Questions on convergence...

Tony Duell ard at
Mon Aug 14 16:35:42 CDT 2006

> On Sunday 13 August 2006 06:05 pm, Tony Duell wrote:
> > > I've done that sort of thing,  on TVs,  and what you describe there ain't
> > > old. TVs that were still using all vacuum tubes when I worked on them
> > > back in the days I had my shop -- now they were OLD!  Convergence used to
> > > be a real PITA, because you had static convergence that was done by
> > > moving magnets on the neck of the CRT (not to be confused with the purity
> > > magnets!) and then you had dynamic convergence,  which was typically
> > > 12-16 separate adjustments, all of which interacted to some extent.
> >
> > Actually, having grown up with delta-gun CRTs, I find them much easier to
> > set up than the in-line type. Tweaking presets is a lot easier than
> > tilting the yoke and hoping....
> As in it's handled in software?  I have not done any convergence adjustments 

No, not at all. AFAIK all monitors with microprocessor control use PIL 
CRTs. I jsut find it easier to turn presets than to shift the yoke.

Now, as to what killed the delta-gun CRT... Well, Think of the effects of
external magnetic fields on the beams, and remember that the electron beam
moves perpendicular to the field. Resolve the earth's magnetic field into 
vertical and horizontal components. The vertical component of said field 
causes the beams to move horizontally, the horizontal component moves 
them vertically.

BOTH components of the field cause convergence errors on a delta-gun CRT. 
But on a PIL tube, the horizontal component of the earth's magnetic field 
just shifts the picture vertically (slightly), and as the phospors are 
put down in vertical strips, it doesn't cause a convergence error. 

Now remember that the vertical component is essentially constant in a 
given area, the effetive horizontal component changes if the CRT is 
turned round, etc.

So a delta-gun tube has to be converged in the place it's going to be 
used. A portable unit is almost impossible. A PIL tube has to be 
converged at apporximately the right latitude, but apart from that, it 
doesn't matter if you move the unit. 

> since those days,  the stuff I'm dealing with any more hardly seems to need 
> it touched.  Perhaps the only thing off in the original poster's monitor is 

You're lucky. I've never seen a modern-ish TV (that is, a TV recent 
enough to be a PIL CRT, but not a plasma or LCD panel) that was 
adequately converged. 

> those static adjustments,  if it suffered some kind of a physical shock,  
> which I would guess is at least worth looking at,  seeing if something is 

If it's suffered a shock, it's possible the 'aperture grille' (the 
equivalent of the shadowmask) has come loose inside the CRT. This would 
cause purity errors, and probably convergence errors too. In fact do we 
know the OP's monitor has good purity (display a screen of each of the 
primary colours -- red, green, blue -- does it look even, with no odd 
coloured patches?)

> obviously out of place.  I didn't see any response to my point about whether 
> this happened all of a sudden or gradually over time,  though.

> > > The other thing is,  to do a proper convergence needs a signal that's
> > > going to give you a stable pattern,  although I suspect that's easier to
> > > do with a computer driving it than not,  I had to buy a little generator
> > > for TVs back
> >
> > Indeed. With a computer, it's trivial to generate a cross-hatch pattern
> > (the main one needed for convergence).
> I always used to start with dots,  myself,  for the center convergence.

I tend to switch between dots and cross-hatch.

> > For TVs, you either buy/build a cross-hatch generator (I remember
> > building one from a kit about 15 years ago, it was only a handful of ICs,
> > one of which was a TV sync generator chain). Or you use a home computer,
> > suitably programmed. There were programs for the BBC micro, C64, etc,
> > published in the magazines.
> The one I bought back in 1974 (!) does include the dot, crosshatch, and color 
> bar patterns as well as a blank raster,  for purity adjustments,  but only 
> fixed sizes of them,  not fewer or single ones,  which I saw featured on some 

Mine will generate a grid of vertical lines, and a similar grid of 
horizontal lines. The logical AND is dots, the logical OR is a 
crosshatch. All 4 of those patterns are switch-selectable, along with an 
8-bar greyscale, the obvious 8 bar colour bar pattern, and a plain 
raster. IIRC, you can enable/disable each of the primary colours for all 
patterns, so the plain raster can be used to give the 3 primary colour 
rasters for purity adjustments.

> instruments.  It uses unijunction transistors as frequency dividers,  and has

Mine at least uses ICs (mostly TTL from what I remember, along with a PAL 
encoder chip and a sync generator).

> > monitor) on a workstation, as the CPU. Therefore, I'd want to repair the
> > monitor if at all possible (even if, say, I have to re-wind a transformer
> > to do so,,)
> Some units are no doubt easier to work on than others.

Yes, like the semi-professional TV-rate Barco I have here. The first nice 
feature (if your workshop is as full as mine) is that it's not 'monitor 
shaped'. It's a cuboidal metal case, so you can stack things (at least 
manuals, etc) on top of it. Normal monitors are a pain to store.

This Barco is cery well made. It's all on plug-in PCBs inside, with an 
externder board neatly stored in a spare slot so you can work on said 
PCBs. Oh yes, the manual is worthy of the name, with full schematics, 
waveforms, testpoints, etc. 

That's a monitoe I'd put a lot of time into to keep it running. I'd spend 
rather less on a generic PC monitor.


More information about the cctalk mailing list