how to measure a signal that bursts packets for 0.64ms

[rubdawg]

burst power mesure

Hi, I am working on Zigbee (802.15.4) int the 2.4GHz band. I am trying to measure the tx burst that takes 0.64ms to transmit a packet, but with the spectrum analyzer the sweep time cannot go below 1ms. I am trying to trigger the Spectrum analyzer (PSA 4443A) but not sure if I just tap from the VCC line on our board to trigger the spectrum analyzer , or what other methods would be possible. Also I am trying to mimic the packet signal with a signal generator but ours does not have O-QPSK signal, but I know i can offset Q which would be similar correct?? Any insight would be grateful. Basically I am trying to measure PSD mask (power spectral density).

Just to give a background, i have worked on GSM and CDMA handsets and usually had a call box that handled measuring

 

[flatulent]

One trick is to use max hold and let the analyzer run for a while. It will eventually catch the signal on for enough times to fill in the spectrum display.

 

[boy]

You may try to used gated mode. I have done a burst signal PSD measurement by using the gated function in the PSA b4. If you need only to see the PSD, gated mode is ok. You can also see the power level as a function of time. This may help -> 5968-2320E.pdf in agilent application note . Actualy, this application note is better 5988-4948EN.pdf

 

[rubdawg]

I am wondering for to measure PSD , the specifications states like this:
The transmitted spectral products shall be less than the limits specified in the table below. For both relaitve and absolute limits, average spectral power shall be measured using a 100kHz bandwidth. For the relative limit, the reference level shall be the highest average spectral power measured within +/- 1Mhz of the carrier frequency.

Frequency relative limit Absolute linit
|F-Fc| > 3.5Mhz -20dB -30dBm.

I am having a hard time trying to visualize what this means and the specifications does not havea diagram to explain this.

 

[martinthorn]

One possible way to measure a signal spectrum is as follows :

i) Set analyser RBW to required value (100kHz)
ii) Trigger analyser from Vcc line as you suggested
iii) Set analyser to zero-span (time on the x-axis)
iv) Set analyser frequency at desired PSD offset : in specification
mentioned this will be any frequency at least 3.5MHz away from
signal/channel centre frequency. A measurement at 3.5MHz offset
perhaps will give worst case result.

The PSD can then be measured : the burst power, at the measured
offset, should appear above the noisefloor.

v) repeat measurement at f=fc to get signal reference power

Take care on the number of averages taken (this may be specified)
and ensure that you measure the power at the correct time : not
during training bits for example, just during the random modulation
part.

Hope this helped !

Martin

 

[boy]

Look at chapter 8 in your "PSA Series Spectrum Analyzers Measurement Guide and Programming Example". There is an example in there. I think in that chapter have everything you need. THey start with how to measure the burst signal , and then the CCDF of the power (you may not need this) and APC measurement (this is for your new question)
good luck

 

[rubdawg]

I triggered off out TX Enable line(this turns high each time tranmitting a packet) and I end up getting this on our PSA. For 802.15.4 the channel is 2MHz wide and right now it is contiunously sending packets of 20 bytes which equal to 160 bits and the bit rate is 250 kb/s so ti takes about 0.64ms to transmit a packet, should i change the packet anyway to easily capture the signal. I haev attached a jpeg of what I se using the external trigger. I have the trigger set up for positive trigger. ANy help would be greatful. Thisi s for the 802.15.4 specification, i can provide more detail if needed. Thanks. Also note that the PSA does not have the 802.15.4 standard preconfigured since it is still not finalized from IEEE.

 

[rubdawg]

Also is the gated measurment a standard function in the PSA, I do not think mine has that option.

 

[eemapir]

hi
you can do it with maatlab
bye

 

[boy]

If your modulation is OFDM based, your PSD looks ok for me. The asymetrical, all i know, may cause by AM-PM conversion. How do you get the baseband signal? Any possible non-linear device between your PSA and the source. You may drive the mixer too hard. Also, the nonlinerarity of multicarrier system (OFDM) is quite complicated. make sure you back off enough form P1 compreesion point of every RF components. To be save 10 dB back off, I would suggest.

 

[rubdawg]

it is O-QPSK

 

[boy]

In that case, your PSD will have the same shape as the pulse shaping filter. Can you access to the baseband signal? if you can just measure tha bandband signal first.

 

[rubdawg]

I wil check to see if I can get to the baseband, but most likel not because the RFIC houses the TX AND RX circuitry inside the chip.