Search This Blog

Friday, 4 October 2013

So, lets get a bit smarter

Well,

Been rummaging some more in the shack today and have decided to make a variable attenuator. This is in addition to the fixed attenuators I made here:

http://g0mgx.blogspot.co.uk/2012/11/calibration-complete.html

and the switch attenuator I made more recently here:

http://g0mgx.blogspot.co.uk/2013/09/time-time-time.html

Why? Because I would like to be able to set the output of my newly constructed signal generator from here:

http://g0mgx.blogspot.co.uk/2013/10/so-hows-sig-gen.html

to be exactly 0dBm, how will I be able to set it so accurately? Using the dBm meter I made back here:

http://g0mgx.blogspot.co.uk/2013/01/calibration-finaly-complete-really.html

So, that's the background, here's the variable attenuator:


And now with it attached to a fixed attenuator at the input:


There are two switched input attenuators in this variable thingamabob, lets call them A and B these are then followed by a variable pad to the output port, so, if I connect my new signal generator to the input and my dBm meter to the output I can try and assess what the attenuation of the two fixed pads and also the variable pads are:


and then if I "do the math" and calculate the raw attenuation rather then the power readings:


So it looks like my switch position A is giving me 3.3 dB attenuation and position B is giving me 10.9 (I was aiming for 4 and 10). When combined with the variable pad I am getting 3.3dB to 29dB attenuation in switch position A and 10.9dB to 36dB in switch position B - the fact that the ranges overlap is good enough for me.

Now, for the sake of it I am going  to now look at the return loss of the input and output of this filter, primarily because I am aiming for an impedance of 50R and it will be interesting to see how close to this design goal I have achieved across the HF spectrum.

So, using my return loss bridge from here:

http://g0mgx.blogspot.co.uk/2013/09/so-whats-happened-to-bridge-then.html

I can sweep the HP spectrum and see what the return loss is, firstly lets try the 30dB attenuator I have here:


Now, In the image above, you can see three traces. The purple trace is the normalised return loss of the bridge in a completely unbalanced state i.e. open circuit. The yellow is the return loss with the device under test connected (in my case a port of the 30dB attenuator), and the green is the maths calculated as the delta between the two i.e. the difference between the open bridge and the bridge with the device connected. The table at the bottom then shows the application of the maths to calculate the return loss and VSWR.

So we can see that at 10MHz my 30dB attenuator has a return loss of 33dB which is as close to maximum bananas as makes no odds - thats a VSWR of very close to 1:1. As the frequency increases you can see that the return loss decreases which in turn means that the VSWR increases. At 40MHz the return loss has reduced to 25dB which is a VSWR of 1.1:1. All in all though this attenuator is an excellent match to 50R.

Now, here's the same trace but with the input port of the variable attenuator connected in switch position A:


And now in switch position B:



Interestingly the return loss is flat across the spectrum, but not such a close match to 50R and hence the VSWR is higher. This probably means that the variable attenuator on the output side of the pad is affecting the input impedance and hence the input pad should probably be of a greater attenuation value; the downside of this would be that the device would have a much larger starting attenuation value and therefore not meet my design goal!

Here's the new ham cat Florrie, she seems to be learning from Geddy Cat to assist very well:


Good, egh?

No comments:

Post a Comment