| Topic | Building a Kiosk Style Appliance with Browser on a Raspberry Pi Zero 2W |
| Subtopic | Installing Browser on a Raspberry Zero 2W as a Kiosk Appliance (or just a OpenHamClock Server) |
| Equipment Required | Computer + Internet |
| Costs | Raspberry Pi Zero 2W |
| Document last reviewed and updated (reviewed each year) | 14th March 2026 |
Please note : This article and its contents are intended for personal use on a local Network with appropriate firewall/router. It is not intended to be available to outside users, and as such no hardening or additional firewalling has been implemented. If you intend to put it on the public facing Internet e.g. field days, that you understand the implications or have the appropriate networking / firewall/ security hardening skills to make the appropriate changes.
In effect, building your own OpenHamClock Server is, in simple terms, replicating the web service infrastructure that is available on openhamclock.com, but instead of pointing your browser to openhamclock.com, you will point it to the IP Address of your local server on your local network.
The Server Component itself runs on Linux which is normally a great choice for speed, low CPU utilisation, and low memory utilisation. In fact the server component runs reasonably well on a Raspberry Pi Zero 2W.
It is when you add the Light Desktop, but more so the browser component that consumes the memory, and starts stretching the Pi Zero 2W. It is for this reason that I recommend that if all you have a available is a Pi Zero 2W available to you, that you run it as a OpenHamClock Server only (without desktop/browser). You can build it with the light desktop and browser if its the only hardware you have available, but it would not be my first choice. For this reason you will find this article, commences with the server component first, and you can stop at this point, or you can proceed and push the boundaries by moving to the next step and adding the light desktop and browser.
Similarly, I have provided the full article (with both parts), as you might like to take it as a challenge in removing components, changing parameters, to see if you can improve on the reducing the CPU utilisation, and memory usage. I have taken the basic steps (e.g. light desktop), not installing some components etc, but I would not call it optimisation by any stretch. It would be fair to say my focus was on providing articles on building OpenHamClock across a few platforms, resolving the install issues, trying to make the articles as accurate as possible, so that they can be reliably replicated by readers. In fact, it would be reasonable to say that each version has had at least 3-4 builds, with the last build on each one being built following the published article.
Finally, you might get excited with building it all (including browser) on a Pi Zero 2W (even based exactly on this article), and it will work, but its when you go to use it you will see that it starts to become sluggish. Indicators include
- Zooming in and out of map – redraws start lagging
- The ticker tape news that runs across the bottom is not smooth – in fact becomes jittery
- The Browser starts using a little more memory as time goes on
But as I mentioned, you might want to take this on as a challenge, where as others may decide to pull out an old Pi 4 or buy a dedicated SBC for this project or even a low end PC.
So let’s get started with the basic hardware, that most have sitting around their shack or can obtain quite easily, which is the Raspberry PI Zero 2W.
The Raspberry Pi Zero 2W has 512mb of memory which is just enough to run the OpenHamClock Server component comfortably as shown in the image below
As you can see, it seems to be stable at 350Mb or just a little above.
One of the prime considerations, for something that might run 24×7 on a Raspberry Pi, is to reduce the amount of constant writing to the SD Card. We can’t avoid the server application writing which should not be too much, but if your server requires utilising the Swap file, this will undoubtably lessen the life of your SD Card, and also have a large impact on the CPU utilisation, which will have an impact on your OpenHamClock implementation.
One other consideration, whenever you are implementing a Raspberry PI Server, is that you do not need a full blown Raspberry PI OS with desktop. I have seen many people who cannot get by without a Desktop GUI, and install it every time they do a Raspberry PI Project. So for these Raspberry PI builds, we are going to use Raspberry PI OS Lite, using the latest (at time of writing), 64 bit Trixie. This will provide us with a Command-line interface only. Don’t panic, its not that hard and another tool in your arsenal.
So lets start with our known details
Operating System : Raspberry Pi OS Lite (64-bit) Debian Trixie
Storage : 16Gb SD Card (8gb will probably be fine, but only have 16Gb+ cards available to me)
Hostname : openhamclock
Username : openhamclock
IP Address of PI : 192.168.1.10 (now this should be a free network IP address on your home/shack network)
This guide’s purpose is not to explain how to setup up networking, as there are thousands of guides on this and many choices on how to this which can include:
- Local Static IP Address (setting and IP address on your PI)
- DHCP Static IP Address (if your router or local DHCP Server supports this)
Just a couple of quick notes :
- For years PI OS imager has been a good tool for quickly creating images for the PI. However the last few releases have been buggy (2.06 being the last one at the time of writing), in particular around the localisation selections. To this end, trying to set some countries/cities, have been screwed up, where for Australian cities, it has no selections. This has been reported but nothing seems to have been done, so you may have to select a similar overeas City, then manually select the Country. In the previous version, it was also impossible to select as the selector was doing stupid things.You will need to fix this manually one the OS is running. It has been reported by other users as well, including Hong Kong and a few other locations. I might be wrong, but I have a suspicion that AI is being used for the development of this product introducing stupid errors and loss of quality control.
- For most of the setup, especially if you don’t want to sit on top of your Pi with keyboard, you are going to need an SSH application, like Putty or other SSH App.
- You will note that we are not enabling Raspberry PI Connect as part of the setup as we are not using a GUI/Desktop. The reason for this, it for most of it we just need via SSH.
So before we start, you need to create an image for your Raspberry Pi SD Card, and you do this with the Raspberry Pi Imager (https://www.raspberrypi.com/software/)
The main things options you need to select when using the Raspberry PI image is the following:
- Device : Raspberry Pi Zero 2W
- Operating System : Raspberry Pi OS Lite (64Bit) Trixie with no Desktop Environment (Found under Raspberry Pi OS (Other)
- Storage Device : Make sure it is your SD Card you inserted that is being used, what ever you select will be overwritten
- Hostname : openhamclock
- Localisation :
- Capital City : Adelaide
- Timezone : Australia/Adelaide
- Keyboard Layout : us
- Username : openhamclock
- Password : {your choice, but write down}
- WI-FI : {enter your wifi details here including your SSID and Passphrase)
- Remote Access : Enable SSH – use password authentication
- Enable Raspberry Pi Connect : disabled
Now you can write to your SD Card.
Ok, if everything went well, you should now have a SD Card with a bootable image. Drop it into your Pi 2 Zero W and boot up.
Login to your Raspberry Pi with a keyboard usin the login and password you set (and wrote down). If you did not have a way of setting the IP Address (e.g. DHCP Static), then at the console prompt, type
ifconfig
which will give you your current IP address (this is assuming your wireless details were set correctly when you built your image). Now you can remain using the keyboard if you prefer, or you can remotely connect to your Pi with SSH (e.g. Putty or your favourite SSH Terminal) and connect to the Pi. Again, your login and password will be what you set. Using SSH to connect to the Pi remotely makes it a lot easier to cut and paste commands.
Ok, whatever method you use, lets get started
The first thing we want to do is a step that could be part of your routine when building Linux systems and that is bringing the O/S up to date wit the lastest updates.
sudo apt update
sudo apt upgrade
That should bring your Raspberry Pi OS up to date.
Login to your Raspberry Pi with a keyboard using the login and password you set (and wrote down). If you did not have a way of setting the IP Address (e.g. DHCP Static), then at the console prompt, type
sudo nmcli
which will display your current interfaces and assigned IP addresses (this is assuming your wireless details were set correctly when you built your image). Now you can remain using the keyboard if you prefer, or you can remotely connect to your Pi with SSH (e.g. Putty or your favourite SSH Terminal) and connect to the Pi. Again, your login and password will be what you set. Using SSH to connect to the Pi remotely makes it a lot easier to cut and paste commands.
You also have access to the interactive network configuration tool as well, so you can set a static IP address for your Pi
sudo nmtui
As I mentioned, setting up a static IP address for your PI, is beyond the scope of this guide, and there are plenty guides on the net.
Next thing we want to do, is set Console Autologin enabled via raspi-config. In simple terms we want the Pi to login automatically when it reboots or is turned on, without needing to find a keyboard, plug it in, and login to get the server going after a power failure. To so this we perform the following command:
sudo raspi-config
You will come up with text based menu, and the main area we want is
System Options >>>>> Auto Login >>>> and select YES to “would you like to automatically login to the console”
Next we need to fix localisation issues, so remain in the raspi-config menu and from the top menu goto. (note : that you may not need this if you are not in one of the affected countries, or you have used a later version of the PI imager where it has been fixed)
Localisation >>>> Locale >>> select en_AU.UTF-8 UTF-8 (if Australia) and set this as the default locale as well.
Timezone >>>> Select your correct Geographic Area >>>> Australia (in my case) >>>>> and then the correct city.
Ok that is all the work we need to do under raspi-config
Now we install the openhamclock server component and its dependencies. We will start from openhamclocks home directory.
cd ~
We need to install git with the following command
sudo apt install git
Now we need download and and execute NVM (Node Version Manager) directly from GitHub
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.4/install.sh | bash
Note the above command may not work in the future as they update the install file. you can go to this url and read https://github.com/nvm-sh/nvm/tree/master?tab=readme-ov-file, it will generally show you install instructions, with a command very similar, and the latest up to date install release number. The above command was correct at the time of writing. When it finishes installing, you can try the command it provides, or just close the SSH terminal window and open it again.
Copy and paste the following command, so you can use NVM immediately
export NVM_DIR=”$HOME/.nvm”
[ -s “$NVM_DIR/nvm.sh” ] && . “$NVM_DIR/nvm.sh” # This loads nvm
[ -s “$NVM_DIR/bash_completion” ] && . “$NVM_DIR/bash_completion” # This loads nvm bash_completion
We are now going to reload our shell.
source ~/.bashrc
Now we are going to install node.js, which includes npm
nvm install 22
We can confirm that versions of node and npm are the latest.
node -v
npm -v
Now we need to use git to obtain the openhamclock code.
git clone https://github.com/accius/openhamclock.git (note we are not using sudo)
cd openhamclock
npm install
npm start
Ctrl – C to break it
now edit the .env file and add your settings
sudo nano .env
# Your amateur radio callsign
CALLSIGN=put your callsign in here
# Your Maidenhead grid locator (4 or 6 character)
LOCATOR=(put your 4 or 6 character maidenhead) in here
# ===========================================
# SERVER SETTINGS (Backend API)
# ===========================================
# Backend API server port
# Dev default: 3001 (see DEV PORTS note above)
PORT=3000
# Host/IP to bind to
# localhost = only accessible from this computer
# 0.0.0.0 = accessible from other devices on your network
HOST=0.0.0.0 (change this from local host)
Change these so it autoupdates
# ===========================================
# AUTO UPDATE (GIT)
# ===========================================
# Enable automatic updates (requires git installation and repo clone)
AUTO_UPDATE_ENABLED=true
# Check interval in minutes
AUTO_UPDATE_INTERVAL_MINUTES=60
# Run a check shortly after startup (true/false)
AUTO_UPDATE_ON_START=true
# Exit after update so a supervisor (systemd/pm2) can restart (true/false)
AUTO_UPDATE_EXIT_AFTER=true
This will sync your settings between your PI Browser and the browser that you are using on your PC - it forces it to use a common JSON config file
Just some basic settings - the main one being Metric instead of imperial
# ===========================================
# DISPLAY PREFERENCES
# ===========================================
# Units: 'imperial' or 'metric'
# Deprecated
UNITS=metric
# Seperated Units ('imperial' or 'metric')
# DISTUNITS - Distance units
DISTUNITS=metric
# TEMPUNITS - Temperature Units
TEMPUNITS=metric
# PRESSUNITS - Pressure Units
PRESSUNITS=metric
Save this file
If you want to see thatit is working, just perform the following command again
npm start
Use your browser to connect back to the IP address of your Pi/ It should show your call sign
If that looks ok, CTRL-C again, so we can finish it off and you can customise it as much as you want.
To finish off, we need to set it to start on boot, so we will set this up as a service, so perform the following commands:
sudo nano /etc/systemd/system/openhamclock.service
add the following to the file
[Unit]
Description=OpenHamClock Server
After=network.target
[Service]
Type=simple
User=openhamclock
WorkingDirectory=/home/openhamclock/openhamclock
Environment=NODE_ENV=production
Environment=HOME=/home/openhamclock
Environment=PATH=/home/openhamclock/.nvm/versions/node/v24.14.0/bin:/usr/bin:/bin
ExecStart=/home/openhamclock/.nvm/versions/node/v24.14.0/bin/npm start
Restart=always
RestartSec=5
[Install]
WantedBy=multi-user.target
save it and now perform the following commands
sudo systemctl daemon-reload
sudo systemctl enable openhamclock
sudo systemctl start openhamclock
Reboot and wait about three minutes (as it takes little while for the server to start up) and then try to access it from the browser
If this fails, a couple of quick commands may help locte the issue
journalctl -u openhamclock -f
This will show you some logs pertaining to the starting of openhamclock.
There is the possibility, that the npm version you have installed differs from the one we have in the service file. You can check what your correct version is and its location with the following command.
which npm
Otherwise enjoy exploring, you have completed the build of the OpenHamClock server on the Pi Zero 2W. If you want to install the light desktop and Browser, you can move onto the next OPTIONAL step
************
OPTIONAL
************
However if you want to run a Graphic screen on this unit, keep moving on. We will now install OpenBox and Chromium and have them start automatically.
If your experience with Linux is limited, you may not recognise the format of the command line below. It is a one line command, so if you cut and paste, take the whole line, not one line at a time.
sudo apt install –no-install-recommends \
xserver-xorg \
xinit \
openbox \
chromium \
x11-xserver-utils \
fonts-liberation \
fonts-freefont-ttf
Now we need to create/edit .xinitrc with the following command
nano ~/.xinitrc
and add the following lines to the file
sleep 30
openbox-session &
xset s off
xset -dpms
xset s noblank
exec chromium \
–kiosk \
–ozone-platform=x11 \
–force-device-scale-factor=1 \
–disable-gpu \
–disable-session-crashed-bubble \
–noerrdialogs \
–disable-infobars \
–app=http://127.0.0.1:3000
Now we probably need to explain these lines a little, as they may not been needed for your setup, and you may want to tweak them.
The sleep 30 line, stops the script running for 30 seconds to give the OpenHamClock Service time to start and settle, otherwise your browser window, will not come up with the page.
You can change this, and there are more elegant ways of doing this, but trying to keep it simple, and you may find that you might want to change this sleep time to say 5 minutes (and I will reveal reasons in the openhamclock tips and tricks later). At least you know you can come to this script, and adjust the time here.
The xset commands, were needed for my monitor, they may not be needed for yours, but it is more around the removing the monitor time out / eco features. Otherwise after 5 minutes my monitor went to sleep even though OpenHamClock was active on the screen. I turn the monitor on and off manually (or at least a remote control) as I want to see the screen whenever I turn to the screen, not have to locate the remote turn it on, lose my train of thought. And when I am finished in the shack, I turn it off.
Finally the Exec Chromium line is what I found works reliably. You might, once you have it running, wish to play with these lines, but these work quite well. Naturally the app line points back to the localhost (where your OpenHamClock Server is running).
One more thing we need to do to this script is make it executable with the following command
chmod +x ~/.xinitrc
and finally the last part, we need to autostart X on login, and add the following commands to your bash_profile
nano ~/.bash_profile
if [ -z “$DISPLAY” ] && [ “$XDG_VTNR” = “1” ]; then
exec startx
fi
guess what, that is it…..reboot and you should see your OpenHamClock displayed on your screen connecting to the OpenHamClock Server running on your Raspberry Pi.
You may notice that on the other “build on a Pi articles, you might see that we have included how to install VNC Server to remotely access your Kiosk browser. The Pi Zero 2W, does not have the memory and CPU to handle this as well, so I will not include this on this article. If you want to try as you have greatly reduced CPU and Memory usage, then the instructions are the same as the Pi 4 article.
as I mention in the other articles, I found the DX Cluster panel variable in its operation, and it is used by another panel (band health), so if DXCluster is not getting a feed, Bandhealth is non-functional as well. So I found the change of the DXCluster feed appeared to resolve the issue. So under settings, you might want to try the following
No reason, you have to use this, but worked for me.
Other articles relating to OpenHamClock
