(image intensive) Re-purposing Radian Load Center for Battery bank and SMA SI Load Center

[callmeburton]

I picked up a Outback radian box (empty) and a Outback Radian Load center off Craigslist earlier this year for $150 and planned on using it but didn't know where ...
Originally I thought the Outback Radian box had all the components in it but found out later it did not so I decided it would likely be a good candidate for my Eve 280ah 48vdc battery box ...

Currently I am wiring this up to be an always on microgrid within my current home that feeds all the circuits off our transferswitch. So the transfer switch would permanently be switched to "GEN" and be fed by the microgrid powered by the SMA SI off the battery bank. We are doing this as our new house we are building is 100% off grid and I wanted to use the batteries now and have the SMA and BMS preconfigured and ready to go and we get a lot of blackouts in winter.

https://diysolarforum.com/threads/sma-sunny-island-wiring-set-up.35605/post-657140 (post with the wire diagram for general top level connections between components) Diagram attachment here

The wiring for the batter to load center to SMA SI will be similar to my early mockup of my original Midnight solar setup I was originally planning to use. https://drive.google.com/file/d/1zBiWcO6m7j04EqA0LuignK4pw0Ez3b8m/view?usp=sharing (full res here, attached capture below)
I have a picture of the battery box conversion before I finished wiring up. I have foam on either side of the batter and backer board under and behind it. Batrium BMS sitting on outside. Even though this is "done" I still have not connected the sensor leads to the balancer and they continue to be connected to the tester just in case I mess something up while assembling it ... after I get a cover setup and secured I will migrate these connections from the tester to the blanacer.



The GS Loadcenter looked like it was configured for flex inverters to power the loads, or the grid to power the loads. The generator as wired wasn't connected to anything. And it appears the remote trip breaker could be powered by 12 or 48 volts.


I have now rewired it but I don't have a diagram to show the end results yet ... only that it closely matches the diagram above. I have the mechanical lockout breakers terminating at the AC2 output bar (which I drilled and placed top left in the image) and the inputs going from GEN or GRID. The remote circuit breakers relay for 12v has been removed and I retained the 48 volt relay and rewired it to hook up to the SMA SI or the BMS to act as a load shed relay. The input to the remote circuit is AC1 and the output goes to one of the other 50amp breaker for the LOADS. And this obviously goes to the loads buss bar.

The pre-charger cicuit sits top left and I connected the contactor directly to the DC breakers (which now have a pin in them so they move together) ... The red L2 buss bars have been removed and a din rail was placed in there for the Batrium BMS and expansion board ... I put 3 extra din rail fuse holders for that circuit bottom left near the existing din relay.

The neutral bar is no longer grounded, and the negative bus bar has been replaced with a short one and rotated to make it fit in the space I had as I had to replace the existing buss bar which was incompatible with the batrium shunt sensor. The following image doesn't show current state as I have everything wired up less the external sources (SMA COMM / Relay / inputs / outputs, Battery inputs / balancer) ... but this image is clearer as I haven't cleaned up wires yet since I need two more cables before continuing.



Will update this as I make progress. I am waiting for a shield cat 8 22awg cable to come so I can chop it up and use it as my coms cable to the SI and for the batrium balancer to the core unit (since it is also shielded)

EDIT: the breakers on this unit say 50 amp but if you flip them over or look at the side (wherever the label is) it says they trip at 63amp ... was happy to see that)

 

[callmeburton]

Updates:
Got everything wired up, was about to turn the ON switch "live" in video when I saw it wouldn't work as I expected because the battery was connected to the contactor then the contactor to the plate ... the issue being nothing would get power upon switching the switch to ON :D
I had previously ziptied everything so I had to undo a lot, pull the DC breakers, switch the plates around and rebundle everything ... Oh, and I created a buss bar out of one of the plates used to extend the DC breakers to the other inputs on the original box (as they have 2 terminals for positive and negative)

Cycled the main switch on after checking connections that should and shouldn't exist and everything spun up as it should. Or so it would appear that way ... the contactor did nothing so I had no voltage from the new contactor output and the shunt. I had to re-configure the batrium core to get on the wifi network which was fun as I previously didn't copy down my pin number ... thankfully I left enough room in this design to pull the board out so it took less than a minute to get the pin.

The watchmon is now configured to have Relay A closed when battery check is OK ... which isn't ideal but is desired. I want it to use the same relay to open at different SOC's and I would love it if the DC breaker is open but the battery system is on that it would also have the contact OPEN and only close after the DC breaker closes again. (I will have to read up on this)

Some interesting findings and I don't know what it will mean moving forward as I haven't dived into the batrium programming for the relays or the SMA SI configurations ... but the remote breaker switches when it gets a power impulse of about 3 seconds. So if it is OFF and you cross the two leads for 3 seconds it will slowly switch to on ... at this point you should not be giving it extra power as it will strain ... If you again cross the leads for 3 seconds it will move the switch to off (more quickly than it did on) and again you shouldn't be supplying power.

I think the SMA SI simply sends power all the time if things are "good" then doesn't when things are configured to open the load shed relay. If I had the batrium do this I imagine I would need to use multiple relays and build in some memory with those relays as to the state of the breaker ... then shield the breaker from human hands so as not to loose that "memory"

Current state of the battery ... I still need to take some of my plexiglass and make a cover for it ... as well as mount the switch to the frame.


Photos of current state of the GS Load center ... will have to make a final wire diagram later.




You can see my lugs simply hanging out on the SI battery cable attachment points ... I had to drill some of the lugs greater than the 1/4" holes they had for the various posts ... when I did this I would attach them to their assigned posts to not miss allocate where they go. I haven't decided the length of the final battery cables yet but I have about 5' of cable left in both black and red.

Batrium on wifi Node Chart (I haven't fully configured the low / high values for the eve 280ah cells yet ... won't have to till I get the SMA SI connected to it.

 

[callmeburton]

Was looking to better understand setting for the batrium this morning and configure the contactor tripping and came across this thread https://diysolarforum.com/threads/batrium-watchman-core-news-warning.50606/ ... going to have to add a DC:DC transformer into the mix above (likely victron) ... the general wire diagram above doesn't include this feature.

 

[callmeburton]

I have been thinking about how I can control the contactor being open / closed based off a AND logic gate ... this is what I think I have to do with what I current have (the expansion board and watchmoncore)


This I believe would be
Battery state OK (watchmon condition controlled relay) ... currently using on relay A
AND
Battery SOC > 20% (watchmon condition controlled relay)
AND
Power at input on Expansion board (this would come from upstream of the DC disconnect)

When all three conditions are met the rec bms should get the signal to switch on. If either is NOT met the Rec should open the contactor.

This would also work

 

[callmeburton]

Just realized I don't want to open the contactor when the SOC is 20% ... as it should still be connected to the SMA SI so it can be recharged :/
So it should go Contactor + > Relay A > Rec BMS for now ...

 

[callmeburton]

UPDATE:
Wired the contactor positive to the BatOk set at Relay A then to recBMS ... it seems silly to also setup a 20% relay right now as the SMA would control for that not the BMS ...
Also got my comm cables made (for SMA SI and cpu cable) ... catabolized a double shielded cat 8 cable to make them. So now just waiting for the DC-DC converter to come in to step my 48vdc down to 24vdc

Also was thinking about the SMA SI Relay settings for "all is good" or "Load shed relay" ... I am assuming it sends a constant signal when everything is "good" then disconnects the signal when it isn't.

The radian 63 amp remote trip breaker from no signal takes in a 2 second signal to switch the breaker ON or OFF then takes no signal. (So 0v, 48v for 2 sec, then 0 v again) To make this "work" with a signal like the SMA provides I am thinking about this relay soup solution using NC 2 second delay Open relays (if I can find them, I know they exist just don't know if in 48vdc ... if they are NOT then I will have to use lower voltage [like 5vdc] relays with the same attributes to control other relays which would allow for the 48vdc signal to get through ... meaning it would get a little messier but still work)

Start position at power on of the battery (labeled 48 below) would be ON so the relay would switch OFF ... OR before turning on the SMA make sure the remote trip breaker is set to OFF and I believe this should "work"

 

[callmeburton]

UPDATE:
Looking at the cost to set this up. ~90$ usd right now. 2x multifunction relays (one set for interval open, the other interval close), and 2 "normal" relays (one 48vdc one 24vdc) ... trying to get multifunction relays at 48vdc gets very pricy quick. So all relays less the one the SMA is connected to would be 24vdc for their coils.

So ... this is getting costly for something that was suppose to not cost so much to make (given the project was to reuse an older Radian Load Center. This brings me to another idea.

I have an old project I never completed ... it required the use of a aircraft 3pdt 28vdc relay ... I have not tried to close it yet with 24vdc but it wasn't cheap, think I paid like 70-90 for it in the past. It is hard to find documentation on it but it was going to be used in a electric motorcycle charging setup on the AC side of things. (images attached below)

It is rated for 50 amps, 115/200 VAC 400hz ... which means it is beefy ... normally you would derate a 60hz system when you run 400hz through it ... the main issue is the 28vdc requirement to close the circuit. But this would only require the purchase of one relay at the expense of not using the cool included Radian Load Center reset relay and not having a "switch" to turn it on/off (which I could technically wiring in if I wanted to I guess)

 

[callmeburton]

Musings. If I use a multifunction relay setup for Interval open on signal ... then I think if I feed the SMA relay into the input on the batrium expansion board and "listen" for an on/off signal I can use TWO relays on the board as switches which would go to a single multifunction relay ... at most I could see using 2 relays if I need to drop down the SMA signal to 24vdc.

Theoretical layout:

SMA SI Relay out (for load sheding) > XBoard Input 1 (might go to relay to step down to 24vdc first)
SSR 7 (close for 200ms when seeing Input 1 ON signal) > multifunction relay set for 2sec interval
SSR 8 (close for 200ms when NOT seeing input 1 Signal) > Same multifunction relay as SSR 7
Multifunction relay setup to close Stock switching breaker.

 

[callmeburton]

Updates (New wire diagram attached):
Got the meanwell 48-24 dc-dc converter installed last night, changed up the wire diagram a bit to fuse the converter going into the supply as well as the outputs ... had to add another 5amp din rail fuse for this.

After I got everything up I noticed my supply voltage to the shunt was on 24v and had to remove jumpers and throw it on the pre breaker side to remove the low shunt voltage error.

Was playing with the remote trip breaker preinstalled in the box ... it is an oddball for sure and might require a redesign of the logic to use it :/

Breaker discoveries (breaker in OFF pos)
Cross leads and hold > 1 sec, could be as long as you want and doesn't seem to hurt the device.
Breaker makes a noise then stops like it is winding up a camming device
Release crossed leads
Breaker switches ON (only way to turn it "on,")
Cross leads momentarily
Breaker switches OFF

Wire diagram below shows the current wired state of the box less the SPD ... haven't hooked it up to the SI yet or ordered extra relays for the SI to Remote Breaker logic.

EDIT: Updated attachment for newer version. Includes more detail and information in the off chance someone tries to implement something like this.

Details on the remote breaker:

https://www.carlingtech.com/sites/default/files/documents/Carling-HM-CB-Time-Delays.pdf

https://www.mouser.com/datasheet/2/240/C_Series_Details__26_COS_040314-3050202.pdf

CA2B026650321CG / RB2130CU1G

 

[callmeburton]

Attached diagram how I think the final system (less inputs from gen / grid) will be laid out ... just ordered the relays required to do this.
The goal would be to set the interval relay to the min amount as possible and it should "work" as expected.

 

[callmeburton]

EDIT:
Updated Wire Diagram (v1.2) to show Relay connection details for this setup more accurately.
Updated Wire Diagram (v1.3) to fix connection to SI Relay from NC to NO

UPDATE:
Batrium expansion can't be used how I thought it could ... ran 48 volts to input 1 then tried all options and couldn't get the functions I wanted to launch on the other relays to trigger. ... That is they are not programable but rather tied to events you cannot fully define :/

Ordered another timed relay ... the new plan is SMA > 48v relay (output battery 48vdc or 24vdc) > both timed relays ... one set ON interveral one OFF interval > load breaker. My testing last night confirmed this should work but it means waiting for a couple more days for the new relay to come in.

The reason for the SMA > 48v relay is I don't want to assume the signal out of the SMA is straight battery bypass or if something else ... so I am going to isolate the SMA relay output by running it to a relay in the load center which will switch either the 48vdc battery voltage or the dc-dc 24vdc meanwell source. It appears the timed relays I got should handle 48vdc so will likely leave it at that and fuse it with another din fuse.

Updated wire diagram:
Added diodes aft of the multifunction relays
Added patterns for multifunction relays
Added another Din fuse for those relays
Removed expansion board only solution
Updated some line styling for readability
Updated multifunction relay to have signal inputs

 

[callmeburton]

updates:
Got the remote breaker setup correctly today, in fact just 15 minutes ago ... will have to tidy up the wires tomorrow but everything is working as expected.

Both timing relays on a new din rail and the 48vdc relay which will receive the SMA SI relay output for load sheading ... this was testing before mounting the new din rail into place.



Next shows where I mounted the new din rail to hold these extra relays as well as the 3 amp fuse din holder on the same rail. Everything is untied here and will need to be retied to make it tidy ... see what I did there?




Next steps, tidy it all up, then move it upstairs next to our main panel and connect it to the SMA SI I have sitting around so I can start to configure it.

 

[callmeburton]

UPDATE:
The load center / battery has been connected to the SI for a couple days now and running. There is an issue with the SI receiving a low SoC signal from the batrium bms which I have posted a separate thread here to discuss. https://diysolarforum.com/threads/batrium-sma-sunny-island-integration-issue.55111/

To "fix" this and to test the batrium bms as the source for this signal (or not) I have required the load shed breaker so the power from the SI battery terminal block feeds the relay to drop voltage down to 24vdc. (previously this would have gone through relay2 on the SI) From the 24vdc output I now run it to relay5 on the expansion board set to the same conditions as the SI for load sheading. (attached modified wire diagram below, note description and version number have also changed ... any bypass SI relay logic versions of the diagram I will put a B after the version number)

In the Batrium:
Control Logic > SoC Limits
SoC% limit 1

• Lo Cutout -5
• SoC% Lo 25
• SoC% hi 101.5
• Hi Cutout -5
• Based On: Critical Ok (this means the relay will be OPEN if there is a critical issue)

Hardware > Expansion
SSR5 SoC% Limit 1