RGB to sync on green

[Malcolm Reeves]

Hi,

I'm looking at various new TVs etc. and it seems like the TV will take
component/RGB and the DVR etc can output RGB (component on some but
not all). However, scart rgb uses the composite for the sync and the
TV expects sync on green. So I can either buy a hugely expensive
converter or knock my own up.

So I think all I need is a LM1881 to get the sync from the composite.
Since the LM1881 is high input impedance and I'm not using the
composite then a lash up should not feed back significant noise to the
RGB lines. The LM1881 sync output is rail so it could be 5 or 12V. I
only need 0.3V so a 1K2 or 3K for 12V from sync output to G should
give me the right sync. The high impedance will attenuate any noise
on signal as well as not disturbing the impedance.

Sounds too simple - am I missing something?


--

Malcolm

Malcolm Reeves BSc CEng MIEE MIRSE, Full Circuit Ltd, Chippenham, UK
([email protected], [email protected] or [email protected]).
Design Service for Analogue/Digital H/W & S/W Railway Signalling and Power
electronics. More details plus freeware, Win95/98 DUN and Pspice tips, see:

http://www.fullcircuit.com or http://www.fullcircuit.co.uk

NEW - www.CharteredConsultant.co.uk - The Consultant A-List

 

[martin griffith]

Hi,

I'm looking at various new TVs etc. and it seems like the TV will take
component/RGB and the DVR etc can output RGB (component on some but
not all). However, scart rgb uses the composite for the sync and the
TV expects sync on green. So I can either buy a hugely expensive
converter or knock my own up.

So I think all I need is a LM1881 to get the sync from the composite.
Since the LM1881 is high input impedance and I'm not using the
composite then a lash up should not feed back significant noise to the
RGB lines. The LM1881 sync output is rail so it could be 5 or 12V. I
only need 0.3V so a 1K2 or 3K for 12V from sync output to G should
give me the right sync. The high impedance will attenuate any noise
on signal as well as not disturbing the impedance.

Sounds too simple - am I missing something?

Havent done this for years, but FWIW
it should work, but video impedances are 75R, and the 1881 can only
manage a couple of mA, so buffer with a inverter (74UC04 I think)and
you might be able to get away with capacitively coupling to the G sig,
there are too many funnies to be predictable. You could generate a -ve
sync with a transisto current source with the correct current into
75R.You may need to put a subcarrier LC trap to get rid of the burst,
since the LPF in the data sheet sucks and you will probably get a
picture shift, since the H phase will have a time difference between
the composite



martin

Serious error.
All shortcuts have disappeared.
Screen. Mind. Both are blank.

 

[Malcolm Reeves]

Havent done this for years, but FWIW
it should work, but video impedances are 75R, and the 1881 can only
manage a couple of mA, so buffer with a inverter (74UC04 I think)and
you might be able to get away with capacitively coupling to the G sig,
there are too many funnies to be predictable. You could generate a -ve
sync with a transisto current source with the correct current into
75R.You may need to put a subcarrier LC trap to get rid of the burst,
since the LPF in the data sheet sucks and you will probably get a
picture shift, since the H phase will have a time difference between
the composite

Good points. I need 4mA so too much for the LM1881. The 74HCU04
sounds good. More than enough there. BTW did you mean the unbuffered
or just 74HC? I would have thought the HC would be better as I want
squarewave not digital.

In terms of space transistors to do constant current is probably more
than an 04 and resistor. Errors in the 75ohm will give similar errors
in the sync level so I don't see there is any advantage to the
constant current method.

The info on the poor LPF in the LM1881 is good to know made me look at
other chips. The EL4581 styles itself as an improved LM1881 and it
has an inbuilt 3 pole filter. That sounds a better option. A bit
more expensive but not that much. Similar output drive to the LM1881
so I still need the 04.

Why would I get a phase shift? I think sync on green just needs the
composite sync so I was going to use that. I don't care about the
error in the H output of the EL4581 (LM1881).


--

Malcolm

Malcolm Reeves BSc CEng MIEE MIRSE, Full Circuit Ltd, Chippenham, UK
([email protected], [email protected] or [email protected]).
Design Service for Analogue/Digital H/W & S/W Railway Signalling and Power
electronics. More details plus freeware, Win95/98 DUN and Pspice tips, see:

http://www.fullcircuit.com or http://www.fullcircuit.co.uk

NEW - www.CharteredConsultant.co.uk - The Consultant A-List

 

[martin griffith]

Good points. I need 4mA so too much for the LM1881. The 74HCU04
sounds good. More than enough there. BTW did you mean the unbuffered
or just 74HC? I would have thought the HC would be better as I want
squarewave not digital.

In terms of space transistors to do constant current is probably more
than an 04 and resistor. Errors in the 75ohm will give similar errors
in the sync level so I don't see there is any advantage to the
constant current method.

The info on the poor LPF in the LM1881 is good to know made me look at
other chips. The EL4581 styles itself as an improved LM1881 and it
has an inbuilt 3 pole filter. That sounds a better option. A bit
more expensive but not that much. Similar output drive to the LM1881
so I still need the 04.

Why would I get a phase shift? I think sync on green just needs the
composite sync so I was going to use that. I don't care about the
error in the H output of the EL4581 (LM1881).

Have a look at the LC cct on
http://www.techmind.org/vd/vidmk2.html
half way down the site, best way of minimising delays

With the const source you dont load the 75R, technically cleaner IMHO,
as for HC04 or U04, use a socket, see if there is any difference!

Phase shift, there is more processing in the signal source (DVD) after
the RGB out to derive the composite, hence, generally some H shift.
I always find it annoying, but then i'm used to looking for this sort
of thing



martin

Serious error.
All shortcuts have disappeared.
Screen. Mind. Both are blank.

 

[Malcolm Reeves]

Have a look at the LC cct on
http://www.techmind.org/vd/vidmk2.html
half way down the site, best way of minimising delays

I think I see what you mean but on the same circuit there is the
EL4581 which composite straight in and sync straight out which is what
I had progressed after find out the LM1881 was not so hot. The EL4581
has a filter inside, and at 3 poles more then the 2 pole LC.

With the const source you dont load the 75R, technically cleaner IMHO,
as for HC04 or U04, use a socket, see if there is any difference!

I was going to argue there wasn't much difference but thinking about
it although the source impedance is 1200 that's still a 3% error which
means my green signal would be 3% less than the R and B. That doesn't
sound very clever! So I'm with you that the constant current source
sounds a better bet.

Phase shift, there is more processing in the signal source (DVD) after
the RGB out to derive the composite, hence, generally some H shift.
I always find it annoying, but then i'm used to looking for this sort
of thing

OK, I follow, but not much can be done about it as CVBS is the only
sync available :-(


--

Malcolm

Malcolm Reeves BSc CEng MIEE MIRSE, Full Circuit Ltd, Chippenham, UK
([email protected], [email protected] or [email protected]).
Design Service for Analogue/Digital H/W & S/W Railway Signalling and Power
electronics. More details plus freeware, Win95/98 DUN and Pspice tips, see:

http://www.fullcircuit.com or http://www.fullcircuit.co.uk

NEW - www.CharteredConsultant.co.uk - The Consultant A-List

 

[martin griffith]

I think I see what you mean but on the same circuit there is the
EL4581 which composite straight in and sync straight out which is what
I had progressed after find out the LM1881 was not so hot. The EL4581
has a filter inside, and at 3 poles more then the 2 pole LC.


I was going to argue there wasn't much difference but thinking about
it although the source impedance is 1200 that's still a 3% error which
means my green signal would be 3% less than the R and B. That doesn't
sound very clever! So I'm with you that the constant current source
sounds a better bet.

The problem is not really a "gain" problem, its DC offsets, any DC
offset will cause an unpleasant colour cast, so blacks will be
slightly green, this is why YUV is prefered for component signal
distibution. If there is a slight DC offset on the Y, blacks are
slightly grey, and the highlights will be slightly hotter or colder
which is far less objectionable.
On the gain, side gain errors are again far less objectionable in YUV
than on RGB, which will cause peak whites to have a colour cast.

This all depends on how and where the TV actually does a black level
clamp internally. Fortunately I dont "do" TV sets, so its a S.E.P (
see Douglas Adams)




martin

Serious error.
All shortcuts have disappeared.
Screen. Mind. Both are blank.