Power considerations for 50 Ohm resistors in power combiners
Quick question about power combiners using: wilkinson, branch line, and rate race. We are working at power levels of 30-35dBm, and I was worried about the power ratings for the resistors in those types of power combiners.
For the branch line and rate race, assuming a pretty decent match to the coupler, the resistor should be isolated and burn no power, right?
However, I dont remember much about a wilkinson. The flash backs to a virtual short under odd mode, make me think the resistor might see some appreciable current?
Any clarification would be much appreciated!
Thanks in advance,
Sami
- Obtain the thermal resistance['C/W] of resistor Rth, determine if there if any change in condition, test spec, thermal conductor or insulation added. and adjust Rth
- compute/ convert to RMS power [W]
- estimate max ambient temp 'C
- Choose a max temp rise that wont' heat up any sensitive parts like your finger to say at 65 'C or perhaps 85'C even though at max power the resistors can tolerance 150'C
- Then derate and choose the power Resistor power rating to solve the equation,
delta T rise (= Rth * Pd) + Ambient T (~30'C) = 85'C or your choice. This usually means derating power to (85-25)/(150-25)= 48% of max power.
Some values depend on material of resistor, so I am using poor memory, so compute yourself. But using only 1/2 the power rating of any resistor is a good Rule of Thumb.
Hello,
Thank you so much for the quick response! EDAboard is always so helpful. If I may follow up with a another question?
Obviously what I did above is the VERY ideal case. It's a simple simulation of a wilkinson power combiner. I am plotting the current though my load and through the 100 Ohm resistor. IF things are matched well, it seems very little current will flow through the resistor. Does it make sense, that even in a power combining scenario of 1-2W would only need a few mW rated 100 Ohm resistor?
I'll make the simulation much more representative of reality, which i'm sure will increase the current though the 100 Ohm branch, but I don't foresee it going too large.
I will freely admit, my simulation may be wrong, but under even mode excitation I wouldn't expect any current to flow through the 100 Ohm res.
Thanks again,
Sami
Consider what happens if there is a short , open or impedance mismatch rather than a perfect match on all ports. Try 10% mismatch.
SIW Wilkinson power divider may help dissipate more power. Maybe 5-10 times more than microstrip line, depending on substrate thickness, line width.
Hello,
Thankfully this power combiner will be nestled between PAs and an antenna. So I don't have to worry about presenting too much of a mismatch to the circuit. Last question before I let the thread die. I made a much more realistic design:
I made a very crude model for a 0201 chip resistor to put between the branches, (some pads for soldering and a sheet resistor of 100 Ohm);
putting 30 dBm on each of the inputs i get close to 33dBm out:
to calculate the power through the 100 ohm resistor, i measure the current through the resistor and calculate: P = I^2*100
that gives me close to 0.5W across the resistor. Does that seem in the ballpark for reality? I have zero real experience designing these types of things, so, I just want to double check I am not missing something obvious.
Thanks again,
Sami
just remember to derate power capacity ~40~60% for reasonable temp rise.