• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

DIY 'Chargenectifier'

Yeah, they're super loud data center style fans. Only reason they don't both me is that mine only run when my generator is also on and making a racket.

I don't know if there's a way to make this setting permanent, but you can switch the fan between auto and always being on full blast;

Python:
fan_auto = True
send_can_message(channel, [0x03, 0xF0, 0x00, 0x33, 0x00, 0x01 if fan_auto else 0x00, 0x00, 0x00])
 
Yeah, they're super loud data center style fans. Only reason they don't both me is that mine only run when my generator is also on and making a racket.

I don't know if there's a way to make this setting permanent, but you can switch the fan between auto and always being on full blast;

Python:
fan_auto = True
send_can_message(channel, [0x03, 0xF0, 0x00, 0x33, 0x00, 0x01 if fan_auto else 0x00, 0x00, 0x00])
this is what I get with the same app when the fans are on full blast, same distance (20cm)
 

Attachments

  • Screenshot_2025-01-10-21-54-12-361_mobacorn.com.decibelmeter.jpg
    Screenshot_2025-01-10-21-54-12-361_mobacorn.com.decibelmeter.jpg
    47.2 KB · Views: 10
65dB is not exactly quiet though. Doesn't help it's a high pitch noise. Maybe some people are just less sensitive to that.
65dB is super loud they are in fact screaming, if they would do this under normal operation I would get a divorce. As I said, they run at below 30dB in normal auto-fan mode at 30A load each.
 
This "beasts" are unbelievable loud! 🤯
7.5Ω resistance connected to 2 units with 52VDC for low current testing (about 2 x 3.5A = 7A total to test if low current limits are working).

this might be the reason, 3.5A is a very unusual current for them, keep in mind the efficiency curve of these units: 30A @53v is 1600Wx0.03=50W of heat, not much. The units get to 70C to reject the heat, so you need virtually no air movement to reject 50W.

Screenshot 2025-01-10 225603.png
 
Last edited:
I've nearly finalized the installation for now. Here are some of the software points, I'd like to mention because of my findings:
  • At first, I misunderstood that the current scaling factor (maxUnitAmps/1024 which is 13.65 for my 75A unit) is only valid for setting/reading the amps limit setting. The factor for the sensor of the actual used amps is just 1024 like for voltages.
  • The amp limit setting and the actual amp sensors are very accurate (checked with a DC clamp meter with 2.5% accuracy)
  • The DC power out (0x73) measurement is NOT very accurate and stays always 0 for power <200W-300W. So I've implemented an own power sensor mechanism which gets triggered by any change of DC amps or DC voltages and calculate and update the DC power out sensor - this is way more accurate.
  • On top of the DC power out triggers I've implemented an energy counter which uses the uptime sensor in seconds (type: seconds, update_interval: 1s). I store the previous values of DC volt and DC amp (per unit) and calculate the energy in "Ws" for this last time frame with the last values of volts and amps. I sum all values up in a (with restore_value: true to make it persistent) global in "kWh". A reset button is able to reset the energy counters to zero. It's nice accurate!
  • If I set "wrong" values for the amps limit (like I did at the beginning of my tests by mistake), the cooling fans in the rectifiers speed up to highest speed as soon as a load (even a small) is connected (maybe a protection mechanism in the units if wrong/illegal CAN values are sent to the unit)! Now while I set correct values for amps limit the cooling fans are only spinning speed depending on the temperature (load?) like they should. Now, they are not very loud for power <300W.
Regarding my complains about noise, it was my own fault! After correcting the software bug (see red section above), the cooling fans are running very slow (and quiet) with low load (<=180W per unit). I think they will speed up for sure for higher loads, but I have to check if I need additional noise reduction if the batteries will be finally connected (I hope I can finish everything this weekend).

Update:
Btw., does your fans speed up if the current limit kicks in (even with low load)?
 
Last edited:
Finally my DIY Chargenectifier project is finished.

Pics: A pair on each battery rack / left pair with CAN controller / right pair / left AC grid feed (same for the right side)

20250112_090542_resized.jpg 20250112_080728_resized_1.jpg 20250112_080737_resized_1.jpg
20250112_081159_resized_1.jpg

Some more information:
  • The fans are only speed up to max. speed if the amp limit kicks in AND/WHILE the voltage has to be reduced very much to hold the amp limit, even with low load. I've tested this with a dummy load on DC out (no battery). Now while the batteries are connected to DC out and the voltage is set between 50-56 volt the fans are NOT speed up if the amp limit kicks in!
  • With 1kW per unit, the fans are running slow and quiet. With an garage ambient of 20°C (68°F) the units heat up to 58-61°C (136-142°F) but the fans are still running slow 😀 I still have to do tests with higher power.
  • I left all AC-in circuit of the AIOs connected but switched the breakers for them off - just in case I need them for service, etc.
  • update: The efficiency of the units are 97% with 1kW 😀
  • update: The values of Solar Assistant are no longer sufficient to describe the entire system. The battery charging values are now only the PV charging values but the charge from the rectifiers is missing in SA (beside that the grid usage showing always 0). I need to combine the rectifier data together with the SA data into the smart home system to get a complete overview of the solar system.

I'm pretty happy with the results so far. Now I have to integrate the controlling of the rectifiers into my smart home (OpenHAB) rules and delete the complicated AC-in rules which I had implemented earlier (see my build thread).

With this online double conversion, I don't need to fear possible grid feedback problems and the inverters never need again to use the bypass relays.
 
Last edited:
I second that! what a build, looks very tidy. So glad you did not have a noise issue after all.
 
I second that! what a build, looks very tidy. So glad you did not have a noise issue after all.
Thanks! I've already started to plan a fan modding by using an 80x80x25 fan outside of the unit and 3D print a pipe to feed the air into the 40x40 original fan opening... luckily it seems it's not required! I have to find out about the noise level when the units are running with than 1kW each - we'll see.
 
Thanks! I've already started to plan a fan modding by using an 80x80x25 fan outside of the unit and 3D print a pipe to feed the air into the 40x40 original fan opening... luckily it seems it's not required! I have to find out about the noise level when the units are running with more than 1kW each - we'll see.
 
Finally my DIY Chargenectifier project is finished.

Pics: A pair on each battery rack / left pair with CAN controller / right pair / left AC grid feed (same for the right side)

View attachment 269596 View attachment 269598 View attachment 269599
View attachment 269600

Some more information:
  • The fans are only speed up to max. speed if the amp limit kicks in AND/WHILE the voltage has to be reduced very much to hold the amp limit, even with low load. I've tested this with a dummy load on DC out (no battery). Now while the batteries are connected to DC out and the voltage is set between 50-56 volt the fans are NOT speed up if the amp limit kicks in!
  • With 1kW per unit, the fans are running slow and quiet. With an garage ambient of 20°C (68°F) the units heat up to 58-61°C (136-142°F) but the fans are still running slow 😀 I still have to do tests with higher power.
  • I left all AC-in circuit connected but switched the breakers for them off - just in case I need them for service, etc.
  • update: The efficiency of the units are 97% with 1kW 😀
  • update: The values of Solar Assistant are no longer sufficient to describe the entire system. The battery charging values are now only the PV charging values but the charge from the rectifiers are missing in SA (beside that the grid usage showing always 0). I need combine the rectifier data together with the SA data into the smart home system to get a complete overview of the solar system.

I'm pretty happy with the results so far. Now I have to integrate the controlling of the rectifiers into my smart home (OpenHAB) rules and delete the complicated AC-in rules which I had implemented earlier (see my build thread).

With this online double conversion, I don't need to fear possible grid feedback problems and the inverters never need again to use the bypass relays.
Spectacular!
 
Some more positive updates about the 4x R4875G1 rectifiers:
  • The cooling fans are very quiet up to 40A current. Between 40A and 50A the fans are speeding up a bit but still acceptable noise level (even after longer running at about 20°C ambient temperature). Above 50A they are really loud. Luckily I usually will use them with <=40A (2kW) because 8kW with 4 units is enough in 99% of all my use cases - so no requirement for fan modding so far :)
  • It looks like the fan speed depends on the current load limit setting (and maybe on the current actual load - but still not sure). I still don't know if they are additionally temperature controlled (I guess they are).
  • I'm very happy that the units are reporting the actual DC output voltages (battery voltages) via CAN bus, even if the units are OFF (hibernate). This enabled me to program an automatic charging behavior just with the ESP32 itself (no external voltage checking rules via smart home required) (y)
    • Usually the units are OFF (hibernate)
    • I've defined 5 working modes
      • mode 0: automatic mode (this is the default) which sets the DC output voltage to 52V and the current limit to 20A. If the DC output voltage drops to or below 51.5V the rectifiers are switched to ON to charge the batteries to 52V (with max. current of 4x 20A). If the 52V has been reached, they switch OFF again. This rule prevents deep discharge and keep the batteries around 20% SOC.
      • mode 1: "4kW mode" - Set 56.3V DC output voltage with 20A limit per unit (80A - 4kW charging)
      • mode 2: "8kW mode" - Set 56.3V DC output voltage with 40A limit per unit (160A - 8kW charging)
      • mode 3: "12kW mode" - Set 56.3V DC output voltage with 60A limit per unit (240A - 12kW charging)
      • mode 4: "16kW mode" - Set 56.3V DC output voltage with 75A limit per unit (300A - 16kW charging)
    • The modes 1-4 can turned on manually via smart home buttons and I plan to add 5 physical buttons (with status LEDs) mounted on the CAN controller case to be able to manually select the modes 0-4 when standing in front of the chargers without having a phone or laptop.
    • If the units are running in manual modes 1-4 and 56V has been reached, the default mode 0 (automatic) gets activated and the units switch to OFF again.
  • In OFF (hibernate) the fans standing completely still (silent).
  • Even with higher output power, the efficiency seems always to be >97% - way better compared to the integrated chargers in the EG4-6500EX AIOs.
  • The voltage and current measurements are very accurate over the whole range (power and energy is calculated by myself in the code based on these measures).
  • The MQTT connection via ESP32 to my smart home system (OpenHAB) works as expected without issues.
I can really recommend these rectifiers - I'm impressed.
 
Some more positive updates about the 4x R4875G1 rectifiers:
  • The cooling fans are very quiet up to 40A current. Between 40A and 50A the fans are speeding up a bit but still acceptable noise level (even after longer running at about 20°C ambient temperature). Above 50A they are really loud. Luckily I usually will use them with <=40A (2kW) because 8kW with 4 units is enough in 99% of all my use cases - so no requirement for fan modding so far :)
  • It looks like the fan speed depends on the current load limit setting (and maybe on the current actual load - but still not sure). I still don't know if they are additionally temperature controlled (I guess they are).
  • I'm very happy that the units are reporting the actual DC output voltages (battery voltages) via CAN bus, even if the units are OFF (hibernate). This enabled me to program an automatic charging behavior just with the ESP32 itself (no external voltage checking rules via smart home required) (y)
    • Usually the units are OFF (hibernate)
    • I've defined 5 working modes
      • mode 0: automatic mode (this is the default) which sets the DC output voltage to 52V and the current limit to 20A. If the DC output voltage drops to or below 51.5V the rectifiers are switched to ON to charge the batteries to 52V (with max. current of 4x 20A). If the 52V has been reached, they switch OFF again. This rule prevents deep discharge and keep the batteries around 20% SOC.
      • mode 1: "4kW mode" - Set 56.3V DC output voltage with 20A limit per unit (80A - 4kW charging)
      • mode 2: "8kW mode" - Set 56.3V DC output voltage with 40A limit per unit (160A - 8kW charging)
      • mode 3: "12kW mode" - Set 56.3V DC output voltage with 60A limit per unit (240A - 12kW charging)
      • mode 4: "16kW mode" - Set 56.3V DC output voltage with 75A limit per unit (300A - 16kW charging)
    • The modes 1-4 can turned on manually via smart home buttons and I plan to add 5 physical buttons (with status LEDs) mounted on the CAN controller case to be able to manually select the modes 0-4 when standing in front of the chargers without having a phone or laptop.
    • If the units are running in manual modes 1-4 and 56V has been reached, the default mode 0 (automatic) gets activated and the units switch to OFF again.
  • In OFF (hibernate) the fans standing completely still (silent).
  • Even with higher output power, the efficiency seems always to be >97% - way better compared to the integrated chargers in the EG4-6500EX AIOs.
  • The voltage and current measurements are very accurate over the whole range (power and energy is calculated by myself in the code based on these measures).
  • The MQTT connection via ESP32 to my smart home system (OpenHAB) works as expected without issues.
I can really recommend these rectifiers - I'm impressed.
Nice, you've got the makings of a good product there, a case and a display/UI and you've got a real competitor to the CVGC. What's the BOM cost for one pair (just a guess is OK)?
 
Nice, you've got the makings of a good product there, a case and a display/UI and you've got a real competitor to the CVGC. What's the BOM cost for one pair (just a guess is OK)?
BOM for 4 units (2 pairs):

Code:
 $519.00 - 4x R4875G1 (4 units incl. $135 shipping and incl. $44 for remote loaction DHL addon)
 $ 51.48 - 4x PCB connector for R4875G1
 $ 30.99 - 5x ESP32 ESP-WROOM-32 board
 $  7.99 - 3x CAN bus tranceiver module (SN65HVD230)
 $ 29.96 - 2x 40A double pole Siemens breakers
 $ 59.98 - 2x Nema SS 2-50R 50Amp Receptacle
 $ 39.98 - 2x 50 Amp Twist Lock Plug
 $ 11.99 - 1x DC 12V 24V 48V to 5V Step Down Converter
 $ 12.99 - 1x (pack of two with 200A fuses) MEGA/AMG Fuse Holder 200A
 $  7.49 - 1x 22 AWG JST SM 2 Pin Plug Male and Female Connector (20 pairs)
 $ 33.98 - 2x pair 3/8" (red & black) Heavy Duty Dual Studs Battery Junction Post Terminal Kit
 $ 63.99 - 25ft Welding Cable 8 AWG 40A (I've cut the plugs - only needed the cable)
 $ 20.99 - 1x (pack of two) Battery Disconnect Switch 275A
 $ 12.99 - 1x USB 3.1 Male Connector with Type C Housing set
 $  9.99 - 1x (pack of two) Junction Box
 $ 13.89 - 66ft flexible 22 AWG Electrical Wire (CAN bus)
 $ 11.98 - 10Pcs 6 Awg Splicing Connector (to connect the 8 AWG 240V to 2x 12 AWG inside the junction box)
 $ 88.00 - THHN wire 8 AWG (red, black, white, green each 25ft)
 $ 40.00 - 4 AWG and 6 AWG copper lugs with 3/8 and 5/16 holes
 $ 70.00 - 20 ft 4 AWG and 20 ft 6 AWG battery cable red & black (e.g. WindyNation)
 $ 20.00 - couple of wire terminals 6.3mm for PCB connector and some ring terminals
 $ 40.00 - wood, screws, zip ties, 3D filament, some resistors, LED's, PCB prototyping board, shrink tubing, etc.
======

About $1,200 for 4 units.

Sources:
R4875G1 from ebay (offer for 4 units)
PCB connectors for R4875G1
ESP32 board
CAN bus tranceiver module
40A Siemens double pole breakers
Nema SS 2-50R 50Amp Receptacle
50 Amp Twist Lock Plug
DC 12V 24V 48V to 5V Step Down Converter
MEGA/AMG Fuse Holder 200A
22 AWG JST SM 2 Pin Plug Male and Female Connector
3/8" Heavy Duty Dual Studs Battery Junction Post Terminal Kit
Welding Cable 25ft 8 AWG 40A
Battery Disconnect Switch
USB 3.1 Male Connector with Type C Housing
Junction Box
22 AWG Electrical Wire
6 Awg Splicing Connector
PCB prototype board 5cm x 7cm
 
Last edited:
BOM for 4 units (2 pairs):

Code:
 $519.00 - 4x R4875G1 (4 units incl. $135 shipping and incl. $44 for remote loaction DHL addon)
 $ 51.48 - 4x PCB connector for R4875G1
 $ 30.99 - 5x ESP32 ESP-WROOM-32 board
 $  7.99 - 3x CAN bus tranceiver module (SN65HVD230)
 $ 29.96 - 2x 40A double pole Siemens breakers
 $ 59.98 - 2x Nema SS 2-50R 50Amp Receptacle
 $ 39.98 - 2x 50 Amp Twist Lock Plug
 $ 11.99 - 1x DC 12V 24V 48V to 5V Step Down Converter
 $ 12.99 - 1x (pack of two with 200A fuses) MEGA/AMG Fuse Holder 200A
 $  7.49 - 1x 22 AWG JST SM 2 Pin Plug Male and Female Connector (20 pairs)
 $ 33.98 - 3/8" 2x pair (red & black) Heavy Duty Dual Studs Battery Junction Post Terminal Kit
 $ 63.99 - 25ft Welding Cable 8 AWG 40A (I've cut the plugs - only needed the cable)
 $ 20.99 - 1x (pack of two) Battery Disconnect Switch 275A
 $ 12.99 - 1x USB 3.1 Male Connector with Type C Housing set
 $  9.99 - 1x (pack of two) Junction Box
 $ 13.89 - 66ft flexible 22 AWG Electrical Wire (CAN bus)
 $ 11.98 - 10Pcs 6 Awg Splicing Connector (to connect the 8 AWG 240V to 2x 12 AWG inside the junction box)
 $ 88.00 - THHN wire 8 AWG (red, black, white, green each 25ft)
 $ 40.00 - 4 AWG and 6 AWG copper lugs with 3/8 and 5/16 holes
 $ 70.00 - 20 ft 4 AWG and 20 ft 6 AWG battery cable red & black (e.g. WindyNation)
 $ 20.00 - couple of wire terminals 6.3mm for PCB connector and some ring terminals
 $ 40.00 - wood, screws, zip ties, 3D filament, some resistors, LED's, PCB prototyping board, shrink tubing, etc.
======

About $1,200 for 4 units.

Sources:
R4875G1 from ebay (offer for 4 units)
PCB connectors for R4875G1
ESP32 board
CAN bus tranceiver module
40A Siemens double pole breakers
Nema SS 2-50R 50Amp Receptacle
50 Amp Twist Lock Plug
DC 12V 24V 48V to 5V Step Down Converter
MEGA/AMG Fuse Holder 200A
22 AWG JST SM 2 Pin Plug Male and Female Connector
3/8" Heavy Duty Dual Studs Battery Junction Post Terminal Kit
Welding Cable 25ft 8 AWG 40A
Battery Disconnect Switch
USB 3.1 Male Connector with Type C Housing
Junction Box
22 AWG Electrical Wire
6 Awg Splicing Connector
Amazing informative post.
I edited my 1st post and added a link for this post.


If anyone else wants posts added let me know the number.
I really should try and index this a bit and at least add links for @upnorthandpersonal 's code.
 
Amazing informative post.
I edited my 1st post and added a link for this post.


If anyone else wants posts added let me know the number.
I really should try and index this a bit and at least add links for @upnorthandpersonal 's code.
If I'll find time, I think about putting the code on github. The maintainer of the original code made the suggestion to create a new repository with my code for 4 units (without the help and the amazing research work of this guy and others in the github discussion this project would not have been successful!). I like more and more the ESPHome environment (even if the declarative programming style with YAML is still a bit unusual for me as a "typical fun functional programmer").

First I want to extend my circuit to add 5 manual buttons with LEDs on the lid to be able to select the charging modes manually with physical buttons on the units itself and LEDs which indicates which mode is currently active...
 
BOM for 4 units (2 pairs):

I just found this repository where somebody has made a general purpose ESP32 board with built-in CAN and a wide range voltage converter. https://github.com/MagnusThome/RejsaCAN-ESP32

It's for cars so you might still need your voltage converter or need to mod the board design so the onboard converter can accept 58+v instead of just 28-ish volts.
 
Nice! If you change the holes in the sides to diamonds they shouldn't need supports when printing.
You're right, but I print circular small holes without support with good enough results. The rear brackets printing orientation is grille on printing bed, so the form of the grille doesn't matter. But in general you're right, a diamond shape would be easier to print. I didn't invest any more time in the design than was absolutely necessary...
If someone is interested in enhancing the design, I can provide the OpenSCAD files also.
 

diy solar

diy solar
Back
Top