High voltage hybrid inverter (GW6000-EH) with EV battery (Chevy Volt)

[yasko]

@powertrack,
I've just checked your project thread (with Google translate), and I'm impressed. That is real hardware with all that power electronics and big Leaf battery. I was thinking about such an approach but then decide to go Goodwe hybrid inverter. It's a pretty good device and indeed has all features for the HV energy storage system.
I agree with you a CAN Converter/Gateway is a core component in these projects. The current solution (Raspberry Pi) is somewhat too complicated, and I'm not sure about its long term reliability. But it's making hacking and debugging a lot easier. I'm planning to make an embedded device (MCU-based) and replace it, but this will take some time.
BYD HV System is an intriguing device supported by many inverters. I have to check how Fronius Symo talks to it, whether CAN bus or MODBUS.

@Cheap 4-life,
I took the battery from a local dealer of used Chevy Volt cars. (located in Bulgaria, Europe). It's from a 2012 car and was staying unused for a long time. I've checked it and found two dead cells (zero voltage). Other cells showed a capacity of around 38.5Ah from the original 45Ah.
I've tested cells with iCharger B208 according to datasheet limits from 4.15V to 3.0V. So the numbers that I gave in my previous post are real-life data, not estimate. Now I'm using 94 of 96 cells, which equals the maximum available capacity of 13.4kWh. Let say that I go with 80% DOD that is 10,7kWh usable capacity from the inverter.

 

[Cheap 4-life]

Some of my batteries also seemed to have dead cells. They weren’t dead, the little fuses for those cells was blown. Had atleast 6 of them blown. David Poz has a video where he shows replacing one of the fuses. For me I didn’t want to deal with more fuses blowing in the future so I connected a wire directly to each cell for the bms. Basically removing all of the fuses. If you checked the cells at the bms connector, instead check the cells at their individual tabs which bypasses the fuse. My apologies if you read all that and already checked the cells at the tabs.

Yes, 4.15 to 3.0v is typically max volt and low volt for the cells to prevent damage. Same for most lithium cells (except Lifepo4). I will find where I seen that there’s only 11kwh of available power to use from a new 17kwh gen1 full pack. And I’ll also try to find the info about which voltages the chevys bms allowed. I’m fairly certain it was 3.5v to 4.05v.

 

[Cheap 4-life]

There ya go. Gen1 17kwh full new pack has usable capacity of 10-11kwh. The bms only allowed a 65% DOD. Your pack is 16kwh so that’s 9-10kwh of usable power if the batteries were new. You said your cells have 86% of original capacity. 86% of 10kwh is 8.6kwh of usable power...

 

[Cheap 4-life]

As for the actual cell voltages the chevys bms allowed... I haven’t been able to find where I read that. I know it was conservative. They did this to mitigate the chance of over-under charging-destroying cell, causing fires. I’m glad they did that because it helped keep these Chevy cells in good shape. I think Nissan Leaf also ended up having people bring their cars back in and reduced the voltage range that could be used. Then people were complaining about no longer being able to drive as far from a charge. I’ll keep trying to find it, but I know it was 4.05 to 3.5. That’s why I set my voltages real close to that at 4.07 to 3.5

 

[yasko]

I agree with the numbers above, and they are correct. But in my case, I'm not using all components of the original BMS, so there are no limits for charge and discharge voltage. I'm working with the so-called BICM (Battery Interface and Control Module). They are attached to the top of the battery sections and provide cell voltage and the temperature inside a battery.
So it's my responsibility to chose proper limits for charge/discharge voltages. Your numbers: 4.07->3.5V sounds reasonable, and I will test them. To make things more clear, attaching a simplified connection diagram

 

[powertrack]

Hi Yasko, thanks for this informations, what I'm really woundering is, does the goodwe not need any kind of precharge?

 

[yasko]

Based on my experience with the inverter can conclude it integrates a precharge circuit but can't reference any document. There is a relay that switch-on after battery connection.
Also, the inverter can work with lead-acid batteries (not recommended by Goodwe) without additional BMS. That is the so-called self-define battery, and it's possible to adjust its parameters by the app.
So I will rely on the inverter precharge circuit, but it's necessary to add a contactor for safety disconnect.

 

[powertrack]

Hmm, I'm wondering regarding that this inverter can handle lead-acid batteries without BMS, because:
Approved Battery Options Statement_2019.pdf (goodwe.com)
The EH Series would not possible to do this, but maybe my English is too bad, can you choose on your GW6000-EH "Self define battery" with min/max Voltage (Battery) and min/max charge/discharge current?
If so, this would be an absolute "unique selling point" for Goodwe, all HV-DIY builders could use this

 

[yasko]

Officially Goodwe doesn't support the usage of lead-acid batteries with the EH inverter series. But unofficially, I can confirm that GW6000-EH can work batteries without BMS with some limitations.
I'm attaching some screenshots from the "PV Master" app. When you choose a battery type, there is an option for a self-define model. Then it comes to a warning about usage of a lead-acid battery, and after that, you can choose battery parameters. The last screenshot shows the inverter using it in a self-define mode.



What are the limitations of self-define mode? The first issue: a SOC estimation doesn't match well for a Li-Ion battery and usually shows a higher value. Maybe LiFePO4 will work better. The second: max charging current is internally limited to 5.5A despite what you enter by the app.
There is SOC protection during discharge, but you can't set a min discharge voltage for the battery. Also, there is no temperature sensing in self-define mode.
I've been using my inverter in that mode for some time together with a Chevy Volt battery, but I think it's not suitable for Li-Ion batteries. I highly recommend BMS usage, and here is a screenshot of when there is a working BMS.

 

[powertrack]

Hey Yasko,
thanks for this detailed investigations!
That means it makes absolutely sense to go for this solution as you are showing, if I can anyhow support you, let me know.
-powertrack-

 

[yasko]

Powertrack, you are welcome!
I'm glad to share what I've learned. A few months ago, when I started that project, there was too little information about these inverters. I've asked some local distributors without success and then ordered mine from Germany. The distributor can't answer all my questions, and I have some doubts about the inverter. But it met my expectations and working well.
If you decide to go in this direction, I think, can help. It will be a fancy project to run such an inverter with a Leaf battery

 

[bikefish]

Hi Yasko,
very nice project. My plan is to do a similar project. HV ESS with 96S configuration with 3 phase inverter. Goodwe is also selected. BT or ET series is not fixed yet. I plan also to extend my solar panels so the selection is 80% to ET series fixed.
Currently I'm in sourcing process for the cells. It should be a LiFePo4 cell with approximateley 100Ah. (90-110Ah).

Another point ist BMS. I preferable plan to use Orion BMS https://www.orionbms.com/products/orion-bms-standard/
It's not cheap but it provides a configurable CAN interface and I hope it could cooperate direct with Goodwe inverter.
Other options for BMS are welcome. I'm interessted in sharing 'CAN' knowledge.
bikefish

 

[yasko]

Hi Bikefish,
That will be a cool project - 96x100Ah LiFePO4 cells, which make around 30kWh energy storage. What cell brand do you plan to use? Now the market is flooded with these cheap blue LiFePO4 cells, so I',m a little bit confused.
The Orion BMS is a good option, and I think it can handle the inverter protocol. You need to send around seven CAN frames with major battery parameters to the inverter. I'm making a brief protocol description and will upload it soon.
Regarding alternatives, I know three: REC-BMS, Batrium, and FoxBMS. AFAIK, all of them handle 96 cells (with expansion modules), but none of them can speak to HV hybrid inverter. The last one is an open HW/SW project and can be customized, but you have to build yourself.
There is an ongoing project here based on FoxBMS, but still at the beginning phase.

 

[Northwind]

Nice work on your system, I'll will be following with great interest.
I just picked up a 50KW 2019 Tesla model 3 battery for a wonderful price. I'm hoping to get my off grid system design figured out and begin building it in about a year.
There is a lot of great information and interesting projects on this forum.

 

[yasko]

Here is a brief description of the BMS CAN bus protocol I'm using in my project with the GW6000-EH inverter (firmware version 030313). I think it will work too on other series HV hybrid inverters (BH, ET, BT) but cannot guarantee that without further testing. It uses settings for the so-called default battery found in the "PV Master" app.
This work is neither supported nor approved by the manufacturer of the inverter (Goodwe). I've contacted them once regarding the CAN bus specification but didn't receive an answer. Their official policy about the integration of new batteries it's found here.
So probably, the usage of the Chevy Volt battery voids the warranty of my inverter. But that are the risks of DIY projects

P.S
There are some explanations about the attached protocol in this post.
P.S.S - 31.03.2021
A minor fix in the BMS CAN protocol regarding battery sub-modules.

 

[prepared1]

Yasko: Your project is looking good! I do not understand all the work you are doing with the communication, because it is beyond my experience level. I've been looking at the Volt batteries lately, and have found a local 2018 Generation 2 battery unit for a good price. Low vehicle miles, front end collision. I'm confused, because on-line I read that the 2018 was 18 kWh, but also see statements that say 31 kWh.

Did you disassemble the complete battery yourself? I've seen the David Poz video, and it doesn't look very difficult. But I'm a bit leeery to buy any battery from a wrecked car, because there is no warranty if you disassemble it. I just ordered a bunch of the Samsung SDI modules from Battery Hookup, and they should arrive in about 2 weeks. No sweat, but more money, of course.

Just a friendly suggestion: you might want to put a stiffening brace under the front edge of your shelves, because over time, that weight can cause significant sag. They look like metal shelves, so maybe a piece of 1" square tubing? I do this on wooden shelves all the time, and generally run a piece of solid wood 2.25" so it creates a hardwood face on the 3/4" plywood shelf, and extends another 1.5" below as a stiffener. That even supports my 100# amplifiers, so should work well for batteries. Of course, it could be done on the back wall too.

Hey, good luck with your project! Keep up the good work!

 

[powertrack]

@yasko:
Thanks for the protocol description, actually from my point: I will follow the concept with the Infini hybrid inverter because:
1. I spend (too) much time & money for my complicated solution
2. I'm hopefully only a few steps behind getting the system running, I will decide going on with my solution this week (waiting for some special parts) after the following week I will make the decision of going on with the infini or switch to goodwe

Anyhow, I think the community should not go to my solution (It will be some bucks cheaper, much more risk but in general to complicated) but I think it could make sense to do following:

Hardware:​

Control Main Conductors​

Control Precharge​

Bulk over and under voltage​

​

Software​

Check internal Pack BMS​

Cell voltages​

temperatures​

SOC and SOH​

​

Software will be able to enable/disable the Main conductors but HW will overwrite it (Software can make wrong decisions I think we should not trust in software (I have to to with safety relevant software in my job...), I would do the hardwarecheck for over und undervoltage single fault safe)

If you want, I can make a block diagram how I would go on.
To be honest, I would stay at a raspberry computer with an Can Should, the big advantage I see there is you can easy connect the battery to your home network (home automation, etc.)

 

[bikefish]

What cell brand do you plan to use?

Hi yasko,
'A grade' cells direct from a manufacturer or licensed trader. Decision is not done yet. (CATL, Calb, EVE)
Why?
96S is a long chain and the storage is good as the worst cell is.
I design the ESS for a service time of approx. 15 years. (min 10 years) --> I need reliable cells.
I do a lot of investigations on different cells and I'm not convinced from this cheap blue cells due to lack of specification.

Thanks for sharing your CAN experience.

 

[yasko]

@prepared1: A 2018 Chevy Volt battery sounds like a good deal. It's an 18kWh, and these batteries are well made. But it is large and heavy and no easy to handle, so I took mine disassembled. The modules alone weigh a lot, and you are right about shelves. There was a noticeable deformation when I put them for the first time. I've done something similar to what you suggest, and now it's OK. Thanks for your comments!

@powertrack: I hope you get your project running. You have done a lot of work. I'm a hardware guy too and liked it
It will be interesting to see your block diagram because I'm also thinking about some hardware protection. The Raspi works fine, but I don't want to rely solely on it. I still need to verify how the inverter responds to various alarms from the battery. Here Raspi can help, and I'm planning to do some logging.

@bikefish: Your design concept looks pretty well. Let us know how things are going on. I am curious whether the CAN BMS protocol can be implemented on Orion BMS. Good luck with the project!

 

[powertrack]

@yasko
It seems to me I will switch to the Goodwe solution, have done a lot of testing the last days with the infini, in principle it is/would work but with some compromises which I'm not sure I will go on with them.
My main question regarding the goodwe is:

• Do you use it with an Modbus powermeter to do regulated feed in control (zero feed in)?
• if yes which powermeter do you use?