Power Supply Considerations
I have put about 10 KiwiSDR into service over the years and here's a few thoughts on PSU
- "wall warts" are varied, as there as older linear ones, simple chargers, switchers with actual filtering for RFI/EMI
- assume warts to be suspect and test them thoroughly for both radiated and conducted noise
- Linear PSU are almost always a safe choice but are inefficient and their power consumption impacts your electric bill and any UPS run-time
- Industrial Switcher Units are usually quiet and usable
- I am using Meanwell DIN Rail mounted PSU in a few installations and they are available on Amazon, reasonably priced, and appear to be quiet.
Comments
I used switched mode supplies nearly all the time these days.
The only problem I occasionally find is that despite lots of filtering it is very difficult to entirely remove the switching noise on the VLF bands, typically <100 kHz that is produced by older designs.
More mosern deigns with switching frequencies in the region of 1 to 2 MHz are a lot easier to filter and are also generally more efficient.
Regards,
Martin
After starting a new receiver on short waves, I received interference from a switching power supply. Partially they were reduced by a power filter in front of the 220 volt outlet. I'm thinking of using a 12 volt transformer power supply and a 5 volt switching converter with a high frequency of 1.5 MHz. I remember someone already sharing this idea here on the forum. But I can't find this thread. Can someone remember? Thank you!
Search the forum for MP1584
Regards,
Martin
Martin, this is exactly what I was looking for. Thank you for experimenting with the power supply!
In most of the ready-made modules, a 100kOhm resistor is installed on the MP1584, which sets the frequency of PWM operation. You can reduce its resistance and raise the frequency from 1 MHz to 1.5 MHz. Perhaps this will further reduce interference.
Regarding discussion on whether the switched power supply is noisier than a good old transformer. A softrock style receiver was noisier on a transformer DC power supply than on an old DANCALL switched power supply made for NMT phones. The explanation is simple: The mains is noisy from all the switched power supplies around the house. There is a filter at the mains side of the DANCALL switched supply that stops the noise coming through to the DC side. Thus a mains filter is likely a good idea even with a transformer type non-switched wall wart.
73, Vojtech OK1IAK
Back when I purchased my Kiwi, I decided I wanted the quietest power supply possible... so I built my own! :) It's 100% linear/analog, dual regulated. It has an AC input filter. The 16V from the transformer is regulated down to 12, which is further regulated down to 5. Unfortunately, my neighborhood has a LOT of RF noise... even on a battery-powered portable radio, I hear a significant amount of static on a large swath of frequencies from 200KHz well into the upper SW bands. On my Kiwi, it shows WIDE swaths of S7+ noise levels.
As for setting a new supply to run at 1.5MHz... that would block out a semi-local AM station I listen to now and then. :(
Do you have any more info on this? Schematic, parts list, pics, or anything like that?
Interested in how you handled some of this, like the double regulating, etc.
-Nate
N8BTR
As Vojtech, OK1IAK, has already commented, the power supply itself is only part of the equation.
Filtering the DC output to provide additional common mode isolation from signals on the AC side of the PSU also seems to be an important factor.
The 'loop' that is formed between the AC mains 'ground' and antenna 'ground' often acts as an additional noise pickup antenna, which ideally should be avoided.
If you terminate the RF input of your KiWi with a 50 ohm screened load, and can't see any significant noise across the spectrum, then the power supply is likely to be quiet enough.
If you then touch just the outer screened connector on your antenna cable, to make contact with the outer screen of the 50 ohm load, and you see noise appear, then there is likely to be a conducted noise and ground loop issue.
Regards,
Martin