Dear all,
After a few weeks playing with my ASHP Samsung Mono Gen 7 R290 Gen7, I’m finally having a nice step done.
Although chilling outside temperature have not reached us (just yet). And therefore I don’t know if my ASHP DIY setup is having a good CoP. I’m willing to share with you my latest dev around the samsung monitoring/controlling.
Here is the repository I’ve started: GitHub - 70p4z/samsung-nasa-mqtt: Bridge for home assistant to rule your EHS
It’s a python tool, meant to run on raspberry (or other SBC), which connects the NASA link on F3/F4 connector, the link normally used to wire remote controllers.
I was not satisfied with the remote controller unability to go wireless. And therefore replaced them with home assistant + mqtt + a few homemade wireless temperature sensors.
The project will soonishly comes with a PCB board to connect the raspberry on one side (probably USB serial converter) and the F3/F4 connectors.
I’m currently completely avoiding MWR controller (they’re disconnected), and replaced them with the python.
I’ve exhibiting many monitorable data from the NASA link (temps/power/settings) but I’m willing to extends if anyone requires changes.
Here is a screen shot of the currently exhibited sensor/controls
The code is not yet that great, but it works nicely.
FIrst time the space heating gets on this year.
Here is a screenshot of HA stats.
No cycling for an hour, even though temperature outside is quite high (13C), and a CoP starting at 5.7, finishing at 4.7, with a very low power input (~400Wh).
This looks very promising - looking for a way to control my Samsung without simple switching it on and off or an external thermostat - how have you currently connected the cabling into the Pi?
Yes it’s now fully connected, I’ve even disconnected the wired remote controller.
Actually I’m finishing a PCB to make the connection even easier (today I asembled a THVD8000 evalutation board, a ftdi uart and a raspberry pi)
If you want to have it a go, I can support you!
Cheers
Is it possible to use this with the wired controller still connected? Or does it require permanently disconnecting the wired controller?
I’ve been communicating with my Samsung ASHP via the Samsung Modbus module (MIM-B19N), but this is quite limited compared to what you can get from F3/F4/:
As a matter of fact I’m pursuing the goal of having the remote disconnected, but I’ve not yet finished to reverse the plug and play procedure for my ‘virtual’ remote controller to be recognized (mainly due to wild power outtage due to experimentations on my solar inverter…)
I mainly use my script to push zone 1 and 2 room temperature using an array of RF sensor to cope with the fact samsung does not provide wireless remote controller. Therefore I disconnect the internal thermistor inside the wires controller from samsung (mwr wg00n IIRC)
If it’s just to configure fsv and grab data, then you won’t have any trouble. I also enable/disable zones and dhw operation with home assistant rules to schedule my house. Not only relying on the pure configuration madness of the wired controller
Ihope this contribution may serve others. although I still hace to smoothen the hardware connection. For now I still use a bunch of wires and multiple board (raspberry, ftdi usb, thvd8000 with hw mods)
Nonetheless, it’s been running real fine since end of august.
Cheers
In a post of yours, you mentioned that you can control PWM 0x40C4.
I would also like to control it, but only to limit it overnight to get the noise down.
Right now I am using esphome_samsung_hvac_bus and ESP32 + RS485 on F1/F2.
Do you have an idea how I can write to this register?
Hi @Kocta,
Sorry for the latency!
I agree that controlling 40C4 would allow you to reduce the noise (it’s impressive how it can reduce).
However, the first drawback is that this value is not saved in eeprom, therefore after any power outage you have to reapply it.
The second drawback is that the mean to change the value requires using a F3/F4 connection (modulated RS485) on the control layer, I’ve not yet tested the F1/F2 connection (pure RS485) on the set layer. I’m willing to try that, but hadn’t yet had time to.
Also, there is another setting you can change to reduce the pump by default to 70% (which from 100% really kills the noise),
From the wired controller user manual:
FSV #4051=2 : Inverter pump use + Output 70%
And that setting is manageable using the wired controller directly IIRC.
Cheers,
Great work. Thank you for sharing your findings.
I am currently using the Modbus gateway (MIM-B19N) connected to an ESP32 using ESHP-Home and Home Assistant to view data from my EHS ASHP.
Seeing your success so far with NASA data on F3/F4 I would love to get something working with an ESP-32 but preferably on the F1/F2 connections, probably using Atom hardware and using Home Assistant to control the heat pump.
You mentioned you have written to 0x40C4 to control the PWM pump speed. Does this allow you full control across the whole PWM range? Does it get overwritten by the heat pump and is there a way to cancel the input to revert pump control back to the heat pump?
I have found the Samsung PWM control crude at best. I would like to automate PWM control with different settings for hot water, radiators, and perhaps ambient temperatures. Even when the speed is limited to 70% through FSV4051, it starts up at 100% before dropping back to 70% after some time. It’s just not a nice implementation.
Actually, the 40c4 does not get overwritten, only the power cycle resets it.
It does not cancel the heatpump value though, the pwm range is still respected but if you play with the control factor , the ashp pwm is void. Home assistant with a reading of the water flow is much more efficient at driving it.
I hope to finish a “packageable” interface so that I can share more than the python!
Stay tuned I’ll probably be able to play on f1/f2 a during the xmas holidays.
Cheers
I’ve not seen the usbc connection . Through f3/f4 I cab pretty much control everything (change FSV, ambient targets etc, DHW temp, complex outing mode etc)
Home assistabt is fairly simple and I’ve made a auto discovered MQQT interface for the samsung, therefore there is close to nothing to configure. A few docker commands and that’s pretty much it.
These days I’m playing on the F1/F2 link to see if I can do the same as F3/F4 to simplify the electronic part.
Cheers
Hey folks,
2024 passed by so quickly. Since I started ground works for installing the heatpump in last june, it went like a breeze.
Anyway, I’ve finally tested what I promised, and couldn’t change configuration (setting of ambient target and current ambient settings, which are my main showstopper since I don’t use wired controller temp control (disconnected temp sensor, but left the wired controller plugged in).
I tried sending frames on rs485 as if I were the 100000 host (outdoor unit) or 500000 (wired controller) on the set layer (F1/F2 connection).
Any packet I’ve sent went into the void. However, information on this link is really interesting regarding consumption and heatpump state. Someone with MITM RS485 link between outdoor unit (100000) and control kit (200000) could even change the set temperature for the water flow to avoid going too high in case of small surface of coil for DHW heating (the control kit asks max 75°c on my gen7 r290, ans this could clearly be reduced to avoid overconsumption while giving time for the heat to be transferred to the DHW, my coil is too small (reused the DHW tank from the previous gas boiler so far).
I’ll make another try on the wifi kit interface instead, to see if I can harness a bus with much less hardware hassles than the F3/F4 link. And if I can’t, then I’ll make a open source PCB for anyone to be able to exploit values from the F3/F4 link to even replace wired controller completely (and grab useful data).
I hadn’t had time to check if my control kit has a USB plug on the board. I’ll figure that later I think.
Interesting that the USB C converter is stated in my version, I’ve dug the picture I took from various parts/schematic to avoid inside/outside round trips.