Physically, why do weak color CRTs "bleed"?
Michael Shell
I have known for sometime that when a color CRT gets weak, colors
tend to bleed to the right, especially when the contrast, saturation
and brightness levels are turned up.
This is mentioned in the CRT FAQ:
http://www.bithose.com/serfaq/REPAIR/F_crtfaq.html#CRTFAQ_047
However, *why* does it do this? I can understand that a low cathode
emission level would result in a dimmer picture. If the lower beam
strength caused things to defocus, then wouldn't that result in bleeding
to the left *and* right (color blurr)?
On a related note, are there any brands (CRT or TV or monitors) that are
resistant to this type of degradation? Are Trinitrons more likely to
do this than those CRTs using Invar shadow masks?
Thanks in advance,
Mike
Andy Cuffe
I think it's due to the electron gun clipping when it's overdriven.
Even a new CRT will bleed if it's driven hard enough, but most TVs are
designed so you can't turn up the contrast that much. Once the CRT goes
into clipping, it must take a short time to start working normally again
after the drive level falls belop clipping. The same thing happens when
cirtain problems develop in the video amp.
All CRTs do it when they get weak enough. Samsung seem to be worse than
most. In general, all CRT manufactuers have been cutting costs. A
larger percentage or 8 year old or newer TVs that I see have bad CRTs
than ones that are more than about 16 years old. I just picked up a
heavily used 1982 Zenith from the side of the road and it has a better
looking CRT than most new TVs.
--
Andy Cuffe
balt...@psu.edu
JURB6006
First of all, realize that most sets do this with a lowered G2 voltage. While
on the surface, lowering the G2 seems to mimic a weak CRT, this is not the
case. Only in some ways it does, apparently.
Some sets have an unusual resistance, and others seem to have an unusual
propensity to "bleed" (we call it flaring at my shop)
I was around before ICs, and have had an opportunity to study the design of the
video output circuit(s) of TVs without ICs. I do understand circuit design, and
have reached the following conclusion:
(this completely excludes sets with AKB)
It depends on how the video output transistors are driven. It seems that if
there is a path for current feedback, the set will flare. This is the worst in
sets that drive the emitters of the outputs.
Other sets, which could be designated as "voltage drive sets" either have such
solid drive to the emitters that CRT load doesn't affect it or they just drive
the base(s) of the outputs.
Engineering wise, I think it basically boils down to the output impedance of
the video output stage, and on a single ended stage, as most are and were, this
impedance is different for negative and positive going slopes in the waveform.
There are other theories, this is only one. I speak of this because the flaring
effect was almost always noticed on some chassis, back when all the CRTs were
interchangable. It followed the set, not the CRT.
I think this could turn out to be an interesting thread. :)
Michael Shell
>
>
> I think it's due to the electron gun clipping when it's overdriven.
> Even a new CRT will bleed if it's driven hard enough, but most TVs are
OK, I agree with what you are getting at here. Consider what happens
when a transistor is overdriven. There are so many excess carriers in
the
device that it takes a long time for recombination to occur. This delay
will
result in the transistor taking a longer than normal time to switch off.
Jurb6006 caught something I should have mentioned: Often bad video
drives
can cause bleeding (flaring) too. These cases could be caused
by overdriven transistors (see above), feedback loops, or some type
of ringing effect. I am interested in this as well, but my real
fascination
is when the CRT is the trouble maker.
Unlike a semiconductor, a CRT is a pure majority carrier device - no
holes,
just electrons flying around in there.
What bothers me is this: Say we have a test pattern consisting of a
solid
red square in the middle of an otherwise black screen. We turn up the
saturation/contrast (and have a weak CRT), we will see bleeding to the
right of the square. Instinctively, we FEEL we know what is going on.
But think about it. The instant the electron beam leaves the square, the
voltages on the CRT grid/cathodes are such (or should be) that the red
gun should be shut off. (It is only like a nanosecond from the gun to
the
screen.) If the CRT cathode is weak, or the G2 voltage is too low, then
I
would expect the beam to cutoff even faster! Yet, the phosphors in the
bleed DO see electrons exciting them! So, what is happening here? Did
charge somehow build up somewhere in the tube? OR has the tube changed
properties in such a way to cause trouble for the video output stages
in a manner which would cause problems like Jurb6006 suggested? In the
later case, it should be possible to design or modify the output stages
to be resistant or immune to this problem. (I am not suggesting it would
be worth it though).
The thing I really despise is that it seems to happen on CRTs that still
have
perfectly acceptable brightness.
>
> All CRTs do it when they get weak enough. Samsung seem to be worse than
> most. In general, all CRT manufactuers have been cutting costs. A
> larger percentage or 8 year old or newer TVs that I see have bad CRTs
> than ones that are more than about 16 years old. I just picked up a
> heavily used 1982 Zenith from the side of the road and it has a better
> looking CRT than most new TVs.
I didn't mention this in my initial post because I didn't want to bias
(pun intended) anybody. I suspect that you may be right. I used to
work a lot on older tube color televisions and I don't recall the
bleeding
problem. The first time I remember seeing it was on a 1978 Sampo. I have
seen it EVERYWHERE since then.
Has anybody seen the problem on a, say 1970, tube type set with good
video
drivers and correct CRT voltages (so you know it to be the CRT).
I wonder if the rise of the inline gun is related?
Mike
--
> Andy Cuffe
> balt...@psu.edu
JURB6006
>when a transistor is overdriven. There are so many excess carriers in
>the
>device that it takes a long time for recombination to occur. This delay
>will
>result in the transistor taking a longer than normal time to switch off.
Very good point. Now I'm thinking it can''t be any kind of "secondary
emmision", because if it was emmitted from anywhere but the cathode the purity
would be off.
Perhaps what your talking about is when they flare, yet the CRT isn't weak.
This usually is due to a low collector suppply voltage to the vid outs.
Unfortunately there is no way to tell on a normal scope whether the effect is
being driven to the CRT or if the effect is IN the CRT.
Acually on what I call "voltage driven" units, you CAN see some type of
clippilng when either the CRT is weak or the G2 is low, but let's say on an
older Sony, it seems like the clippilg is omnidirectional.
Yet when this happens, the purity is not extremely affected.
On the sets that flare profisely, is it possible that the designers stumbled on
a rudimentary form of AKB ?
OK I owe you 2 cents and you owe me 2 cents, so we're even so far. Everybody
in this thread, wait till it's allll over. . . .
Asimov
--- on the topic of "Physically, why do weak color CRTs "bleed"?"
It's analogous to how speakers and amps distort when the clip. When the
clipping occurs in the crt it's bandwidth falls to zero and you see then
a type of ringing or smear. It's like the beam gets cutoff then
saturates repeatedly very fast at a video rate. Also the electrostatic
voltages which set the convergence get all thrown out of whack when this
happens. This last is why a crt with weak emission will also show poor
purity, bad convergence, and a loss of tracking.
MS> I have known for sometime that when a color CRT gets weak, colors
MS> tend to bleed to the right, especially when the contrast, saturation
MS> and brightness levels are turned up.
MS> This is mentioned in the CRT FAQ:
MS> http://www.bithose.com/serfaq/REPAIR/F_crtfaq.html#CRTFAQ_047
MS> However, *why* does it do this? I can understand that a low cathode
MS> emission level would result in a dimmer picture. If the lower beam
MS> strength caused things to defocus, then wouldn't that result in
MS> bleeding to the left *and* right (color blurr)?
MS> On a related note, are there any brands (CRT or TV or monitors) that
MS> are resistant to this type of degradation? Are Trinitrons more likely
MS> to do this than those CRTs using Invar shadow masks?
MS> Thanks in advance,
MS> Mike
... If all else fails, hurl it across the room a few times!
Maarten A.E.Bakker
> I think it's due to the electron gun clipping when it's overdriven.
> Even a new CRT will bleed if it's driven hard enough, but most TVs are
> designed so you can't turn up the contrast that much. Once the CRT goes
I have a 17 year old Grundig TV with the A66-540X thick-neck inline picture
tube. When it is overdriven, the picture goes 'double vision' but I can never
get it to bleed. Now I come to think of it, I saw some terrible out of focus
and low emission thick-neck pictures (can you spell
Standard Elektrik Lorenz / ITT) but I never saw one bleed...
--
Greetings, Maarten.
Maarten A.E.Bakker
> There are other theories, this is only one. I speak of this because the flaring
> effect was almost always noticed on some chassis, back when all the CRTs were
> interchangable. It followed the set, not the CRT.
This is certainly applicable for European Philips 2B chassis. They
did have a rudimentary AKB circuit, along with a picture mute when
the guns are not within range. Cutting a diode on the drive panel helps
for this, but also seems to encourage the bleeding effect some more.
--
Greetings, Maarten.
Cliff Raeihle
by the increased brightness/contrast/color settings the user or the ABL
circuit is using to try to compensate for a soft CRT. When the transistor(s)
saturate, they are slow to come out of saturation - thus the bleeding to the
right.
Increasing the G2 setting may help, by reducing the amount of drive needed
to attain a given brightness.
Some poorly designed sets will do this even when new, if the
brightness/contrast/color are advanced beyond normal.
Of course, if the CRT is soft, the best remedy is to replace it.
Cliff
"Michael Shell" <mike...@yahoo.com> wrote in message
news:3A32BCFF...@yahoo.com...
Michael Shell
>
> The video driver transistor(s) are going into saturation, being
> overdriven by the increased brightness/contrast/color settings
> the user or the ABL circuit is using to try to compensate for
> a soft CRT. When the transistor(s) saturate, they are slow
> to come out of saturation - thus the bleeding to the right.
> Increasing the G2 setting may help, by reducing the amount
> of drive needed to attain a given brightness.
> Some poorly designed sets will do this even when new, if the
> brightness/contrast/color are advanced beyond normal.
> Of course, if the CRT is soft, the best remedy is to
> replace it.
>
> Cliff
>
OK, this is getting interesting. If you're right, it would
explain a lot of things. I'll bet anything that the old
tube type video output stages don't take so long
to come out of saturation. This may explain why we don't
see the problem on ancient sets.
My estimate for the length of a flare is 1 to 10usecs.
I know we are dealing with power electronics, but that
seems to me to be a pretty dang long time till cutoff!
I need to lookup some specs to refresh my memory.
If what you are suggesting is true, then it may be possible
to modify the drives so that we may get clipping, but
not a delayed cutoff. This may be much less objectionable
when viewing. It may even be cost effective if you have
a discrete final drive stage (pop in three new super
spec ECG's vs new CRT)- which raises another question:
Who usually saturates first? The final stage? The stage next
to the final? 5us is a long time, only a high power stage
should suffer this badly, right?
BTW, MOSFETS, like tubes, are majority carrier devices and
might be the thing to use as final drives (if you are
a designer or doing heavy hacking- you may have to use
a BJT as a constant current source and two MOSFETS in
a differential configuration).
As far as a ABL circuit compenstating things- How would it
"know" that the CRT is getting weak? (1) because the user
is at the last 10% of the settings? (2) because grid
currents are falling? (3) because cathode voltages are
falling?
If the modern ABL is that smart, maybe that's another
reason we see it more in modern TVs.
What really bugs me is that many, if not most, of these sets
that flare like mad seem to still have adequate contrast and
brightness. If the drive circuits would not behave in this
manner, the set would be usable for a long, long time. The
video electronics should GRACEFULLY handle an aging CRT.
And those poorly designed new sets that flare right out of
the box (when you turn the brightness up a little too much),
they'll be doing it on normal settings within 2 years!
Mike
David M. Brodbeck
> What really bugs me is that many, if not most, of these sets
> that flare like mad seem to still have adequate contrast and
> brightness. If the drive circuits would not behave in this
> manner, the set would be usable for a long, long time. The
> video electronics should GRACEFULLY handle an aging CRT.
--> That would be inconsistant with selling more TVs.
---------------------------------------------------------------------------
David Brodbeck, N8SRE dmbr...@mtu.edu
finger gu...@cyberspace.org for my public key block.
"He's so dull, he ordered Irish coffee during a layover at Shannon
Airport with the instruction, 'Hold the whiskey, and make it decaf.'"
-- Margaret Carlson, writing about Warren Christopher.
nospam
poisoning.
Michael Shell
>
> As far as a ABL circuit compenstating things- How would it
> "know" that the CRT is getting weak? (1) because the user
> is at the last 10% of the settings? (2) because grid
> currents are falling? (3) because cathode voltages are
> falling?
> If the modern ABL is that smart, maybe that's another
> reason we see it more in modern TVs.
>
(I probably should have used AKB instead)
I think I have answered my own question.
Check out the AKB IC at:
http://www.semicon.toshiba.co.jp/noseek/us/td/08frame.htm
Specifically the example circuit (showing a discrete output stage) in:
http://www.semicon.toshiba.co.jp/pdf_e/docweb123/e003967.pdf
They monitor the cathode current.
Do sets with AKB IC's like this, limit their drive, or do they still
flare?
Mike
Michael Shell
>
> I allways thought that the tube "bled" as a result of cathode
> poisoning.
I am not saying that you are wrong, but this is what I want
to know:
Explain to me how this allows the electron gun to
continue to fire for 5us after the video output drives have
acted to shut it off!
Mike
Tom MacIntyre
Sony's use AKB, and they bleed.
Tom
>
>
> Mike
Andy Cuffe
>
> The video driver transistor(s) are going into saturation, being overdriven
> by the increased brightness/contrast/color settings the user or the ABL
> circuit is using to try to compensate for a soft CRT. When the transistor(s)
> saturate, they are slow to come out of saturation - thus the bleeding to the
> right.
> Increasing the G2 setting may help, by reducing the amount of drive needed
> to attain a given brightness.
> Some poorly designed sets will do this even when new, if the
> brightness/contrast/color are advanced beyond normal.
> Of course, if the CRT is soft, the best remedy is to replace it.
>
> Cliff
>
I have been thinking about this and I think you're right about the
bleeding being caused by the video driver, not the CRT, but I don't
think it's could be caused by a simple overdrive. No new TV will bleed
much, if at all, when the brightness and contrast are at maximum, but
with a weak CRT, they bleed even at mid contrast. Assuming the TV
wasn't adjusted internally, the CRT is being driven with the same
signals as when new.
My theory is that a weak CRT (or a good one with reduced G2) represents
a higher impedance load to the video output transistor. The biasing of
the video outputs would have to be designed for the load created by a
good CRT. When the output load impedance goes high enough, the voltage
can go high enough to saturate the transistor (the CRT isn't pulling
enough current to keep the C-E voltage from going close to 0). You will
notice that all detail in the area that's bleeding is missing,
indicating that the transistor is in saturation. The transistor then
takes a longer than normal time to turn off after the drive signal goes
back down. A quote from a circiut theory book I have says this is also
the reason BJT's aren't used in digital circuits any more (because a
saturated transistor takes a long time to turn off). Basically, the
output is no longer operating in the linear region. This could probably
be confirmed by putting a large resistor in series with one of the
cathodes of a good CRT.
This is confirmed by an Emeraon TV I saw a couple of years ago. A bad
cap in the 200v supply to the video outputs was causing severe
bleeding. I was given the set because it was diagnosed as having a bad
CRT, but I looked more closely after seeing an otherwise bright and
sharp raster. The reduced 200v supply must have been allowing the
outputs to saturate at normal drive and load levels.
Tube TVs were before my time, but I have to admit never seeing one that
bled, even though most of the ones I've seen had weak CRTs. I guess
tubes, being like FETs (in that they are majority carrier devices) which
are used in high speed digital circuits because don't have any delay in
getting out of saturation.
I wonder if it would be possible to modify the video output circuit to
eliminate the bleeding in a TV with a weak CRT that's not too far gone
to be usable?
--
Andy Cuffe
balt...@psu.edu
Thomas Tornblom
Hi Jeff.
Sorry to use this forum, but mail to you bounces.
Please contact me at tho...@hax.se.
--
Thomas Tornblom Tel: +46 8 631 2100 E-mail: Thomas....@Sun.SE
Sun Microsystems AB Fax: +46 8 631 1102
David Carlstrom
>Andy Cuffe wrote:
>OK, I agree with what you are getting at here. Consider what happens
>when a transistor is overdriven. There are so many excess carriers in
>the device that it takes a long time for recombination to occur. This delay
>will result in the transistor taking a longer than normal time to switch
off.
>by overdriven transistors (see above), feedback loops, or some type
>of ringing effect.
>video drivers and correct CRT voltages (so you know it to be the CRT).
>I wonder if the rise of the inline gun is related?
I have a 1966 all tube set. The CRT is 30 years old and a bit tired. It
does not bleed at all.
Michael Shell
Andy Cuffe wrote:
>
> My theory is that a weak CRT (or a good one with reduced G2) represents
> a higher impedance load to the video output transistor. The biasing of
> the video outputs would have to be designed for the load created by a
> good CRT. When the output load impedance goes high enough, the voltage
> can go high enough to saturate the transistor (the CRT isn't pulling
> enough current to keep the C-E voltage from going close to 0).
I think we may have a winner. A scope on the cathode should be able
to confirm it.
> tubes, being like FETs (in that they are majority carrier devices)
> which are used in high speed digital circuits because don't have
> any delay in getting out of saturation.
SIDE NOTE:
They found a way around the BJT saturation problem. In the Schottky and
Low Power Schottky TTL families 74Sxx, 74LSxx) a small diode was
connected across the BJTs to help them recover faster. The
problem with TTL is (1) relatively high power consumption at even low
clock speeds (2) it is not as easy to make them small for VLSI chips
(MOS transistors have a "self aligning gate"). CMOS logic has almost
zero power consumption at low clock speeds, but as the speed goes up,
the power consumed can exceed BJT designs (The is caused by the current
required to charge and discharge all those MOS gates). To keep 1GHZ
Athlons from burning up, you gotta make those MOS transistors really,
really small.
For the ultimate speed trip, there is a logic family known as ECL
(Emitter Coupled Logic). In ECL, the BJTs aways stay in the linear
region. Power consumption is enormous (0.2W for 4 OR gates, unloaded).
ECL can clear 1GHz without breaking a sweat (and drive a 50ohm load all
the way). Use GaAs instead of Si, and it'll clear 10GHz.
END of SIDE NOTE
>
> I wonder if it would be possible to modify the video output circuit to
> eliminate the bleeding in a TV with a weak CRT that's not too far gone
> to be usable?
>
I plan to try this in the next month or so. If I have any luck, I will
post the results. In order to do this, voltages are gonna have to climb
to keep that transistor out of saturation. This is going to result in
more heat. Could it be possible that energy conservation mandates from
the government resulted in flaring? If so, the same thing that causes my
CRTs to flare also causes my toilet and shower to lose power.
From the point of view of the cathode, the CRT is a current controlled
device like a BJT. From the point of view of the grid, it is a voltage
controlled device, like a MOSFET. I can't remember why the video drive
is applied to the cathodes rather than the grids, but I know there was/is
a good reason.
So, barring a radical change, such as CRT grid drive, I think we want to
use BJTs, but we need total control over Ik, regardless of how
Mr. CRT feels. So, we would drive the CRT cathode with the NPN BJT
collector, and have a fairly large Vce - which implies a bigger transistor
with more heat sinking. I had thought of an alternative - using a
transistor that comes out of saturation faster, but as you mentioned
before, we lose detail in saturation and thus would still not like the
resulting picture. Another simple solution would be to "tweak a few V's".
Instinctively, I feel we can do this. Then why haven't the OEMs?
I thought up some humorous explanations for flaring during the course
of this thread. Maybe somebody will get a laugh.
1. A crust spot on the cathode, which only emits at high drive
levels, causes the electrons that pass through it to have a reduced
velocity - by a factor of a 10000. These delayed electrons form the
flare.
2. The front of the cathode starts to wear out. Under high drive levels,
electrons are emitted from the BACK of the cathode. The G2 voltage
then pulls them around, but the result is a corkscrew spiral path
to the screen whose total linear length is on the order of a 1000ft.
Hence the delay. Purity is not affected due to "circular symmetry".
3. A crust forms on the cathode. Under high drive levels, electrons are
trapped within the layers of this crust. Even when the video drive
cuts off current to the cathode, the trapped electrons continue to
leak to the surface resulting in a flare. This is known as the
"cathode becomes a capacitor" theory.
4. The phosphor of old CRTs become "flaky". When hit by electrons,
ionic radiation is emitted radially outward. This radiation makes
other adjacent phosphors super sensitive to future electron exposure.
Hence, we are able to see a nearly zero electron beam which results
in the flare.
and finally:
5. It's all in your head. Buying a new TV raises endorphins in the human
brain, fixing the problem for a couple years. This also explains why
more expensive, feature rich TVs have the problem less often. The CRT
restorer should be applied to the USER not the TV. In skillful hands,
it can also cure one of "color vision" allowing the use of much more
inexpensive BW sets.
;)
Mike
Andy Cuffe
> From the point of view of the cathode, the CRT is a current controlled
> device like a BJT. From the point of view of the grid, it is a voltage
> controlled device, like a MOSFET. I can't remember why the video drive
> is applied to the cathodes rather than the grids, but I know there was/is
> a good reason.
I think it was due to a change in the design of the electron guns.
Delta gun CRTs and some early inline CRTs had the Y signal driveing the
cathodes and R-Y, B-Y and G-Y driving the G1's (I may have reversed
that). Modern CRTs have a common G1, so there's no way to use this
arrangement any more. The signals are usualyl combined on the CRT
board. I guess this had the advantage that a brightly colored object
would just result in higher grid voltage causing no bleeding.
> So, barring a radical change, such as CRT grid drive, I think we want to
> use BJTs, but we need total control over Ik, regardless of how
> Mr. CRT feels. So, we would drive the CRT cathode with the NPN BJT
> collector, and have a fairly large Vce - which implies a bigger transistor
> with more heat sinking. I had thought of an alternative - using a
> transistor that comes out of saturation faster, but as you mentioned
> before, we lose detail in saturation and thus would still not like the
> resulting picture. Another simple solution would be to "tweak a few V's".
>
> Instinctively, I feel we can do this. Then why haven't the OEMs?
I guess they don't care about how the set will perform as the CRT ages.
Some designs do seem to be better than others. I've seen some pretty
sad looking sony's from the 70's, but none of them bled much.
--
Andy Cuffe
balt...@psu.edu
JURB6006
This thing is flaring from the RIGHT !
Maarten A.E.Bakker
> Tube TVs were before my time, but I have to admit never seeing one that
> bled, even though most of the ones I've seen had weak CRTs. I guess
> tubes, being like FETs (in that they are majority carrier devices) which
> are used in high speed digital circuits because don't have any delay in
> getting out of saturation.
Hmm... I just looked it up... The TEA5101 driver IC uses MOSFET
outputs... Does anyone have experience with this IC in combination with
ageing CRT's?
--
Greetings, Maarten.