Gentlemen: thanks for your posts; useful and informative. A lot to digest - but that's just fine!
Aside.
Now with all due respects to Paul Stenning and the mod. team, I was a bit doubtful that I would receive any responses to my OP, since - in my opinion - my post is a 'bit on the fringe side' of topics for a vintage radio forum. But I suspected that if I
did receive any such replies, it would be from you two guys!
Quote:
Originally Posted by G0HZU_JMR
. . . a step attenuator and if you can get hold of a healthy Hatfield 2105 attenuator that would be a good tool to use.
|
I hadn't thought of that.
But I do have such an item - and an HP 335D switched attenuator as well. So, time to make some more measurements . . . see further down . . .
Quote:
Originally Posted by G0HZU_JMR
. . . do it the way I did it decades ago at 10MHz using metal film resistors as the test loads. e.g. a 100R resistor has a 2:1 VSWR or 9.54dB return loss (and so on). This method is slow and a bit tedious but it would only take half an hour assuming you have an E12 or E24 range of resistors to use as test loads across a few ohms through maybe 1000R.
|
Which is exactly what I did between my OP and now, and plotted a graph of Vout (d.c.) against various mis-match loads (40Ω to 65Ω). Energizing signal was my HP 8640B: about 700 mV r.m.s. I had previously determined that the linear dynamic range of the 'RHO-Tector' was from about 500 mV to at least 1000 mV (8640B voltmeter). The 'RHO-Tector' claims a B/W from 500 kHz to 2.5 GHz at 0.5 watt. (Is this the device max. internal dissipation or max. power throughput?)
I repeated same tests at different freqs. (same drive level): 20, 50 and 80 MHz. For loads < 50Ω, responses were same at all freqs., but for loads > 50Ω, there was an increasing reduction in the V d.c. out as freq. and magnitude of mismatch was increased (140 mV as opposed to 150 mV - which is only a 7% error, and at high mismatched loads too). I suspect that could have been due using metal oxide Rs and / or 'invisible' impedances at the BNC sockets where the mismatch loads were soldered.
I also found - when plotting the graph Vout vs. load R ( and consequent RC & RL) - a few points which were well away from the implied curve (which looks approx. parabolic). In one case, replacing the load R by another exactly same, made all the difference and produced a 'believable' result!
But not so in two other cases: so I ignored those 'erroneous' points when I sketched in the resultant curve.
Last measurement was same as above, but over a smaller range of load Rs.: 46Ω to 54Ω. That range is probably appropriate for what I hope to use this RHO-Tector for. Vout d.c. was 1 mV max., so some d.c. amplification will be required.
Having said all
that, above results were with 700 mV drive, which is almost certain to be far too high for things I hope to be using this 'RHO' device with. Overall then, I need to do a bit more experimentation & development with this idea, but I do feel that there are encouraging signs so far . . . .
Al.