Awesome troubleshooting! Thanks for sharing it.
Have you contacted your elected officials and asked them to ban the
import of those junk supplies?
Hams won't be the only services impacted - the military and others have
an enlightened self-interest
in stopping this problem from multiplying.
Just a thought.
On 3/24/26 05:27, vk3od--- via RFI wrote:
My HF operation was halted recently by heavy HF interference on all HF ham
bands.
I tracked the interference down to a neighbor 750 meters away. Lucky for me we
were on good speaking terms.
After investigation and H field probing, we determined that it was his bank of
Bitcoin mining GPU power supplies that had no filtering that was causing the
interference. The WD8DSB loop made finding and pin pointing the noise very easy.
10, 12, 17 and 20 meters had S7 birdies across the bands.
15 meters was the worst effected and appeared to be the resonant frequency of
the interference. S9 plus 20 DB of oscillating birdies and hash.
All the low bands had S9 birdies every few Khz that were drifting around.
I built up some test filters. 2 X #31 FT240 cores, 17 turns on one and 7 turns
on the other. I also used a inline commercial EMI filter as per Don WD8DSB and
the W9RE Interference case success story.
The result was rather spectacular. It killed the interference almost entirely.
I used a Chinese EMI filter with documented specs. It appears that most of the
EMI filters including brands like Schaffner are now made in China and the
savings are really good ordering directly from China.
I did measure the performance using the TDK App note procedure. I verified the
filters on the VNWA. The filter is a genuine common and differential mode
filter with attenuation greater than 50Db in both modes of the conducted
interference paths.
The only issue is that I will need to build 10 to 15 of these filters since
each GPU power supply runs anywhere from 900 watts to 1300 watts when its
operating at peak performance.
I worked out and compared the economics of buying Type #52 FT240 cores which
were double the cost of Type 31 cores from my distributor. I will need series
chokes so may chose to use to 7 turns on 2 cores and 7 turns on 1 core as per
K9YC using #31. The G3TXQ Type 52 filters looks very impressive wound on 3 X
Type #52 cores that will just about cover most of the HF spectrum with high
choking impedance. It will however double my cost.
It appears that these compact bitcoin GPU mining power supplies are being built
to be compact and they are leaving out the filtering. The small bitcoin rigs
are using smaller PC power supplies with filtering in them. My neighbor has
agreed to contribute to the cost of all the hardware. That is so kind of him!
This type of interference is very similar to the grow light ballast
interference fingerprint I will also be trying filters similar to those that
W0IVJ and W0QE have built.
https://www.w0qe.com/RF_Interference/grow_light_electronic_ballasts.html
https://www.w0qe.com/RF_Interference/grow_light_electronic_ballasts.html
http://tomthompson.com/radio/GrowLight/GrowLightBallastFilter.html
http://tomthompson.com/radio/GrowLight/GrowLightBallastFilter.html
Interference from Bitcoin rig power supplies is similar in intensity to the
grow light interference and it might be a bitcoin mining operation that is
causing you trouble rather than grow light interference.
A lucky day for me.
73
Craig
VK3OD
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|