diy solar

diy solar

NEWS: New Midnite Solar DIY Series - SCC's & Inverters (great $ too)

Pick one protocol on CANbus and support that. Then offer an interface module to translate protocols.
You do that to connect Midnight Classic to Sunny Island.
Firmware options or changes are another way to go, but simple plug-together hardware is easier and worth the cost for someone installing a single unit.
Software compatibility issues and updates are something I hate dealing with. I don't want any good equipment to go obsolete because not supported by software anymore.
 
Supporting the SMA Sunny Island is a bit different here as we were selling SI's and SI E-Panels as complete wired systems.
One engineer took one full year developing that CANBUS <---> MODBUS converter system and even then, SMA did not acknowledge this.
I hear that MidNite solar maybe 2 dozen of those boxes. Unfortunately we cannot just willy nilly spend a lot of engineering time on something that may not be profitable or useful.

We could start with just one BMS canbus protocol but would prefer to work with the most popular one first.


RVC is the most popular canbus protocol for the RV industry by far but may not have lithium battery charging included ?
Not exactly sure which BMS protocol is most popular for R.E.

There are also lots of MODBUS BMS protocols as well. Maybe more modbus than canbus at this time ?

If you know of any resources for this type of development, please let us know. Engineers are very hard to find which is why we spend too much time of product development as it is. It's not easy especially these days to find good help.

I sure do like some of the hobbyists work that has been done on some of these smart-phone apps though !
 
Last edited:
As far as supporting MANY different CANBUS BMS's, (Steve_S) I don't think that they all use a "standard" for making them work.
Victron and others publish their stack information. I don't think there is a "Standard" as such. JK, JBD and other BMS makers have Canbus Enabled products that a lot of folks can use. I would imagine the companies which produce these ESS Telecom Boxes would also either make available their stack info or work with a Company like Midnite happily.

The alternative is to go the route of Rolls Surette who use a Xanbus comm port provides plug and play integration with SE Conext XW+, SW, SCP, ComBox and Solar Charge Controllers.

Ideally, like with the Classics, the ability to access the data either with the provided software or in the Classics case ModBus.

This may also be helpful, I dunno.
 
Kind of sounds like a job for a Raspberry Pi with a canbus hat and hobbyists (DIYers) ?

There are a some of those people out there, like yourselves that aren't afraid to get their hands dirty.

Maybe if we were to work with just one of these BMS protocols to start with, then that might make a good example where others could add to the list of protocols in open source style.

In fact, something like this would be in the spirit of DIY which is why we are all here
 
SMA dropped DC coupling (was MSTE's product rebranded by them), also wind and hydro.
I'm happy with the AC coupled systems, like how they let battery charge current be regulated separately.
DC coupled can do that as well, but only some systems have that feature.
AC coupling adjusts slowly, a few seconds to command different power output. Battery has to source/sink the difference.
Not a problem for my AGM, but lithium would have to be kept below 100% SoC to accept the charge current.

I can appreciate the difficulty of getting firmware to work correctly.
I used to write self-test microcode and point-use CAD tools, but no released software products.
I like being able to wire alternative energy equipment together, would like analog interfaces for zero software.
The older SMA family of equipment has worked for me, but I've told them about bugs and they had to update compatibility chart.
New models, of course they're using more networking. I'm not using those, don't know if any issues or not.
I also use old GPIB instruments at home and past jobs. The new USB and Ethernet ones I use at work cause me more grief.

I picked up some Sunny Island Charger and bought the MSTE interface from it to Sunny Island, so I haven't used the Midnight interface to classic. I like having a 4x enclosure. Your higher wattage would be better for many applications.

I think communication with a BMS will make a better system (although I'm not using lithium at this time.)
We understand it is useful to slow charging so no cells trigger over-voltage shutdown, and leave time for balancing.
It appears if charging voltage is set at a moderate level, this can be accomplished without communication, if cells remain reasonably balanced.

But the suggestion I gave you is something I haven't seen addressed anywhere (probably is handled by the auto makers.)
We understand that lithium battery capacity will be degraded, cells eventually destroyed, if charged at excessive current while cold.
Charge controller puts out some current. Battery gets all of that if not drawn by inverter.
Some BMS provide low-temperature disconnect to prevent charging.
If somebody sets up a system with 0.5C charge capability and zero degrees C disconnect, I think they're going to wreck their battery in cold weather.
That's where you can help, without even speaking "BMS". Just offer programmable current vs. temperature.
You also have Whiz-Bang Jr., but I haven't read in your documentation what is done with the current reading. It should be used to adjust charge controller output to deliver desired charge current, while providing more current when consumed by loads such as inverter.
That may be able to respond very fast in your DC coupled system, less likely to cause BMS to trip for over-voltage.
 
All good stuff. Let's just use GPIB instead of all the new-fangled stuff !:) I have a USB to HPIB/GPIB converter for one of my old HP analyzers.

I actually saw one of those MSTE chargers several years ago. I can see why they wanted the see the Classic used instead at least in the US.

Regarding the current limit vs. lithium battery temperature, it was brought up here that you won't have access to the batteries for that. There should be a current limit specified for the batteries and their BMS also which can be done now. Maybe some Li batteries have communications that allow for variable limits and fast charging. That by itself would be a good reason for comms I suppose.

Best thing for lithium battery communications would be for those manufacturers to write the code themselves. I do kind of like the idea of an open source hardware and hardware hat for the rPi though. MidNite will be publishing our protocol so those DIYers and battery BMS manufacturers can use it.

The WB Jr. is used for battery SOC% and maximum battery current limit information.
 
thanks for committing to publishing the spec

it increases the trustworthiness rating!

and engenders respect from those who are “hands on”

??

it’s a burden on every company, but an open standard for communicating battery variables and needs based on those variables, would profoundly increase safety and efficiency across the world

cheers!
 
The Midnite LiFePO4 B17 modules should be ready for some testing in the next 2-3 weeks. Looking forward to playing around with them. 96 of the 26650 cells per module...:)
 

Attachments

  • b17batt2.jpg
    b17batt2.jpg
    290.1 KB · Views: 27
  • b17batt3.jpg
    b17batt3.jpg
    317.7 KB · Views: 27
  • B17BatteryWeight1.jpg
    B17BatteryWeight1.jpg
    213.3 KB · Views: 27
Regarding the current limit vs. lithium battery temperature, it was brought up here that you won't have access to the batteries for that. There should be a current limit specified for the batteries and their BMS also which can be done now. Maybe some Li batteries have communications that allow for variable limits and fast charging. That by itself would be a good reason for comms I suppose.

Best thing for lithium battery communications would be for those manufacturers to write the code themselves. I do kind of like the idea of an open source hardware and hardware hat for the rPi though. MidNite will be publishing our protocol so those DIYers and battery BMS manufacturers can use it.

The WB Jr. is used for battery SOC% and maximum battery current limit information.

BMS has no ability to control charge current, unless with communications. All it can do is disconnect if unhappy with voltage, current, temperature.

What you could offer in firmware of your Midnight Classic or other SCC is to read the temperature sensor and roll off current output from programmed maximum to zero at some upper and lower temperature limits. That should support lithium across a range of temperature.

So WB Jr. tracks (or lets you track) SOC, and it reports battery current. Does your charge controller use that to limit current?
For instance, I have one AGM bank that is 104 Ah 48V. It wants 20A charge current. Could I used the Classic 150 and WB Jr. to deliver up to 96A to DC bus (if inverter draws 76A or more), while maintaining battery charge current at 20A?

With temperature sensor and WB Jr, we could configure for a 100 Ah LiFePO4 battery so it receives 50A charge while at 25 degrees C, tapering down to 0A at 0 degrees C and some high temperature.

Right now I think the only thing we can do is set SCC for a charge current (say 30A) and set BMS to disconnect at some temperature below which 30A would be excessive (say 15 degrees C.)
 
The WB Jr. and Classic does allow battery charge current limiting, yes. Charging current over the battery limit causes it/them to charge less so that amount isn't exceeded.

Charge current proportional to temperature I'm sure would be great but the BMS would be the best place I think for sensing battery temperature...

After all, the BMS has access to the battery's internal temperature more accurately than a sensor stuck onto the side of that box I would think.

The BMS can just communicate that in some way to the rest of the system. As good as they want to.. The BMS or and aggregate of BMS current info can become the WB Jr. too in some cases if that is built into their design
 
@JoeHam Good article. I just sent the link to my stepson who has just bought an off grid property.

Here is another excellent read :


Much longer but well worth the time.
 
Charge current proportional to temperature I'm sure would be great but the BMS would be the best place I think for sensing battery temperature...

After all, the BMS has access to the battery's internal temperature more accurately than a sensor stuck onto the side of that box I would think.

I think you're brushing off a great opportunity to provide a feature that protects batteries and enhances performance, which I think no-one else offers.
Sure, use temperature (or target current) from BMS if communication is available. But offer settings based on your own temperature sensor too. You can do that right now, and it would perform very well for anyone who places the temperature sensor in contact with a cell.

The option people have today, setting max current and setting low-temperature disconnect, aren't sufficient imho.
I expect batteries to be prematurely degraded unless those are set at some single point, which prevents colder charging that would have been OK at low current, and higher current charging that would have been OK at moderate temperature.

I think many users will get damaged batteries in cold weather. My recommendation has been to limit charging to 0.15C, and set temperature accordingly. But that limits power production if they don't have a shunt and SCC which can control charge current.
 

i use the MCP9808 in my projects lately because its in stock, it’s cheap at 5usd, digital, factory calibrated, accuracy ±0.25°C over the sensor's -40°C to +125°C range and precision of +0.0625°C.

5 dollars for this level of nuance!

anyways, i will be using it to monitor each individual cell in my builds, most likely affixed to the side of the cell and insulated over. (DIY cell build so access is easy)

i’m only an amateur so i could be off base but, since i don’t know of a manufacturer that allows me to load a temperature vs current curve into their device to treat my battery right, i’m just gonna build it myself :)

even a small 20 point Look Up Table that the user can enter in twenty XY pairs. X=temperature. Y=allowedCurrent.

it might be wrong for the company but it’s definitely easy to implement in software in my humble opinion.

people say “never charge LFP while freezing” but it’s more like “don’t try to charge LFP at full rate while freezing”. there are even papers that observe longer cell life when cycled at -5 degrees celsius.


1634758014537.png

just some random thoughts from a diy amateur!

good luck!
 
TL;DR

the electrolyte and materials inside LiFePO4 cells degrade faster at higher temperature and current density therefore a thermometer and a look up table provided by user allow proper management of cell degradation and longer cell life.

to limit charge current based on max non damaging rate according to battery brand
 
What you could offer in firmware of your Midnight Classic or other SCC is to read the temperature sensor and roll off current output from programmed maximum to zero at some upper and lower temperature limits. That should support lithium across a range of temperature.
[…]
With temperature sensor and WB Jr, we could configure for a 100 Ah LiFePO4 battery so it receives 50A charge while at 25 degrees C, tapering down to 0A at 0 degrees C and some high temperature.

Right now I think the only thing we can do is set SCC for a charge current (say 30A) and set BMS to disconnect at some temperature below which 30A would be excessive (say 15 degrees C.)

High and Low Taper Off sound good too.

Either Look Up Table or High Low Taper Off would profoundly increase the value to me.

For now I will continue look into communicating with my existing MPPT using its HEX protocol on UART wired data port. By measuring temperature myself, I can simply change the MPPT’s current limit parameter on the fly. ?? arduino is the best.

Publicly documented data port interfaces allow value maximization!
 
And of course a company like Midnite has to weigh how much demand there might be for each added “feature”.

If enough folks don’t want it you’re just driving up cost which is a common spec everybody looks at.

Case in point, see Will’s YT video from yesterday where he explains how to easily have a Growatt AIO communicate with an EG4 battery.

And then he admits he will never use that feature himself.

Same here for me with the same battery. I’ve hooked up a computer to look at cell level data but have no other use for the communication. The BMS and battery must fend for themselves in my world.

Diffrent strokes for diffrent folks and all that. ?
 
Might I add in here to the Midnite folks (yes I have a classic 150 and E-Panel for an SW and you guys have helped me out immensely on the Midnite forum)... But might I also suggest an E-Panel for the Growatt 12kw. I was Going to buy another SW E-Panel and modify it for the Growatt as they are almost exactly the same size (see below) , but the 250a breaker is to small. Midnite needs an E-Panel for the Growatt 12kw with either the appropriate breaker or a battery switch. ?

Growatt - Epanel 1.JPG
 
Last edited:
Just stumbled on this thread (too many threads to follow). Nothing useful to contribute here, but wanted to jump in and thank boB (@SpongeboB Sinewave) and Robin (@Robin Gudgel) for jumping in here. You guys are legendary in the solar community. Thanks for everything you have done at Outback and Midnite. I know whatever comes next under your banner will be good.
 
Just an update .
I friend of mine decided to gamble with a midnite diy charge controller.
Upon unboxing I must admit it seems well built. In a week time I would share on its performance.
I wouldn't be cracking it open to seem what's inside because of warranty. But I do hope the warranty couldn't least be 2 years on all the midnite diy stuff.
And yes I support the idea of Will testing at least one of midnite diy inverter.
Well an update on the midnite diy charge controller.
As i mentioned the unit seems solid.
Installation was easy compared to even a classic. It has nice space for wires with good quality brass studs.
Performance and accuracy when compared to a fluke meter both in Amps and voltage were good.
User interface for me needs improvement.Its a bit difficult as compared to a classic or other charge controllers in the same price and probably quality range.
Buttons were my major dislike .They are flimzy and responds a bit too slow or too fast.They were also noisy when making changes and adjustments.
Overall you can't complain for that price but I still do recommend at least 2years warranty
 
Back
Top