EMRFD Message Archive 11722
Message Date From Subject 11722 2015-10-03 17:25:31 kerrypwr Bridge-T Diplexer Report Just a report on the final version of my mixer that includes a bridge-T diplexer.
I reduced the diplexer Q to 5; that gave just-more practical values, notably for the small-value inductor in the parallel-tuned section;
The 124nH inductor was wound on a Teflon toroid about the size of a T37; the toroid required only a few minutes work in the lathe;
The following test results are with the SBL-1 mixer disconnected; I will do VNA and spectrum analyser tests of the complete module later.
The amplifier is the "standard" post-mixer type; bias is 45mA. It is followed by a 6dB attenuator.
The diplexer is designed for 12.8MHz; the as-yet-unbuilt filter will use 12.8MHz crystals.
The responses of the amplifier + attenuator and of the diplexer + amplifier + attenuator are shown here;
The 3dB bandwidth is about 2MHz.
11723 2015-10-03 18:22:13 Dave Re: Bridge-T Diplexer Report Have you done a return loss plot?
Dave - WB6DHW
11724 2015-10-03 19:17:31 kerrypwr Re: Bridge-T Diplexer Report Good idea Dave; after all, SWR/RL is a major objective of the diplexer. :)
In fact I did but forgot to include it in my initial post; this is just the diplexer, nothing else
11726 2015-10-03 21:46:43 K5ESS Re: Bridge-T Diplexer Report
Seems like making a Teflon toroid core is a lot of unnecessary work. Most of us don’t have a lathe. What does it provide over just a cylindrical Teflon or air core? Maybe I’m missing something. Open to enlightenment.
11727 2015-10-03 22:05:03 kerrypwr Re: Bridge-T Diplexer Report Compactness and stability were the main reasons I used a toroid; I did consider a cylinder though.
Material #O toroids are available (I think DIZ has them); they are phenolic with no magnetic properties.
11728 2015-10-04 06:11:37 DuWayne Schmidlko... Re: Bridge-T Diplexer Report You can get nylon washers in assorted sizes. Early BITX builds used faucet washers.
11751 2015-10-09 01:38:29 Weddig, Henning-C... Re: Bridge-T Diplexer Report Kerry,
you can improve the VSWR in the passband (i.e. aroung 12 MHz) if You
tune both resonators !
I have seen that using a too high loaded Q of the diplexer will make it
more difficult to obtain a good match (low VSWR) within the passband of
Am 04.10.2015 um 04:17 schrieb email@example.com [emrfd]:
> Good idea Dave; after all, SWR/RL is a major objective of the diplexer. :)
> In fact I did but forgot to include it in my initial post; this is just
> the diplexer, nothing else
> Kerry VK2TIL.
11752 2015-10-09 02:56:08 kerrypwr Re: Bridge-T Diplexer Report Thanks Hennig; that is a very insightful observation.
I'm trying to balance SWR with selectivity in the diplexer; it's clear that there are "trade-offs" in this.
Perhaps my idea of obtaining some "free" selectivity at this early point in the chain is not necessarily a good one.
I'm fairly happy with the result; SWR of 1.8:1 in the passband is not too bad; I will have to think about whether I want to trade that for less selectivity.
Your post ties the selectivity -v- SWR relationship very neatly; thanks.
11753 2015-10-09 06:07:26 kb1gmx Re: Bridge-T Diplexer Report There is an error there.That is that the diplexer, bridged-T is intended to provide selectivity, this is not the case.The circuit is there and functions to rout out of band signals to a dump load (resistors)and desired signals to the succeeding stage.Over some frequency range that can be achieved but as you go higher the values becomeimpractical for higher Q and the losses greater.If selectivity is required then a highpass, bandpass, lowpass triad is needed.Since in most cases the diplexer is followed by gain or selectivity (usual is both in that order)the need for selectivity is low and the design goal is to effectively terminate the mixer for widebandwidth.Attempting to achieve termination and selectivity at low loss is a conflicting set of goals.Allison 11754 2015-10-09 06:56:23 Bill Carver Re: Bridge-T Diplexer Report Hennig's post was correct. Tune both series and parallel tuned circuits to the i.f. frequency. There's no reason to accept 1.8:1