[1000mp] Fwd: Re: Mark V key click/noise blanker modification

[Pete Smith]

This was in response to my questions to Tom about OH8NJ's data.  To be fair 
to Jaakko, he acknowledged that his NB ON data were probably compromised by 
the fact that he used steady signals for his IMDR measurements rather than 
something intermittent.  Seems to me this is essentially what Tom is saying 
below...


>Return-Path: <w8ji@contesting.com>
>Reply-To: "Tom Rauch" <w8ji@contesting.com>
>From: "Tom Rauch" <w8ji@contesting.com>
>To: "Pete Smith" <n4zr@contesting.com>
>Subject: Re: Mark V key click/noise blanker modification
>Date: Fri, 25 Jul 2003 10:14:43 -0400
>
>You are welcome to post this Pete if you think it will help anyone.
>
> > That's what I and a number of other people had not understood.  And that
>is
> > also why I was hoping you would publish an explanation.
>
>I didn't understand what needed explaining, and now I think I do.
>
>Measurements by the other fellow actually agree with my data except for his
>inclusion of NB ON data. NB ON data is useless with a two-tone test of equal
>steady signals, I'll explain why later.
>
>The first thing everyone should understand is that every measurement has
>tolerances. When we see data carried out to 0.1 dB, we assume it is accurate
>to that amount. Unfortunately that is almost never the case. We see examples
>of overstated accuracy everywhere, where the data might have a 10dB
>tolerance but the data is given to hundredths of a dB. For example the
>accuracy of a properly calibrated Bird 43 meter is + 5% or - 5% of full
>scale anywhere on the scale. The scale accuracy of an old spectrum analyzer
>is considerably more than one dB. Yet one commonly accepted report using
>those very instruments sorted products based on tiny fractions of a dB, when
>the system likely had 2-4 dB measurement accuracy!
>
>In the case of IM3 DR, measurement tolerances with good equipment can be
>over one dB. That is with exceptional care and modern equipment. I use
>calibration certified newer gear for critical measurements. For example my
>signal level-measuring equipment resolves to 0.01dB, is less than two years
>old, and is calibrated yearly and always double checked before use. Despite
>that, I can not claim better than a few dB tolerance in my IM3 data. Even
>though I COULD publish something like "89.31 dB", I would be fortunate if it
>were within one dB.
>
>Another thing, even the same model receiver will have some differences.
>Components and adjustments have tolerances, especially when operated near
>overload.
>
>Despite this, his measurements and my measurements closely agree. Certainly
>they are well within method and equipment tolerances. Both sets of
>measurements show about 10dB of dynamic range degradation in narrow-spaced
>signals with the stock Yaesu NB system when the system is OFF.
>
>The only major flaw I see, other than the common tendency to over-imply
>accuracy by carrying data out to useless resolution (which almost everyone
>does anyway), is he published data that includes NB ON. That will not work,
>any data taken in that condition is useless.
>
>NB ON data is useless because of the design and operation of noise blanker
>systems. Noise blankers contain an internal AGC loop that reduces gain to
>set trigger threshold at a level just above the short term input signal
>level. This is done so peaks above that level disable a gate, turning the
>receiver off. When a two-tone carrier test (like the IM3 DR test) is used,
>the NB AGC automatically corrects NB gain settings. Obviously with a steady
>carrier level the NB won't affect the IM3, if it is working properly when
>on. We should expect no IM3 problems with a properly operating NB when
>signal levels are CONSTANT.
>
>The problem comes in when signals vary in level, such as a SSB or CW signal
>or group of signals. In this case the NB punches a hole in weak desired
>signals whenever a strong signal appears nearby. The result is what sounds
>like clicks, desense, IM, and combinations of other receiver distortions or
>problems. That's why most experienced operators know to NEVER use a NB when
>someone else is on the band anywhere near them, and why novice operators
>sometimes use the NB and whine and complain about other signals being
>"wide". If you doubt this, leave your NB on throughout the next contest and
>see how well the receiver works. See if it has the superior IM3 performance
>indicated by the fixed and equal level signal test!
>
>The IM3 data with NB "on" is totally useless, and has no bearing on anything
>because it is a steady-tone measurement of a system designed to
>intentionally distort transitions in signal level. If we wanted to test NB
>ON IM3, we would have to use a third pulsed tone to trigger the NB.
>
>So all the data agrees closely. When the NB gain is UP, and when the NB is
>ON, the FT1000MKV has serious receiver degradation. It can be 10dB or more
>degradation. The MKV is the same as in the MP, except gain adjustment is now
>hidden in a menu instead of being on a front panel control. Even with the NB
>gain fully down, the receiver has degradation from poor NB input design that
>Yaesu has copied from rig to rig. When the NB gain is up, the degradation is
>worse.
>
>The mod I suggest has absolutely NO effect on wide spaced signal performance
>(outside roofing filter passband), and offers the BEST close-spaced
>performance possible with the receiver. The one or two dB "disagreement" by
>the other data is due to normal measurement tolerances and complicated by
>over-implied accuracy. It is NOT an actual change. I'm absolutely positive
>of this.
>
>73 Tom

73, Pete N4ZR
The World HF Contest Station Database was updated 17 June 03.
Are you current? www.pvrc.org/wcsd/wcsdsearch.htm