50 ohm impedance matching

What is the best way of matching a 100MHz rf output from the BH1415 chip? It has an output of 75 Ohms, and I want to put it into an amplifier with an input of 50 Ohms. What do you suggest I do to match the impedances? Thanks

YOu should consider the following notes:

The voltage reflection off of your 75 ohm input is ρ = (50-75)/(50+75)= 0.2
So the voltage transmission into your 75 ohm input is T = 1 - (0.22) = 0.96

in dB's, that means you will loose 20 Log (0.96) = 0.35 dB of power due to the mismatch.

It would probably be hard to find a broadband transformer for a 50:75 ohm match that had less than 0.4 dB of loss!

but the following site may help you for finding good transformers
h**p://www.maurymw.com/products/rfac/cxadptrs/8882/8882.htm

Is there a way of doing it without using a transformer? I hear the BH1415 burns out after about an hour if I put the 75 Ohm output directly to a 50 Ohm load.

I am thinking of amplifying it using a Mar-8ASM chip.

If you can tolerate the losses in your path a simple resistive pad can be used

I would probably use the pi attenuator as the 1.57 Ohm resistor would be difficult to come by, unless I use some resistors in parallel.

If I put the -7dB output into the Pi attenuator, then I'll have say -13dB, which gets fed iinto the Mar-8asm chip. The gain on the Mar-8asm chip is about 30db. This would mean I have a gain of 17dB.

What would the milliwatts be for this? About 5 mWatt?

Could I then take the output from the one mar-8asm and amplify it again using another mar-8asm, or should I use a different amplifier?

This all seems a bit too simple - I know I'm going to run into some practical problems.. Do you know what problems I may encounter with this? Thanks

E-design, what software is that?

The MAR8ASM will compress at around 12.5dBm so that is the maximum output you will get. That is a tad over 10mW. Even if you add another MAR8ASM you will be limited at the compression point. Also the MAR can only handle +13dBm input without damage (about 20mW)

The program is RFcalc free from Milmega

http://www.milmega.co.uk/downloads.php

What Monolithic or other amplifier would you recommend for the final stage?

Added after 2 minutes:

Sorry, My desired output is 100mW

It depends on the output level you desire. If you want 10mW out then the MAR will do fine. I will increase the pad a bit to prevent compression in the MAR. (12dB)

If you want more like 100mW you need to find another amp with a compression around +20dBm with enough gain or design a discrete transistor output stage.

It is 100mW desired output. What do you mean by increasing the Pad?

I guess that you're using special problem , Can u plz upload it

Thanks .

The pad shown is for 6dB and with the MAR8 not to run it into compression you should use around 12dB cause of the 30dB gain of the MAR

ehh - you mean 10 * log10(0.96) = -0.177 dB ie. square 0.2 voltage ratio give 0.04 in power ratio.

10 * log10(0.04) = -13.94 ie. 14 dB in return loss.

or, i thinking wrong ?

Added after 30 minutes:


14 dB in return loss (4% reflected power compare to outgoing) is not so much.

is two factor can kill RF-drivers.

1. to much output power reflected back and amplifier structure is not build to handle absorbed heat of this - many big power amplifier build on this way and using circulators on output for protection.

2. outgoing wave together reflected wave can give so high voltage or current (depend of phase angle) and break down amplifier structure (FET and MOS channel). this situation can exist on days very optomized low voltage power amplifier for mobile equipment.

Most case of RF-output designed for unconditional stability ie. load impedance can place every point inside the Smithchart circle inclucive open and shorted output without damage or oscillating. 50 <-> 75 Ohm step is not problem in most case.

In bigger RF-system you cannot calculate better return loss than 14 dB and needs handle this well - 50 <-> 75 Ohm step give around 14 dB return loss.