Sanwizard
Solar Wizard
- Joined
- Feb 2, 2021
- Messages
- 2,714
They are in the business of selling batteries for solar inverters. It is a wise investment to code for the most prevelant inverters being used by your supposed customer base. In order to support that code in the future, you need the hardware/firmware to test with, and more importantly, keep up to date. Since there is currently no standards available in the industry, many have defaulted to the Pylontech protocol for inverter to BMS communications. The issue for Seplos is they support that protocol over CAN bus today, and used the RS485 ports for inter-BMS communications. They need to support the protocol over the RS485 port in order for the Voltronics variants to work properly with their BMS.When software is written, it has to be tested to see if it works correctly, and if the rest of the code still functions correctly.
Communicating between two devices, each has to be stimulated to produce various message. Especially when one isn't their own, so can't trigger the messages or detect response artificially. Then there are timing issues, and how corruption of messages are handled.
I'm surprised they would do this even for 40 units, unless they're hoping there is a market for it. Which they may be.
Software generally costs much more than hardware to design, test, debug, install, repair. Remote updates can be done, but get it wrong and you "brick" the system. Nowadays our cell phones are updated over the air, but we use to take them in to the Sprint store. Cost of labor in the US vs. overseas assembly, and you can imagine the cost difference. 15 minutes of a store clerk's time cost as much as an assembler in China for a day, back then.
Generally, software probably costs 10x as much as the hardware it runs on, but at some high volume it becomes cheap. 20+ years ago, a cell phone contained 100 man-years of code. Once phones reached 100,000,000 units sold of a given model, that doesn't seem so bad. Software development is often assigned to the usual suspects, where an engineer's salary is 1/10th what it is in the US, but having hardware and software on different continents certainly doesn't help get it right. The fact firmware does get updated so often in the PV field tells you something.
A different protocol for battery SoC and charging parameters is just a portion of the code in a BMS, but not trivial.
Are they actually going to program this just based on documentation, or will they purchase LV6548 and try it themselves?
They may need an additional port for that, or update firmware to use the existing port for Inverter comms vs BMS to BMS comms.
I think Seplos may have a shot at becoming a standard in the industry if they make the investment.