| Products > Test Equipment |
| Measuring Oscope Probe Bandwidth with an RF Sig Gen |
| (1/4) > >> |
| killingtime:
Hi, Bought some Oscope probes and wanted to measure their -3dB bandwidth. I know some people use a square wave generator and measure the edge rise time, but I don't have one, so I wondered if I could use an RF sign gen instead (sine wave) with a 50 Ohm o/p impedance and an oscope. Wind the frequency up until I see the voltage drop by x0.707. I've tried this using a probe to BNC adapter and I'm getting weird results. The plots don't look flat at all. Attached. Two probes. A genuine R&S 300MHz one and the new one. The RF sig gen (50 Ohms) has been tested on a spectrum analyzer. It's flat (into 50 ohms) across the freq range (+/- 0.5 dBm). No issues there. . The scope has a 300MHz i/p bandwidth. . The scope probe used as a 'control' has a 300 MHz bandwidth. 12pF tip capacitance. R&S RT-ZP03. Bought new. No reason to suspect it's faulty. . All probes set to x10. . I'm testing up to 200 MHz, so I'm well away from the -3db point on the mfr spec sheet of the control probe and oscope. . I'm measuring the peak to peak voltage displayed on the scope. . I'm aware scope probe tip capacitance will present a reactance that goes down as freq goes up, and that this will reduce the voltage across the probe given there's a 50 Ohm source resistance in the sig gen. I've run a sim in LTSpiceIV of the equiv circuit. Attached. The probe tip voltage is 20% down at 200MHz on account of resistive divider circuit formed with the sig gen o/p impedance. I think this is why the plots aren't that flat. Even before we get into bandwidth limitations (probe design, oscope front end etc) a lot of amplitude is lost to the sig gen o/p impedance. Would it be as simple as just subtracting the theoretical divider loss from measured Vpk-pk results to get the real probe -3dB bandwidth (Vpk down by 0.707)? Something else I've noticed is that the probe tip capacitance isn't constant. Not too sure on why though. I've swept a nanoVNA into the probe front end and I get 12pF back on the smith chart but it goes up with frequency to about 25pF at 175 MHz Thanks. |
| TimFox:
The usual rule-of-thumb that RiseTime = 0.35/Bandwidth is not true in general for oscilloscope amplifiers, but is commonly used. Measuring the bandwidth directly is accurate, so long as you know that the output of the generator is truly flat. Even then, there will be an unwanted lpf at the input, with a time-constant given by 25 ohms (50 ohm source into 50 ohm termination) x Input Capacitance (typically, 15 pF), or 424 MHz, that will have a small effect. |
| tautech:
A probe is never part of a scope BW test system. 50 ohm source terminated into 50 Ohms via coax, sinewave @ 1V and BW -3dB = 0.707 on the display. |
| noisyee:
Hi, You would need a test fixture to proper test probe BW. It's an easy one with only a 50 ohm termination resistor and proper transmission line (or even a 50 ohm resistor bridging at a connector pin). Better measure its S11 before use. Don't be surprised if the result is weird. High BW passive probes have quite complex compensation network to compensate cable reflection and FR. The compensation network will only fit particular scopes. I had tested a claimed 500 MHz BW probe only have <200 MHz BW on my scope. Some hack needed to adjust the HF compensation (it has adjustable HF compensation but sealed), but the FR never become flat enough. That's why scope manufactures recommend particular probe for particular scope. If you don't use it, it's your own responsibility to verify the system performance. As for probe capacitance, manufactures specific it at LF. If you are interested in HF property, use a VNA to measure its S11, you can easily converted it into impedance curve. The curve should be closely but not perfectly match to an ideal capacitor. The impedance variation maybe the cause to non-constant capacitance measured on the smith chart. |
| joeqsmith:
One thing you can count on, all scope probes load your circuit. https://www.eevblog.com/forum/testgear/12-ghz-active-probe-project/msg4973839/#msg4973839 |
| Navigation |
| Message Index |
| Next page |