Use a basic, manually tuned RF source, feed it to the OPS, check the forward path response with my dual trace CRO. One trace input reference, one trace output to see amplitude and phase.
Move to a faster DSO if I need to.
Not as flash as a network analyser, like Ovidiu Poppa, but should be adequate.
This is only 30 MHz after all, not some microwave project.
Some of the information should be clear well below 30 MHz.
Anyone with expertise in this area like to comment?
Best wishes
David
Hi,
Usually is not always, and it can always be done by
overdoing it, by definition. Generally less is more
in terms of compensation around the VAS than
a lot more which is poor for the inclusive case.
In terms of simple single pole comensation.
rgds, sreten.
Last edited:
Yes, not difficult.
Sure not as nice as an HP 89410A computerised Vector Network Analyser.
But not too bad.
Best wishes
David.
This touches a key point.
To the extent that a system is minimum phase then correction for Gain will inherently correct the Phase.
So the limit is placed by non-Minimum Phase Behaviour.
It looks possible to practically eliminate non-MP from the VAS/TIS.
Non-MP is typically small anyway, and can be substantially reduced quite easily.
There is no inversion in the Emitter Follower OPS so it should be close to MP.
I have previously asked about this but had little sensible response.
So I plan to measure it.
Once I have a few data points I can then do it the easy way and model.
Best wishes
David
How do you plan to decouple the output stage from the rest of the amplifier in order to ascertain its frequency response?
I have considered to build a prototype OPS as a separate board anyway, with short links to connect to the front end.
This was mainly for flexibility but would be convenient for connection to test equipment. Reminds me to put accessible test points on the board anyway.
Can always pull front end transistors if tests are not done before their installation. Worse case is cut a few traces and fix later.
Still need to maintain the quiescent current bias of course.
That depends on exact circuit details, at the moment I plan a push/pull VAS/TIS rather than a constant current source load so, off the top of my head, maybe just a couple of resistors to supply the bias circuit with current.
The test oscillator should drive that without problems I think.
Best wishes
David
A non-minimum phase system is one with right half plane singularities (poles or zeros).
The only stage in the Thompson topology which generates a right half plane singularity (a zero in fact) is the transimpedance stage (TIS).
The RHP zero is only of concern in amplifiers with a MOSFET TIS; otherwise, with a BJT arrangement, it is orders of magnitude removed from the unity loop gain frequency, and is, therefore, of no concern.
I am intending to work on an amplifier with a nest of major loop and localised two pole compensation (viz. so-called "TMC").
The amplifier features an existing and unusual topology for each of its three gain-generating blocks, but uses these gain stages in a novel fashion. Watch this space.
The amplifier features an existing and unusual topology for each of its three gain-generating blocks, but uses these gain stages in a novel fashion. Watch this space.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.