{ "numMessagesInTopic": 6, "nextInTime": 1283, "senderId": "JtwyRd4BVPCt0NmOyNohhZBbFPJgQMxUq56CuHRREl7qvX2gRf-7uqR5V5WT7KVFijUcPHbCAIAYutXBoJMd0srjpfs-wI_xU7QIV8gF8w", "systemMessage": false, "subject": "Re: technical monitor question", "from": ""fluppeteer" <yahoo@...>", "authorName": "fluppeteer", "msgSnippet": "Thank you again for the reply. Sorry to drag this out - I m just trying to make sure my understanding is correct. ... *Nod*. I d established the duration of", "msgId": 1282, "profile": "fluppeteer", "topicId": 1277, "spamInfo": { "reason": "12", "isSpam": false }, "replyTo": "LIST", "userId": 192443393, "messageBody": "
--- In IBM_T2X_LCD@yahoogroups.com, "af2412ripe" <af2412ripe@...> wrote:
\n>
\n> > From what point does vertical blanking begin
\n> It is a multiple of H lines. The vertical sync period is simply a
\n> fixed number of H lines.
\n
\n*Nod*. I'd established the duration of the vsync signal,
\nit was only its start point which was troubling me.
\n
\n> >When vsync fires, the beam "resets" from the bottom
\n> >of the screen to the top.
\n>
\n> Not exactly... It takes an amount of time equal to several
\n> H-lines for V-sync to retrace back to the top.
\n
\nSorry, my poor communication skills again. By "when vsync fires",
\nI mean "when vsync fires the beam *starts* to reset from the
\nbottom of the screen, and this transition finishes when vsync
\nis no longer active". Since I know the duration is a whole number
\nof lines and standard monitor EDIDs, GTF, etc. can say how many
\nlines vsync lasts for, my concern is only with the exact timing
\nof the start of the vsync signal.
\n
\n> Retracing to the top in less than one H-line requires a
\n> HUGE frequency range for the oscillator. (There are separate
\n> electromagnetic coils for H and V deflection of the electron beam).
\n> While it may be possible today, it was not in 194n.
\n
\n*Nod*. I was considering a vsync of one line duration solely
\nfor simplification; I accept that the actual duration is usually
\nlonger (and that an accurate ASCII-art diagram would therefore
\nbe unintelligible). Since we know vsync is a whole number of
\nlines in duration the actual number doesn't matter from the
\npoint of view of this particular question. Sorry that I wasn't
\nclearer that I was trying to simplify things.
\n
\n> In an interlaced scan system with an odd number of total H-lines,
\n> the electron beam moves upwards from the lower right before the
\n> first field, and upwards from the center before the second field.
\n> It ends in the upper left, or upper center, respectively.
\n>
\n> Progressive CRT scan is from lower right to upper left.
\n
\n> The V-sync signal begins to
\n> go low at the start, rather than the end, of the H-sync retrace
\n> interval (except between the first and second fields of interlaced
\n> video).
\n
\n[From the follow-up message]
\n
\n> On further thought, the vertical retrace of the physical beam on an
\n> old television would be from the lower left to the upper left. This
\n> would be because the V-deflection circuit does not know that V-sync
\n> has happened until after the completion of an H-sync period, during
\n> which time the H-retrace has been completed. This is basic state-
\n> machine logic. It has been so long since I have worked on CRT
\n> circuitry, I had to think about it more.
\n
\n[End follow-up message]
\n
\nI think this is what I was looking for. May I just restate this
\nto ensure I understand?
\n
\nThe vsync signal is considered to (start to) transition to
\nactive at the point that hsync becomes active. The receiver
\nwill not "notice" until hsync deactivates again, in a CRT
\ntelevision (as the second reply stated)? So the path of the
\nelectron beam during vsync is from bottom left to top left
\nbecause it's effectively delayed by an hsync duration,
\nbut if the vertical movement coincided exactly with the input
\nvsync signal (and there was no internal state machine in the
\ndisplay) then the beam would travel from bottom right to top
\nright?
\n
\nIs there a reference anywhere which dictates this behaviour?
\n(Not that I don't trust you, just because it would be nice
\nto know whence the standards come.) Or should I just be
\nlooking at a composite television signal, and assuming that
\nthe same applies to all CRTs?
\n
\nI may have confused matters by talking about the electron
\nbeam when I'm only actually interested in the signal. It
\nsounds as though the correct timing for vsync to become
\nactive in the signal is the duration of the horizontal front
\nporch after the active display area, plus an integer multiple
\nof the horizontal line frequency according to the vertical
\nfront porch. An "old television" would notice a period
\nlater corresponding to the duration of hsync.
\n
\nI appreciate that there's some analogue ramping going on
\nin an analogue television. However, since (in DVI at least)
\nhsync and vsync are digital signals, a change in timing
\nbecomes detectable.
\n
\n> Ironically, it does not matter where the V-retrace
\n> physically begins, since the V-deflection circuit is a
\n> low frequency ramp generator, and it simply cannot maintain
\n> a high precision tolerance to begin the V- retrace in exactly
\n> the same place every time. This is why the "vertical interval"
\n> includes several black H-lines. This is a difference between
\n> analog and digital thinking.
\n
\n:-) Unfortunately, the signals which are meaningful in the
\nanalogue world need to be interpreted in the digital one. It
\nsounds as though the reason I can't find a specification for
\nthis is that it's not supposed to matter. Which probably means
\nthat if I'm generating a signal I should do as mentioned above
\n(vsync activates as hsync activates), but that if I'm receiving
\na signal I need to be tolerant of other timings.
\n
\nThanks very much, hopefully this means I'm now un-confused!
\n
\n--
\nFluppeteer