Rectifier Huawei R4850G2 48V 42~58V 3000w

[560SEC]

[/quote]

2. The computer serial port debugging tools, see figure set parameter writing CAN control board (USB serial port module.)
[/quote]

Hi,

Where did you get the can controllers?

 

[dominik h]

Do you know how the R24/4850 is switched between 24V and 48V or is this done via CAN?
I'm still looking for a 24V rectifier.

I can answer my question now. The voltage change is done with the two port wires, the wires left side next to the can

 

[Michael20]

Are there revisions of R4850 or r4875 which incompatible with this mod? Is it ok, to always connect the Huawei to the battery?

Anyone optimized the Fan to cool more and be more silent?

 

[TIOK]

Are there revisions of R4850 or r4875 which incompatible with this mod?

Which mod do you mean?

Is it ok, to always connect the Huawei to the battery?

How to understand your question? These units are designed for charging batteries.

 

[Michael20]

Are there revisions of R4850 or r4875 which incompatible with this mod?

Which mod do you mean?
Sorry the CAN steering.

Is it ok, to always connect the Huawei to the battery?

How to understand your question? These units are designed for charging batteries.

Want to charge a 48V battery, when enough sun(pv) is here, when sun is weak Stop charging(maybe switch off AC?) till next day, when sun is strong enough start charging again.

So is it ok that this Rectifier is connected always to battery? And also a pv charger parallel to the Huawei?

 

[dominik h]

They are designed for charging batteries and working in paralell with others of them. As long as you do not exceed it's voltage limit it will work and survive.

 

[snejky]

Hey guys,

Currently looking for a power supply for charging my 20s/72v battery through one of my Nucular controllers. Sort of set on the Eltek Flatpack2 but haven't made a final decision. There was a brief discussion (with very little details) on the Nucular telegram chat a while ago where someone tried using this Huawei R4850G2 and was having issues. Does or has anyone used or know for sure if these units will work with Nucular controller converter mode? With the Flatpack2 (or really any power supply) there is no need for voltage or current manipulation/regulation because the controller in converter mode regulates the current and steps up the voltage. The only limitation is that the power supply voltage needs to be lower than the battery voltage at any/all state of charge so the motor phase can do it's job of "stepping up" the voltage.
Any info regarding this Rectifier unit and Nucular converter mode would be appreciated!

Yes it works without problems. If you have big motor(qs205), you can charge at maximum power into battery. I've used Botox can Arduino to set highest output voltage and save that as default. And hw power on mod mentioned before. Do not forget to unplug can after settings is done, it was cased some problems. 22s battery.

 

[andrea6]

I have SMU02B. There are no fan settings in it.
You can only see the temperature of each unit.

May I ask how do you connect an SMU02B to R4850G2? Which frame do you use, or you just use direct wiring? Can you post a photo? Thanks a lot!

 

[bogus105]

Hi guys. I have Huawei R4830N2 power supply unit. Do you think that N2 models can work the same in terms of communication via CAN? I've managed to buy those famous small modules from taobao.com in China meant to control R4850G2 PSUs - got one R4850G2. Will try to connect R4830N2 to it. Why not now? - i'm away from home till next year. That's why i'm asking if you know if lower power version R4830N2 can be controlled the same way as R4850G2. Thanks!

 

[analog8484]

Hey everyone,

I have been testing 4 R4850G2s in parallel to charge a 16s LiFePO4 pack with some great results in terms of amperage and voltage control. After receiving the units I used a USB-CAN analyzer to change the Off-line default voltage to 50V and the Off-line default current to 30A. During a charge cycle I set the On-line voltage to 55V and On-line current to 30A - Repeat packets every 45 seconds until the charge is complete.

The problem is the fans on the R4850G2s don't change speed and the units are generating a huge amount of heat during charging. They hit between 60-73c after only 20mins.
temp.jpg

I noticed a few people have had problems with fans running full speed with no load, the only condition I could find this happened in one unit dated 2015 spins up full when no CAN communication is detected, as soon as you connect it to the CAN bus it drops back down. Also noticed fans ramp up slightly when the battery is disconnected from the DC output. Once you reconnect the battery it ramps back down.

Clearly the fans do work to some degree but oddly not during extreme heat conditions when you need them to be 100%. Also would have expected the one R4850G2 that reached over 70c to shutdown or throttle but it continues to run regardless.

Could be some flag that needs setting to automatically control the fans, perhaps they are set in manual mode and are expecting a SMU02 module to change the speed depending on temp. Observing Unknown frame 83 during a charge cycle shows 01 83 00 00 10 00 00 00 for a few minutes before returning back to 01 83 00 00 00 00 00 00. I believe this could be a flag to indicate a fan error.

Does anyone have any idea why they refuse to ramp up? anymore have experience with a SMU02 controller, does it have any settings for fan control?

Any help would be much appreciated.

Thanks for looking and stay safe.

Thanks for sharing the test results. Were you able to figure out the the high temp issue? It doesn't seem practical to charge with 70C temp. I don't see the such high temp using only one charger so I wonder if the parallel setup is causing the issue. How do you physically parallel the chargers? Do you use the edge card to screw terminals adapters? I would like to know how the pre-charge works for parallel chargers.

 

[dominik h]

The Huawei change their fan speed with increasing the load and internal temperature.I ran them 8 hours with 3kW output with nearly the same case temperatures as written above.
So no need to worry, Huawei knows what the psu can do.

Also the others, Eltek, Tyco Lineage, Vertiv, Megmeet do not run cool. Some of them rate their full load output up to 55C ambient temperature. So case temperatures must be definitly hotter.

 

[analog8484]

The Huawei change their fan speed with increasing the load and internal temperature.I ran them 8 hours with 3kW output with nearly the same case temperatures as written above.
So no need to worry, Huawei knows what the psu can do.

Thanks, that's good to know. Do you run them in parallel?

 

[analog8484]

Thank you jairbj!! Problem now solved!! 8) 8)

I took off the rear adapter and connected AC voltage lines directly, and guess what...the power supply did NOT turn on, which is exactly the same situation as you reported. Logically, I started thinking the adapter must have some design error that is somehow causing the unexpected POWER ON behavior, and my testing described below confirmed this was, in fact, the case...the adapter has a critical design flaw! At least is does for the version of my power supply...R4850G2

I looked closely at +48v and -48v PCB traces on R4850G2, and I noticed the PCB design has changed (over time) compared to photos from other users that have previously been uploaded to this thread. Apparently, HUAWEI has changed the traces in the PCB from previous versions, but unfortunately the adapter that I purchased was not matched and suitable to the revised traces on the PCB.

I was able to find a simple workaround so I can still use the adapter, but also re-enable the START PIN function. I applied a KAPTON tape on BOTH SIDES of PCB traces on the power supply, actually on the top of the -48v traces.

After tape was applied, I reconnected the adapter to the power supply, and when I apply AC voltage the adapter works correctly. :thumb: To turn on the power supply is now very simple: GROUND only 1 of the 2 wires from the JST connector (the black wire). Note: Grounding the other JST wire (white wire) has no effect, so I leave it floating. The photo below will hopefully help others who have found this same problem.

Thank you again jairbj for your help and ideas, PROBLEM IS NOW SOLVED!! 8) 8)

Power_supply_PCB.JPG

PCB_power_supply_3.JPG

Thanks for the solution. It worked as described but I am wondering if it disables the pre-charge function?

 

[dominik h]

I tried three paralell and up to 4 serial.

 

[analog8484]

I tried three paralell and up to 4 serial.

Thank you. That's great to know. Is there any special charging procedure for parallel setup?

I have been using a single unit with the edge card adapter and it seems I have to power on the unit before connecting to the battery (LiFePO4) to avoid in-rush spark. Is this typical for you?

I am guessing all parallel units should have the same voltage setting and turned on before connecting to the battery?

 

[dominik h]

In serial use I power the units before connecting them to the battery. If you try the other way they can die when the output capacitors charge not exactly in the same way/time.

In paralell use there is no need to power them up before connecting to the batterie, in real life they get always hot plugged, when they are inseret in their 19" racks. But i also do power mine before connection. I used a Zivan NG3 before and it has massive sparks during connection , when it was not powered up before.

The paralell units should have nearly the same voltage settings.
But I also testet different voltage settings. For example in a three phase 9kW setup for 16S Lifepo4 , phase 1 55,6V , phase 2 55,8V and phase 3 56V, so the BMS has a little more time to equalize the cells.

The inrush current is exactly the same 17A.
It only lasts a little longer when the rectifier is connected to the battery.

 

[TIOK]

I took off the rear adapter and connected AC voltage lines directly, and guess what...the power supply did NOT turn on, which is exactly the same situation as you reported. Logically, I started thinking the adapter must have some design error that is somehow causing the unexpected POWER ON behavior, and my testing described below confirmed this was, in fact, the case...the adapter has a critical design flaw! At least is does for the version of my power supply...R4850G2

All adapters are arranged the same way. Since 2018, contact "6" has been added to all blocks. If you don't need it, you can just cut the path coming from it.

 

[analog8484]

The inrush current is exactly the same 17A.
It only lasts a little longer when the rectifier is connected to the battery.

That seems different from my experience. I see a pretty big spark that suggests a much higher in-rush current than 17A. Perhaps it's because I added the Kapton tape mod (https://endless-sphere.com/forums/viewtopic.php?f=14&t=86038&start=175#p1723367) provided by OhmSchool? I would like to keep the ability to enable/disable output by connecting/disconnecting the slot detect signals to ground.

 

[analog8484]

I took off the rear adapter and connected AC voltage lines directly, and guess what...the power supply did NOT turn on, which is exactly the same situation as you reported. Logically, I started thinking the adapter must have some design error that is somehow causing the unexpected POWER ON behavior, and my testing described below confirmed this was, in fact, the case...the adapter has a critical design flaw! At least is does for the version of my power supply...R4850G2

All adapters are arranged the same way. Since 2018, contact "6" has been added to all blocks. If you don't need it, you can just cut the path coming from it.
R4850G2.png

Does the contact "6" control output enable/disable?

 

[dominik h]

The can code works also with the R4875G5.
I had a New one to test.
But you also need the special 4kW plug.

 

[TIOK]

Does the contact "6" control output enable/disable?

When you insert the connector, pin "6" closes to the negative output, and the unit starts.
If you don't need it, you can cut off the track coming from the contact, in the place indicated by the arrow on my photo. And start the unit as before, closing contact 10 on a common wire. And you don't need to glue anything.

 

[dominik h]

All adapters are arranged the same way. Since 2018, contact "6" has been added to all blocks. If you don't need it, you can just cut the path coming from it.

Do you have a picture/port description for the R4875 plugs .
I got plugs with two aditional contacts.
I think the one next to the DC+ is for switching between 3kw and 4kW mode
The second one is port or slot detection
The third one is Can
What is the one near the AC input for?

 

[TIOK]

Do you have a picture/port description for the R4875 plugs .
I got plugs with two aditional contacts.
I think the one next to the DC+ is for switching between 3kw and 4kW mode
The second one is port or slot detection
The third one is Can
What is the one near the AC input for?

The port is here on page 9. The R4875 units do not have a fourth pair of contacts. I have not seen such. And where did you buy it?

 

[dominik h]

Bought it on Ebay , from China.
On the picture it looked like it has 3 plugs and I had already bought there two times the 4kW plugs.

I will try the plug in a few days , when my the next 4kW Huawei arrive.

 

[dominik h]

The port is here on page 9. The R4875 units do not have a fourth pair of contacts.

4 months ago and I have already forgotten the picture. I'm getting old.