• 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'

The R4875G1 units were tested at 40°C ambient temperature and have their MTBF set at 500k hours under those conditions. This figure halves for every 10°C increase above that ambient temperature.

The units are hot-swappable, meaning they slide in and out of server racks that offer little to no additional cooling on the sides—certainly no thermal compound! (In fact, they are stacked right next to each other.) Their bottom plate also has a thick plastic insulator. In other words, the fan is all they have to stay cool.

As you observed, the fan responsible for cooling the internal components (some of which sit directly in the airstream, while others do not) operates very quietly at room temperature, even as the unit appears to get "hot." However, to effectively transfer heat via air cooling at 60°C ambient air temperature, the components need to reach around 85°C. The hotter the unit, the less air needs to be moved to achieve heat transfer.

I have conducted some limited testing of the R4875G1 at a 75A output. The units allow the "output temperature" to approach 80°C before the fan significantly increases its speed. I find that using the maximum fan setting (perhaps something to try when you're home alone) substantially reduces the output temperature, indicating that there is considerable margin built in for higher ambient temperatures. I suspect these units could reliably power a cell tower in the middle of the desert without additional cooling.

That said, many mobile operators control ambient temperature and air quality with air conditioning units and air filters to protect their radios and rectifiers from excessive dust and heat. This is why many used rectifiers appear so clean. In other words, as long as sandstorms are kept out, these units will likely last a lifetime without modifications. This is industrial-grade equipment—not the "designed-to-fail" consumer rubbish we’ve come to expect. I trust its quality.
Thanks for the details and explanations!

The more I learn about these devices, the more impressed I am by them (y)

These devices are proof that it is possible for manufacturers to build devices for moderate prices that still have high quality - this is not very often seen today...
 
Thanks for the details and explanations!

The more I learn about these devices, the more impressed I am by them (y)

These devices are proof that it is possible for manufacturers to build devices for moderate prices that still have high quality - this is not very often seen today...
Some of it is volume, instead of thousands of solar products they are building millions of data center products, and they can afford to do the engineering and testing and revision to get products that run at 85° C all day long.
 
150V max is a little limiting gonna need a bunch of parallel strings.
yep, good for partly shaded installs, 9 panels (3x3) per inverter I would say, I stumbled across it when I was trying to find out if anyone had put an MPPT controller into one of these rectifiers... the drop from 150V to 54V is small - the key to make them 98% efficient I guess.
 
My Emerson works fine with 120V, I would assume others are the same.

If you DIY make sure you use the appropriate sized wire...
Post in thread 'DIY 'Chargenectifier'' https://diysolarforum.com/threads/diy-chargenectifier.56329/post-1134200
Thnks, if I got one it would be one of these preconfigured, I see one with Bluetooth control doesn't get into details would be nice to be able to set current limits so I can run it on a baby 120V inverter gen @15A
as well as flatout on a big boy 240V gen.
 
I know that y'all just touched on this a few posts ago, but how the hell even can they go for that price?

I admit I have not fully read this whole thread, but those are basically ready to go with just AC input wiring and DC output wiring, right?
It sounds like you just tell them what voltage perimeters you want and they send it already set up.


*Edit, that was before I saw the $85 shipping.
But still that's a damn good deal for what they are.
 
I know that y'all just touched on this a few posts ago, but how the hell even can they go for that price?

I admit I have not fully read this whole thread, but those are basically ready to go with just AC input wiring and DC output wiring, right?
It sounds like you just tell them what voltage perimeters you want and they send it already set up.


*Edit, that was before I saw the $85 shipping.
But still that's a damn good deal for what they are.
I'm not sure if the possibility to adjust the volts and amps on the device with LCD display is worth the additional costs. A standard R4875G1 will cost $80 + $49 shipping via eBay from China (total: $129) which is still $100 less compared to $143.30 + $85.73 shipping (total: $229.03). The model from the 2nd link even more expensive.

If you just want to use these units as a LiFePO4 charger I would go for the standard R4875G1 with the additional CAN overhead to implement to program them. In case you would add a 2nd unit, the CAN controller could be the same to control more than one unit which makes it even cheaper.

Usually you set (program) a fix output voltage once (e.g. 51V) which will keep the SOC for a LFP battery at about 20% and protect your battery from undervoltage situations with the help from grid or generator.
You even don't need to program. It's possible to buy a USB-CAN-Bus adapter, plug it into a PC, download a free CAN-Bus tool (e.g. SavvyCAN) and program the output voltage (and amps limit) and the fallback volts and amps and you are good to go. After programming the values, you can unplug the CAN communication and the unit will persist your changes.

In case you are looking for an "adjustable laboratory power supply" in a voltage range from 50-103 (or 48-58.8V for the other model) because of other use cases for a power supply, it may be useful.
 
In the meantime I've change one pair of my DIY chargers to get rid of the "cheap" battery switch and the MEGA fuse and replaced them by a 160A DC breaker (this thing is huge!). Also I've changed my previous 6 AWG wire (from the units to the studs) to 4 AWG and the previous 4 AWG wire (from the studs to the battery bus bars) to 1/0 AWG to get rid of my high temperature problem. Now it should be no problem to use 75A per unit and 150A combined. With all four units together I can feed 300A to my battery bus (if required).

It works nice! Now everything stays cool (I don't know why I've used 6 AWG and 4 AWG wires at all at the beginning - even I'd already planned to use 4 AWG and 1/0 AWG initially and documented it in my original plan... maybe I was a bit sleepy at the time I've ordered the wires :fp2).

Next weekend I plan to change my second pair to do the same changes (it's a lot of work to disassemble, re-crimp all custom wires, etc. but it's worth to stay on the safe side).

Btw. I've enhanced my ESPHome CAN-BUS application in "AUTO" mode with the following rules which is working like hoped:
  • Set the output voltage (and fallback voltage) to 51V (about 20-25% SOC), turn the units "OFF" (hibernate)
  • If the voltage drops below/equals 51V the units turn "ON" with 40A current limiting to help keep the batteries from completely discharging (for sure, it could be set even lower - I still need to adjust and play around with different values...).
  • If the voltage climbs above 51.8V the units turn "OFF" again (fans are off, no noise) this is what usually happens if PV production starts (if enough solar is available every day, the units will be usually always OFF).
  • If no voltage of >=55.5 has been seen for at least 3 days, the units are automatically set to 56V output to start a battery balancing and let the batteries reset the SOC counter (after 3 days without 100% SOC my SOC values are like rolling a dice....). If the voltage reaches 54.4V in this situation, I reduce the current to 20A per unit to give the batteries more time to balance. After the 56V has been reached the units go back to 51V as usual.

with-160A-breaker_resized.jpg
 
I'm not sure if the possibility to adjust the volts and amps on the device with LCD display is worth the additional costs. A standard R4875G1 will cost $80 + $49 shipping via eBay from China (total: $129) which is still $100 less compared to $143.30 + $85.73 shipping (total: $229.03). The model from the 2nd link even more expensive.

If you just want to use these units as a LiFePO4 charger I would go for the standard R4875G1 with the additional CAN overhead to implement to program them. In case you would add a 2nd unit, the CAN controller could be the same to control more than one unit which makes it even cheaper.

Usually you set (program) a fix output voltage once (e.g. 51V) which will keep the SOC for a LFP battery at about 20% and protect your battery from undervoltage situations with the help from grid or generator.
You even don't need to program. It's possible to buy a USB-CAN-Bus adapter, plug it into a PC, download a free CAN-Bus tool (e.g. SavvyCAN) and program the output voltage (and amps limit) and the fallback volts and amps and you are good to go. After programming the values, you can unplug the CAN communication and the unit will persist your changes.

In case you are looking for an "adjustable laboratory power supply" in a voltage range from 50-103 (or 48-58.8V for the other model) because of other use cases for a power supply, it may be useful.
I could see the variable voltage being useful if for example you've got a big storm system coming and you want to top up to full ahead of that, versus more of a 20% low end protection in general day to day operation.
 
I could see the variable voltage being useful if for example you've got a big storm system coming and you want to top up to full ahead of that, versus more of a 20% low end protection in general day to day operation.
That's right, in case you are not want to implement such a feature with a CAN-Bus controller by yourself e.g. with an ESP32 and ESPHome - which is a lot cheaper and you'll have the possibility to connect it to e.g. Home Assistant or OpenHAB, etc. for further controlling/rules, etc.
 
That's right, in case you are not want to implement such a feature with a CAN-Bus controller by yourself e.g. with an ESP32 and ESPHome - which is a lot cheaper and you'll have the possibility to connect it to e.g. Home Assistant or OpenHAB, etc. for further controlling/rules, etc.
yea, and if you really don't want to get involved in automation or coding, 3-5 dollars worth of hardware and this free ESP32 firmware, will give you a web-app to control the R4875G1 unit or any R48xxGx unit from a phone:). Zero coding required, just bring a soldering iron and solder six connections, bingo.
 
yea, and if you really don't want to get involved in automation or coding, 3-5 dollars worth of hardware and this free ESP32 firmware, will give you a web-app to control the R4875G1 unit or any R48xxGx unit from a phone:). Zero coding required, just bring a soldering iron and solder six connections, bingo.
Awesome, I'm not much of a programmer(though I don't really know how in-depth it would be to actually"program" something like this), but I do know how to solder 👌
 
In the meantime I've found an additional way to save a lot of grid energy costs when using CAN Bus controlled rectifiers.

In my situation I need quite often (at least a bit) grid support. Only 2-3 month per year I don't need any grid support (I definitively need more PV panels!).

Because of using the Chargeverters now, I'm not longer using the AC-in and the AIOs are only and always running in inverter mode (Solar/Battery mode - SBU mode but never using the "U").
With the free programmable possibilities of the DIY Chargeverters I can now choose a "Demand Time-of-Use" plan from my energy provider. This plan offers half of the kWh price compared to the "Basic" plan when only using the grid at off-peak hours (¢6.88 instead of ¢12.6 per kWh).

I just need to program the ESPHome to not use the Chargeverters (put them to hibernate) while on-peak time frames are active. To prepare for these on-peak times, the program can use the Chargeverters to fill up the batteries a bit from the grid (if required) before the on-peak time frame starts - the required SOC to make it over the next on-peak time frame could be determined by previous usages depending on the time in the year and time of day.

My POCO has in summer (May-Oct) one on-peak time frame between 3-7pm (4h) and in winter (Nov-Apr) two on-peak time frames between 6-9am and 6-9pm. Weekends and major holidays don't have on-peak times.

I hope I can reduce my grid cost up to 50% while using the Chargeverters with this Demand TOU plan and some programming logic. By just using the AC-in's of the AIOs it would be nearly impossible to implement such a feature (beside all the other problems which I've solved by using the Chargeverters now).

Just requested a plan change online with my POCO - they say it could take up to 60 days to activate the new plan and a change of the meter would be required !!?? Unbelievable!

As soon as I'll have the first results, I'll let you know.
 
In the meantime I've found an additional way to save a lot of grid energy costs when using CAN Bus controlled rectifiers.

In my situation I need quite often (at least a bit) grid support. Only 2-3 month per year I don't need any grid support (I definitively need more PV panels!).

Because of using the Chargeverters now, I'm not longer using the AC-in and the AIOs are only and always running in inverter mode (Solar/Battery mode - SBU mode but never using the "U").
With the free programmable possibilities of the DIY Chargeverters I can now choose a "Demand Time-of-Use" plan from my energy provider. This plan offers half of the kWh price compared to the "Basic" plan when only using the grid at off-peak hours (¢6.88 instead of ¢12.6 per kWh).

I just need to program the ESPHome to not use the Chargeverters (put them to hibernate) while on-peak time frames are active. To prepare for these on-peak times, the program can use the Chargeverters to fill up the batteries a bit from the grid (if required) before the on-peak time frame starts - the required SOC to make it over the next on-peak time frame could be determined by previous usages depending on the time in the year and time of day.

My POCO has in summer (May-Oct) one on-peak time frame between 3-7pm (4h) and in winter (Nov-Apr) two on-peak time frames between 6-9am and 6-9pm. Weekends and major holidays don't have on-peak times.

I hope I can reduce my grid cost up to 50% while using the Chargeverters with this Demand TOU plan and some programming logic. By just using the AC-in's of the AIOs it would be nearly impossible to implement such a feature (beside all the other problems which I've solved by using the Chargeverters now).

Just requested a plan change online with my POCO - they say it could take up to 60 days to activate the new plan and a change of the meter would be required !!?? Unbelievable!

As soon as I'll have the first results, I'll let you know.
I am no I-T guy and tend to gravitate to the 'easiest solution I am capable of' low tech approach.

Like you, I have my inverters with no incoming grid connection at all. The inverters just invert.
I bought two EG4 chargverters, and supply them with grid power via 240 40A relays. The control side of these relays is 120v AC.
I bought two smart plugs and ran the control relay of each EG4 CV to each smart plug with a chunk of old lamp cord with a 2-prong plug end.

The smart plugs are simple to program for the time of day - or many times of day - time slots you are looking for. If you wanted to add in another relay between the smart plug and the 240v CV relay - to control for "need to charge" you could do that too. - ie so the smart plug supplies power to the coil only if the main system is also calling for battery charging.

I didn't go that far, I just play with my phone and decide if I need any charge tonight, and if so how many hours I would like.
 
Some feedback of how the 4 chargeverters work to prevent the batteries from being completely discharged (in winter time I can't get enough PV and need grid support every day).

I've set the chargeverters output to 51V. They are in OFF (hibernate) until <=51V will be reached (~20% SOC while discharging) and then they turn ON (at around 9pm in the evening) with 40A max. per unit current limit (8,160 Watt max.). SOC drops over night and approach and never fall under 8% SOC. They will turn OFF if the voltage jumps over 51.8V (at around 10:30am the next day at about 10% SOC while PV is re-charging).

The cyclic peaks in the charts are the central 5 ton heat pump consumption spikes.

Here are some SA diagrams over the last 3 days:
1738333900821.png
1738332926228.png

And the corresponding diagram for the provided chargeverters power (all 4 added):
1738333050161.png
The peak grid usage is between 5-8am because the chargeverters needs to provide the complete house load to prevent the batteries from draining even further (prevent them from dropping below 51V).


I'm happy about how accurate the load is shared between all 4 chargeverters even if the first pair is connected to a different battery rack compared to the second pair which feeds the other battery rack (equal length wires are a must have!).

Here is a actual snapshot from my temporary OpenHAB dashboard which shows all chargeverter data (btw. I've reset the energy counters yesterday, so the energy values in the picture are not for the last 3 days, just since yesterday evening):
1738334224765.png
This setup provide a smooth mixing of grid and battery energy while preventing the batteries to drop below 8% SOC.
It looks like (at least for me with 16S batteries) the 51V charging voltage is the sweet spot for my situation.
 

diy solar

diy solar
Back
Top