How to Reduce Noise When Measuring 20MHz Frequency Using Relay

Thread Starter

wertghrtsgr

Joined Mar 14, 2025
5
Hello everyone,
I'm currently making a test machine for testing (voltage, frequency) PCBs. There are 5 test points on PCB where I need to measure frequency 20MHz. So, I'm using relay to switch around test point and connect to the probe of my oscilloscope (Keysight DSOX1202a). I've found out that relay causes a lot of noise and I'm wondering if I could get help on how to reduce noise and measure 20MHz with stability.
I've tried using SSR type relay, but noise still occurs.

Do you guys know any other relay types that are stable for measuring 20MHz or something else that I could use instead of relay?

Thanks in advance..
 

Thread Starter

wertghrtsgr

Joined Mar 14, 2025
5
We've tested by directly connecting test point -> relay -> probe
and found out that relay is causing the noise somehow. there's more noise when I connect multiple test pointsfor testing frequency.
As you can see in the picture, I'm getting 20MHz which is what I'm supposed to get, but when there's a noise, the value drops significantly and even loses connection (shows No Signal). The values are basically keep on changing
and do not maintain 20Mhz with stability, and that's why we concluded that relay is causing noise.KakaoTalk_20250315_114513157.jpg
 

Thread Starter

wertghrtsgr

Joined Mar 14, 2025
5
I'm not able to test this right now. I'll reply this as soon as I test this.
However, are you able to assume anything about this issue for now?
I really appreciate your help.
 

BobTPH

Joined Jun 5, 2013
11,463
approximately 4 meters
I'm at home so I cannot take a picture of the test set up right now.
Hopefully, you are using shielded coax for the connections, and correct terminations, right? If not, the noise is likely a combination of reflections and and antenna pickup.

Edited: I have no expertise in handling 20MHz signals, but I do understand a bit of the physics here. It may be necessary to tune the length of your cabling to avoid distortion due to reflections. The wavelength if 20MHz is 15 m, so your 4 meter cabling is significant, with a round trip being about half the wavelength.
 
Last edited:

schmitt trigger

Joined Jul 12, 2010
2,027
Too many unknowns in the problem description.
Besides the obvious coax cable, what sort of relays are used, make and model? Switching board layout? Source and termination impedance? Any power cables running parallel? Ground loops?
Etc, etc, etc.
 

MisterBill2

Joined Jan 23, 2018
27,159
Post#11 is totally correct!! What it looks like is some other frequency being picked up by some inadequate shielding, OR the AC voltage powering the relay coil being picked up by the signal wiring. That is a common problem in small signal switching systems. So attention to the coil supply voltage may be in order.
 
Top