Transmission Lines and Op-Amps

[hurliehoo]

My understanding is that the capacitance of a typical transmission line will increase proportionally to it's length. So in the case of a simple passive LPF circuit, consisting of just the LPF and the transmission line to the load (of infinite impedance), this would in turn mean the cut-off frequency would decrease as the line gets longer.

My question is this : what kind of op-amp could be best used to negate this effect so that the cut-off frequency would be independent of the length of the transmission lines capacitance?

*EDIT I'm guessing a voltage follower? Although not quite sure exactly how they work tbh...

 

[jsgruszynski]

Inductance is also increasing with length as well. A transmission line is approximated as an LC ladder. If the transmission line is operating as a transmission line then it will have a characteristic impedance which when interacting with the load at the end of the transmission line will be the Op Amp's effective load.

A voltage follower is often used with transmission lines specifically because transmission lines of 50, 75 or 300 ohms are typically the characteristic impedance. Of the 3 types of transistor circuits, a voltage follower has the lowest effective output impedance so it will likely match the load better. Strictly the effective Thevinen R of a voltage follower stage is ~Re + (Rb + Rpi)/(beta + 1). Voltage followers are common for transmission line drivers for this reason.

Most op amps have low output impedance already because they are designed to resemble "ideal voltage sources", which would have a low Thevinen resistance.

What may be at issue is loop gain and phase which can cause oscillation with capacitive loads. Normally there is a compensation procedure described by the manufacturer that can address this. Adding additional inductance is sometimes a work around if compensation does not work.

Be aware that if the gain-bandwidth product of the op amp is below your targeted operating gain-bandwidth, none of this will matter and you won't be able to push beyond it with any circuit tricks.

 

[Bob S]

As mentioned above, transmission lines look like an LC ladder. The most common impedances are 50 ohms (coax), 75 ohms (coax), 90 ohms (coax), and 300 ohms (twinax only). The propagation velocity ranges from about 67% to 90% speed of light, determined by amount of dielectric. If you do not terminate the coax in its characteristic impedance, some of your signal will be reflected. See
http://en.wikipedia.org/wiki/Transmission_line

 

[berkeman]

My understanding is that the capacitance of a typical transmission line will increase proportionally to it's length. So in the case of a simple passive LPF circuit, consisting of just the LPF and the transmission line to the load (of infinite impedance), this would in turn mean the cut-off frequency would decrease as the line gets longer.

My question is this : what kind of op-amp could be best used to negate this effect so that the cut-off frequency would be independent of the length of the transmission lines capacitance?

*EDIT I'm guessing a voltage follower? Although not quite sure exactly how they work tbh...

You've gotten good advice so far. But I have to ask, what is the application that you have in mind? I'm guessing that we can give you even better advice if we know what you are asking about. For example, op-amps do not generally do well driving capacitive loads (Quiz Question -- why?). And filtering and transmission lines do not generally interact with each other, unless there is a misunderstanding, or a novel use of transmission lines.

If you can post more about your application, we can probably help you more, and clarify how you should be thinking about the situation.

 

[TheAnalogKid83]

Aren't most opamp bandwidths insignificant to the frequencies where you really need to start looking at wires and traces as transmission lines rather than lumped models? Maybe tell us what cutoff frequency you're considering?

 

[hurliehoo]

Not sure exactly what I'll be doing yet, looking at a few different projects along the lines of power engineering (simple as poss for now lol - in fact any ideas there would be welcome but don't want to trouble everyone too much). Will be sure to post here when I get started and if I run into serious difficulties though, thanks guys you've been great help.

 

[Bob S]

I have gotten 1 nanosecond risetime out of RG-8 (50 ohms) with a 50-ohm termination on the end, so don't worry about cable capacitance. If everything is matched, the op-amp out sees only 50 ohms, not a load capacitance.. All you have to do is match the load to the cable impedance, and an amplifier can be made to match. Sometimes a resistor should be put between the output emitter and the cable to give a good reflected signal match. Is your load resistive? It is easy to put an emitter follower on the output of an op amp.

 

[TheAnalogKid83]

For example, op-amps do not generally do well driving capacitive loads (Quiz Question -- why?)

If I take a guess, will you give the answer? My guess would be because the low output impedance of the opamp combined with the capacitive load would cause some kind of over current and voltage clipping of the opamp, and in use with a feedback loop this might make the opamp try to compensate and act strangely before it reaches a steady state.