17 May
2010
17 May
'10
7:50 p.m.
I have the need to optically isolate a composite video signal to compensate for the
different ground reference between the devices. Digging through my stash of parts, I have
some optos, but the ones I have are more suited to digital applications. Having not tried
to pass an analog signal through one before, I'm not quite sure what sort of part
I'm looking for...
Anyone have any suggestions?
-Ian
17 May
17 May
8:13 p.m.
On 17 May 2010 at 19:50, Mr Ian Primus wrote:
I have the need to optically isolate a composite video
signal to
compensate for the different ground reference between the devices.
Digging through my stash of parts, I have some optos, but the ones I
have are more suited to digital applications. Having not tried to pass
an analog signal through one before, I'm not quite sure what sort of
part I'm looking for...
Most garden-variety optos are hideously non-linear. You can use a
matched pair and use one to provide feedback to compensate for the
nonlinearity of the other, but you may not end up with enough
bandwidth.
I'm assuming that ADC->opto->DAC is out of the question.
Why not use a made-for-purpose part such as the HP/Agilent HCPL4562?
--Chuck
8:29 p.m.
On Mon, May 17, 2010 at 23:13, Chuck Guzis <cclist at sydex.com> wrote:
On 17 May 2010 at 19:50, Mr Ian Primus wrote:
But given the nonlinearity would only affect the luminance component,
it might not be very noticeable. I would just try it. :-)
I have the need to optically isolate a composite
video signal to
compensate for the different ground reference between the devices.
Most garden-variety optos are hideously non-linear. I'm assuming that ADC->opto->DAC is out of
the question.
Regular optoisolators would probably not be fast enough, but a simple
one-bit sigma-delta A/D-D/A chain is not too hard to do...(at least in
theory)
Joe.
--
Joachim Thiemann :: http://www.tsp.ece.mcgill.ca/~jthiem
8:31 p.m.
Date: Mon, 17 May 2010 19:50:59 -0700
From: ian_primus at yahoo.com
Subject: Semi-OT - Using optoisolators with composite video signals
To: cctech at classiccmp.org
I have the need to optically isolate a composite video signal to compensate for the
different ground reference between the devices. Digging through my stash of parts, I have
some optos, but the ones I have are more suited to digital applications. Having not tried
to pass an analog signal through one before, I'm not quite sure what sort of part
I'm looking for...
Anyone have any suggestions?
-Ian
Hi Ian,
Lets get a bit deeper into your requirement:
Is it really composite NTSC, picture data or a composite digital signal (computer game
etc.)
I ask because perhaps just isolating digital RGB from the source by opto may be the way to
go, and regenerate the composite NTSC on the other side of the opto barrier with the
National LM1886 / LM1889 combination.
If you really have to isolate a NTSC signal, here are your options:
RF Modulator, demod with some fidelity loss
Broadband transformer with possibly a gain stage on the output side (think a 75 ohm 1:1
transformer) depending on how forgiving the monitor is on signal levels, and front porch
back porch, sink sag during the essentially DC time of the signal. A quick test of this
idea would be to wire 2 coax to 300 ohm flatlead baluns back to back (I know they are
expecting RF, not DC-4.5MHz but a cheap and quick experiment just the same)
And, finally, as you suggested, a linear optocoupler. They are inherently non-linear, but
avoid all the DC problems at the sync pulses. If you go this route, the topology will use
2 optocouplers, one one carying the signal, and one in a feedback path to linearize it.
This was a popular circuit, and I bet we can find it in a single IC package by now...
Let me know some more specifics of your two devices you are mating, and we can come up
with something.
It may be just simpler to add an isolation transformer to the monitor, or if its a
'hot' chasis just be sure the supplies of both systems have no difference in hot
and neutral. This was the original IBM PC solution, remember the special polarized plug
for the monitor, that plugged into the CPU?
Cheers,
Randy
_________________________________________________________________
The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail.
http://www.windowslive.com/campaign/thenewbusy?tile=multiaccount&ocid=P…
9:20 p.m.
Date: Tue, 18 May 2010 00:11:10 -0400
Subject: Re: Semi-OT - Using optoisolators with composite video signals
From: wdonzelli at gmail.com
To: cctalk at classiccmp.org
Aww heck Will. Why spend a buck when you can spend so much more, and countless hours
making some obsolete technology perform?
Thats the credo of classicmp.
;)
Randy
_________________________________________________________________
The New Busy is not the too busy. Combine all your e-mail accounts with Hotmail.
http://www.windowslive.com/campaign/thenewbusy?tile=multiaccount&ocid=P…
If you really have to isolate a NTSC signal, here
are your options:
Or just get an isolating video DA at a hamfest for a buck. At least on
the East coast, broadcast quality video stuff is very common.
--
Will 
18 May
18 May
12:36 p.m.
> > If you really have to isolate a NTSC
signal=2C here are your options:
>=20
> Or just get an isolating video DA at a hamfest for a buck. At least on
> the East coast=2C broadcast quality video stuff is very common.
But not, AFAIK, in the UK...
Aww heck Will. Why spend a buck when you can spend so
much more=2C and cou=
ntless hours making some obsolete technology perform?
I can think of one very good reason. It's called 'Education'. If you jsut
use a black-box solution, you don;'t understnad how it works, you don't
know how to debug it if it dorsn't work. If you design the thing, then
you (hopefuilly) understand what you were trying to do, so when it
doesn't work, you know what to test, you can then redesign it to work
properly.
I get verry worried by jsut throwing black boxes at a problem. Somebody
had to design said black box in the first place. And if you don't
understnad how it works, who is going to design the next generation of
modules?
Put it this way : A so-called electronic design (in any area of
electronics) who can't work with discrete tranisitors is going to get no
respsect from me.
-tony
5:21 a.m.
--- On Mon, 5/17/10, Randy Dawson <rdawson16 at hotmail.com> wrote:
It may be just simpler to add an isolation transformer to
the monitor, or if its a 'hot' chasis just be sure the
supplies of both systems have no difference in hot and
neutral.?
This is essentially the situation. I've modified a black and white television set to
accept composite video. It was surprisingly simple to do, and the picture is acceptable,
considering the quality of the set. I'm currently running the TV on an isolation
transformer. It's a hot chassis set with a polarized cord. Even though the neutral is
at the same reference as the computer input, you'd run into problems on circuits with
a GFI outlet.
The isolation transformer is ideal in terms of safety, but not in terms of size or cost.
I'm using my bench iso for testing, but I want something I can mount inside the
television, so I won't need extra equipment to use it. Arcade machine isolation
transformers are too big - I'm looking for a smaller/cheaper one, since the power
requirements are minimal, around 30 watts.
Optically isolating the video signal is another option, eliminating the need for the
isolation transformer. The advantage here is low cost and small parts. I was unaware of
the single-chip video isolation solutions, hence why I asked :)
-Ian
12:31 p.m.
It may be just simpler to add an isolation transformer
to the monitor=2C or=
if its a 'hot' chasis just be sure the supplies of both systems have no di=
I wold not recomend that. As you know, I have nasty things to say about
health-n-safety, but I will not have any exposed metalwork directly
connected to either side of the mains supply. Period.
fference in hot and neutral. This was the original
IBM PC solution=2C reme=
mber the special polarized plug for the monitor=2C that plugged into the CP=
U?
No. The IBM MDA monitor had a mains isolating transoformer built-in. Im
fact the mains was stepped down to 12V to run the monitor electronics.
The CGA nonitor (which plugged into the wall, not into the PC's PSU) had
an isolating SMPSU.
The power output socket on an IBM PC PSU is just switched mains. It is
not isolated in any way.
I can think of 2 hot chasiss monitors used with computer-related
equipment. Both were converted TVs. One was a Philips used witha Philips
Viewdata set. The montor was a modified Philips colour TV which was
powered via a special socket on the Viewdata set. This connected to a
secondary winding on the Viewdata set's mains transformer (giving about
100V IIRC), this provided the isolation.
The other, of couse was the US version of the origianl TRS-80 Model 1
monitor. This was a converted RCA portable TV. The modification involved
replacing the IF amplifier PCB with an opto-isolated board. The monitor
chassis remained 'hot'. This monitor needed a 5V suppy from the computer
to power the LED side of the opto-isolator.
FOr the record, the European version of this monitor was not hot-chassis.
It had the same chassis run from a 250V - 115V _isolating_ transformer
inside the case. The IF PCB was replaced by a simple video amplifier.
-tony
5333
days inactive
5333
days old
8 comments
6 participants
participants (6)
-
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
ian_primus＠yahoo.com -
joachim.thiemann＠gmail.com -
rdawson16＠hotmail.com -
wdonzelli＠gmail.com