My experiences finding the right black-start UPS for an off-grid solar system

[400bird]

How much time and money do you have into making those EG4 inverters work for you? I love a DIY project (probably more than most) but this seems like quite a lot just to improve EG4's crap design.

You've already got your main battery bank there, why build a separate small UPS instead of pulling power (before the contactors) to run the RSD and contactors?

 

[fmeili1]

How much time and money do you have into making those EG4 inverters work for you? I love a DIY project (probably more than most) but this seems like quite a lot just to improve EG4's crap design.

I'm sure you know that, sometimes a project takes on a life of its own and becomes much more extensive than originally planned. Also electronic/electric and computer controlling is a hobby of mine so the whole project is a lot of fun (beside the roof work with the concrete roof tiles... I hate this type of work...). I don't have time pressure and because it's a one man DIY project it was clear for me right at the beginning that it would be a lot of work and will take long time to finish it.

If I would have known all the details of the project before, I'm sure I would have changed something right at the beginning (but not very much). But after starting the project I need to bring it to an end. The most components were sitting for more than 6 month in my garage until I found time to start the project, so no real chance to replace something.

I'm still in <10 years for the system to pay back - which was one of my goals. Even the 2nd mini spit to actively cooling down the AIO's, which was not planned at the beginning, the ROI is still working. If the electrical energy prices will rise more than 1% per year over the next 10 years, it will be payed back sooner.

After EG4 discontinued the 6500EX-48 I would have chosen a different model if known before. But overall the AIO's are working reliable and I kind of still like them - even with some of the known design flaws. Over the month my understanding of some of the details how they work in real life gets better and better. In the meantime I found a very reliable setup with my smart home integration to monitor and automatic control the whole system. Now I hope it will last for the next 2 decades... I'm an optimist

You've already got your main battery bank there, why build a separate small UPS instead of pulling power (before the contactors) to run the RSD and contactors?

Yes, this was one first idea. But I was able to get a 12V Victron inverter from a fried for a good price and a small 12V battery does not cost very much, I've decided to not using the exisiting 48 LFP batteries for it (I needed an inverter for my blackstart situation because the controller needs 120VAC to drive the AC-in contactors)

 

[400bird]

I'm sure you know that, sometimes a project takes on a life of its own and becomes much more extensive than originally planned. Also electronic/electric and computer controlling is a hobby of mine so the whole project is a lot of fun (beside the roof work with the concrete roof tiles... I hate this type of work...).

I'm with you on all of that! I ended up building a solar tracker just because, so I get it.

I'm still in <10 years for the system to pay back - which was one of my goals. Even the 2nd mini spit to actively cooling down the AIO's, which was not planned at the beginning, the ROI is still working. If the electrical energy prices will rise more than 1% per year over the next 10 years, it will be payed back sooner.

Hahaha, 1% per year!? If your power company is anything like what I'm reading here about different companies all around the country, your time to payback will be shrinking more than that every year.

(I needed an inverter for my blackstart situation because the controller needs 120VAC to drive the AC-in contactors)

Interesting, what's this controller?

 

[fmeili1]

Interesting, what's this controller?

Right from the beginning I've decided to make the system as safe as possible. Beside the passive things to do like

• mount everything on cement board
• using quality breakers on all relevant locations (PV, battery, AC-in, AC-out)
• oversize the cable cross sections - at least a bit
• be very accurate about battery screw torques
• using LiFePO4 batteries and no other lithium based chemistry because they can't thermally runaway
• using a thermal camera while a (nearly) full load test with batteries was running to check for bad connections (I was not able to do a full load test with the possible 39kW inverter power... I just don't have enough load for testing - but I've tested with 24 kW)
• ...

I also wanted to integrate a security shutdown mechanism which should be triggered if one of the following situations occur:

• smoke detectors over the inverter wall will be activated
• an inside located emergency shutdown button was pushed
• an outside located emergency shutdown button was pushed
• remote via WiFi relay (via integration into my existing OpenHAB smart home system)
• if a camera will detect smoke or fire in the picture analysis (this was a later addon because it was easy to implement with my existing smart home and camera solutions)

I don't want to look paranoid, but nearly all of these AIO's are produced usually with cheap parts and unknown build quality. Our master bedroom is behind this inverter wall. We are living in the Mohave desert with usually less than 10 days rain per year and many days above >115 and even >120 Fahrenheit. In the event of a fire, a house here would burn down within minutes.

In case of an emergency (automatically detected or manually initiated) my goal was to cut all possible energy sources of the AIO's, which are

• solar PV
• batteries
• grid for optional charging the batteries (just an an backuP)

To disconnect the solar PV, I'm using the module based Tigo PVRSS system. The batteries are connected via 600A/80VDC relays (one per battery rack) which are normally open (NO relays). The grid is connected via two 3-pole 110A/240VAC contactors (3-pole is required because I'm using 3 AIO's per phase) and these contactors are also NO.
I also wanted to be able to disconnect all three of these energy sources remote via WiFi in case I would like to do this for whatever reason. Last but not least, the controller enclosing should be able to integrate the Rasperry Pi with SolarAssistant for monitoring of the batteries and remote controlling and monitoring of each individual AIO.

For all these requirements, I've designed and build a controller circuit. Because both the Kidde smoke detector controlling module and the Raspberry Pi getting pretty warm (beside some other parts), I've also installed a temperature controlled fan to cool the compartment of the controller. At the end, the controller consists out of about 13 relays (SSR-, WifFi and mechanical relays). The 600A battery relays need 48VDC to drive, the 110A AC-in contactors need 120VAC to drive, some other relays need 12VDC to drive, the Raspberry Pi needs 5VDC, so the controller has a 48VDC power supply and two buck converters create the 12VDC and 5VDC from it.

Because all 3 power sources are connected via NO relays (even the Tigo PVRSS), I need a blackstart UPS to power at least the controller circuit for a short time to "boot" the system.

It was a lot of work but also fun to design and build it. Now I can sleep with a better feeling and I know I've done a lot to reduce the risk of a disaster.

 

[zanydroid]

If you were starting with a higher level of UL listing on the inverters and batteries, where would you have focused the effort for bang for the buck? Or do you think you would have done a similar level of support equipment.

I think the supplemental cooling is a good idea. Wall mounting on fire rated material and smoke detector / heat sensor are required by code when you have indoor batteries

I am pretty sure UL will let system go out the door with much flakier battery shutdown than what you have

 

[400bird]

Wow, I applaud your work.
Sounds like belt and suspenders. And a girdle, and duct tape, and rope, and a smoke detector, and I obviously ran out of ways to hold up pants. Duct tape sounds too hacked together, which obviously isn't this, well done. Just far more layers of safety than I feel necessary.

My BMS disconnects the battery. PV monitors for arc fault and ground fault.
Outside of that, the inverter can monitor for it's own faults and shut down as needed. I did go with tier one equipment, so I've got less concern than I would with relabel inverters.

 

[fmeili1]

If you were starting with a higher level of UL listing on the inverters and batteries, where would you have focused the effort for bang for the buck? Or do you think you would have done a similar level of support equipment.

At the beginning of this project it was hard for me to decide to go with no Tier 1 equipment. But for the performance data which I would like to achieve, it was nearly impossible from the price point to go with Tier 1.

Because it's a fully off-grid solution and a 200A serviced house (everything electrical), my goal was to have at least 30kW inverter power. One goal was that I don't want to have any loss of comfort (I've promised my wife ti get her ok for the budget ) which is a bit unusual for an off-grid solution where people usually change their behavior of using electrical loads (e.g. use the dryer if the sun is shining or don't use the oven/cooktop in parallel with the dryer, etc.) - which makes total sense - but I didn't want to break my promise . In the future I may own an EV and a wall box would be required, so the system should have enough power reserve for the next years to come. As a result the system is way oversized for the current situation by intention.

Also because I have 8 PV strings the solution needed at least 8 MPPT chargers (with possible options to extend). The MPPT's need to be able to handle up to 400Voc in my situation. An optional grid charging possibility, because I'm not allowed to get disconnected from the grid (because of city code), I would like to use it as a last resort backup if the batteries/solar will not be enough. So I need to pay the base fee ($15) anyway I would like to use it if required (maybe 1-2 days per year for just a few hours).

To reach all these goals without blowing the budget, I've decided to go with six AIO's (EG4-6500EX-48) running in parallel - with a tear in my eye because it's not Tier 1 equipment. But I have

• 39kW off-grid max. continuous inverter power
• 12 possible MPPT strings with 500Voc each and max. 39kWp
• 39kW max. grid power (optional and very rare) if the AIO's need to switch to bypass mode because of empty batteries and/or no solar power
• because I use 6 units the system is a bit of failsafe in case one or two of the AIO's may break in future

With Tier 1 equipment it would have cost a fortune to get this. The additional costs for wiring 6 AIO' in parallel and the additional cooling a very low compared to Tier 1 equipment - but it was time consuming.

I think the supplemental cooling is a good idea. Wall mounting on fire rated material and smoke detector / heat sensor are required by code when you have indoor batteries

Yes, cooling helps any electronic anyway - on top of that, it was a requirement to get the noise levels of the AIO's down...

I live in an area where still the old NEC2008 is active and there was not requirement for smoke detectors and rapid shutdown when I've get the approval for my DIY plans from the energy provider and from the cities building department - the even didn't ask about that.

I am pretty sure UL will let system go out the door with much flakier battery shutdown than what you have

I hope I will pass inspection easier if they realize the additional effort which I've implemented for security and fire safety features... I will find out in the next 2 month...

 

[fmeili1]

Wow, I applaud your work.
Sounds like belt and suspenders. And a girdle, and duct tape, and rope, and a smoke detector, and I obviously ran out of ways to hold up pants. Duct tape sounds too hacked together, which obviously isn't this, well done. Just far more layers of safety than I feel necessary.

My BMS disconnects the battery. PV monitors for arc fault and ground fault.
Outside of that, the inverter can monitor for it's own faults and shut down as needed. I did go with tier one equipment, so I've got less concern than I would with relabel inverters.

I know, with Tier 1 equipment I could have saved a lot of additional time to work around some things - but for my required/whished performance data I would have not been able to pay for it (see my previous answer).

I've done a lot of research about the EG4-6500EX-48 and this device is based on a "Axpert Max (MKS2-6500)" design and is produced by Voltronic. I know, the design has some weaknesses like it's an HF and not an LF inverter (less inductive surge capacity, but I compensate this with more units in parallel and a soft start for my central 4 ton heat pump and flickering lights issues because of the HF design, but I can live with it and it's workarounds).
Beside this, the Axpert Max design is "more or less" mature and Voltronic is a big inverter manufacturer with quality standards established. So I've decided to try it out. The worst thing now is that EG4 discontinued this AIO and there will be no more firmware updates available in future
So far and for the price of the units, I am quite happy with these AIO's.