Digital capacitors and NRF24L01+

[Joe Wright]

Hi, possibly a bit of a stupid question but here we go:

I'm developing a prototype system, which includes my own design of sensor modules that transmit their data via nrf24l01+'s to a base station. One is located in a racket handle and, while is generally working, sometimes suffers from radio link problems. I read that having your device in an enclosure, and interaction of the body (hand in this case) can throw the tuning off.

When I've finished my prototyping I will take along what I've done to some professionals to make it market ready, including antenna and matching network tuning, as I don't have the knowledge or equipment to do a really good job. However, I was wondering if I could do a better prototype for field testing, using digitally tunable capacitors to try improving the link. I was inspired by reading pages like:

www.psemi.com/.../digitally-tunable-capacitors-dtc

electronicdesign.com/.../back-basics-impedance-matching-part-3

My thoughts would be to arrange the sensor/racket in various positions, chuck a load of data out to test link reliability (and counting number of retries, lost packets, etc...), with automation to test different DTC settings.

So my questions are:

1. Would this work/be worth doing?

2. If yes, is there any reference circuit designs I could look at that would be suitable for using the DTCs?

I'm thinking of including an RFX2401C for PA/LNA, and currently I'm using a chip antenna (Wurth 7488940245). In the racket sensor I could use a much longer antenna but I don't know what would be suitable.

So far I've been following as much as possible the Nordic reference designs.

Thanks for any help, Joe

[Hung Bui]

Hi Joe,

1. It would possibly work. But I guess you are using a PI tuning network here ? What if the tuning require one of the component be a inductor ?

2. No we don't have one. I would suggest you to try using a PCB antenna instead, because it's easier to tune by simply use 1 shunt component and the length of the antenna as the 2nd factor to tune. We provided a tuning guide here.

Side question, how is the radio performance currently on your board ? How bad is the "radio link problems" ?

[Joe Wright]

Thanks for the reply.

1. Yes I'm using a pi network. I spotted the issue of the inductors as well but wondered if another network could be used. It seems the dtc's are being promoted for tuning so I wondered if it could be done.

2. Wouldn't length of PCB antenna tuning require getting more PCBs manufactured? I get my PCBs from China so it would be a lengthy process. Also I don't have network analyser. Really all I can do is digital testing by sending a load of packets and measure number of retries, etc... If I had a digital method of altering the tuning and matching then I could automate this process. Otherwise, I guess I'll have to wait until I can hand over to someone to do the final tuning/matching for me....

[Joe Wright]

Performance is variable at the moment. The first prototype was ok, second is a bit more flakey. There's been some changes in the design but I'm not sure how to explain the difference. However, it was tested in a different environment so maybe there were interference issues. For the next field test I be doing some channel scanning and better selection of channels (previously I just had the option of 2 different presets in my firmware).

I have also now got a receiver with a better antenna so that will help. Performance still is an issue in my workshop if I hold the enclosed sensor behind my back but I'm not sure how relevant that is.

Thanks, Joe

[Hung Bui]

Actually it's simpler than you thought. You can simply make the length of the antenna longer than calculated and then cut it off gradually with a knife. But I agree, it's harder if you don't have a network analyzer to check. But you can use RSSI and number of packet missing to verify the strength.

We can also do radio tuning for you for free. However, you need to come up with a final design, and send us all the final product with housing.

Human body can really affect the RF performance as 2.4GHz is absorbed greatly by water.

[Joe Wright]

Fantastic. Thanks very much for the support. Apologies for the delay in replying but I plan to get on with things this weekend. I did another field test a few days ago. It was better - perhaps because I implemented a channel scanner to pick the channel for the sensor, and a different environment. I'm encouraged by you message to have a go at the antenna.

I was wondering about a different frequency range, but I would require quite a bit more than the 50kbps than the nRF905 can do unfortunately.