Hey,
It occurred to me this morning that I had measured the output
peak-to-peak voltage of the DDS in open circuit and calculated
the power like if it was 50 ohm terminated...
Sure enough with a 50 ohm termination (after the L-match), the
output power of the circuit is -5.9 dBm and the characteristic of
my filter much more normal:
2 dB bandwidth: 4.50 khz
Insertion loss: -5 dB
https://docs.google.com/spreadsheet/ccc?key=0As9CZnZ-A5a2dEZaM1dlWVdVX3pXaUlJMXd0cHNscHc#gid=1
lesson learnt!
thomas
On Thu, Jan 23, 2014 at 3:56 AM, Mark G0MGX <mark.g0mgx@xxxxxxxxx
<mailto:mark.g0mgx@xxxxxxxxx>> wrote:
Hi Gang
I just pulled 8 crystals out of the bag and built another
crystal filter without attempting to match them - just out of
curiosity.
The sweep is attached, the -2dB bandwidth is 3.5KHz and the
insertion loss 4dB (the tracking generator is at -20dBm).
The sides of the filter don't look quite as steep but it
still looks pretty good without any matching - maybe I got
lucky? Certainly can't explain why you are seeing a 10dB
insertion loss - that's very high.
Mark
G0MGX
On 23/01/2014 10:42, Thomas Sarlandie wrote:
Clark: I had a face-palm moment when I got your email. This
is obviously a much simpler solution. Thanks a lot.
Farhan/Jerry: I read EMRFD on return loss bridge and this
page too: http://www.qrp.pops.net/RF-workbench-3.asp but I
am still not sure how I would measure the output impedance
of the dds. If you have some more links or quick suggestion,
that would be very helpful. Thank you.
I used Clark technique to measure the output impedance of
the DDS and found 100 ohm again (not a big surprise). I then
built an L circuit to match it to 50 ohms. I did not have
the perfect parts available so I ended up with 39+i.14
calculated (66 ohms measured with the pot).
I took the time to graph the filter response:
https://docs.google.com/spreadsheet/ccc?key=0As9CZnZ-A5a2dEZaM1dlWVdVX3pXaUlJMXd0cHNscHc#gid=1
I get a center freq of 20,002.5 kHz, a 3dB bw of 5 kHz and a
ripple of 2.2 dB. If you have feedbacks on my method or if
you think I goofed somewhere please feel free to point it
out. I will probably try again with matched crystals ... and
a better idea of what I am aiming for.
One thing that still puzzles me is the insertion loss. I get
about 10dB when Mark reported 3 dB. I am not sure what could
explain such a big difference.
thomas
On Wed, Jan 22, 2014 at 2:18 AM, <jerry@xxxxxxxxxxxxxxxx
<mailto:jerry@xxxxxxxxxxxxxxxx>> wrote:
The return loss bridge is one of the handiest tools on
my workbench. I
got my design from chapter 7 of EMRFD.
-------- Original Message --------
Subject: [minima] Re: Experimenting with the crystal filter
From: Ashhar Farhan <farhanbox@xxxxxxxxx
<mailto:farhanbox@xxxxxxxxx>>
Date: Wed, January 22, 2014 4:04 am
To: minima@xxxxxxxxxxxxx <mailto:minima@xxxxxxxxxxxxx>
the correct way, is really to measure it with an RLB.
six resistors and
a balun will go a long way in helping us measure these
impedances. oh,
btw, u can measure the input and output impedances with
the RLB.
- f
On Wed, Jan 22, 2014 at 3:08 PM, Clark Martin
<cmmac@xxxxxxxxx <mailto:cmmac@xxxxxxxxx>> wrote:
Sent from an iPhone, don't ask whose.
> On Jan 22, 2014, at 1:34 AM, Thomas Sarlandie
<thomas@xxxxxxxxxxxxx <mailto:thomas@xxxxxxxxxxxxx>>
wrote:
>
> I am using an ad9850 and I measured the output
impedance to be 100
ohms. I simply measured the output voltage in open
circuit (519 mV rms)
and the output current in short circuit (5.1 mA). This
is as simple as
it gets but I think this is correct.
A better way is to measure the open circuit voltage then
apply a
resistive load to it and adjust the resistance till the
voltage is one
half the open circuit voltage. The load resistance will
be equal to the
source impedance.
