UPS RF shield – quick test results

Yesterday I wrote a post about my latest try to level down the interference caused by computers, power supplies and other peripherals using simple Faraday’s cage.

The UPS is huge noise generator – or it should I say, it was!

In picture below, on the left hand side this is the noise generated by the UPS when running, on the right hand side UPS is inside a very simple metal mesh cage built from the stuff from the attic in half an hour. Receiver antenna was 34cm away from the UPS case and SDR was set to listen from DC to 10MHz (note the scale in the picture, I’m using HamItUp converter with 125MHz base frequency). I consider this test to be successfull 🙂

This is my second metal mesh I have built and based on the test I will cover all my computers, peripherals and other with similar cage during the next field day at my remote site.

More details what I did and how I found this can be read from my earlier post.

Testing the BASE interference for remote RIG update

While waiting my trip to my remote site I started to plan a new RF shielding and interference protection. After all I’m going to replace 99% of the computer HW I have at my remote site and it is known fact that computers are the worst noise generator one can have.

The new system is built using the latest version of Intel’s NUC computer and RaspberryPi 4th gen. The NUC is well protected by default but of course it will cause a lot of noise to VLF and HF bands, but the worst is the RPi.

Next step is to level down the interference SDR HW causes, for that I needed to measure the base interference of all the RTL-SDR dongles I’m gonna use and also the RSPdx.

I’ve had a great experience how to level down RF noice by wrapping cheap power supplies with metal mesh, so I’ll continue with that idea with the computers and external devices that are going to be installed to the remote site.

Measuring the BASE interference

As with the power supplies, I used Airspy Mini as a receiver and HamItUp up-converter so that I can have measurement from the DC to 10MHz. Software used was SpectrumSpy and during the measurement the antenna was kept in place and all the devices under measurement were placed inside the antenna loop one by one.

“Under the Loop” went:

NooElec SDR “Mini” – without it’s cheap plastic case (removed for cooling)
NooElec SDR “NESDR SMArt – Original metal case intact
NooElec SDR “Nano three” – Original metal case intact
RSPlay’s RSPdx – Original metal case intact

RaspberryPi 4 – Aluminum cooling grid installed, but no RF shield
External USB 3 HDD – Samsung 2TB 2,5″ HDD with plastic case
External USB 3 SSD – Kingston 256GB 2,5″ SSD with aluminum case
External USB 3 SSD – Samsung 256GB 2,5″ SSD with aluminum case


Yes, I need to have metal mesh especially around RPi and external HDD’s, but it would not be that bad idea to have a metal mesh also around RTL-SDR dongles and even the RSPdx too as they seem to be quite noisy ones too.

Measurements were done so close to the device as possible – do note, that noise levels went dramatically down when I moved the antenna loop 20cm away from the device, so the metal mesh does not have to be a “military” grade EMP protection 🙂

1: Open USB 3 cable, no connected device
2: Open USB 3 cable, no connected device, cable wrapped around the loop
3: External 2,5″ HDD, plastic case, Drive software eject
4: External 2,5″ HDD, plastic case, Drive connected, spinning up and active
5: External 2,5″ SSD, aluminum case, Drive connected and active, model Kingston
6: External 2,5″ SSD, aluminum case, Drive connected and active, model Samsung
7: External 2,5″ SSD, aluminum case, Drive software eject, model Kingston
8: External 2,5″ SSD, aluminum case, Drive software eject, model Samsung
9: RPi 4 max load shutting down to standby
10: RPi 4 from bootup to running, no load
11: RPi 4 from standby to power cutoff
12: RSPdx 10M mode, max USB bandwith in use
13: RSPdx UNO started device active (Play)
14: RSPdx UNO device not active (stopped) uno closed
15: RSPdx UNO started, SDR activation
16: RSPdx plugged in, no UNO running
17: 1st Gen. NooElec dongle no plastic case
18: 2nd Gen. NooElec Dongle no plastic case
19: 3rd Gen. NooElec Dongle original metal case
20: 4th Gen. NooElec Dongle original metal case

And the bonus measurement – The UPS

I need to install an UPS (Uninterrupted power supply) to my remote site as I need to have better protection against lightning, power grid interference and mains power loss. The internet connection is via 4G broadband radio modem, so I am able to connect to the remote site as long as there is still power on the operators radio tower and in my UPS battery.

UPS devices are terrible noise generators. What they do is that they rectify the mains 220v AC to DC, store spare power to battery for later use and for devices. To be able to react to the power line fluctuation, un-clean power grid electricity and total power loss in few milliseconds the UPS devices usually re-create the output to consumption devices using AC–>DC–>AC conversion. To achieve this these UPS are generally powerfull DC to AC inverters – and inverter is huge noise generator what comes to the VLF, HF bands. Usually the switching power system used in cheap inverters works around 100KHz up to 600KHz switch frequency. If you live near an solar plant you know what kind of interference an inverter can spread around if not well insulated and RF protected.

My UPS is a cheap Eaton model 3S 550 with plastic case – absolutely great device for uninterrupted power and noise! 🙂

So, how noisy the UPS is, lets find out! Loop next to the UPS and …

Well, yes – maybe I need to first protect the UPS and then go with the rest 🙂

The bottom part of the picture above is when the small loop antenna was 1 meter away from the UPS. I then slowly moved the loop closer to the UPS and the upper part of the picture is where the loop rest against the UPS case.

Final words

The system used in my measurement is not meant to be high tech absolute value deep level measurement, but a hobby level visualization of the RF interference present in the hobby level devices for remote site operator. Also it should be noted, that the interference of these devices goes dramatically down after some few meter away from the device itself.

In DX hobby the thing is to try to hear very fade signals all around the world, often signals that are even under the base noise of the band. With a good antenna and a good radio one is able to receive the signal of a “flies fart” from Japan, so when put into that perspective a noisy power supply few tens of meters away of the antenna could totally disable your DX station. SSTV DX Gallery

I got so excited about SSTV that when I do not use my system to other DX’ng I have it on 24/7 and tuned to 14230KHz and let RX-SSTV to automatically decode and save received SSTV transmission to my Google album.

To see the album, head to the album page on my Google Photos!

And a warning!!! Some of the pictures are graphical in nature – for some HAM’s sending X-rated material over the SSTV seems to be the thing 🙂

System used:
Radio: SDRDuo
Antenna: City Windom (40m)
SDR software: SDRUno, Digital mode with 3KHz bandwidth
Location: KP01tn, Finland

The All New remote RIG – Part 1 “The POC”

I had a plan to create all new remote Rig already during 2018, but due to some problems to get parts to it (the main computer) this has now been moved to be “The Project – 2019”.

The Idea:

* Fully remote controlled – power supplies included
* Windows 10 based main computer
* RaspberryPi based controll computers
* Video monitored, all led-lights, displays, others
* Enough computing power for realtime AirSpy R2, 10MHz bandwith
* Fully internal networking solution for standalone use
* Enough USB slots for:

– Four USB3 connected radios
– Six USB2 connected radios
– External keyboard and mouse
– External USB3 SSD for full system clone
– Spare USB3 & USB2 slots for later use

* At least 20/20mbit internet connection

After several iterations, tests and re-planning I finally had a system for real-life PoC (Proof of Concept) that I could use for testing all the aspects of my idea.
There are still some parts missing from the Rig, like display and another USB-hub, but with the current configuration it was possible to test the raw computing power and usage of the system to find out if there are huge problems, or not!

Here is the system under testing and some details about the components used and for what are they for:

1: 10/100mbit standard ethernet switch
2: 3G/4G/Broadband router with WiFi base station
3: USB3 HUB with external power supply (3xUSB3 & 4xUSB2)
4: External USB3 240GB SSD for full system clone/backup
5: RaspberryPi 3 for Remote access, Webcams & Power control
6: RaspberryPi Zero W, Remote keyboard and Remote access backup
7: Webcams for remote video feed
8: Lattepanda Alpha 864 Intel based SoC computer


Internet connection is deliverd via either 3G/4G mobile broadband or via other broadband connection (ethernet). The router has fully configured SOHO router system with firewalls, access control and other basic services.

The USB HUB is capable of hosting three USB3 devices and four USB2 devices. There is also USB2 & USB-C power output (no data) available in case there is a need to power some equipments, like fan’s, lights or so. In this PoC there is only one HUB connected, but in the final configuration there will be two of them.

The external SSD drive is mainly for full clone of the Lattepanda Alpha Windows 10 system as when this system is finally installed to the remote location there is no easy physical access to the system, so if something bad happens one must be able to remotely restore the system – by having a full clone of the system it is possible to boot from that clone and fix the main SDD with fresh copy of the system.

The RaspberryPi 3 handles the most important part of the remote access. Via it all connections will be opened so that one can remote access and control all the local computers safely via Internet using SSH. The RaspberryPi 3 also handles the power supplies of the USB HUB’s and Lattepanda Alpha with the Relay Card installed to it.

The RaspberryPi Zero W handles the remote keyboard functions for Lattepanda Alpha. For full remote control one must be able to enter commands via keyboard even if the system is fully halted – or access to Lattepanda BIOS is needed. With the RaspberryPi Zero W it is possible to mimic external keyboard connected to the Lattepanda Alpha and because it is a linux computer one can SSH into it and remotely give commands as “a remote keyboard”. Very crusial for this project as one must have access to the Lattepanda Alpha what ever happens.

Four webcams are for “remote eyes” to the system.
Using VLC one can see the system remotely to check the status of the system/led’s and also to see the screen of the main computer enough to handle possible system problems when there is no remote access available to the main computer via RealVNC, Anydesk or Microsoft Remote Desktop (RDP) – even if the connection to the Windows 10 is lost because a BSOD, Windows crashed, system down for other reason or whole computer halted while booting.

And the main computer, a Lattepanda Alpha 864. A very powerfull SoC (System on Chip) computer capable of handling the load of a new MacBook Pro. It has a Windows 10 OS installed as that is the most supported OS with SDR softwares and devices currently available.

Specs for LPA864 are: