Designing a wideband RF splitter and amplifier
i want to design wideband rf splitter and amplifier(active splitter) any one have a schematic or pcb using minicircuit device?
What is your desired bandwidth, ripple, P1dB, etc...?
BW=100MHz-1GHz
and ripple or P1dB , .... not important(any)
i want to use mar-xxx for this circuit
If you want to have a 2-way splitter, then the easiest way and also very-well proven for this frequency range is a normal 6 dB passive splitter followed by MAR amplifiers in each branch.
Use 0805- or even 0603-sized 51 Ohm (or even better two 100 Ohm in parallel) for 6 dB splitter. All AC-coupling capacitors should be a parallel combination of 100pF and 10nF (also 0805 or 0603).
Remember to use a RF-choke (plus series resistor) at the output of the MAR amplifier for bias, since a resistor-only biasing affects the gain - you could get a 1-2 dB loss of gain easily.
Also, watch for power levels, so that you won't bring any of the MAR amplifiers into saturation.
Hope this helps,
rfmw
Hi,
The link below points to the construction details of a wide band CATV splitter:
https://www.edaboard.com/viewtopic.p...832&highlight=
it is for 75 ohm (Catv), but you can use it as a guide for 50 ohm (if is necessary)
regards,
Al
Hello AL0117.
Could you tell me what ferrite type you have used?
IF possible manufacture type number, etc...
Thanks a lot.
S.
Hi Sinatra,
We use some custom made Toko parts (B4F-617PS1063)
http://www.tokingtec.com.tw/toko/pro...mers/b4fl.html
I will look for details tomorrow. I think that we have the specification sheets.
We also used some double aperture cores from Neosid
http://www.neosid.de/DWL/Katalogseiten/Teil%201_25.pdf
regards,
Al
Thanks already now Al.
Normally I like to make my own stuff instead of buying. That's the best way to learn new things and to have pleasure with electronics.
Till now I didn't get good ferrites that could give good bandwidth.
That is a good news from you. That's great, now I can probably with those ferrites make better mixers, splitters and maybe frequency doublers....I was very limited till now....
I wait ansious for your specs. :')
and thanks again in advance.
Greetings.
S.
Ps:Feel free to ask any help from me too.
i use microstrip for splitter and don't use ferrite core.
Hi Carrier,
For this frequency range using a microstrip splitter will be impracticable. The bandwidth of such a single section splitter is less then an octave (1:2), and you will have serious missmatch. Also think about the dimention of this splitter. The ferrit based splitter needs about 4x4 mm surface on your board. If you want to increase the bandwidth and you will use multiple sections, your dimensions will be huge.
regards,
Al
ALO117,
carrier wrote twice that he won't use ferite-based splitter, but microstrip-based resistor (6dB) splitter with the addition of some MMIC to boost the power. Isn't that right carrier?
Carrier,RFMW,
Sorry, RFMW you recomended the resistor based splitter (this is called resistive power divider, and has 6 dB loss as you mentioned). Carrier says in his last post, that he will use microstrip and no ferrite core. As I know, the microstrip power dividers has to do something with the wavelength?Am I right? This will influence your splitter dimensions. Of course if he use resistive power divider, and the 6 dB is not a concern, he still can include in his simulation the microstrip lines between the resistors of the divider.But it is still a resistive divider. Maybe the question was not exact.
Dear fellows, sorry to bother you.
Moderators, please remove my last posts.
10x,
Al
Hello,
You can find a complete description of a wideband passive splitter in RF Design Guide (Vizmuller - Artech House) pag 149.
It will work well if you only want to build several units, but if you want it for mass production you have to find a very fine and reliable manufacturer !
You can try the circuit in the atachment.
Lower transistor is a INFINEON BFP650 and upper transistor is a PHILIPS BFM520. It is similar to the cascode configuration but with two transistors, I know it is not very orthodox but it will work for many uses, one of the main disvantages is the poor output return loss (- 5 dB) at the final end (1 GHz), but you can optimize it reducing the general gain (10 dB).
Bye
ALO117,
you didn't bother anyone and your suggestions are surely appreciated. Isn't this forum about knowledge from different people? Like you said, user carrier didn't exactly specify what he needs, so we all suggested some solutions...
Best regards,
rfmw
i want to use mar-xxx for amplification and don't use ferrite for splitting
I guess this is something that should work for you... Of course, this circuit uses MMIC INA-10386, therefore you should modify biasing and coupling caps for your MAR-xxx parts.
This can be very convenient for you if the cost is not restrictive issue..
http://www.anadigics.com/products/ad...04_Rev_1.0.pdf
If so, you may use also ferrite double aperture core ( for instance from epcos) to split the signals and then amplify them.
thanks rfmw
but what lenght of 50 ohm line in your circuit?
Length of the 50 Ohm transmission lines are unimportant, since the circuit uses resistor splitter and is therefore broadband. I'm using this circuit for 4-way RF active splitter from 100MHz to 2.5 GHz. Length of the transmission lines would matter in the case of narrowband Wilkinson-type passive splitter.
EDIT: attached the image of the 4-way PCB
At the input on the left is a PI-type attenuator, then two resistive splitters and finally 4 MMICs to boost the power. All ground connections are made like this: drill a hole approx 2.5-3mm in diameter and at the bottom close it with copper foil. Then at the components side (upper side as shown) fill this closed holes with tin. Finally solder the components ground connection pin to this prepared tin bubble ;)