EMRFD Message Archive 10385

Message Date From Subject
10385 2014-10-26 10:20:26 Steve VK2SJA Losses in Diplexer?
Hi All,

Newbie again. Seems like every time I get stuck and need to cry "Help!"
it's about losses. This time in a Bridge-Tee Diplexer I built for 20Mhz.

Built one at Q1 - too wide. Needs to be Q10. Then built Q10 version and it
is exactly where it should be in frequency and just a little over 2Mhz
wide. Very pleased with myself, I have it under control.

Except that I have a 6dB insertion loss. This article says that with good
quality components I should be looking at 0.5dB or less!

http://www.qsl.net/g3oou/mixerterminations.html

I used the following online calculator:-

http://www.changpuak.ch/electronics/calc_16a.php

And then manually verified the answers against the equations in the paper
linked above.

I played around with it trying a few things and did manage to get the loss
down to 4.5dB but its still a long way to 0.5dB

So questions. Is 0.5dB truly a realistic home-brew target before I go
knocking myself out. And can someone suggest where the common losses
occur. As the circuit has so few parts is it in the capacitors used?
Brown, round, disc-ceramic, Modern. One side NPO the other not. If so what
type of capacitors should I be using here?

73, Steve.
10386 2014-10-26 10:29:41 Lasse Moell Re: Losses in Diplexer?
Use RFsim99 and vary the Q for each component and you should see what you need to acheive a acceptable  loss. 6 dB sounds like an awful lot though...

/Lasse SM5GLC

26 oktober 2014, 'Steve VK2SJA' vk2sja@bigpond.com [emrfd] skrev:

 

Hi All,

Newbie again. Seems like every time I get stuck and need to cry "Help!"
it's about losses. This time in a Bridge-Tee Diplexer I built for 20Mhz.

Built one at Q1 - too wide. Needs to be Q10. Then built Q10 version and it
is exactly where it should be in frequency and just a little over 2Mhz
wide. Very pleased with myself, I have it under control.

Except that I have a 6dB insertion loss. This article says that with good
quality components I should be looking at 0.5dB or less!

http://www.qsl.net/g3oou/mixerterminations.html

I used the following online calculator:-

http://www.changpuak.ch/electronics/calc_16a.php

And then manually verified the answers against the equations in the paper
linked above.

I played around with it trying a few things and did manage to get the loss
down to 4.5dB but its still a long way to 0.5dB

So questions. Is 0.5dB truly a realistic home-brew target before I go
knocking myself out. And can someone suggest where the common losses
occur. As the circuit has so few parts is it in the capacitors used?
Brown, round, disc-ceramic, Modern. One side NPO the other not. If so what
type of capacitors should I be using here?

73, Steve.

10414 2014-11-05 07:06:05 kb1gmx Re: Losses in Diplexer?
For circuits that have a Q requirement or low loss I tend to use NPO or better caps.
the common disk ceramic is a poor part usually.  Same for coils, at 20mhz skin effect
is significant so larger wire and low loss cores(toroids) are the order of the day.
In the end the measured performance vs the design performance speaks directly 
to the part quality or assembly technique.

If the diplexer is used as a post DBM filter then the need for low loss is critical
for a high performance system.  Getting under 1db is not that hard and .5 is 
achievable.  Also you may look at a high-pass low-pass design as a less 
critical design that bridge T.

Also build on a ground plane (dead bug) not a logic style bread board.
for filters what you call ground is critical to performance of the filter.


Let use know here how you did.

Allison

10416 2014-11-05 10:04:07 Clint Re: Losses in Diplexer?
Interesting comments about the quality of capacitors.  I have been building out a simple 3rd order bandpass filter.  The coils were toroids.  The filter had about a 3 db insertion loss.  I had what I though was a high quality 1500 pf capacitor (square body).  If was probably not ceramic.  I replaced these caps (two each) with standard 1500 pf ceramic disks and the insertion loss dropped to about 0.5 db.
 
73
Clint
W7KEC