NRF24E1 output impedence matchment quetion

[yellowtooth]

We use the chip NRF24E1.



Now we have a problem of the impedance matching at the RF output port of ANT1 (pin 19) and ANT2 (pin20).

I wonder how to design the impedance match circuits (L1, C11, C12, C13) form ANT1 and ANT2 to the antenna.



For the datasheet of NRF24E1 does not have the impedance of the ANT1 and ANT2 at all.



I think the ANT1 and ANT2 are differential bi-directional port, and the antenna is a single ended port.



So the matching network (L1, C11, C12, C13) should be a balun.



Or it is just a matching network without balun function at all, and just put one port to the antenna and the other port to the ground.



Would you please tell me the detail of the network?

As we know, the shunt cap and the series cap make different direction on the smith chart.



Why place a shunt cap (C13) after a series cap (C11 or C12) ?





Thank U very much!

 

[pisoiu]

Hi,
The datasheet of nRF24E1 is not very clear regarding the required load for maximum transfer, but if you look in the datasheet of nRF2401A, you will see "The load impedance seen between the ANT1/ANT2 outputs should be in the range of 200-700ohm. A load of 100ohm+j175ohm is recommended for maximum output power (0dBm).". Since the RF cores are identical, the same should apply to nRF24E1.
I've used the matching network from the picture, but with 3,3nH inductor ( I was unable to find 3,6nH in small qty), C11=C12=1,5p and without C13. With these values, the output was around 0dBm, measured with an R3131 SA. The measurements were performed without modulation (bursts of carrier only). The components are all in 0603, the only exception is the 3,3nH, which is 0402.

/pisoiu

 

[yellowtooth]

I have also seen the 2401's datasheet.


The problem is that "100ohm+j175ohm " cannot be transformed to 100ohm by L1,C11,C12 adn C13.


U can check it by Smith chart tool.


Do U ?

 

[yellowtooth]

And Do U know the input impedence of the antenna?


U just test the output power from the port, do U use any antenna then?

 

[pisoiu]

.....
U just test the output power from the port, do U use any antenna then?

In my design, the matching network is connected to a switch, then PA/LNA, then other switch, then antenna. For testing purposes, I've connected the output of the matching network into the SA, using SMA connector, and of course, the RX/TX switch was decoupled.

/pisoiu

 

[yellowtooth]

I also want to do just like U wiht 2 switchs.

Would U please tell me , which to conect the switch control signal?

Just to pin18(VDD PA) ?


Thank U.

 

[pisoiu]

It is not only the switch which must be controlled. The PA and LNA must be controlled, if both turned on at the same time, they may oscillate. I use a separate signal from the CPU to control the direction. Theoretically, the VDD PA should do the same, but I use the CPU output signal for other test purposes as well. I will try these days to controll the RF stage with the VDD PA, and I'll let you know.

/pisoiu

 

[yellowtooth]

Thank U very much.


I donot want to use another CPU other than the one in the 24E1.


And what the method did U use to control PA and LNA not to work at the same time?

Just control the power on and off ?


I upload the smith chart of the impedence matchment.

 

[yellowtooth]

"100ohm+j175ohm " transformed to 100ohm can just use only one capacitor.


But if use the circuit the datasheet give, it is just as follows.

 

[yellowtooth]

So I think that either the 24E1 datasheet is wrong, or the 2401 is wrong.


Nordic is cheating !

 

[pisoiu]

The PA and LNA are controller through an enable pin, they both have such feature. About matching, I did not used Smith to simulate it, I just used the values indicated in a previous post, and I measured the output, which is roughly at 0dBm. Initially, the results were quite bad, with the values exactely as in the datasheet, but with some trials with diffrent values, I've obtained 0dBm. I think this difference is because of the PCB, which is FR4 and the parasitics of the components. However, at the PCB level, the traces are minimal, good ground plane, but the layout is diffrent than as it is indicated in nordic app notes.

/pisoiu

 

[pisoiu]

I've tried to control the whole power stage with the VDD PA of the CPU. Unfortunatelly, it was unsuccesfull, mainly because the PA (RF2172 from RFMD) enable pin is sourcing few mA when pulled down, current which the nRF is not capable to sink through the VDD PA pin, it remains always with approx. 1,1V level. This is not ok, because it will keep RX side off when in receive mode. The only choice I have is to use a pin from nRF CPU to controll the PA/LNA/switches, or to insert a push/pull stage (either p+n mosfet or a logic gate) between VDD PA pin and the control pins for RF part.

/pisoiu

 

[yellowtooth]

Thank U very much!

Followings R from Nordic reply:

It looks like you might have an old version of the datasheet. The newest,
revision 1.2, is from june 2004, and can be downloaded from our web-page.

The impedance at the antenna is 100 + j 175 ohms.

ANT1 and ANT2 are bidirectional. The nRF24E1 is both transmitter and receiver.

The matching network is a balun which converts the RF-signal from differential
to single ended. The new matching network is a bit different from the one I
think you are using, and it has a better suppression of harmonic spurs. The
shunt capacitor is no longer used.

If you are going to use a 50 ohm antenna (most antennas are 50 ohms), you just
use the matching network we provide.

 

[mikal]

Not to bring up old wounds, etc... but I'm wondering if anyone could recommend a balun to replace the 0603 components used for matching...