power supplies and radio interference

[johnk]

From time to time I come across discussions where radio restorers [domestic and mil] want a low current valve-voltage (tube-voltage) power supply.

I often suggest that they look at the various Nixie supplies discussed here.

 

The usual response is along the lines that ANY “switching” or not-linear-regulator supply will cause too much “noise” on the radio bands involved.

 

Are there any active ticketed amateur radio guys here who have had first-hand experience of this topic and are willing to share their wisdom?

 

John K

Australia.

[David Forbes]

I work in radio astronomy, and I wear a Nixie watch. No problem. 

But seriously, it is possible to build a low-emission switching supply. It takes time and care and shielding. High voltage is more of a challenge. 

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/001501d553ea%24e0530ed0%24a0f92c70%24%40internode.on.net.

[johnk]

Thanks for the reply David – I thought of you when I posted.

I am keen to get some replies from HF listeners where particular “brand”/circuit supplies get commented on.

Any idea how your watch [supply] performs [when displaying] around HF comms?

 

John K

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAPbqtvdqOmNa8sMGF%3Du5A-DnCHpB33FjoBWFy%3DiSJCZfV1mxSg%40mail.gmail.com.

[Bill Notfaded]

You can always build a little tempest facility around it. Like a small Faraday cage.

Bill

[johnk]

Sigh, Tempest. "We" used to service some Tempest spec sites.
Btw, you need a security clearance to be able to access/read the spec too :-))
And, some of our guys just would not keep the spares cabinet locked.

I agree that shielding/filtering the supply is not impossible, but is so much harder than having to do nothing when it comes to the linear regulator supplies. Apparently the users aren't bothered by the level of noise from diode steps [eg zeners].

John K

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/897d7d9a-88e5-40d0-bbec-7bc9eae61da5%40googlegroups.com.

[Paul Andrews]

I have several AM and FM radios and I have not experienced any noticeable interference from the various nixie clocks. On the other hand, I have experi nc d quite a lot from various LED lights and laptops. All of it is typically short range (several feet/meters).

This article discusses a radio battery replacement (these batteries were 90V!): https://www.dos4ever.com/battery/battery.html

[jf...@my-deja.com]

My three B7971 clocks all produced RFI in the AM and SW radio bands, and I kept my radios at least 5m away.  They were based on the CT7001 clock chip, so the MUX rate was about 2 KHZ.  The RFI (and the acoustic singing from the tubes) seemed to change with the digits displayed.  I assumed that it could be blamed on the switching of the HV on the tubes.  Someone once suggested that the RFI came from the SMPS, but that is unlikely since I used old-fashioned linear supplies.   In hindsight, I realize that I did not put a choke on the power cord, which might have helped.  The moral of the story is that not all RFI necessarily comes from the power supply.

[gregebert]

Oscillations and switching signals are not enough on their own to create RFI. They have to be radiated. 

Shielding is a common way to reduce emissions (and also to reduce susceptibility to incoming interference). I always use twisted-pairs for my power cabling, and when I have to use ribbon cable for signals, I try to straddle the clock signals with a static signal, or ground, when possible. In extreme cases, which I doubt exist in nixie clocks, you need to use differential pairs.

Ground loops are another source of trouble, which is why you want to bring all grounds together at a single point. If you have ground carried thru 2 or more paths, you may have introduced a ground loop if the paths are not identical.

[Robert L]

Sniffing around with E- and H-field probes, the largest contributor I can find are  fast edges going over SPI lines on ribbon cable between multi-tube carrier PCB... quite obvious up to and beyond 1.5 GHz. I can see the digit fade activity on the spectrum analyzer... Noise for the first half second or so while cross fading between digits. Relative quiet the second half second. This with the probe right on the cable. All the noise significantly reduced just a small distance from the ribbon cable. Not much coming from the pcb proper with 4 layers including power and ground.

I'll add a small value series resistor where high rep rate signals launch out onto cabling... slow the edges for clock and data lines. 

All of this while PWM dimming the display tubes at 15 kHz or so. Lots of HV switching that doesn't seem to be a large contributor.

By *FAR*, the strongest signal I noticed in my shop was 10 MHz from a BNC cable running between my 10 MHz reference and a counter input.

Hmmm... All the clocks, computers and assorted electronic gadgets may be part of the S7 noise floor I observe on 40M... will have to see what happens with the clocks and computers turned off!

Bob

[Paul Andrews]

I was just playing with an AM radio in my bedroom. I found that it was picking up interference when it was trying to tune in to a weak station (which is most of them). One Nixie clock was a culprit, though I guess it could also be the wall wart I am using for that. Another was a USB wall wart.

There is also something spraying wide band clicking noises, but that is a different story.

[Bill Notfaded]

Yeah cheap switch mode power supplies can really be culprits.  Check out some of the EEVblog videos on it.  He shows it on a scope.

R,

Bill

On Sat, Aug 31, 2019, 6:55 AM Paul Andrews <pa...@nixies.us> wrote:

I was just playing with an AM radio in my bedroom. I found that it was picking up interference when it was trying to tune in to a weak station (which is most of them). One Nixie clock was a culprit, though I guess it could also be the wall wart I am using for that. Another was a USB wall wart.

There is also something spraying wide band clicking noises, but that is a different story.

--
You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/BMTi2VroZ5Y/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/d00b4796-fe11-4007-bb9e-f3db743e835b%40googlegroups.com.

[Jim KO5V]

HI John,

I just saw this posting, so sorry for the delay. I have a two-digit B7971 clock that I built from a kit about 15 years ago. It runs directly off of the 110V mains, so there is no switching supply involved here - just a transformer, rectifier and some filter caps on the clock's power supply board.

The clock displays the hour for one second, then the minute for another second. Before the hour is shown again, the clock does a "flick-flack" routine for another second. It makes two clockwise "rotations" by sequentially lighting up each tube segment that is connected to the center of each tube. There is a very annoying  a buzzing sound in sync with the flick-flack that I could hear on 40M and 80M.

I have been lazy, and just put the clock in a room on the other side of the house from where the ham equipment is kept. I have also improved my grounding and antennas, so I no longer hear the clock.

Maybe some day, I'll chase down the clock's noise, or just try a toroid on the clock's power cord (which is a long one - so I'm sure it acts as a decent antenna), but for now, I don't think the clock is as much of a problem as all of the cheap wall-wart supplies that power all of my wonderful electronic gear - my house's noise level is about S4. This stuff has all of the required stickers, but I think most of them mean absolutely nothing. I believe the manufacturers just stick them on without thorough testing, knowing that the FCC does not have the resources to do anything about it. Anyway, I have a lot of other noise to sniff out before diagnosing the clock.

Good luck. 73, Jim

[gregebert]

Linear supplies can generate switching harmonics  as the diodes turn on and off, but the harmonics taper-off pretty quickly so I'm doubtful they have much energy in the RF region (AM radio band starts at 530kHz).

I'd place my bets on the switching-on and off of tube segments, along with the associated wiring. Then the microcontroller.

[Bill Notfaded]

I was thinking more noise in general that you can see blatantly on a scope.  Another culprit in the videos that was really bad was caused by LED lighting.  True it's different if it falls in the HAM radio bands somewhere.

R,

Bill