great technical writing

[William Sommerwerck]

I believe we discussing (arguing) the other day about what comprised good
technical writing. Here's a example of outstanding writing:

http://www.onsemi.com/pub_link/Collateral/AND9075-D.PDF

[William Sommerwerck]

"Don Pearce" wrote in message
...
On Fri, 12 Jul 2013 19:14:52 -0700, "William Sommerwerck"
> wrote:

>> I believe we discussing (arguing) the other day about what comprised
>> good technical writing. Here's a example of outstanding writing:
>> http://www.onsemi.com/pub_link/Collateral/AND9075-D.PDF

> It fails in the very first sentence "The data eye diagram is a
> methodology to represent and analyze a high speed data signal".

> First, it is a method, not a methodology. Methodology is the study
> and analysis of method.

> Second, it constrains itself to high speed data. No, it [IS] equally valid
> at any data rate.

> A couple of lines later it refers to folding the data signal to
> produce the diagram. This is never done as it would result in the time
> line reversing direction at each fold. No, the data is cut and shifted
> by a fixed number of cycles of the recovered clock.

> So no, from the start this is poor technical writing. It is typical of
> the lazy, sloppy writing that results when somebody with a poor but
> inflated grasp of English translates a badly written technical note
> into turgid prose.

I appreciate your taking the time to criticize, but you as you have repeatedly
shown you don't like me, I'm not surprised at what you wrote. (Only your point
about folding seems to be a criticism worth making. "Overlapping" would have
been a better choice.)

I'm editing a magazine article, and discovered I didn't properly understand
eye diagrams. I found this, and was startled that it was a coherent,
comprehensible piece. Your claims about its poor quality are beyond absurd.
You're making them simply to figuratively kick me in the nuts.

PS: The article I'm editing is one of those pieces where the author starts by
trying to make all his points in the first paragraph, then gives the details
as they come to mind (without any apparent organization development). And did
I mention that he repeats the same things over and over? I had to figure out
what he really wanted to say, then reorganize and (largely) rewrite the piece.
I challenge you to do as well.


> And finally, of course, I have to comment on your introduction -- very
> Street, almost Ghetto to omit the verb (I believe we discussing). And
> you made it a example instead of an example. Are you a Hip hop artist
> in your spare time, by any chance?

Both errors are typos. (You made the same "dropped verb" error in your third
paragraph, and omitted a space in the second.)

[Don Pearce[_3_]]

On Sat, 13 Jul 2013 06:33:09 -0700, "William Sommerwerck"
> wrote:

>"Don Pearce" wrote in message
...
>On Fri, 12 Jul 2013 19:14:52 -0700, "William Sommerwerck"
> wrote:
>[i]
>>> I believe we discussing (arguing) the other day about what comprised
>>> good technical writing. Here's a example of outstanding writing:
>>> http://www.onsemi.com/pub_link/Collateral/AND9075-D.PDF
>
>> It fails in the very first sentence "The data eye diagram is a
>> methodology to represent and analyze a high speed data signal".
>
>> First, it is a method, not a methodology. Methodology is the study
>> and analysis of method.
>
>> Second, it constrains itself to high speed data. No, it equally valid
>> at any data rate.
>
>> A couple of lines later it refers to folding the data signal to
>> produce the diagram. This is never done as it would result in the time
>> line reversing direction at each fold. No, the data is cut and shifted
>> by a fixed number of cycles of the recovered clock.
>
>> So no, from the start this is poor technical writing. It is typical of
>> the lazy, sloppy writing that results when somebody with a poor but
>> inflated grasp of English translates a badly written technical note
>> into turgid prose.
>
>I appreciate your taking the time to criticize, but you as you have repeatedly
>shown you don't like me, I'm not surprised at what you wrote. (Only your point
>about folding seems to be a criticism worth making. "Overlapping" would have
>been a better choice.)
>
>I'm editing a magazine article, and discovered I didn't properly understand
>eye diagrams. I found this, and was startled that it was a coherent,
>comprehensible piece. Your claims about its poor quality are beyond absurd.
>You're making them simply to figuratively kick me in the nuts.
>
>PS: The article I'm editing is one of those pieces where the author starts by
>trying to make all his points in the first paragraph, then gives the details
>as they come to mind (without any apparent organization development). And did
>I mention that he repeats the same things over and over? I had to figure out
>what he really wanted to say, then reorganize and (largely) rewrite the piece.
>I challenge you to do as well.
>

I've never met you, so liking or not liking doesn't come into it. My
points are serious. This is by an author who only barely grasps the
subject, and it shows. Better say nothing than teach nonsense - I
presume the idea is that the reader will learn from this.

What is really missed here, though, is that the eye diagram is no more
than a visual sanity check - quick and dirty. The actual figure of
merit comes from the MER measurement, in which the RMS amplitude of
the wanted vectors is compared to that of the noise vectors. The
answer, expressed in dB, is what you actually use in system
assessment.
>
>> And finally, of course, I have to comment on your introduction -- very
>> Street, almost Ghetto to omit the verb (I believe we discussing). And
>> you made it a example instead of an example. Are you a Hip hop artist
>> in your spare time, by any chance?
>
>Both errors are typos. (You made the same "dropped verb" error in your third
>paragraph, and omitted a space in the second.)

Yeah, that last bit was just for fun. Usenet is rife with typos. I've
probably dropped some more into this.

d

[Scott Dorsey]

Don Pearce > wrote:
>
>What is really missed here, though, is that the eye diagram is no more
>than a visual sanity check - quick and dirty. The actual figure of
>merit comes from the MER measurement, in which the RMS amplitude of
>the wanted vectors is compared to that of the noise vectors. The
>answer, expressed in dB, is what you actually use in system
>assessment.

It's kind of more than that. The eye pattern is entirely a qualitative
measurement and can't be used to compare different equipment in any
meaningful way. It's important to say that.

BUT, the useful thing about the eye pattern is that it provides an
instantaneous visual indication as to the current quality of the
signal. Therefore it is very useful for "tweak and peak" alignment
work where a technician is adjusting a control or a mechanical setting
and watching the pattern on a scope.

But it _is_ important to point out that it is a qualitative measurement
only. Not to demean qualitative measurements in any way; they are often
most useful.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Don Pearce[_3_]]

On 13 Jul 2013 10:03:57 -0400, (Scott Dorsey) wrote:

>Don Pearce > wrote:
>>
>>What is really missed here, though, is that the eye diagram is no more
>>than a visual sanity check - quick and dirty. The actual figure of
>>merit comes from the MER measurement, in which the RMS amplitude of
>>the wanted vectors is compared to that of the noise vectors. The
>>answer, expressed in dB, is what you actually use in system
>>assessment.
>
>It's kind of more than that. The eye pattern is entirely a qualitative
>measurement and can't be used to compare different equipment in any
>meaningful way. It's important to say that.
>
>BUT, the useful thing about the eye pattern is that it provides an
>instantaneous visual indication as to the current quality of the
>signal. Therefore it is very useful for "tweak and peak" alignment
>work where a technician is adjusting a control or a mechanical setting
>and watching the pattern on a scope.
>
>But it _is_ important to point out that it is a qualitative measurement
>only. Not to demean qualitative measurements in any way; they are often
>most useful.
>--scott

Peak and tweak is an important addition. Otherwise, pretty much
agreed.

d

[William Sommerwerck]

"Don Pearce" wrote in message
...

> What is really missed here, though, is that the eye diagram is no
> more than a visual sanity check - quick and dirty. The actual figure
> of merit comes from the MER measurement, in which the RMS
> amplitude of the wanted vectors is compared to that of the noise
> vectors. The answer, expressed in dB, is what you actually use in
> system assessment.

I will backtrack -- slightly -- and agree with you to the extent that an eye
pattern gives a rough -- but not necessarily definitive -- assessment of
signal quality.

I was looking for something that explained what an eye pattern actually was --
and this piece did it very well. Although the Wikipedia article -- which I
read first -- pointed out the importance of the error rate -- it did not
supply me with a sense that I understood much about eye patterns.

[Mike Rivers[_2_]]

On 7/13/2013 9:33 AM, William Sommerwerck wrote:

> I'm editing a magazine article, and discovered I didn't properly
> understand eye diagrams. I found this, and was startled that it was a
> coherent, comprehensible piece.

I was looking for something practical in there. I've always wanted to
play with eye patterns, but never knew where to hook up the scope
(assuming that I had some product documentation). This gives a clue, but
an example with real hardware would have been welcome. I realize it'
wouldn't be universal, but would be something to start from.


--
For a good time, call http://mikeriversaudio.wordpress.com

[Scott Dorsey]

In article >, Mike Rivers > wrote:
>
>I was looking for something practical in there. I've always wanted to
>play with eye patterns, but never knew where to hook up the scope
>(assuming that I had some product documentation). This gives a clue, but
>an example with real hardware would have been welcome. I realize it'
>wouldn't be universal, but would be something to start from.

Mike, get a Philips CD player and get the service manuals. They have
detailed directions about what the signals are at each of the test points,
including the eye pattern, and what you should see on the scope. Just doing
a CD player line up is a good introduction to the whole digital audio
technology.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[William Sommerwerck]

"Mike Rivers" wrote in message ...
On 7/13/2013 9:33 AM, William Sommerwerck wrote:

>> I'm editing a magazine article, and discovered I didn't properly
>> understand eye diagrams. I found this, and was startled that it
>> was a coherent, comprehensible piece.

> I was looking for something practical in there. I've always wanted to
> play with eye patterns, but never knew where to hook up the scope
> (assuming that I had some product documentation). This gives a clue,
> but an example with real hardware would have been welcome. I realize
> it wouldn't be universal, but would be something to start from.

I don't think the article is intended to be "practical" in the sense you mean.

The issue -- as I understand it -- is not where you connect the 'scope (which
is more or less obvious from the description) but how you "exercise" the
transmission path to produce the wide range of data that's needed for a useful
eye display. As the article indicates, a special pulse generator is required.

[Don Pearce[_3_]]

On Sat, 13 Jul 2013 12:10:59 -0700, "William Sommerwerck"
> wrote:

>"Mike Rivers" wrote in message ...
>On 7/13/2013 9:33 AM, William Sommerwerck wrote:
>
>>> I'm editing a magazine article, and discovered I didn't properly
>>> understand eye diagrams. I found this, and was startled that it
>>> was a coherent, comprehensible piece.
>
>> I was looking for something practical in there. I've always wanted to
>> play with eye patterns, but never knew where to hook up the scope
>> (assuming that I had some product documentation). This gives a clue,
>> but an example with real hardware would have been welcome. I realize
>> it wouldn't be universal, but would be something to start from.
>
>I don't think the article is intended to be "practical" in the sense you mean.
>
>The issue -- as I understand it -- is not where you connect the 'scope (which
>is more or less obvious from the description) but how you "exercise" the
>transmission path to produce the wide range of data that's needed for a useful
>eye display. As the article indicates, a special pulse generator is required.

Not really necessary, pretty much any data stream will do nicely. Data
is usually pre-conditioned with an XORed code to prevent long
sequences of contiguous 1s or 0s. As long as there is a reasonable
mix, the eye diagram will build itself. A dozen or so passes will
usually be stored on top of each other to generate the diagram.

Specific bit sequences are used for bit error rate measurements, and
it is probably convenient to examine the eye while using these.

d

[William Sommerwerck]

"Don Pearce" wrote in message
...
On Sat, 13 Jul 2013 12:10:59 -0700, "William Sommerwerck"
> wrote:

>> The issue -- as I understand it -- is not where you connect the 'scope
>> (which is more or less obvious from the description) but how you
>> "exercise" the transmission path to produce the wide range of data
>> that's needed for a useful eye display.

> Not really necessary, pretty much any data stream will do nicely.
> Data is usually pre-conditioned with an XORed code to prevent
> long sequences of contiguous 1s or 0s.

This is certainly true in the case of NRZ coding, to prevent the creation of
substantial DC offsets.

[Mike Rivers[_2_]]

On 7/13/2013 2:17 PM, Scott Dorsey wrote:

> Mike, get a Philips CD player and get the service manuals. They have
> detailed directions about what the signals are at each of the test points,
> including the eye pattern, and what you should see on the scope.

Never thought about that. I have an old Magnavox CD player. I guess
that's Phillips. I'll have to see if I can find a service manual for it.


--
For a good time, call http://mikeriversaudio.wordpress.com

[Scott Dorsey]

In article >, Mike Rivers > wrote:
>On 7/13/2013 2:17 PM, Scott Dorsey wrote:
>
>> Mike, get a Philips CD player and get the service manuals. They have
>> detailed directions about what the signals are at each of the test points,
>> including the eye pattern, and what you should see on the scope.
>
>Never thought about that. I have an old Magnavox CD player. I guess
>that's Phillips. I'll have to see if I can find a service manual for it.

Should be. If it's a CDB460, the manual is great.

The absolute best are the original 14-bit machines because you can actually
follow stage by stage with the scope and watch oversampling happening. It
totally blew my mind in 1983 or so.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Mike Rivers[_2_]]

On 7/13/2013 8:02 PM, Scott Dorsey wrote:

> If it's a CDB460, the manual is great.

That's what I have. I haven't turned up a source for a manual yet,
though. Neither the Magnavox web site nor the on-line sources for
manuals that I have bookmarked have it listed, nor does a Google search
turn up anything useful.

--
For a good time, call http://mikeriversaudio.wordpress.com