I am glad the instructions now show the Thermal Pad Placement. My kit did not come with 3 pads so I am wondering where I can buy the correct thickness pad for the 3rd IC.
Moving to the rpiboot section; I downloaded the SW on my Windows 11 PC. I connected the USB-C Port on the HA Yellow to the PC as explained and teh Jumper set to USB + holding USB-C Recovery.
I ran the rpiboot software and saw the LEDs flash. Since I am color blind, I assume D4 is Red, D5 is Green? and D6 is Yellow?
It was more than a couple seconds as outlined in the instructions but the D5 LED is now solid. I assume it is done.
Next I loaded Rasberry Pi Imager v1.9.4.
Now I am wondering if I need to move the Jumper back to UART (but there were no instructions saying to do so, so I left it on USB. I guess that makes sense since I Raspberry Pi Imager would be controlling it through the USB port .
I have a 1TB NVMe installed that came with my kit that I purchased from Ameridroid.
When running the Imager, I selected Device=Raspberry Pi 5, OS=Home Assistant OS 15.2 (Yellow). When I select Storage, I only see the drives on my PC and not the NVMe or other storage on the HA Yellow or CM5.
Any idea what step I missed or what I may be doing wrong?
OK, here is what Google Gemini said when I supplied the text from my first post. I am not sure about it’s accuracy but it seems like rpiboot may not have worked properly, causing my 1TB NVMe not to be visible.
From Gemini…
Let’s break this down step-by-step to get your Home Assistant Yellow (HA Yellow) set up correctly. You’re very close!
Thermal Pad for the 3rd IC
First, regarding the thermal pads:
Missing 3rd Pad: It’s concerning that your kit didn’t come with all three thermal pads as shown in the updated instructions. The third pad is for the PMIC (Power Management Integrated Circuit) on the CM5, which can generate heat.
Where to Buy: You’ll need a thermal pad that’s typically 1mm thick. You can purchase these online from electronics retailers like:
Amazon: Search for “thermal pad 1mm”
Digi-Key / Mouser: More specialized electronics suppliers, good for specific brands.
eBay: Can be an option for smaller quantities.
Recommendation: While you can proceed with the setup, it’s highly recommended to get that third thermal pad. Proper thermal management ensures the longevity and stability of your Compute Module, especially under load. A 1mm thick, high-thermal-conductivity pad (often specified in W/mK - higher is better) would be ideal.
rpiboot and LED Status
You’re correct on the LEDs:
D4: Red (Power)
D5: Green (Activity)
D6: Yellow (Home Assistant System Health)
Regarding the rpiboot process:
D5 Solid Green: When rpiboot finishes successfully and the CM5 is recognized as a USB mass storage device, the D5 (Green Activity LED) should indeed become solid. This indicates that the CM5 is in USB boot mode and is ready to be accessed by your PC. So, you’re on the right track here! The “couple of seconds” might be for the initial detection, but the solid green indicates it’s ready for flashing.
Raspberry Pi Imager and Jumper Setting
This is where the key misunderstanding often lies, and it’s a common point of confusion.
Jumper Position (Crucial!):
For rpiboot and Initial Flashing: The jumper MUST be set to USB. This puts the CM5 into a mode where it acts like a USB mass storage device, allowing rpiboot to find it and for tools like Raspberry Pi Imager to write to its storage (the NVMe).
For Normal Operation (After Flashing): Once you are done flashing the OS, you MUST move the jumper back to UART (or completely remove it, though UART is the default recommended position for normal boot). If the jumper remains on USB, the HA Yellow will try to boot from the USB connection again instead of booting from the NVMe drive.
Why You Don’t See the NVMe in Raspberry Pi Imager:
You correctly selected “Device=Raspberry Pi 5” (which is actually a generic selection, not specifically for the Yellow with CM5, but that’s okay) and “OS=Home Assistant OS 15.2 (Yellow)”.
The problem is likely with the “Select Storage” step. When rpiboot successfully runs, it makes the internal storage of the CM5 (or the NVMe connected to it) appear as a drive directly connected to your PC.
It will NOT show up as “NVMe” or “Home Assistant Yellow”. It will appear as a generic USB drive or uninitialized storage device.
Common Names:
Look for a drive that matches the size of your NVMe (1TB in your case). It might be labeled something generic like “USB Device,” “Mass Storage Device,” or just show its capacity with no name.
Be extremely careful to select the correct drive. If you select your PC’s main drive or another important drive, you will erase it!
Sometimes, Windows might prompt you to initialize or format the drive when it detects it. You can generally ignore these prompts if you’re about to flash an OS to it with Raspberry Pi Imager, as the imager will handle the formatting.
What to do now:
Confirm Jumper: Ensure your jumper is still set to USB on the HA Yellow.
Rerun Raspberry Pi Imager:
Select “Choose OS” → “Home Assistant” → “Home Assistant Yellow (Raspberry Pi CM4/CM5)” (or whatever the exact name is for Yellow).
Select “Choose Storage”. Carefully examine the list. You should see an option that corresponds to your 1TB NVMe drive. It won’t have a friendly name like “My NVMe” or “HA Yellow.” It will likely just show its capacity (e.g., “Generic Mass Storage Device - 1TB” or similar).
If you have other external drives connected, disconnect them to reduce confusion.
Double-check the size to ensure it matches your 1TB NVMe.
Flash the OS: Proceed with flashing. This process will take some time.
Crucial Post-Flashing Step: Move the Jumper!
Once Raspberry Pi Imager completes the flashing process, immediately move the jumper on your HA Yellow from USB to UART.
Disconnect the USB-C cable from your PC.
Connect the HA Yellow to network and power, and it should boot from the NVMe.
Troubleshooting if you still don’t see the NVMe in Imager:
Restart rpiboot: Sometimes rpiboot needs to be run again if the connection drops. Close Raspberry Pi Imager, disconnect/reconnect the USB-C cable (ensure jumper is on USB and you’re holding the recovery button if needed), and rerun rpiboot. Wait for D5 to be solid. Then open Raspberry Pi Imager again.
Driver Issues: Though less common with rpiboot, ensure your Windows 11 PC has up-to-date drivers.
Different USB Port: Try a different USB port on your PC.
Another USB-C Cable: Some USB-C cables are “charge-only” and don’t support data. Ensure you’re using a data-capable USB-C cable.
You’re on the right track with the rpiboot and Imager. The main hurdle is correctly identifying the NVMe in the “Select Storage” menu of Raspberry Pi Imager and remembering to move that jumper back!
This time I got success as I still had the Raspberry Pi Imager v1.9.4 open and found that I could now see both the 62.5 GB on the CM5 and the 1024.2 GB on the NVMe.
I selected the NVMe
Then ran into an “access denied error while writing file to disk. Controlled Folder Access seems to be enabled. Please add both rpi-imager.exe and fat32format.exe to the list of allowed apps and try again.” error
I allowed both rpi-imager.exe and fat32format.exe using the following steps in W11
Steps to Allow Apps Through Controlled Folder Access:
Open Windows Security:
Click on the Start Menu.
Type “Windows Security” and press Enter, or select “Windows Security” from the search results.
Go to Virus & Threat Protection:
In the Windows Security window, click on “Virus & threat protection” in the left-hand navigation pane.
Manage Ransomware Protection:
Scroll down to the “Ransomware protection” section.
Click on “Manage ransomware protection”.
Allow an App Through Controlled Folder Access:
Under “Controlled folder access,” ensure it’s turned “On” (if it’s off, you could try temporarily turning it off, but it’s better to add exceptions for security).
Click on “Allow an app through Controlled folder access”.
You might be prompted by User Account Control (UAC) to confirm. Click “Yes.”
Add Raspberry Pi Imager:
Click on “+ Add an allowed app”.
Select “Browse all apps”.
Navigate to the location where you installed or extracted Raspberry Pi Imager. This is typically:
C:\Program Files\Raspberry Pi Imager\rpi-imager.exe
Select rpi-imager.exe and click “Open.”
Add fat32format.exe:
Click on “+ Add an allowed app” again.
Select “Browse all apps”.
Navigate to the same folder where rpi-imager.exe is located (e.g., C:\Program Files\Raspberry Pi Imager\).
Look for fat32format.exe in that directory.
Select fat32format.exe and click “Open.”
Close Windows Security: You can now close the Windows Security window.
I then proceeded with Raspberry Pi Imager and successfully installed the Home Assistant image on the NVMe.
Nice detailed writeup! Thank you for sharing it with the community.
For me, the simpler solution is to simply place the m.2 MVME SSD in an external USB enclosure, plug it into a computer, and then use the RPi Imager to install the proper HAOS image on it. Pop the NVME SSD back into the HA Yellow and power it up! It does require an extra piece of hardware, though. But it can save many hours of frustration!
I have an external enclosure and didn’t know enough about what I was doing to do it that way. I was also nervous as the Raspberry Pi imager may have killed two USB sticks in a row that I can no longer reformat. I didn’t want to break my new 1TB Memory.
Thanks for your compliment on the write-up. The instructions seem to be lacking for anyone new to HA that just purchased a HA Yellow Kit with CM5. Most of the detailed instructions are for those upgrading from CM4 to CM5. Thats why I tried to be specific about the steps I took to get the HA Image installed.
