433MHz transmitter ATA8402 Application note

Has anyone here come across this application note for the ATA8402 (T5754) Atmel 433MHz transmitter?

formatting link

I have very little RF experience and I might be mistaken, but this AN is very dodgy indeed. I'm trying to make sense of the impedance-matching capacitor calculation.

The main problem is that the equations are not accompanied by numeric examples for the demo board handled in the same AN. Apart from this, the equations given don't take into account the RF choke, nor the added Lx,Cx filter in the demo board.

But OK, I don't even understand the matching-capacitor equations as they are presented. If I understood correctly, we get Z|| from Eq. 8, then use Eq. 9 to calculate r, which we later use to get the matching capacitors. To get r, we need to use the ideal Zload, but this is given as

166+j223 Ohm, i.e. a complex number, while Z|| and r are obviously real numbers. So, I'm stuck at Eq. 9 and have no idea how to proceed. I have asked Atmel support what Zload is and received a short and (for me) useless response stating that Zload is complex, but giving no hint as to how the quotient of a real and a complex number should result in r, a real number. I asked again and have received no response.

I have since read application notes for competitors' TX chips, some of which are well written and comprehensible even for an RF beginner like me, but their IC's do have real output impedances so this problem doesn't arise.

Thanks for any suggestions.

Viktor

Reply to
Viktor
Loading thread data ...

I haven't used these chips. I decided that, without proper test gear, it really wasn't worth my while getting involved in creating impedance matching networks and all the design problems that throws up, so I used an RFM12B (from

formatting link
instead which is a small surface mountable complete radio module based on the SiLabs Si4421 chip. These are very popular modules and lots of application notes and user projects on the 'Net. They are also very cheap (around $2.20 in volume). I would highly recommend this module, I have used it very successfully and it has excellent range and sensitivity.

Mark.

Reply to
markp

I have not used the RFM12B but the OP's post immediately made me think of it as I've been considering it for a project. It seems to be all you claim for it, and is stated to have automatic impedance matching of the antenna.

Reply to
pimpom

Yes, it really is a nice module (the Si4421 is a great little chip). It even has a wakeup timer which is very low current, 1.5uA if I remember, and that was even lower than my micro, plus the resolution of that timer can be programmable to much finer resolution than my micro. I used that to achieve pseudo-random wakeup times. It runs from 3.8V down to 2.2V too which is ideal for 2 alkaline cells.

Reply to
markp

.

of

aim

na.

ven

at

ve

Thanks for the pointer about the Silabs range of RF IC's. I don't think the modules actually solve the antenna matching problem since they don't have the antennas on them and I suspect the IC's RF pins are just connected to the module header. (Correct me if I'm wrong here.)

What kind of antenna do you use and how did you match the impedance?

There's some vague reference to an auto-tuning circuit in the IC datasheet, but the optimum load impedance is also given as 52+j152 Ohms for 433MHz, so it seems you have to achieve that for optimum power output. I looked at Silabs AN369 and saw that they took a different approach with the PCB loop antenna matching problem. They used an EM modelling software to tweak a rather complicated loop antenna to have just the right impedance for the IC's PA output. This approach seems to me in the million_monkeys_eventually_writing_Shakespeare category.

Anyway, I'm stuck with Atmel's chips for now. Both the receiver and transmitter are past the prototype stage. I'll just have to tweak the caps by trial and error or settle for less than maximum attainable range.

Thanks again,

Viktor

Reply to
Viktor

The RFM12B is a module that uses the Si4421 chip. This module is available in 3 versions, 433MHz, 868MHz and 915MHz. The difference is that the antenna matching components are slightly different for each version. The Si4421 chip itself is capable of using all of the RF range with minor firmware changes (in fact if you look on page 39 of the Si4421 spec, you'll see the matching circuit the RFM12B uses and the changes in components for the three ranges).

The antenna interface on the RFM12B is designed to directly connect to a 1/4 wave whip antenna or 1/2 wave dipole, so I just use 16.8cm of wire and the ground plane of the board acts as the ground plane itself (1/4 wave whip). The Si4421 actually impedance matches to a degree itself to get the best out of it.

So, no impedance matching problems at all, and no additional components. As long as your antenna looks like a whip or dipole electrically it can be a printed PCB antenna or anything else.

Mark.

Reply to
markp

m...

AN

ric

s,

s
o
g

ld

e

er

r,

nd

nk

t
s
e

nna

hip

s

ng

s).

1/4
e
.

out

As

Yes, matching whip antennas is no problem. They are around 50Ohms and you always get a schematic with recommended component values. There's one on my receiver and I had no problems at all with it. The transmitter, however, is one of those car-key sized thingies and must have a PCB loop antenna.

I haven't yet come across an application note that explains the calculation of matching capacitors for a PCB loop antenna where the PA output impedance is complex. Still, the truth is out there..

Viktor

Reply to
Viktor

I use to work with small UHF Tx devices above 300MHz, we put a trimmer in them, tuned against Motorola spectrum analyser, this was in the '80s.

Antenna was either copper wire loop onto PCB, or PCB track, depending on product size, battery type.

I suppose that's gone out of date, people don't want a trimmer in there these days?

Grant.

Reply to
Grant

Well surely the antenna should be in resonance at the desired frequency, ie. real not complex? The beauty of the Si4421 is that it contains circuitry to change its capacitance internally so that any detuning of the external resonant antenna, for example by hand effects or component/mechanical tolerances, are compensated for. This would I presume change the complex output impedance to match that of the detuned circuit, and it does so dynamically.

If you really want to take all the hassle out of it, use this:

formatting link

I've used that Splatch antennae and it is very good. It is less immune to hand effects than others. It requires a ground plane (I used a PCB 70mm x

35mm with the Splatch at the top and the lower part of the PCB flooded copper pour for the ground)

Mark.

Reply to
markp

ie.

to

PCB loop antennas at resonance have a relatively high, real impedance but the output transistor of the TX IC has a small output capacitance, which at 433Mhz, comes to -j50... -j1k (depending on the IC). That's why you need to anull it with an inductive load. The Si4421 would need to lower its own output capacitance to zero to be able to adapt itself to a resonant PCB loop antenna without any matching components. (One would still need impedance matching to adapt to the loop's high resonant impedance, though.) I didn't find much about this automatic impedance matching feature, though it seems fantastic and they should be making a great fuss about it. How does it work? What's the range of antenna impedances it can cope with? Will it still work when the 50Ohm whip-antenna matching circuit is present (as on the modules)?

m x

This looks great. You can use the matching circuit for a regular whip antenna. Unfortunately it's too big for this application, especially with the ground plane.

Viktor

Reply to
Viktor

OK, I don't really know how the antenna matching part works.

My undestanding though is that the matching network they use for the RFM12B (i.e. Si4421) is such that it is designed so that you connect a 50ohm antenna directly to it, such as a 1/4 wave whip. The hassle of calculating the impedance matching components for the driver for that is done for you. This I assume means that if you can arrange your PCB loop antenna to be electrically equivalent (by placing inductors or capacitors or by physical layout) then you're done. Whether this is the most efficient way of doing it I don't know, it may be that a different matching network specifically to match the driver output impedance directly to the PCB loop would be. However, I feel this may impact on the ability of the driver circuitry to actively match because it may well be designed to use an external matching network that interfaces to a 50ohm antenna. Given that the matching networks they suggest must be designed to be efficient at matching the 50ohm load, your best bet might be to use a balun type network to interface with your PCB antenna.

Mark.

Reply to
markp

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.