Thank you for taking the time to do this comparison and put it up. I bought several of those "generic" probes off Ebay a while back and was curious as to how well/not-so-well they worked. Considering the price, I fully understood what I was going to get. But it's still really interesting to see how they actually are, particularly from someone who knows what they are doing. It isn't critical for me, I use them only for very general troubleshooting type stuff on an old HP 54602B. But it's still an incredibly interesting bit of info to have. That was very cool, thank you.
One thing I was unable to see or mentioned is if the signal generator and/or the spectrum analyser have correct loading, a 10x probe cannot probe or terminate directly into a spectrum analyser or a 50 ohm signal source. The 500MHz probe was only scanned to 200MHz so there may have been some dipping further up the spectrum that could have shown a similar response.
This is not a valid test. The scope probe is designed to run into 1 meg Ohms in parallel ~15 to 20p, the input impedance of your spectrum analyzer it 50 Ohms. Adjusting it on your scope sets up the frequency response on your scope's input impedance, not 50 Ohms. You need a wide-band buffer amplifier with an input impedance like your o-scope that can drive a 50 Ohm load.
I wonder if the characterization of the low cost probes would have been a bit different (better?) if you'd been using a probe to BNC adapter since the ground lead may have been introducing some "error"; don't get me wrong, they performed about as expected but may have done a bit better with a different I/O configuration. Thanks for sharing, I have a 200MHz Hantek portable that has more capable probes as I recall.
I wonder if you play with the compensation you could flatten it out a bit more. I know you tuned to 1K but if you had a higher bandwidth square wave you might be able to tune the response a bit on the higher ranges.
He used the calibration signal wrong. You don't set the scope to show multiple periods of the signal. You only need the rising edge on screen. There is no need for a higher frequency calibration signal (in this frequency range), but you need one with very fast rise times which the calibration output of a scope usually has (because that's what it's there for). Also it is pointless to calibrate the probe on the scope and then connect it to a different device. You calibrate the probe for the input it is connected to. If you're serious about it you should even recalibrate the probe when moving it from one channel to another on the same device.
Not saying that it will change the results much, but you don't use the calibration signal like that. You don't set the scope to show multiple periods of the signal. You only need the rising edge on screen, especially if you are trying to be calibrated at higher frequencies. The important part is the fast rising (or falling edge) of the signal and the calibration output of a scope usually provides that (because that's the whole point of that output). Also it is pointless to calibrate the probe on the scope and then connect it to a different device. You calibrate the probe for the input it is connected to. If you're serious about it you should even recalibrate the probe when moving it from one channel to another on the same device.
Maybe remove that long cable and ground the tip. I assume your SA can be set for 1M input. You may need to compensate the probe for the SA. I would expect you to be able to switch the SA to 1dB/div. I'll post a sweep of my PP150 for you to compare. www.eevblog.com/forum/testgear/pp-150-scope-probes/
i noticed on the first probe when you showed it, you were using the ground lead/clip rather than the spring ground, did you try that and if you did was there any improvement?
Thank you for taking the time to do this comparison and put it up. I bought several of those "generic" probes off Ebay a while back and was curious as to how well/not-so-well they worked. Considering the price, I fully understood what I was going to get. But it's still really interesting to see how they actually are, particularly from someone who knows what they are doing. It isn't critical for me, I use them only for very general troubleshooting type stuff on an old HP 54602B. But it's still an incredibly interesting bit of info to have. That was very cool, thank you.
One thing I was unable to see or mentioned is if the signal generator and/or the spectrum analyser have correct loading, a 10x probe cannot probe or terminate directly into a spectrum analyser or a 50 ohm signal source. The 500MHz probe was only scanned to 200MHz so there may have been some dipping further up the spectrum that could have shown a similar response.
This is not a valid test. The scope probe is designed to run into 1 meg Ohms in parallel ~15 to 20p, the input impedance of your spectrum analyzer it 50 Ohms. Adjusting it on your scope sets up the frequency response on your scope's input impedance, not 50 Ohms. You need a wide-band buffer amplifier with an input impedance like your o-scope that can drive a 50 Ohm load.
Thank you for the explanation. Just ordered on eBay a China made P6500B 500MHz probe for USD24, just to try.
I wonder if the characterization of the low cost probes would have been a bit different (better?) if you'd been using a probe to BNC adapter since the ground lead may have been introducing some "error"; don't get me wrong, they performed about as expected but may have done a bit better with a different I/O configuration. Thanks for sharing, I have a 200MHz Hantek portable that has more capable probes as I recall.
A lot has to do with the little ground clip. Most of them are even crimped on wire.
I wonder if you play with the compensation you could flatten it out a bit more. I know you tuned to 1K but if you had a higher bandwidth square wave you might be able to tune the response a bit on the higher ranges.
He used the calibration signal wrong. You don't set the scope to show multiple periods of the signal. You only need the rising edge on screen. There is no need for a higher frequency calibration signal (in this frequency range), but you need one with very fast rise times which the calibration output of a scope usually has (because that's what it's there for).
Also it is pointless to calibrate the probe on the scope and then connect it to a different device. You calibrate the probe for the input it is connected to. If you're serious about it you should even recalibrate the probe when moving it from one channel to another on the same device.
Great video and very usefull, thanks!
I really love your work EU.
Not saying that it will change the results much, but you don't use the calibration signal like that. You don't set the scope to show multiple periods of the signal. You only need the rising edge on screen, especially if you are trying to be calibrated at higher frequencies. The important part is the fast rising (or falling edge) of the signal and the calibration output of a scope usually provides that (because that's the whole point of that output).
Also it is pointless to calibrate the probe on the scope and then connect it to a different device. You calibrate the probe for the input it is connected to. If you're serious about it you should even recalibrate the probe when moving it from one channel to another on the same device.
You have to tear apart the P6139A and the cheap ones to show us what is different.
I think a good tear down would be the new ikea led smart bulbs and battery chargers.
Maybe remove that long cable and ground the tip. I assume your SA can be set for 1M input. You may need to compensate the probe for the SA. I would expect you to be able to switch the SA to 1dB/div. I'll post a sweep of my PP150 for you to compare. www.eevblog.com/forum/testgear/pp-150-scope-probes/
Useful video
i noticed on the first probe when you showed it, you were using the ground lead/clip rather than the spring ground, did you try that and if you did was there any improvement?
hi can I use 150 MHz probe on 50MHz oscilloscope?
yes
@@electronupdate thank you...