Author Topic: Multiplex Power accuracy  (Read 13905 times)

Robert Sack

  • Guest
Multiplex Power accuracy
« on: January 02, 2014, 09:18:04 pm »
Hello Mr. Kolar,
I would ask you to check the accuracy in relation to the linearity of the measured value.
Let´s assume the measurement is correct for 0.0dbr, than it isn´t correct for e.g. -0.2dBr because the showed value is -0.3dbr.
I know it´s less but we here in germany do like good accuracy around 0dbr.
Do you have the possibility to change the allocation of the value of the right in the brackets to the dbr- Value?
In the Moment it is in this way:
0.79           = -1.0dbr
0.80 - 0.82 = -0.9dbr
0.82           = -0.8dbr
0.84 - 0.86 = -0.7dbr
0.87           = -0.6dbr
0.88 - 0.90 = -0.5dbr
0.91           = -0.4dbr
0.92 - 0.95 = -0.3dbr
0.96           = -0.2dbr
0.97-0.99   = -0.1dbr

It should be in this way:

0.50           = -3.0dbr
0.51           = -2.9dbr
0.52 - 0.53 = -2.8dbr
0.54           = -2.7dbr
0.55           = -2.6dbr
0.56           = -2.5dbr
0.57 - 0.58 = -2.4dbr
0.59           = -2.3dbr
0.60           = -2.2dbr
0.61 - 0.62 = -2.1dbr
0.63           = -2.0dbr
0.64 - 0.65 = -1.9dbr
0.66           = -1.8dbr
0.67 - 0.68 = -1.7dbr
0.69           = -1.6dbr
0.70 - 0.71 = -1.5dbr
0.72 - 0.73 = -1.4dbr
0.74           = -1.3dbr
0.75 - 0.76 = -1.2dbr
0.77 - 0.78 = -1.1dbr
0.79 - 0.80 = -1.0dbr
0.81 - 0.82 = -0.9dbr
0.83 - 0.84 = -0.8dbr
0.85 - 0.86 = -0.7dbr
0.87 - 0.88 = -0.6dbr
0.89 - 0.90 = -0.5dbr
0.91 - 0.92 = -0.4dbr
0.93 - 0.94 = -0.3dbr
0.95 - 0.96 = -0.2dbr
0.97 - 0.98 = -0.1dbr
0.99 - 1.01 = 0.0dbr
1.02 - 1.03 = 0.1dbr
1.04 - 1.05 = 0.2dbr
1.06 - 1.08 = 0.3dbr
1.09 - 1.10 = 0.4dbr
1.11 - 1.13 = 0.5dbr
1.14 - 1.16 = 0.6dbr
1.17 - 1.18 = 0.7dbr
1.19 - 1.21 = 0.8dbr
1.22 - 1.24 = 0.9dbr
1.25 - 1.27 = 1.0dbr

Could you change this in an update?
Thank you very much in advance!
Best Regards,
Robert Sack

Jan

  • Hero Member
  • *****
  • Posts: 1128
Re: Multiplex Power accuracy
« Reply #1 on: January 03, 2014, 09:18:17 am »
This is a known rounding error of simplified enhanced range log10 function that is used for conversion from internal linear power ratio to dBr representation. The rounding error is much less than the system deviation (+/- 0.2 dBr). In your example with -0.25 dBr threshold the rounding error is only -0.027 dBr. At 0.0 dBr (0.99-1.01) the rounding error is zero.
Here in CZ the 0 dBr accuracy is important as well. After comparison to other devices, the P175 has been chosen by our national radio as a reference for setting up their transmitters.

Robert Sack

  • Guest
Re: Multiplex Power accuracy
« Reply #2 on: January 03, 2014, 12:40:30 pm »
0.99 - 1.01 would be exact for 0dbr. But the actual range for 0.0dbr is 1.00-1.02.
And is it possible to increase the internal resolution for the Multiplex power from 1/10dB to 1/100dB?
If you look at the graph, it would be a "rounder" course of the line an you would see more details.
Is the rounding error the same in the Pira175 as in the Pira75?
Thank you for reply!
Regards,
Robert Sack

Jan

  • Hero Member
  • *****
  • Posts: 1128
Re: Multiplex Power accuracy
« Reply #3 on: January 03, 2014, 01:44:39 pm »
You probably do not use the latest firmware version (1.4 r3). These characteristics are independent of the device type (P75/P175), only the firmware is relevant.

Update is available at http://pira.cz/fm_broadcast_analyzer/?art=support#fu
Note: If you still use the original firmware (2009) in your P75, switch on the unit after you click on the Start button.

I'll look what can be done with the rounding. Unfortunately the output format cannot be changed because existing software and systems may be unable to read the data correctly after the modification. Thus the power must stay in 1/10 resolution. This appeared to be optimal in 2006 when the device has been designed as in past the MPX power was not restricted here so it flied up to 9 dBr and the 0.1 dBr steps were not noticeable in the graphs.

Robert Sack

  • Guest
Re: Multiplex Power accuracy
« Reply #4 on: July 07, 2014, 05:50:57 pm »
Hello Jan,
Could you take a further look into the accuracy issue?
Thanks & Kind Regards,
Robert Sack

Jan

  • Hero Member
  • *****
  • Posts: 1128
Re: Multiplex Power accuracy
« Reply #5 on: July 07, 2014, 06:45:30 pm »
It's on TODO list. Since it is rather just an esthetic issue, solution is considered within regular update.

Robert Sack

  • Guest
Re: Multiplex Power accuracy
« Reply #6 on: July 07, 2014, 10:21:17 pm »
Thank you much in advance, that´s fine :-)
Kind Regards,
Robert Sack

Jan

  • Hero Member
  • *****
  • Posts: 1128
Re: Multiplex Power accuracy
« Reply #7 on: January 11, 2015, 08:17:56 pm »
The issue has been finally solved in actual fw version 1.5

RobertSack

  • Guest
Re: Multiplex Power accuracy
« Reply #8 on: March 11, 2015, 03:03:04 pm »
Hello Jan,
Thank you very much for the correction of this!
The curve looks so much smoother now around 0dbr, great!
Now I'm happy with it.
Kind Regards,
Robert Sack