View Single Post
Old 10th Aug 2018, 11:10 pm   #49
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Back terminated cable?

Quote:
Now I do not wish to sound like an 'arrogant tech. know-it-all' in this matter, but to me, I would have thought that anyone using a transmission line in a typical application would also have the elementary knowledge to realise the need for that source Z to match the cable's Zo, therefore rendering the need for the expression 'back-terminated' to be simply redundant. In conclusion, to me, the use of that phrase is simply 'tech. over-kill'.
A simple and practical example where you might not want to match the source to the cable Zo is given below. You might have already made one yourself!

Up at VHF through UHF it's hard to make a high impedance probe for probing low Z circuits without loading the circuit too much. It's also desirable to want a flat frequency response. So a typical scope probe is no good because of the high capacitance up at VHF/UHF. Lets keep it simple and say the probe is for probing typical RF circuit boards where there might be 50R points in the circuit.

If you simply used 50R cable with a 50R spectrum analyser at the far end then you would double terminate and cause loading and loss in the circuit under test. This can sometimes be OK but it generally isn't a good idea.

But if you fitted a 470R series resistor at the feedpoint of the 50R coax cable as a form of probe tip, the circuit under test would see 470R in series with 50R when touched by this 470R resistor + coax. So it won't be loaded as much. As long as the 470R resistor has low parasitic capacitance and inductance and the cable is tightly connected to ground at the test PCB then the above probe should give a flat response up past 2GHz. However, if you just use a small metal film leaded resistor it might only be good to about 1GHz. But the probe should give about a 20dB loss with a flat response and it will load the 50R with just over 500R so this is about ten times. So there will be minimal loss in the circuit as it gets probed.

But the relevant point here is that the source impedance feeding the 50R cable is now about 500 ohms. But it still gives a flat -20dB response as long as the far end is terminated correctly with 50R. 470R isn't the ideal resistance for a flat -20dB response from the probe but it is quite close

Of course, this is really just a form of voltage divider so the -20dB coupling only applies if the 50R test point really is 50R. So it isn't really a -20dB probe for real world 50R circuits where the circuit might be 58R or 39R. But it is still a very useful tool to have on the bench
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline