Author Topic: inAir Module Oscillator  (Read 1136 times)

wireless_enthusiast

  • Newbie
  • *
  • Posts: 2
    • View Profile
inAir Module Oscillator
« on: October 07, 2015, 06:03:59 PM »
The sensitivity of the inAir modules is shown as -148 dBm, but only with a TCXO.  Do you plan to offer a TCXO option for these modules?  What type of oscillator is included on the modules that are being sold? 10ppm crystal?  Could the modules be easily modified to use a TCXO, or would this require an entirely new board?  Thanks.

modtro2

  • Administrator
  • Hero Member
  • *****
  • Posts: 564
    • View Profile
Re: inAir Module Oscillator
« Reply #1 on: October 08, 2015, 02:56:20 PM »
Hi, we state the sensitivity as -139, and not -148. From the product page:
Quote
High sensitivity: down to -139 dBm (-148 only possible with TCXO)
With the current hardware -148 is not possible. It would require a hardware redesign using a TCXO, which we do not have planned.

The baud rate for -148 would be very, very low with not too many applications. It is only about 18 bits per second! For -139dBm is is also very low at 146 bits per second, but still usable. The higher the dBm, the higher the speed. For example, at -120dBm, it is 12500 bits/second, and at -112dBm it is 37500bits/seconds. The dBm is configurable via software.

Regarding the crystal, we use one of the crystals recommended by Semtech. We get these specially manufactured, and had a couple of months lead time just to get the crystals. They are Tolerance=10ppm and G-Senstivity=2ppb/G max.

wireless_enthusiast

  • Newbie
  • *
  • Posts: 2
    • View Profile
Re: inAir Module Oscillator
« Reply #2 on: October 09, 2015, 05:59:02 PM »
Thanks for the quick reply.  Much appreciated.  I would like to clarify my question.  My interest is in amateur radio UHF QRP communications.  All I would need to transmit is a call sign and signal report.  18 bps is still highly usable for this purpose and is, in fact, much faster than popular alternatives such as slow speed CW (morse).  My ultimate goal is to achieve the greatest possible sensitivity at any data rate.  I understand that you have no plans to redesign the hardware to include a TCXO.  However, would it be theoretically possible for me to modify your existing hardware to accomplish my goals?  It looks like all I would need to do is remove the crystal, connect a 32 MHz TCXO to the XTA pin on the Semtech chip, and then change the RegTcxo register value?  Looking through the Semtech datasheet it appears that this should enable the lower bandwidths that yield the greatest sensitivity.  How stable (ppm) do you suppose the TCXO would need to be?  The datasheet does not specifically address this.  What considerations am I not considering (trace length, decoupling caps, clock supply filtering, etc.)?  Any advice would be much appreciated.  Thanks!