90° bend ?? best way to take a feedback from a microstrip ??

[actra]

hello again...

the input is comming from the VCO .. 1.3 GHZ

then it divided to two parts .. one to RF amplifiers ...
two to prescalar as a feed-back
the line is a microstrip 50ohm matched..

my questions
1- is it good what i did in the attached image ? if not is there any better suggestions ?

2-is this considered 90° bend to microstrip ?

 

[AL0117]

Re: 90° bend ?? best way to take a feedback from a microstri

Hi,

Take a look at the Zcomm appnotes (AN102 about the layout), as for the simple power splitter you can use 3pcs of 15 ohm resistors (as used in AN105 for summing the 2 VCO). For 1.3 GHz 0805 package will be fine for the resistors and capacitors.

**broken link removed**

Your discontinuity is a 3 port TEE. Port1 input from the VCO, port 2 and 3 the outputs to the amplifier and to the prescaler. If you want to model exactly the TEE, use a planar EM simulator, (Sonnet Lite is more then enough for doing this, also you can include your R,L,C elements in the netlist and simulate them too). But at 1.3 GHz, is not so critical.

As concerns the rounded end of your lines, PCAD (I am using also this pcb layout program) can't do in other way. You have an alternative, to draw these lines as poligons, and in this way the ends will be right angled. Or (and I am doing all my rf layout so) simulate the layout of your design in a microwave simulation package (in this way you can include all transitions and discontinuities), export the layout as DXF or GERBER files, and import it in PCAD. In this way the pcb in PCAD will be identical to the layout simulated.

rgds,
Al

 

[g579]

Re: 90° bend ?? best way to take a feedback from a microstri

Just one remark.
Don't you think that the prescaler not only do his job, but also outputs other signals (subharmonic, multiplied and mixed subharmonics) too ? So the (amplified ) VCO output will not be a clean signal what I suppose you want.
I think you should separate the prescaler by an amplifier or attenuator from the real output of VCO signal too.
g579

 

[actra]

yes ..the output from the VCO is first passed to pie attenuator rep. by resistors shown...

Added after 3 minutes:

AL0117...can u suggest me a suitable easy to use microwave simulator ?

Added after 5 minutes:

about the rounded end of the lines... can i just leave them like this then remove them using any gerber editor ?

 

[rfmw]

Re: 90° bend ?? best way to take a feedback from a microstri

your layout is good for 1.3GHz. But the problem, mentioned by g579 will remain.

You have to insert an attenuator (or another amplifier in prescalar trace) between RF amp and prescalar inputs. Why didn't you connect RF amp directly to VCO output and an attenuator from VCO output to the prescalar?

 

[actra]

ive put an pie type attenuation pad after the output of the VCO to reduce the total gain by 2dB.... so it isolate and attenuate the output to be suitable for the prescalar

do i have to put another attenuator @ the prescalar input ????

what if i changed the vie to be like this


VCO ---------->RF STAGE
.
.
.
v
Prescalar

instead of
RF STAGE
.
.
.
VCO ------
.
.
.
v
Prescalar

 

[g579]

Re: 90° bend ?? best way to take a feedback from a microstri

If you need a clean output signal, then the prescaler should be separated from the other parts of the unit, placing an attenuator or (rather) an amplifier and an attenuator before the prescaler.
At 1.3 GHz it is easy to place a MAR or similar gain block with 10 dB gain and a 10 dB attenuator, which will give you about 30 dB or more isolation before the prescaler.
But if the spurious signals at the real RF output do not disturb you, you can miss this whole thing out.
g579

 

[AL0117]

Re: 90° bend ?? best way to take a feedback from a microstri

One more remark: indeed a good separation between the VCO/prescaler and output is important. The swithching of the dividers inside the prescaler will produce a lot of unwanted products. First you should calculate a power balance: what is the output level of the VCO, and what level you have to provide to the prescaler. Usually the prescalers can accept levels between -20 and +3 dBm, but you have to check the datasheet for your case. So if you have enough power (for ex. +9 dBm VCO output power), you can use a directional coupler (with the highest coupling factor possible, but not too high to provide the level for the prescaler). In some pratical circuits microstrip couplers are used, in some others smd directional couplers (Minicircuits and TOKO (among others)produce these kind of couplers which works up to 2 GHz ).

rgds,
Al

PS. Actra: check your Private Messages.