Linear approximatio...
 
Notifications
Clear all

Linear approximation of PBS2 probe set  

  RSS

ee6055
(@ee6055)
New Member
Joined: 8 years ago
Posts: 2
06/01/2014 4:30 pm  

Hello,

I am using the PBS2 probe set along with the 40db amplifier with a NF-5030.

In the directory "EnglishManualsProbe-Set" of the accompanied CD there is an excel sheet (Aaronia_PBS1_Probe-Converter.xls) which can be used to convert the readings from dbm to dBV/m.
In the cell B42 I found the following equation which is used to perform the conversion for the 3mm electric field probe
B42 = B40 + 113.2 - 20*LOG10(B39) where B40 is the power in dbm, B39 is the frequency in MHz and B42 the result in dBV/m
dBV/m = dBm + 113.2 - 20log10(MHz)

Setting V/m = 1 => dBV/m = 0
The equation becomes dBm = -113.2 + 20log10(MHz)
With this matlab script I calculated and plotted the ebove equation for frequency values 0.1MHz to 10000MHz.

MHz = 0.1:1000:10000;
dBm = -113.2 + 20*log10(MHz);
semilogx(MHz,dBm)

The plot I get is shown in this attachment

Since I don't have any other calibration data for the 3mm E-field probe I resized the above graph to match the size of the graph for the e-field probe shown in the file RF-Near-Field-Probe-Set.pdf
I then used photoshop to superimpose my graph on Aaronia's graph and you can see the result in this attachment.

I can't locate any mistake in my train of thought and if you can find any please point it out.
It seems that this linear approximation favors the range of frequencies from 1GHz to 6GHz which is completely understandable but as I am using these probes for frequencies up to 1MHz, I could use a better approximation.

Can you provide me, and any other probe set user who might read this thread, with the data used to construct all the graphs in the RF-Near-Field-Probe-Set.pdf file?
So I can make my own equations depending on the frequency range I will be using each probe.

Thanks in advance,
George Psarras


Quote
Moderator
(@moderator)
Member Admin
Joined: 11 years ago
Posts: 1048
07/01/2014 8:08 am  

Sorry i dont understand your posting. Your formular works perfekt, as you can see at your graph, at any frequency, so where is the problem?

Snip


ReplyQuote
ee6055
(@ee6055)
New Member
Joined: 8 years ago
Posts: 2
07/01/2014 4:13 pm  

You can easily see that there are some frequencies where using the linear equation results in an error of 3db.
Here is an example I thought to support this claim.

Suppose that there is a time varying electric field of unknown magnitude but with a known frequency of 300KHz.
Let's say that someone uses the 3mm probe to measure this electric field's magnitude and gets a reading of -120dbm at 300KHz.
By using the equation on the excel sheet, the result for the field's magnitude would be 3.66dBV/m or 1.52V/m.
But that is obviously wrong since according to the Aaronia's graph, the required magnitude for a -120dbm reading is 1V/m.

If I use the probe to measure an electric field of 100V/m (40dbV/m) at 300KHz the reading I should get on my MCS screen will be -80dbm*, right?
But if enter these two numbers (300KHz and -80dbm) at the excel sheet the result I would get for the field's magnitude would be 152V/m.

I am constantly mentioning the frequency of 300KHz because it's a frequency I will be working on. As I can see on the graph there is also a 3db error
at 200MHz and other frequencies. That is why I would like your original data used to plot the probe's response, to construct a different approximation
depending on which frequency range I will be using.

If you locate any mistake anywhere above, point it to me so I can correct myself and I thank you for your time again.

* -120dbm(@1V/m) + 40dbV/m = -80dbm


ReplyQuote
Moderator
(@moderator)
Member Admin
Joined: 11 years ago
Posts: 1048
07/01/2014 6:13 pm  

The graph starts going "out of the perfect line" not because of the real probe readout but of the measurement test-setup (that is why the graph stops around 9GHz) so the calculation via the Excell sheet is perfect, dont care about the graph.

Snip


ReplyQuote
Anonymous
 Anonymous
(@Anonymous)
Guest
Joined: 1 second ago
Posts: 0
21/07/2014 10:14 am  

By using i'll be which will demonstrate a example of methods.


ReplyQuote
Moderator
(@moderator)
Member Admin
Joined: 11 years ago
Posts: 1048
12/08/2014 11:02 am  

the thread was updated

Snip


ReplyQuote
Share: