SparkFun Forums

[bill951]

I know that about 112 mils is correct for 50 ohms below 1 GHz.
A few mils either way is not critical. The FR4 dielectric may change
a little at 2.4 GHz but 112 mils should still be close enough unless
you have a long run.

He may not appreciate me giving out his name but I've received
good information from:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jon Gunnar Sponaas, M.Sc.E.E.
Field Application Engineer, Wireless Communication Nordic VLSI ASA Vestre Rosten 81
N-7075 Tiller
NORWAY
jon.gunnar.sponaas@nvlsi.no
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[bill951]

I've got almost all of the pieces parts spec'd out for the nRF2401 circuit in their documentation, except for the inductors b/c I can't seem to find parts with the same tolerances.

Can anyone suggest a supplier for the pieces?
Jeff

Having worked with the nRF2401, the crystal is the most critical part.
I don't know about the Abracon crystal SparkFun is selling without
a complete part number (can someone give us that?). My biggest
concern would be whether the load capacitance is too large. I do know
that the MA03057 which can be purchased directly from Golledge:
http://www.golledge.com/docs/products/xtl_sm/gsx433.htm
works extremely well. Remember that 16 MHz is being multiplied up to
2.4 GHz so a little bad at 16 MHz is a lot bad at 2.4 GHz. If you use the
50-ohm matching network of Figure 20 on the nRF2401A data sheet I
would recommend capacitors as Mouser (AVX microwave size 0603):
581-06035J4R7BBT, 581-06035J1R0BBT, 581-06035J2R2BBT
These are +/- 0.1 pF which in some cases is better than specd.
For inductors I don't think you can beat coilcraft and I would use
0603CS-5N6XJBU and 0603CS-10NXJBU which are 5% tolerance.
If you want 2% tolerance the numbers end in GBU instead of JBU.
You can purchase directly from coilcraft and they have free samples:
http://www.coilcraft.com/0603cs.cfm

If anyone tries the chip antenna from DigiKey with this network, please
post and let us know how it works.

[sparky]

Abrcon Part # : ABMM2-16.000MHZ-E2-T

We've had good sucess with this SMD xtal. You're right, the load capacitance is high, but we could not find a SMD xtal with the Nordic recommended max 12pF load. Indded, the one you spec'd is still 16pF.

Nice part though. Better stability than the abracon part. Any Xtal can be procured at any spec. The main thing for us at SFE is price. The abracon part at 1-$2 is a steal. A SMD xtal can easily run 5-$20 depending on tolerances.

My main point - the abracon part works. I imagine a better xtal, tighter tolerance parts, and a better matching network would increase the range nominally. Then again, a 2.4ghz SMA antenna would do wonders as well. The PCB etched antenna on the SFE boards works, but it's a cost thing again.

What I'd love to see is a nordic setup with a Power Amplifier and an external Antenna. Talk about range....

-Nathan

[MGP]

Hey Sparky -- I've done a couple prototypes of the nRF905 and nRF2401 with SMA connecotrs and Linx 1/4W whip antennas for a couple of my clients. They work great and definitely extend the range. Although we weren't really looking for extended range, the best effect I've observed is that the packet error rate goes down quite a bit using whip antennas.

Using a "real" whip antenna is obviously more expensive than a PCB strip antenna but in my case we were after the best signal we could get with what was already there and cost wasn't an issue.

[Pete-O]

Better packet error rate, eh? I wonder if it's a BW thing...doubt it, but I'd love to get one of there on a network analyzer. If ever I could afford one of those.

Pete

[SOI_Sentinel]

Hmmm... I'm personally looking into creating an 802.15.4 design based on either the Motorola MC13192 or Chipcon CC2420. I'm leaning towards the Chipcon due to the integrated antenna switch (not looking to throw in amps to the legs or use separate antennae), so smaller PCB and simpler design for the same results. I'm definitely going the route of the printed folded dipole as cost is important I want to hobbyist-mass-produce them eventually.

I have another suggestion for crystals:

http://www.mouser.com/index.cfm?handler ... deid=69510

Digikey also carries them IIRC.

[bill951]

Indded, the one you spec'd is still 16pF.

What I'd love to see is a nordic setup with a Power Amplifier and an external Antenna. Talk about range....

-Nathan

GSX-433 defaults as option D which is 16 pF. However, the
MA03057 variant is 9 pF load capacitance.

Throw on a Mouser 551-UPG2301TQ-A power amp chip and a Peregrine
switch or two:
http://www.peregrine-semi.com/prd_switch.html
Nordic has an appnote on adding a power amp.:

http://www.nordicsemi.no/files/Product/ ... -feb03.pdf

[monzie]

I recently bought 2 NRF2401A Mirf modules. I have them hooked up to a pic and used the example source code as a starting point. Reception seems quite poor ( seconds between each time Data Ready goes high). I don't believe that is just noise as if I turn off the transmitter the DR never goes high.

I have added big delays between the data and clock when writing to the nrf2401 thinking that my timing was too fast. This doesn't seem to help.

I noticed in the comments in the Sparkfun source code that a large percentage of packets were being dropped. Is this a problem with the NRF2401A? Likely it is some error on my part but I have worked with rf chips before and haven't encountered this type of problem. I am using the shockburst data transmission mode at 1MB.

[sparky]

Read through some of the other nRF posts. The configuration of the nRF plays heavily into the number of dropped packets. 250mb, 16bit CRC, and 32bit addresses seem to be the sweet spot for us. It probably changes on Thursdays.

Btw - nice crystal find SOI. We'll drop a couple onto the next digikey order.

-Nathan