Amplifying nRF24L01P signal

[kwas]

Hi guys,

I am trying to use a TM2001 amplifier (Taiwan Microelectronics, google returns the PDF) to amplify the signal from an nRF24L01P.

The TM2001 seems to accept a 50ohm signal. I use the "application example" on the nordic chip on page 66.

I simply connect the TM2001 RADIO input to the 50 ohm line instead of a chip antenna. Then TM2001 has its own output which goes to an antenna.

I have used TM2001 before with no problems. However now it is giving very strange effects.

There is way too much power consumption from it. If I touch the antenna it affects how much power is drawn. No radio output is observed unless I artificially raise the resistance of the battery supply (by adding a 10 ohm resistor). Is it delusional to think that I can just hook up the TM2001 like I described? The only part of this circuit I don't understand is the inductors/capacitors between ANT1 ANT2 and VDD_PA. Are the inductors causing this problem? Should I just connect ANT1 or ANT2 to the TM2001?

thanks

[Andreas]

Hi,

 

What you describe is the correct way to go, the RADIO input should be connected to the 50Ohm trace in the application example, it should not be delusional at all.

 

The high current consumption should indicate that it is on at least. For reference what kind of current consumption are you seeing? The fact that it changes if you touch the antenna is probably because touching the antenna will change its impedance, thus changing the load impedance of the PA. Have you tuned the antenna?

 

The inductors/capacitors between ANT1/ANT2 and VDD_PA should be as it is, power is supplied to the nRF radio from VDD_PA through the ANT pins. You can use VDD_PA to toggle TX on, but you can not draw power from it. I am not able to decipher from the product description I found if this is the case or not.

 

Best regards,

Andreas

[kwas]

Hi,

I've tracked down the issue to the NRF24 forgetting register settings.

For instance, I set channel to #10. The first packet is sent on channel 10. Any subsequent packets are sent on channel 2 (default).

I have two boards. On one, powering the board from Pickit (bypassing the regulator) solves the problem. On the other one it does not. I tried soldering, resoldering all ICs, substituting new ICs thinking that they were broken, but it did not help. The power supply is good at 3.3V.

Under what circumstances will the NRF reset itself and lose register information?

[Andreas]

Hi,

 

Have you verified that the supply is stable at 3.3V throughout the transmission, not dropping or anything due to rapid jumps in current from the FEM? It does sound a little strange though, seeing as the first transmission makes it, but the fact that bypassing the regulator solves it seems to point towards the regulator being the culprit.

 

Best regards,

Andreas

[kwas]

I did more testing.

1. if I leave TM2001 on (no duty cycling) the board works

2. If I duty cycle TM2001 at each transmission the radio stops working. SPI communication into NRF seems to be unaltered, excluding mcu problems.

3. if I duty cycle TM2001 but with RADIO pin (connection between tm2001 and nrf24) disconnected from tm2001, the board works with no amplification, provided there is a piece of wire attached to the NRF antenna out pin (after the passives) instead of nothing.

I have already replaced every IC on the board.

I also have another board which was showing the same signs and then healed magically after I replaced a TM2001; but replacing TM2001 does nothing on this board.

Driving me crazy, 2 days no rest and no solution :(

I can provide the PCB and SCH files, this is just a small test board with nothing else on it to test NRF24.

[kwas]

Tried several TM2001, there is only one that works when dropped into either board. All these TM2001 chips all work when used with another RF module (not Nordic). Something tells me that something about TM2001 (which may have some inter-part variation) is interfering with the ANT circuit, either inside NRF24 or in the balun/matching circuit from the application example :(