Hello EriK,
???? Thank you for the reply, and thank you also for those very helpful comments about the SI4432 modules.
I have ordered the filter elements from the Ebay site that you pointed me to. Unfortunately, having to come from Australia, I won't get them until early January, which is going to be a bit frustrating. But what is great is that I know that they are the right bandwidth. Being a direct match to 50 ohms is also potentially quite useful.
Tony (G3PTD).

Will investigate how to do an installer
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

开云体育

For a 868MHz IF you would need? a LO module with 1736MHz maximum frequency in order to be able to exploit the whole range.
You could use the SI4432 with a 868MHz IF but as its maximum output frequency is about 960MHz you would only be able to scan from 0MHz till below 100MHz. Or you would have to switch to below IF scanning creating all kind of mirror problems.?I could not find a cost advantage either.
The special thing about the SI4432 is its continuous sweep range from 260MHz till 960MHz and the SSI with 120dB range. most other designs are banded creating gaps in the range they can cover or have less range in the SSI

Of course, if you are willing to pay more you can use very different components with very different performance, but that is not the subject now.
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

Have you given any thought to a version with 868MHz IF?
As thats also a common ISM band and filters for that are available.

I found a bug causing the tinySA SW to crash on certain PC configurations during startup
The bug is solved in the updated SA.zip file in the files areas
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

YES, same here.
If anyone figures out how, please post.
I don't want to buy the $300 software to run a $35 device...
Have tried downloading two .dll files it needed, installing free VisualStudio, Windows7,
nothing..? Current problem is complaint about entry point to the crt in the downloaded dll...
73,
Gary
WB6OGD

开云体育

Adding a 110MHz low pass filter at the input (almost) solved the problem
Here is the FM band using a small antenna


Still a bit of a hump below 90MHz
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

Extended build instructions uploaded including schematics and a picture of the filter
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

I think this is the filter to be used: Epcos B39431-B3555-U310-M1498

You can find the filter on ebay.

A seller from Poland has a lot? (and you can find a seller that sells two for an exorbitant price,=

Niels
OZ9NY

Hi Tony,

Filter:?
You probably have to build the 434MHz filter. The SAW filters are very cheap as these are used in remote control devices
EBay


Mouser:


I will post some pictures of how to build
There are only 3 critical spec points:
- Bandwith: Should be best 1MHz
- Impedance: In the datasheets there should be mention of 50 ohms and the matching circuit (if needed) should contain 33nH or 39nH inductors)
- Insertion loss: Mentioned in the data sheet and should be maximum 3dB as you need to use two and losing more than 6dB will decrease sensitivity


SI4432 module:
eBay has many in all different colors starting below 2$ (I bought 3 for 5$)

Apart from the SI4432 the module should have a clock generator IC (silver top) and an antenna switch (very small 6 pin black IC) close to the antenna connection
These are all based on the SI4432 reference design.
Make sure 433MHz is mentioned as there are some green modules that are actually 300MHz module and these won't work without removing the output filters on the LO SI4432 module.
The antenna switch is driven by the GPIO outputs. Without the antenna switch the SI4432 can not be used in high power output mode needed to drive the mixer.

--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

Hi there,
?? I would very much like to build and experiment with this exciting project and I have established that I can get all the bits needed to do it - except one.
??? I have not found a filter module seeller that gives the? bandwidth. Indeed most advertisers don't mention bamdwidth at all. Obviously Erik made his own so I decided to do the same. But I have looked at every list of EPCOS SAW filters that I can find, and none of those include bandwidth either!
??? So I am hoping that somebody can come up with a part number of a 1MHz bandwith version.

?? Secondly, the pictured Si4432 modules are on blue PCBs. I can only find black ones. Does it matter?

???? Of course I must finish by thanking Erik (If he reads this) for sharing this with us.
Tony
(G3PTD)

Ah, yes. Will correct
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

Hi Erik,
Re. the build instructions.

Connect the receiver module to D1and the LO module to D5 of the Arduino(WARNING: You MUST use a 3.3V compatible Arduino)

Should that be 'Connect the receiver module to D0' rather than D1 ?? (const int SI_nSEL[2] = { 0,5 };)

Thanks !
Jim

And here is an example where the tinySA fails miserably
Sweeping the signal from a small antenna between 0 and 200MHz?



The DAB transmission at 190MHz is recognizable but the FM band around 84MHz is very bad. As the tinySA still has no input low pass filter we are actually seeing an alias of the 950MHz signal from a nearby mobile phone base station. The same is true for the hump around 30MHz
Time to build a low-pass filter
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

When trying to measure the spurs of a SI5351 the first measurement caused some doubt on the performance of the tinySA.
Is the output of the SI5351 really that bad?
The output of the SI5351 is connected to the input of the tinySA using a 20dB attenuator
Output frequency of the SI535 is set to 135MHz.
A signal 346.6 MHz appeared and this should not be there because the LO low pass filter is still present and it has no direct harmonic relation to 135MHz



Then I remembered my own advice. Always use a 10dB attenuator to test if the harmonics you see are really part of the input.
This led to a much improved measurement, all harmonics above 150MHz are now suppressed.



The spurs around 135MHz are caused by the fractional divider in the SI5351 but what is causing the 76.6MHz and 211.6MHz signals?
Time to switch on the spur reduction. And that made a big difference.



The?76.6MHz and 211.6MHz signals are (almost) gone and this measurement is very much like when using a much better SA.
So keep in mind when you build and use the tinySA: it is easy to get wrong measurement results!
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

The CC1101 has the same problem as the SI4462. Its has a banded design so you can not sweep over a broad frequency range.?
The nrf24l01 is limited in dynamic range (only above -60dBm) and frequency range. It's only useful to scan the 2.4GHz band
The original RF explorer based on a single SI4432 had some serious problems. One is the bad image suppression of the SI4432, the other the inability to scan below 260MHz
The goal of the tinySA to get the simplest, cheapest zero till? above 100MHz SA.

Here you have an indication of the input frequency range without modifying the LO SI4432 (e.g. removing the TX low pass filter) by measuring the harmonics of a 5MHz square wave. You can clearly see the sensitivity drops quickly above 150MHz. Removing the TX low pass filter solves this problem.



The mixer and the 434MHz filter used in the tinySA solve the range and performance problems of the SI4432
Without the 434MHz filter you get this:



The peak at 4.4MHz is a mirror caused by the bad mirror suppression inside the receive SI4432.
By using a 1MHz wide 434MHz filter this mirror is eliminated

--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK

Awesome to see someomeone building this.
I have had the the? about using a cc1101 transceiver module as a logarithmic detector and even have some hardware built, but I had problems with the software.
Si4432 is a legitime detector choise the orginal RF Explorer handheld spectrum analyzer used it!
It's schematics should also be available for study.
I might have to build one of these for playing around with it, the PC side software really makes it special.?
Other modules of interest might be cc1101 as those ate even cheaper than si4432 and nrf24l01, which has a cw testmode and rssi receive register too.
The nrf might be usefull as a 2.4GHz test source.
There is a 2.4GHz spectrum analyzer project that uses it as the main device.