Compiled From The Internet


In article 
Gordon Pritchard wrote:

I've only heard of this 6dB/S-unit by way of *rumour*, too. I've followed this discussion for a bit, but don't know if anyone has mentioned it already. 

The 6dB/S unit figure is covered in the 3B question pool.  It's never been clear to me which text is used as a refrence for any of the questions, but I presume (hope, actually) they are taken from some sort of reliable engineering text.  I think the pool needs a Bibliography!
Jim --KE6JPO

Greetings to All that read this from VE7TMA:

The ARRL handbook refers to an old standard of  S9 calibrated at 50 microvolts RF input and 6 dB per S unit down from that reference point.

The section also clarifies this by stating in reality there is seldom two rigs that follow the same absolute calibration curve considering different designs and gain variations from band-to-band.

It is stated that the S meter is useful for relative comparisons on the same rig operating on the same band.

Although the accurate S meter calibration subject seems somewhat of a dead issue for the transceivers that have been manufactured in the past it is interesting to play with the numbers considering the sensitivity of a typical modern rig and the calibration reference of S9 for a 50uV RF input.

I ran a little program on my calculator to calculate the RF input voltage levels for a  S meter scales of 1 through 7 dB per S unit  with S9 set to 50 microVolts as a reference for each calibration curve.
The following table gives the results:

                           dB PER S UNIT                                                                             

S mtr|    1dB       2dB      3dB      4dB      5dB       6dB      7dB
==== |    ==      ===      ===      ===      ===       ===     ====

S9   |    50.0      50.0       50.0     50.0      50.0      50.0       50.0
S8   |    44.6      39.7       35.4     31.5      28.1      25.1       22.3    
S7   |    39.7      31.5       25.1     19.9      15.8      12.6       9.98   
S6   |    35.4      25.1       17.7     12.6      8.89      6.29       4.46
S5   |    31.6      19.9       12.6     7.92      5.00      3.15       1.99
S4   |    28.1      15.8       8.89     5.00      2.81      1.58       0.89
S3   |    25.1      12.6       6.29     3.15      1.58      0.79       0.40
S2   |    22.3      9.98       4.46     1.99      0.89      0.40       0.18
S1   |    19.9      7.92       3.15     1.26      0.50      0.20       0.79
S0   |    17.7      6.29       2.23     0.80      0.28      0.099     0.035

        \________________________________     _________________________/

                           RF  INPUT LEVEL IN MICROVOLTS RMS

It is interesting to consider the sensitivity of a typical modern HF transceiver. I checked the sensitivity specification  for three modern SSB transceivers and found it varied from 0.20 to 0.25 uV for a 10 dB Signal\Signal + noise ratio.

If one were to consider the sensitivity specification as roughly equivalent to an S zero threshold then it would appear that a calibration factor of 5 dB per S unit is the closest round number fit to the 10 dB S/S+N sensitivity  with a calibration reference of S9=50 uV.  The 6dB per S unit curve would result in a signal to noise ratio of about 3 dB S/S+N which most likely would be below the receiver AGC threshold (which is typically utilized to drive the S meter circuit) which would tend to be impractical to implement.

A  nice quality of a 6dB/SU curve is that the RF input voltage doubles for each S unit increase of signal strength (representing increase of input power of 4 times for each additional S unit) which would be a
practical rule-of-thumb to remember.

I included calibrations in the table down to 1 dB per S unit as the original message posted on this subject thread made mention of a rig with a 1 dB/SU S meter response. It is interesting to note that if S9 were equal to 50 uV on that scale (which it probably is not) the S/S+N ratio of an S0 signal would be about 48 dB which is almost equivalent to the noise performance of a telephone circuit!

That S meter scale would be of little practical value. I have seen similar calibration characteristics in a multi-mode 2 meter transceiver that I own. It is reasonably well calibrated in the SSB mode but pretty hopeless during NBFM operation.

This is a case were the S meter calibration characteristics change drastically on the same rig when switched between operating modes!

I think the current technology could support an improvement in the calibration of S-meters.

For example the National Semiconductor NE604 IF amplifier IC provides an accurate signal strength logarithmic output that closely tracks the input signal level over a wide dynamic range that could possibly be used for driving an S meter circuit. Considering the consistency of surface mount and wide band technologies coupled with the ability to program multi band calibration factors into the microprocessor S meter firmware it should be practical for manufactures to now provide reasonable S meter calibration accuracy.

An S meter tracking accuracy specification should be included as part of  the overall specifications. This specification would tend to take marketing biases out of the S meter calibration scale.

I hope the manufactures take steps to clear up this age old guess meter confusion  so that radio amateurs have the ability to routinely measure signal strengths in a reasonably accurate manner in the near future.

Possibly the amateur radio community may need to lobby the manufactures to give the issue some priority.

73 de VE7TMA 
S-Meter microVolts vs S-Units
Based on the accepted standard of 50uV = S9

MicroVolts      S-Units
   0.2                 S1
   0.4                 S2
   0.8                 S3
   1.6                 S4
   3.2                 S5
   6.3                 S6
  12.5                S7
  25.0                S8
  50.0                S9
 158.0              S9+10dB

On The Other Hand -- Here is what some others measured
Readings on my IC746 are:
S1-2 1dB
S2-3 2dB
S3-4 2dB
S4-5 3dB
S5-6 3dB
S6-7 5dB
S7-8 5dB
S8-9 5dB
s9-9+10 10dB
+10-+20 10dB
To nearest dB - checked overall by dropping from S9 to S1 - total attenuation is within 1dB of the sum of the increments.

Pre-amp seems to have about 12dB gain - at least on 17M.

Here are measurements of my Icom 730 S meter on 40 meters with preamp out:

S1 - 2 1.4 dB
S2 - 3 1.3 dB
S3 - 4 1.6 dB
S4 - 5 2.3 dB
S5 - 6 1.8 dB
S6 - 7 3.2 dB
S7 - 8 3.1 dB
S8 - 9 4.0 dB
S9 - S9+10dB 5.6 dB
S9+10dB - S9+20dB 7.3 dB
S9+20dB - S9+30dB 6.6 dB
S9+30dB - S9+40dB 10.5 dB
S9+40dB - S9+50dB 11.3 dB
S9+50dB - S9+60dB 13.5 dB

Another reports:

There is NO standard for S units, or even for S-9. There are a bunch of suggestions, but that's all. The suggestion the ARRL makes varies with frequency.

Most of my receivers, when designed, tried to use about 5dB per S unit. They have various S-9 calibration points. Very few receivers attempt to use 6 dB per S unit compared to the one's using 5dB or less.

My dozen or so receivers (Drake R4C's, IC-751A's, Yaesu FT1000D, Collins 75S and KWM-2, etc) all range from about 1 or less dB per S unit at S1, to maybe 3 or 5dB per S unit near S-9. 

Maybe time for National Semiconductor NE604 IF amplifier IC Project!





Last Update:July 31, 2002