How to correlate Spectrum analyzer dBm with Oscillosc mvPP?

[JohnG300c]

mvpp to dbm

I'm trying to determine if my test instruments (bought on Ebay) are accurate. I'm feeding a signal from my Marconi 2022 signal generator into a HP 8590B Spectrum Analyzer and the generated and read out amplitudes are not the same. When checking the amplitude with my TDS-220 Oscilloscope the peak-to-peak voltage on the scope is nowhere near the value displayed by the Spectrum Analyzer!

Various resources on the internet shows how to convert dBm to volts (50 ohm load) but i am reading much higher values on the scope (see for example http://ifmaxp1.ifm.uni-hamburg.de/dbm.shtml).

I'm hoping that someone can shed some light on how i should correlate the readings on the Spectrum Analyzer with what i see on the Oscilloscope. I'm using pure sine waves.

Thanks,
/John.

 

[khouly]

convert oscilloscope to dbm

what kind of ur oscillator probe , is it active probe ?

khouly

 

[JohnG300c]

oscilloscope to dbm

No probe. I'm just using a straight 50 ohm BNC cable between my Marconi 2022 Signal generator and the spectrum analyzer's input. I am also using a T-connector to tap off the signal to the oscilloscope.

I believe the question is: "Should the amplitude measured on the oscilloscope be the same as measured on the spectrum analyzer"? Perhaps this is NOT so since the spectrum analyzer perhaps displays the energy in the signal 'spread out' which could explain the much lower reading (the spectrum analyzer shows around 2mV for a peak-to-peak signal of around 200mV on the scope).

/John.

 

[abc123]

converting mvpp to dbm

see the previous post--

 

[PaulHolland]

cable vswr ripple

Hi, You should read -10dBm for 200mVpp input @ 50 Ohm. and errors of 3dB +/- are normal.... Frequency reading error should be small, check this. and make sure your settings are correct. Also NEVER CONNECT a device directly on your RF IN port since one small error and your analyzer is dead !. Always use an 30dB attenuator and if needed a 10dB... buy them too at e-bay ... Analyzers are NO power meters for that you have to buy a power meter !. Just like a scope is NOT a voltmeter.

 

[Mazz]

spectum analyzer dbm

Try this:
connect Marconi directly to Spectrum (no other things than cable) and read.
connect Marconi directly to Scope (no other things than cable), set Scope input impedance to 50 Ohm (I hope the one you have can do this) and read.
Compare the two readings using the formula dBm -> Volt.

I hope this can help.
Mazz

 

[biff44]

dbm peak formula 50 ohms

I tend to agree with the jist of the last post. You have to make sure the oscilloscope is set to have a 50 ohm input. If the measuring oscilloscope has a high Z input, then the actual voltage measured on the scope face will not correspond to the value displayed on the Marconi signal source.

That said, it is not unusual for a spectrum analyzer to have a blown-out front end. Sometimes they work but with 10 dB or so of excess loss. Sometimes there is a lot of VSWR ripple, so as you sweep frequency the spectrum anlyzer shows varying power levels even though you know your signal source is flat.

 

[JohnG300c]

spectrum analyzer dbm, voltage

Thanks Everyone. I will do some more research tonight. I just picked up a new RF Power Meter (mW meter) on Ebay so i should be able to use it as a reliable reference.

And of course! My scope is high-Z so the signal is not properly terminated which likely results in the high voltage seen on the scope. I will do some experiments with a T-connector, a 50 ohm terminator to see if this helps (transmission line stuff my dear watson).

 

[biff44]

spectrum analyzer dbm

Yeah, a high Z scope can give you a lot of errors. First of all the marconi might put out way too much power, because it is not loaded down with 50 ohms. But the Marconi internal power meter may not realize this.

Then at the end of the coax line running into a high z load at the scope, theoretically the voltage there will double (you have a forward wave, and a reverse wave or equal amplitude and the same phase on the line, so at the termination open circuit, they linearly add to double).

Put the T junction and a 50 ohm load on the unused port and you will get a true reading.