Single-Board Universal Module

I believe you can find all the relevant sources at bitbucket.org/PhracturedBlue/universaltx . However, I've not been working on this (don't have a clue how), so I'm not positive about that.

If those are the sources, and any of you would like a shared repository to work in, let me know and we can create one and then add you to team deviationTx.

I have 3 boards which I assembled quite some time ago. I had offered to send one to PB, but he seems to be taking a bit of a break. If someone else wants to take this on, I'd be happy to send one of my boards to them. It's cool that we're now up to a few people who can help build prototype hardware.

PB committed his test code this morning so I could pick up on hardware testing. I've spent time on the module this evening, and can confirm that the signal loss is at the switch. I'm seeing about ~20dB of loss through the switch. With the switch bypassed, the PA is working fine with ~24-25dB of gain. These were measured with an RF probe, so they're not going to be terribly accurate.. but as relative measurements are good enough to show that the PA is functioning properly. I've verified that the switch control lines are being switched by the MCU correctly.

The PA TX_EN and RX_EN seem to be inverted for the NRF24L01, but that's not stopping testing. I think I remember this being an issue with the main repository anyways, and may already be fixed.


Up next is debugging the loss at the switch.

Hi and welcome back
Could 4:1 switch + amp be moved to a daughter board?

Remind me, did you insert the blocking cap between switch output and PA/LNA input? The schematics does not have it, and these things can be sensitive to DC component. SKY13384 datasheet recommends 100pF blocking caps on all inputs/output, and while all inputs seems to not have DC connection, the output can.

Yeah, I've got the blocking cap between the switch and the PA/LNA.

That's good news that it'll fit. I haven't really had a chance to dig into why the current switch isn't working yet.. But I will soon. There is a drop-in replacement available from Hittite as well, but it's super expensive relative to the Skyworks. Which 2:1 switches were you looking at? I've used a couple of the Peregrine switches in the past with good luck.

WB PB!
The only difference I can see between the PE42421 and PE4259 specs is the advertised ESD tolerance, on the front page.

I'm using the 4-way PE42641MLBC-Z with no problems, for another project. These do not require DC blocking caps, unless you put DC on them. Pin-outs are also quite similar but different, but the same general layout would work. The truth table are also different for RF2 and RF3.
Using 3 SPDT switches seems more difficult than necessary.
The control lines should be bypassed with some caps, besides RF-leakage will get into the signal here (not a problem for a TX), the max switch rate is 25 KHz and limiting the bandwidth to keep RF out of these points can't hurt. In your setup the switch rate is super low so caps her could keep you out of trouble.

Neither PE nor SKY advertise RF isolation to control and DC lines.

which RF switch did you use before?

Skyworks SKY13384?

SKY13384-350LF

Not sure if this was the problem...but on this chip you have the single antenna input and 4 RF outputs.
In a transmitter should be the other way around,...4 inputs and one output.
This signal is unidirectional from Antenna pin to each output pin...not opposite.
So I think it will work for receiving but not for transmitting.

the switch does not "know" how the RF signal passed thru it is used, it just conducts, or not.

How do you know that?In the data sheet it is not specified that...... but is specified clearly that antenna input and the other 4RF pins as outputs.I think it should be specified if is bidirectional.

Actually I don't think is true,
Symetrical in RF world means the input and output have equal impedance,..in this case 50ohms.
I think the chip is unidirectional and for this reason not working.

The note below in the datasheet seemed unusual to me. Does it also imply the RF inputs should not be driven when changing the control inputs?

The recommended startup sequence is:
Step 1: Apply VDD
Step 2: Apply VC1 and VC2
Step 3: Apply RF input
The device must be turned off in the reverse order