I have seen many spectrum analyzers with the warning on the RF input port: “!!! Avoid electrostatic discharge and DC voltage !!!”.

Out of precaution I always connected my signal source with a DC block but now I find the DC block filters my signal too much.

The signal comes from a line driver, there should generally be no DC voltage but I can’t completely guarantee that.

Since I have seen this multiple times I assume that this is something generic.

I understand that excessive ESD might be bad ... but DC voltage?

Does anyone know the reason for it and how seriously to take the DC voltage?

I mean, some amount of DC voltage (uVs) may always present.

Am I too over cautious? Is it safe to have a low DC voltage, say +/-1V on the signal?

  • \$\begingroup\$ I'm voting to close this question as off-topic because the answer is in the user manual or possibly printed on the instrument. \$\endgroup\$ – Vladimir Cravero Jan 8 '18 at 8:36
  • 1
    \$\begingroup\$ I think it's an interesting question. Why is it so picky, and why can't it block DC internally like any old scope? But it would really be good to know the exact model. \$\endgroup\$ – pipe Jan 8 '18 at 8:47
  • \$\begingroup\$ Since I saw this multiple times I assume it is something generic and hence my question. I updated the question to reflect that. I will also check my particular model, just in case. \$\endgroup\$ – divB Jan 8 '18 at 21:05
  • \$\begingroup\$ +1 You are definitely not being over cautious, DC voltages (even very small ones) are very harmfull to spectrum analyzers if they don't include an internal dc-block. The very sensitive input stages/filters of the analyzer are not designed to have dc voltages on them. I think it is an interesting question though and I would love for someone with more knowledge hereof to explain in great detail what typically happens when you put dc on the input of a spectrum analyzer \$\endgroup\$ – Vinzent Jan 9 '18 at 0:37
  • \$\begingroup\$ Spectrum analyzers are actually so sensitive to dc that you can destroy them even though you have a dc-block on; if the dc-block is big enough and you have something with a high enough dv/dt, this can actually cause something that looks like a small momentary dc on the input even though you have a dc-block in on it. That happened to one of my previous colleagues, he burned off two spectrum analyzers this way before he realized that his choise of a very big dc-block ment that a dc that was being swichen on coupled though the block to the input.. \$\endgroup\$ – Vinzent Jan 9 '18 at 0:46

in your example, the input power level is 10dBm max. A DC signal could be considered to be an RF signal at 0Hz. 10dBm into 50 Ohms is 0.707V RMS, so 1 VDC is above the +10dBm maximum. 1 VDC is +13dBm into 50 Ohms.

This huge signal at 0Hz can cause measurement accuracy problems with signals at other frequencies, for example compression due to front-end overload. To have DC and still have measurement accuracy, 125mV is probably the absolute maximum that I would ever put on the front end, and I would stay under 70.7mV to make a good measurement. But you should just use a DC block unless there is a good reason not to. The rating is 0 VDC for a good reason, so don't put DC on it. This is not generic advice.

I don't speak for Keysight Technologies.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.