I have replied to similar messages in the past. Peter¡¯s reply is right on. To get the best performance, higher bandwidth scopes with 1 Mohm inputs need the probe to be designed to counteract the effect of parasitics in the front end. Most of these effects are seen in the first 2 nanoseconds, so the practical breakpoint where you should really stick with the suggested probe is about 200 MHz.
The effect shows up as an overshoot and/or ringing in this area of a pulse after a fast clean transition. However, another rule I like to say is that passive probes are really only practical for looking at frequencies below 200 MHz, right at the breakpoint where you should use the suggested probe for the particular scope model. Regardless of how good the probe¡¯s performance is, frequency components greater than 200 MHz will probably be greatly distorted by the circuit under test acting on the severe loading the probe is imposing. The load comes from the capacitive element, not the resistive.
For lower frequencies below a GHz or so, the probe load can be estimated by calculating the capacitive reactance of the specified input C at the frequency of interest. (capacitive reactance, Xc, = 1/(2*pi*f*c) . So, for a high quality 200 MHz probe with an input C of 10 pF, the load would be about 80 ohms at 200 MHz. So ask yourself, would the circuit I am probing change its properties if I placed an 80 ohm resistor from the point I am probing to ground? Well, that is what you are effectively doing, but only for frequency components around 200 MHz. As the frequency goes down, the loading decreases until you hit the resistive limit, which is usually 10 Mohm.
So if you really need to see the high frequency components in the circuit, you should use either an active probe (which has less than one tenth the input C), or a transmission line probe. The latter has a higher DC resistance, but very little capacitance.
With higher frequency active probes (BW > 1 GHz), simple capacitive reactance may not represent the loading, as the inductive parasitics enter as well ¨C forming LC resonate circuits.
------
In general, I am not a fan of switchable 1X/10X probes. In my opinion, all of these l have crappy high frequency performance. The reason is that the attenuation switch is in the wrong place. Putting it in the probe head adds several parasitic elements in the sensitive very high impedance portion of the probe circuit. This would not be a problem if the switchable element were on the lower end of the attenuator, with the switch itself on the grounded side. But this would require the switch be placed in the comp box at the scope input, which is inconvenient for the user. Thus probe designers compromise and put the switch in the probe head, which really screws up the high frequency performance.
BTW, in case you are wondering, I worked as an engineering manager for 4 years in the probe design group at Tek.
- Steve
---In TekScopes@..., <valvesruleok@...> wrote :
All scope input attenuators/amplifiers have unique transmission characteristics due to parasitics that are (or should be) compensated for by the manufacturers recommended probe. A third-party probe cannot do this so will not be as accurate over the passband because it can only be designed for a theoretically perfect R//C input over a range of C. But how much difference this makes in practice would take quite some testing and gear to discover. For hobby purposes it's probably not worth worrying about.
Peter
