I've encountered an issue on my bench where something is creating intermittent interference at exactly 20kHz and 25kHz, strong enough to be picked up by a high-gain input stage, even with a 68k grid stopper right at the terminal. My first thought was some nearby SMPS, but the only SMPS devices on my bench are a Fluke 45 and a Hantek DSO4072 - the rest are all linear. I'm about 10 feet away from a 5kW solar inverter, but I'm still seeing the interference when that's quiescent. There's also a transmitter for an invisible fence system, but unplugging that doesn't affect it. So I'm stumped.
Besides, aren't 20kHz/25kHz awfully low frequencies for SMPS? What kind of device might be behind this?
Besides, aren't 20kHz/25kHz awfully low frequencies for SMPS? What kind of device might be behind this?
25kHz isn't too low for a switching power supply.
I tracked one down a couple of months ago that was at 28kHz.
It was causing some detectable interference with a (very simple) circuit I was playing with. The circuit was a voltage divider attached to a 9V battery.
I was trying to get a look at the noise generated by the voltage divider. What I found was 28kHz interference from a powersupply plugged in under the workbench. Just the oscilloscope probe and a couple of resistors was enough to pick it up. I ended up with just the two resistors in series connected between the probe tip and the probe ground, and could still detect and trace the interference.
Once I unplugged the powersupply, I found interference at 80kHz from the monitor above the workbench. Much weaker, but still detectable with just the resistors.
Lots of things can cause that kind of interference.
You can try to track it down with your oscilloscope.
Make a coil of about 15 meters of 22 gauge wire. Make the coil about 10 centimeters across. Put a 5nF capacitor in parallel with the coil. That should be resonant somewhere around your 25kHz signal. (That's if I didn't bungle the math somewhere.)
Clip your oscilloscope probe and ground to each end of the coil.
With your oscilloscope set to its most sensitive range (2mV per division for your Hantek DSO4072,) wave the coil around your work area and look for places where the voltage is higher. Move the coil around your equipment and around the power cables. Even the lights can be the source.
The place with the highest voltage should be closest to the source.
-----
I suggest using the resonant coil because the DSO4072 isn't as sensitive as my old scope. My scope has a 100 microvolt per division setting - I can track really weak sources of interference with it. My ancient Telequipment D43 is about 20 times more sensitive than your Hantek.
I tracked one down a couple of months ago that was at 28kHz.
It was causing some detectable interference with a (very simple) circuit I was playing with. The circuit was a voltage divider attached to a 9V battery.
I was trying to get a look at the noise generated by the voltage divider. What I found was 28kHz interference from a powersupply plugged in under the workbench. Just the oscilloscope probe and a couple of resistors was enough to pick it up. I ended up with just the two resistors in series connected between the probe tip and the probe ground, and could still detect and trace the interference.
Once I unplugged the powersupply, I found interference at 80kHz from the monitor above the workbench. Much weaker, but still detectable with just the resistors.
Lots of things can cause that kind of interference.
- Your phone
- The monitor on your PC
- The PC
- A phone charger plugged in to an outlet somewhere in the area
- The oscilloscope
You can try to track it down with your oscilloscope.
Make a coil of about 15 meters of 22 gauge wire. Make the coil about 10 centimeters across. Put a 5nF capacitor in parallel with the coil. That should be resonant somewhere around your 25kHz signal. (That's if I didn't bungle the math somewhere.)
Clip your oscilloscope probe and ground to each end of the coil.
With your oscilloscope set to its most sensitive range (2mV per division for your Hantek DSO4072,) wave the coil around your work area and look for places where the voltage is higher. Move the coil around your equipment and around the power cables. Even the lights can be the source.
The place with the highest voltage should be closest to the source.
-----
I suggest using the resonant coil because the DSO4072 isn't as sensitive as my old scope. My scope has a 100 microvolt per division setting - I can track really weak sources of interference with it. My ancient Telequipment D43 is about 20 times more sensitive than your Hantek.
Last edited:
Not for IGBT high voltage inverters, no, 20kHz is actually very high for these. IGBTs are slower than MOSFETs (but much more robust at high voltages). If its precise it sounds like crystal-derived which won't be a normal small SMPS, very likely to be solar inverter - queiscent clearly isn't the same as off!
My shop computer is about 15 feet from my workbench, and it can radiate enough interference over to my bench whether it's booted up or not.
So I simply disconnect its power cord when I'm doing some testing.
So I simply disconnect its power cord when I'm doing some testing.
- I pur ferrite cores on every wire goes to/from my PC and now I don't see any harm of it. Neither at 1 meter from PC, nor at ARTA measurements with this PC. And noise floor of sound card became super low. (But besides I have a good ATX power suply in it).
If in short - now I started to put ferrite cores on almost every power wire (or long > 1 m signal wire) I see.
Last edited:
Ferrite cores on wiring do not always remedy some radiated interferences.
It all depends on the type of interference, and the equipment involved.
In some cases it may only minimize radiated interference, in others the interference is unit-based, radiating from the actual (unshielded) circuitry.
It all depends on the type of interference, and the equipment involved.
In some cases it may only minimize radiated interference, in others the interference is unit-based, radiating from the actual (unshielded) circuitry.
As ripping a DVD-A "Hotel California" 96khz, I asked some studio man's, the response was signal from the neon lamps.
Also, as measure the high res free given wave files, here too. Contacted the owner, and origin of the signals unknown.
Measure at first, before recording, would be a to do task
Hp
How are you seeing the tones? Via FFT with a sound card? If so, could be 'idle tones' in its ADC (though being exactly 20k and 25k would seem to make that unlikely).