I have been working on a commercial MSF 60kHz receiver PCB connected to a PIC with an LCD display running on a linear wall wart.
I noticed that it worked all night but then stopped receiving early this morning - no pulses from the receiver at all.
I tried an MSF clock that I have and that could set the time so the transmission is still there.

After a lot of fiddling about, moving the aerial, LCD, etc. I discovered that if I touch either supply line the receiver springs to life. So I bunged on extra supply decoupling wherever I could - makes no difference. I tried it on a battery and it works fine.

Given that it is a linear supply I was at a loss - especially as it worked solidly all night.

Then the sun went behind a cloud and it started working again. Then the sun came out again and it stopped. Repeat, consistently.

Now I know that radio reception is influenced by day and night but I don't think that applies to 60kHz and definitely not by a local cloud!

Then it occurred to me that I have solar panels on the roof connected to an inverter feeding power back into the mains and that that must be the source of the problem.

It can't be radiated interference as that would still be a problem when running on the battery so, presumably, as the extra decoupling didn't help it must be common mode getting through the interwinding capacitance in the wall wart. Touching one of the supply pins with my finger puts an extra 'load' on the common mode interference and reduces it enough to remove the problem.

The questions:
1. Do you think my conclusions above are correct?
2. How do I stop this without putting my finger on it?

 

Your analysis sounds OK to me. I agree, a cloud should not affect reception distance, but day/night definitely does at 60 kHz.

Is it a homemade receiver or one of the commercial ones?

John

Edit: Here's a link to reception distances for WWVB on 60 kHz in the US: https://tf.nist.gov/stations/wwvbcoverage.htm To get daytime reception, I needed to build a much better (100 ft) shielded loop antenna. Also, there are lots of interference sources, including my TEK TDS210 scope's power supply.

 

The receiver is one of these: https://www.ebay.co.uk/itm/WWVB-60k...e=STRK:MEBIDX:IT&_trksid=p2057872.m2749.l2649

I might also mention that I am close to the transmitter, 17 miles by road says google maps, so signal strength should be excellent

 

The sun is now moving away from the panels and the generated power has been dropping. There were some inconsistent pulses and then pulses every second but by eye on the LED they were of varying widths. As the power dropped, the pulses got better and at the moment they are being decoded correctly. I think that confirms the source of the problem and any experiments will have to wait until tomorrow (assuming it's sunny tomorrow!).

 

hi A,
Have you tried connecting the 0V to Gnd via say a 10k resistor.
When you are touching it with your finger you maybe getting grounding.???
E

 

I am 1200 miles from WWVB. If I were as close as you are, I could light LED's directly with my antenna. And, what you show seems to be identical to a receiver I bought from a supplier in the UK. It seems to work at least as well as the older version.

 

hi A,
Have you tried connecting the 0V to Gnd via say a 10k resistor.
When you are touching it with your finger you maybe getting grounding.???
E

By ground you mean mains earth?
I will try that tomorrow sun permitting.

 

What is there between your PV array and the battery? Any sort of charge controller that isn't a bang-bang (hysteritic low-rate ON/OFF) type can generate lots of RFI. Some of that stuff scrimps on filtering and internal conducted RFI becomes radiated when it leaves the box on the input and output cables.

I'm a long way from WWVB, so I generally can't get reception during the day at all, but I've noticed that moving a clock a metre can determine whether it is able to receive or not. You're so close to the transmitter I'm surprised you'd have any problems.

As I recall, the chippies for those receivers have an RSSI output that might be helpful in debugging. It would likely be useless for discriminating in-band noise from signal, but might give you some clues with regard to what the touching is actually doing.
Small amounts of strategically placed inductance can be much more effective than just capacitance for dealing with crud on supply lines.

===
I had intentions of building a clock I could connect to my LAN using such a receiver, but gave it up as a lost cause because I couldn't find the receiver chips anymore. GPS receiver modules are unbelievably cheap and perform incredibly well, but they are power hogs compared with the 60 kHz receivers. NMEA sentences aren't hard to parse and many of the modules are programmable so the sentences contain only what you want.

 

What is there between your PV array and the battery? Any sort of charge controller that isn't a bang-bang (hysteritic low-rate ON/OFF) type can generate lots of RFI. Some of that stuff scrimps on filtering and internal conducted RFI becomes radiated when it leaves the box on the input and output cables.

This an online system so no battery. A large box in the attic takes the DC from the panels and converts it to 50Hz AC sine back into the mains supply. I have no idea what is inside the box.

The receiver doesn't have an RSSI output.
Its signal output is low when the carrier is present and high during the gaps in transmission.
When the interference is present this output is low continuously.

 

Could the inverter on the output of the solar panels be generating interference at 60 Khz ? The inverter would probably shut down when the panels were in shade.

Les.

 

It seems to be definitely linked to the panels generating power so it is likely to be the inverter generating interference. The forecast is for sun again tomorrow so will get the scope measuring between mains earth and the wall wart output and see what I see. Then I can try the 10k resistor to earth suggested above. The wall wart has a plastic earth pin so there is no earth available there.

 

I've seen wall warts that are really badly designed and assembled. I won't use anything that plugs into AC mains from China unless I buy it from a reputable importer/distrubutor and I'm convinced the safety agency markings are genuine and not counterfeit. Amazon "fulfills" orders for dangerous goods sold by their marketplace vendors.

A decent switching WW will have an RF capacitor connected between one side of the AC line and one side of the DC output. The purpose of this is to provide a local path for RF junk on the output to return to the primary side, which is usually responsible for generating most of the RF in the first place. This same capacitor can propagate trash on the AC line to the DC side - it depends on what sort of filtering is present on the AC input. The filters are designed to keep the noise inside the WW, not keep noise out. Of course the filters do help with the latter, too, but sometimes filtering adequate to meet regulations is a simple choke in only one side of the AC line (plus a cap or two). It isn't unusual for a good filter to take up a third of the circuit board, so they usually aren't any more generous than absolutely necessary.

If you have some small (1 - 1.5 cm OD range) high permeability ferrite toroids kicking around (doesn't everyone? or am I odd in that I have a few hundred in assorted sizes and materials?) a few bifilar turns can make a common mode choke for the DC output of the WW. A ferrite "bead" intended to go over a video or data cable or the like can also be used. 3 or 4 turns on something like that might be sufficient.

 

The wart I am using ws made in the UK, is linear and regulated. It feeds another regulator on the PIC board.
Remember that the interference isn't being generated by the wart as everything is just peachy as long as the sun isn't shining.
I will try a ferrite tomorrow. The forecast says sun again

 

Today it is sunny and the receiver is on strike.
Three turns round a ferrite ring makes no difference.
Connecting mains earth directly or via 10k makes no difference.

I can hold the insulated output wire of the wart in my hand to fix it!
I have ordered a 22mH common mode choke which should give very good attenuation at 60kHz. It is due to arrive Wednesday or Thursday.

[EDIT] I have also tried a variety of warts, both linear and smps and all give similar results.

 

What, you're not willing to stand there clutching the wire? You shouldn't have been so quick to declare Igor was redundant.

It's weird, but it does sound like conducted noise is probably the issue. When you touch the receiver or hold the wire you are probably acting as an antenna, increasing the signal to noise ratio.

Iron core transformers can couple noise due to inter-winding capacitance. Transformers with the primary and secondary wound in different places on the core (rectangular "ring" core with 4 separate windings is common) are better, but I've never see such in a wart. I have seen some with split bobbins on EE cores where one end of the bobbin is the primary and the other the secondary, but you still have capacitance between them.

Noise has been the bane of my existence. On one occasion, if I'd had a valid passport at the time I'd have been sent off to troubleshoot noise issues in ultrasonic pipe weld inspection equipment on a ship in the North Sea at Christmas. Someone else went & I coached him via telephone. He got the problem solved, then very shortly after that they broke the ship because of failure to drop the pipe and run from an approaching storm, It cost several million to fix the ship (called the Solitaire - there used to be info about it on the web - refitted grain superhauler; it had had some sisters but they broke in two and sank). Many people have been killed because storms hit before the pipe could be dropped.

 

Morning Albert,
Did you manage to solve this problem.?

E

 

I am still waiting for the common mode choke which is due today or tomorrow.