Running Large Inverters w/ Relays: Not a good idea? I have 3 alternative solutions

[Will Prowse]

Really great discussion going on about my recent "relay for large inverters" video (click here).

But after thinking about it and seeing everyone's input, I am going to redesign my raw cell blueprint to avoid the use of relays and FET based switches entirely. Both work well if sized properly for application, but there are too many potential problems:

1. Some relays are not rated for continuous duty
2. Relay kickback issue, which can destroy BMS if no fly back diode
3. Adds unnecessary complexity
4. Capacitive load issues. If relay is sized properly, it shouldn't be an issue. But you need to make sure that relay is huge.
5. Increased idle consumption

And FET based switching mechanisms, such as Battery Protect, have their own problems:

1. Inrush surge can cause damage to fets. Victron recommends not using them with inverters anymore for this reason
2. FET's running capacitive loads can't be that great (from what I am reading right now. those spikes will damage them eventually. even if they do not fail, you may get a leaky fet)

New and Improved Design:
3 safe methods for controlling low voltage disconnect of large inverters:

1. Using a high quality inverter with relay input. This way you will never need to worry about switching large currents. Use a BMS or other form of logic control system to control the inverter.
2. Increase the voltage of your battery bank to reduce the current requirement of your inverter, then run a standard 100-300A BMS. A 100A 48v BMS is small and can easily power a 4500W inverter. A 300A 48V BMS is a bit larger, but can run a 13,500W inverter.
3. Run BMS in parallel. But! Not in a single bank. That can cause other problems. You will need to make multiple battery banks with their own BMS, then parallel the battery banks together for a larger current capacity.

Would love to hear your thoughts on this. I will start making a new blueprint for all this unless someone disagrees with the new designs.

 

[Will Prowse]

I just started updating the schematics:https://www.mobile-solarpower.com/design-your-own-12v-lifepo4-system.html

Will finish them up in the morning

 

[Steve_S]

Will I suggest you look at the Samlex EVO Inverters, the have comm ports to accept a remote off command from a BMS (any 12v signaller). This is mentioned in threads here now.

Next I suggest you put in a disclaimer that these are suggestions and people are using this info at their own risk. I can see lawyers in the future if you don't. No need to feed those trolls ! Miss one safety step / point which results in loss, injury or death and the "stuff" hits the fan... Damned litigative societies.

Parallel Bus Systems are barely touched on and do need more detail. They can be complex & problematic if things go awry. They don't have to be though if the rules of KISS are applied. ** Some clarity needed too, a few folks "think" they can share one BMS with two separate packs... oivey… ( 1 pack + 1 BMS = 1 complete separate entity.)

I know you have no experience with the Chargery BMS', I've pointed Jason Wang from Chargery at you & this forum, I suggested they send you a couple of examples to play with. Hopefully you've been contacted but given the 2 week break starting now, I'll remind him with my next order for BMS'. He did have the Firmware updated along with the manuals to include Low Temp Cutoff after I suggested them to do it. Yes, it IS a company that listens, also suggested a BT or PC interconnection module which is being pondered over there ATM.

Last suggestion, safeties & protections need to be considered, not only for the RV/Mobile world but for actual ground based installations, including addressing Lightning Protection (it's not pleasant if/when it happens). Not only do panels & their racks need protection but so does all the DC & AC stuff as well.

BTW: I have seen some very disturbing comments form a few which border on the edge of wanton malice, almost as if they want things to fail or cause damage. I pull away and make no comment, not sure what to do , I don't need to be fighting with trolls and / or ding-a-lings, too old for that shyte. Given that, I suggest you protect yourself & this forum + website with appropriate disclaimers and try to address the general safety issues and problems with "known good practices" . As to how to deal with the wives tales, mis-information (unintentional lack of knowledge / understanding) or dis-information (willful information to confuse, obfuscate or to create fear) is something that needs to be addressed (like many other places, not unique to here).

Hope that helps Will,
Thanks for all the Great Work & Effort to make this place happen and help people gain some freedom & independence.
Steve

EDIT:
I just looked at the WIP, certainly headed the right way. I saw you posted one diagram of a single 4s2p config. Would it not be possible to draw up a Pack Layout Sheet showing the common Series, Series/Parallel configuration for 12,24 & 48V packs. Thinking along the lines of what they have posted here but better: https://www.ev-power.eu/BatteryPacks/ but including S/P packs as well. Many seems to get a bit confused when starting to link up within a pack, especially larger ones and how to BMS them properly. Imagine how useless a BMS16 would be on an 8s2p the P's would be dupes and essentially false values. Just a thought.

 

[Darkstar]

I don't suppose the MPP solar LV5048 has any sort of relay input or remote on/off function controllable by BMS?

I was thinking of going that route in my new system, but...

 

[lmwilco1]

Can you add a shunt to all of the diagrams as an option for power monitoring? Still wondering what impact it will have if the main BMS ground goes through the shunt and the balance leads do not.

 

[atatistcheff]

I also keep going back to the question if a LV cutoff is really necessary when powering an inverter? Seems to me that most 12V inverters cut out around 11V anyway which is above the typical 10V you'd see the BMS cut off. Of course this assumes cells are bottom balanced so you don't have one that's terribly low. I'm not sure about 24V inverters, do they also typically cut out before the BMS LV protection?

 

[Steve_S]

Can you add a shunt to all of the diagrams as an option for power monitoring? Still wondering what impact it will have if the main BMS ground goes through the shunt and the balance leads do not.

Correction: NOT Ground, it is Negative Line. As it is on NEG it has no impact on power off/on. A shunt does add a little bit f resistance & drop but not significant on it's own BUT every connection, every terminal adds resistance / drops through the whole circuit. It is essential to take volt reading at all points and correct for deviations on the Inverter / Charger & the SCC so they match actual readings. The balance Leads do not carry current to/fro that is internal to the system.

 

[Will Prowse]

I don't suppose the MPP solar LV5048 has any sort of relay input or remote on/off function controllable by BMS?

I was thinking of going that route in my new system, but...

For that I would just set the lvd threshold to 12v. That would work great

 

[Will Prowse]

I also keep going back to the question if a LV cutoff is really necessary when powering an inverter? Seems to me that most 12V inverters cut out around 11V anyway which is above the typical 10V you'd see the BMS cut off. Of course this assumes cells are bottom balanced so you don't have one that's terribly low. I'm not sure about 24V inverters, do they also typically cut out before the BMS LV protection?

Yeah that's the main issue. Most BMS top balance. If one cell has lower capacity than others and cuts off a bit late (which I had happen a year ago), then the inverter won't see it and will over discharge a cell. Individual cell voltage sensing is required for this cut off system.

 

[Will Prowse]

Can you add a shunt to all of the diagrams as an option for power monitoring? Still wondering what impact it will have if the main BMS ground goes through the shunt and the balance leads do not.

Put a shunt between p- and loads/chargers. Don't put it between the main negative and b-

 

[Will Prowse]

Yikes those are expensive. I agree, people should spend more and use one with remote off command with a bms, but it sure is hard to convince people to spend 2-3x the amount of a similar sized PSW inverter. But yes, looks cool. I need to test one

No I am well aware of the risks so there are disclaimers on literally every website I own. The first businesses and books I started were covering medical advice, which required a lot of disclaimers.

Did you read my disclaimers? Do you have suggestions on what to add or change? They seem good to me.

With the cell configuration in schematic, its 100% fine to put cell packs in parallel and then series. Not sure what you are referring to here. Maybe communication ports between bms when they are in parallel? Not sure what you mean

It did not have low temp cut off on the chargery before? that's concerning... how long have you used them for? Really hard for me to be curious of their bms if they are just now adding that feature. what voltage rating are the FET's?

Yeah I havent covered that at all mainly because code for small arrays does not need it. Same with dissipating excess charge from a small array. Completely unnecessary. I haven't released information on making large arrays yet for this reason. I know the code and what is needed, but making a suggestion for people to follow can be problematic. Still thinking about this right now. Also, code is different in various locations on this topic.

Wanton malice? Like what? Can you link me some of these instances?

Yes, I stated site is under construction at the top and I am adding 24v/48v cell configurations today. I am still pretty sick and trying to crank them out as fast as possible. they are coming

Will I suggest you look at the Samlex EVO Inverters, the have comm ports to accept a remote off command from a BMS (any 12v signaller). This is mentioned in threads here now.

Next I suggest you put in a disclaimer that these are suggestions and people are using this info at their own risk. I can see lawyers in the future if you don't. No need to feed those trolls ! Miss one safety step / point which results in loss, injury or death and the "stuff" hits the fan... Damned litigative societies.

Parallel Bus Systems are barely touched on and do need more detail. They can be complex & problematic if things go awry. They don't have to be though if the rules of KISS are applied. ** Some clarity needed too, a few folks "think" they can share one BMS with two separate packs... oivey… ( 1 pack + 1 BMS = 1 complete separate entity.)

I know you have no experience with the Chargery BMS', I've pointed Jason Wang from Chargery at you & this forum, I suggested they send you a couple of examples to play with. Hopefully you've been contacted but given the 2 week break starting now, I'll remind him with my next order for BMS'. He did have the Firmware updated along with the manuals to include Low Temp Cutoff after I suggested them to do it. Yes, it IS a company that listens, also suggested a BT or PC interconnection module which is being pondered over there ATM.

Last suggestion, safeties & protections need to be considered, not only for the RV/Mobile world but for actual ground based installations, including addressing Lightning Protection (it's not pleasant if/when it happens). Not only do panels & their racks need protection but so does all the DC & AC stuff as well.

BTW: I have seen some very disturbing comments form a few which border on the edge of wanton malice, almost as if they want things to fail or cause damage. I pull away and make no comment, not sure what to do , I don't need to be fighting with trolls and / or ding-a-lings, too old for that shyte. Given that, I suggest you protect yourself & this forum + website with appropriate disclaimers and try to address the general safety issues and problems with "known good practices" . As to how to deal with the wives tales, mis-information (unintentional lack of knowledge / understanding) or dis-information (willful information to confuse, obfuscate or to create fear) is something that needs to be addressed (like many other places, not unique to here).

Hope that helps Will,
Thanks for all the Great Work & Effort to make this place happen and help people gain some freedom & independence.
Steve

EDIT:
I just looked at the WIP, certainly headed the right way. I saw you posted one diagram of a single 4s2p config. Would it not be possible to draw up a Pack Layout Sheet showing the common Series, Series/Parallel configuration for 12,24 & 48V packs. Thinking along the lines of what they have posted here but better: https://www.ev-power.eu/BatteryPacks/ but including S/P packs as well. Many seems to get a bit confused when starting to link up within a pack, especially larger ones and how to BMS them properly. Imagine how useless a BMS16 would be on an 8s2p the P's would be dupes and essentially false values. Just a thought.

 

[Will Prowse]

Correction: NOT Ground, it is Negative Line. As it is on NEG it has no impact on power off/on. A shunt does add a little bit f resistance & drop but not significant on it's own BUT every connection, every terminal adds resistance / drops through the whole circuit. It is essential to take volt reading at all points and correct for deviations on the Inverter / Charger & the SCC so they match actual readings. The balance Leads do not carry current to/fro that is internal to the system.

They are still voltage sensing wires involved, so I personally like to keep the shunt on the p- side of bms regardless. Thats how anyone would do it in drop in lead acid replacement lifepo4 battery system, so I am sticking with that.

 

[John Frum]

Are the doubled blue bus bars necessary or are they just for extra beefiness?

 

[Steve_S]

Wanton malice? Like what? Can you link me some of these instances?

I have seen a few and moved passed them quickly as I am not into getting "into it" with people. If I come across anymore I'll PM you as there is no point to go public as that only stirs stuff up. no one wants / needs that.... kinda like politics & religion in a forum is toxic.

What's expensive ? The Samlex ?? 1400 +/- USD compared to a Yiyen or a car inverter sure but it's a different beast.
I use a Midnite Classic 200 which sells for 7-800 USD, sure there is cheaper but....

Chargery has had the electronics there, it was just a mater of creating new firmware, they are responsive and do support folks.

on the pack design: Layout of cells like 8 series or 8 series X 2 for a 8s2p ad how to bridge them together with the bus bars in diff formats. Also though, where is the best place to put the BMS leads on the cells in different configs. It can get quite confusing for new people trying to absorb all this info and a picture image works wonders and no language barriers.

Example of a doodle I did for my documentation when I put rebuilt the ShunBin pack.
I have to document everything seriously because in a short time I won't be around to explain / manage all of this and whoever ends up owning this place IF Maggie decides she can't stay here, they'll need to know what, where, how & why details.

 

[Will Prowse]

I am in love with the Samlex. Reading the massive manual right now. Beautiful little piece of engineering

Yes, midnite classic is quite nice. And very true

I have seen a few and moved passed them quickly as I am not into getting "into it" with people. If I come across anymore I'll PM you as there is no point to go public as that only stirs stuff up. no one wants / needs that.... kinda like politics & religion in a forum is toxic.

What's expensive ? The Samlex ?? 1400 +/- USD compared to a Yiyen or a car inverter sure but it's a different beast.
I use a Midnite Classic 200 which sells for 7-800 USD, sure there is cheaper but....

Chargery has had the electronics there, it was just a mater of creating new firmware, they are responsive and do support folks.

on the pack design: Layout of cells like 8 series or 8 series X 2 for a 8s2p ad how to bridge them together with the bus bars in diff formats. Also though, where is the best place to put the BMS leads on the cells in different configs. It can get quite confusing for new people trying to absorb all this info and a picture image works wonders and no language barriers.

Example of a doodle I did for my documentation when I put rebuilt the ShunBin pack.
I have to document everything seriously because in a short time I won't be around to explain / manage all of this and whoever ends up owning this place IF Maggie decides she can't stay here, they'll need to know what, where, how & why details.
View attachment 5353

 

[Steve_S]

I am in love with the Samlex. Reading the massive manual right now. Beautiful little piece of engineering

Yes, midnite classic is quite nice. And very true

BTW, I have gotten to know the Samlex tech engineers and THEY know their shit like no one else ! and really good support too. Even Mike the company president is a tech head and not afraid of it. We've tossed some things around and percolated some ideas.... ( I was part of the Protocol Standards & Communications group for the Canadian Gov, so... leads to things) Really good company to deal with IMO and I do NO say that lightly, I'm a real "prick" when it comes to companies & their support (see shunbin thread for an example of me getting prickly).

 

[lmwilco1]

Put a shunt between p- and loads/chargers. Don't put it between the main negative and b-

The inverter needs 200A and the BMS is only 100A so I need to use either the relay to bypass the BMS or control the inverter on/off with the BMS output. Either of these options require bypassing the BMS for the negative feed to the inverter which is why I am asking about placement of the shunt. If I connect it after the BMS then it won't see any of the inverter current.

 

[Power4life]

1. Using a high quality inverter with relay input. This way you will never need to worry about switching large currents. Use a BMS or other form of logic control system to control the inverter.

So if were to use this method of connecting my inverter (which I prefer the most) and I want to have a shunt for capacity monitoring, wouldn't I want to place the shunt between my main negative and the b- of the bms with the inverter (-) connected to the load side of the shunt
Here you said.

Put a shunt between p- and loads/chargers. Don't put it between the main negative and b-

If I were to do this wouldn't my capacity reading be false since the inverter is bypassing the bms thus bypassing the shunt monitor all together.

 

[Will Prowse]

Ohh I see what you guys are saying. So bypassing it altogether and using external communication method. Then put it on the battery main negative then. That works.

Honestly put the shunt wherever you need to on your system. Really depends on what you're doing. I think shunt location should be up to individual system because variation in busbar/bms/inverter wiring. Just stick it wherever you want to measure it.

 

[Maast]

Really great discussion going on about my recent "relay for large inverters" video (click here).

But after thinking about it and seeing everyone's input, I am going to redesign my raw cell blueprint to avoid the use of relays and FET based switches entirely. Both work well if sized properly for application, but there are too many potential problems: .................
Would love to hear your thoughts on this. I will start making a new blueprint for all this unless someone disagrees with the new designs.

I would argue that for the DIY audience on this forum using large relays to isolate the battery bank on fault or low/high cell voltage is the BEST way to go about it. I would even argue its the best way to go about it period - because simplicity is usually better and this is pretty simple.

HOWEVER you can NOT cheap out on components: As mentioned you have to use a continuous duty relay, and you have to oversize the relay by a factor of 2 or 3 and you have to have a flyback diode. None of this is particularly expensive.

The reason being that most of the people here are looking for a best-value solution and simply will not be willing to pay the extra for inverters that will interact with BMSs and you need cell-level monitoring. Furthermore if you start designing solutions around those inverters you get locked into a narrow (and expensive) range of solutions to talk to those inverters.

From a reliability standpoint, having large currents going through a external mechanical relay will greatly outperform large currents going through inexpensive BMS. A good relay thats properly sized for its application will last just about forever.

Also, you can greatly reduce the power requirements of a relay by using a resistor/capacitor circuit but thats getting a little complicated for here and a good quality relays doesnt use much power anyway - in fact you can judge the quality of a relay by how much power it uses and how hot it gets.