stability testing/feedback.

Something as simple as... " We've had some issues with xxxxx, working with LuxP to see how we can rectify by xxxdate so keep an eye out for this...."
Or even designate an affiliated beta tester to post such progress.

In a perfect world that would be ideal but in the real world companies don't engage in an honest discussion about problems because what they say to the customers can be used against them in court.
In my experience you will be hard pressed to find a company that states that there is a serious issue unless that issue can cause physical harm and even then they tend to cover it up as long as possible.
Even the righteous Apple corporation denied for months almost every "Gate issue" they ever had.
Remember Antenna Gate? Oh your just holding the phone the wrong way
Then there was Batterygate and Bendgate

The most you can hope for is someone with contacts like @Quattrohead giving you some unofficial info.

 

[chill633]

I did the FlexBoss LCD Firmware update today. Here was the process:

1. Purchase new 16 Gb USB flash drive from Walmart ($3.88)
2. Mount drive (Linux laptop). Already FAT32.
3. Unzip FlexBOSS LCD Kit.zip and copy DWIN_SET folder to mounted flash drive.
4. Safely remove drive
5. As my FlexBOSS was already powered down, I just plugged the flash drive into the USB port on the LCD screen and powered on the inverter
6. The LCD screen awoke and displayed the EG4 logo in the center, a sun icon in the lower left, and a battery icon above it.
7. Time passes...just over a minute and a half, then the unit reboots.
8. The LCD screen awoke and fully booted -- many more icons. Once finished, in the lower right corner it now says "LCD Version 23".
9. The Tao is Good. All my firmware are in harmony.

 

[alvitdk]

Just checking with the "Pioneers" on here before updating the FW on my GB and both FB21 units... did it all go well or better waiting a bit longer? My system works fine right now, so there might be no reason to update but if I do, I see some conflicting information in all those posts on how to go about it.

Did you update all at once via the web interface, and the GB first and then the FB one at a time afterwards, or just all at once (AI seems to think so)?

 

[Quattrohead]

Hit the manual bypass on the Gridboss and then push the updates, does not matter which order.
When all done and running, restart again one more time.

 

[chill633]

Just checking with the "Pioneers" on here before updating the FW on my GB and both FB21 units... did it all go well or better waiting a bit longer? My system works fine right now, so there might be no reason to update but if I do, I see some conflicting information in all those posts on how to go about it.

Did you update all at once via the web interface, and the GB first and then the FB one at a time afterwards, or just all at once (AI seems to think so)?

I had zero issues using the web interface. FB21 first, then GB, then finally the FB LCD unit via USB stick.

The units managed their own reboots. I did nothing but ignore it for 10 to 15 minutes and then checked back.

 

[TacoMeat]

Hit the manual bypass on the Gridboss and then push the updates, does not matter which order.
When all done and running, restart again one more time.

When you say "restart again"

Remove ALL power from the system? E.G. Disconnect Grid power + complete battery / inverter power removal?

I looked in the Maintenance section and did not see a "reboot / restart" setting for the Inverter or Gridboss. If there is an official steps could you provide that documentation?

If not, no big deal however I'd ask that EG4 address this with some kind of best practices doc for upgrades. Its a great way to set expectations for post upgrades and a simple way to inform the end users of the proper steps to a upgrade.

 

[geordi]

I looked in the Maintenance section and did not see a "reboot / restart" setting for the Inverter or Gridboss. If there is an official steps could you provide that documentation?

The lack of clarity in the documentation is a sticking point that I have been bringing up for a while.

This is new technology to most of us, so having a stated procedure somewhere to assist with best practices for basic procedures, seems like a good idea. A firmware update is definitely a basic procedure, so there should be instructions and also definitions for things like a "dark start" which I have seen referenced here on the forum - but not fully defined.

For me, a "dark start" is powering EVERYTHING off - like they just came out of the cardboard before installation - and then bring up the devices in an orderly fashion. Since the Grid Boss will panic and dump loads when it loses communication with the master inverter, the sequencing can be important to avoid multiple power losses on the loads. Also presumably since most of us probably have our internet access ALSO powered by the loads, having multiple power failures on the load side is annoying or possibly damaging to the tech so we want to avoid that where possible.

I did not go to manual bypass on the GridBoss before updating - perhaps I should have. But I updated the FlexBoss first, and experienced 5-6 load power losses as it rebooted itself and worked through the update cycles. Each of these was initiated by the GridBoss losing communication from the FlexBoss and dropping loads entirely, rather than switching to grid. Others suggested going the other way (updating the GridBoss first) may have prevented these power losses to the loads.

Enduring that over the 15 minutes or so it took to get all the updates applied, I wanted to FULLY power everything off and bring it up in a controlled (lol as if) fashion. I started at the inverter: PV off, battery breakers off (I did on the batteries themselves), and then hybrid breaker off at the GridBoss. Then all the smart load breakers, and finally the grid breaker. Now I'm standing in the dark. Wait five minutes - which I had some wiring changes to make so I did that while it was all cold.

Powering up: I'm not sure that the order here mattered much, but I started with the battery breakers which woke up the inverter, then the hybrid breaker which woke up the GridBoss, grid breaker, and finally the loads and PV was last. Once the internet finally came back up, everything seemed to be operating and I have not seen any strange behavior, and I have not had to re-apply my settings to either unit. So far I'm satisfied with this firmware and I do NOT anticipate updating for a while unless there's a good reason.