or Connect
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › OLED TVs: Technology Advancements Thread
New Posts  All Forums:Forum Nav:

OLED TVs: Technology Advancements Thread - Page 7

post #181 of 9448
Thread Starter 
Indeed, I replied a bit too harshly, probably because your latest comments brought to mind your previous comments along the same line. In them I could clearly hear the whine of an idealist: personally, I'm not opposed to such philosophical bents, but am first and foremost a pragmatist. Cost and feasability are foremost on my mind in this area.

Your comments seem to be phrased as a kind of argument for a particular design, or perhaps the lament of one who senses that no design in existence can provide the sought-after ideal qualities. In this you are most certainly correct.

Passive-matrix driving schemes do not inherently require short-duration high-power operation. Rather, such a scheme is chosen to minimize the driver circuitry cost - remember that passive-matrix driver circuits are more costly because the switching elements must carry the full drive current. To minimize the cost, underpowered switching elements are used, but only carry current for a portion of a frametime, thus allowing them to dissipate heat which, if run continuously, would overheat devices of such density.

The practical implementation of this method is frequently row-at-once or column-at-once drive, ie. alternate rows or columns are activated sequentially. At low frequency, this can sometimes give the visual impression of an interlaced display.

As far as the technology of passive-matrix driver design, if circuit cost is no object it can be made to drive the entire panel at once, but it is still limited to small displays, due to I^2R losses in the ITO conductors connecting each pixel. These losses increase with both panel size and resolution (longer and thinner traces, respectively).

All large low-voltage displays use active-matrix, for both cost and technical reasons as I outlined previously. It can be guaranteed that this will not be changing in the future. So if we are talking about TV-sized displays, the discussion ends there: no passive-matrix will ever be used at such sizes. As for smaller devices, they're out of the purview of the thread's topic, which is OLED TVs (nominally, devices of 20" and larger).

Regarding emission times and lifetime, as noted previously, it is up to manufacturers to determine how to compromise between on/off ratio and lifetime. Perhaps you may be able to have some influence on their plans if you can design a regime which shows advantageous performance over current ones.

To finish on a positive note, driving schemes can enjoy significantly enhanced flexibility compared to those for plasma cells, due to the unnecessary sustain and priming pulses, and the need to use PWM to simulate n-bit grayscale emission levels. The nonlinear property of gas ionization requires on-off switching, at least in the plasma display cell.

SED, LCD and OLED can all be driven using continuously-variable voltage or current sources, which is a boon to those who are disturbed by the visual impression of noise generated by plasma PWM. This effect is particularly evident at the viewing distance required for the immersion effect (30 degree FOV or greater), and worse in darker scenes than lighter.
post #182 of 9448
Thread Starter 
After carefully re-reading your last post, I'm mystified as to where you get the 'several milliseconds' for OLED response time. From drive circuit to emission, the switching time should be no more than a handfull of microseconds.

I think you're confusing sample-and-hold motion smear with other properties of the circuit.
"Ask yourself how an AM panel is addressed and controlled. Why the transistors are there and why SAH is synonymous with AM. Furthermore, why LCDs use AM (which is not the same reason why OLED is using them)" Rather than ask myself, I'll ask you instead (since I've already provided the technical explanation for why transistors are there). Liquid crystals require less power to hold in a particular position than emissive OLED cells require to emit; the primary reason for active-matrix in LCD is cheaper drive circuitry. Still, trace resistance is a significant issue in large panels of both types.

Regarding the on-off ratio per unit time that manufacturers may use in standalone OLED displays, that has yet to be decided as none are in production. Depending on manufacturer decisions, initial panels may have either shorter or longer on-times than CRT; they may very well have continuous emission. This will probably change with time, as newer OLED emitter materials are introduced, assuming the market for such devices does succeed.

Finally, my statement regarding OLED response time being better than CRT (and any other display device currently in existence, excepting ILED) was not speaking about motion smear or the sample-and-hold effect. I was speaking about response time in the normal sense of the word, which is the time taken for a pixel to reach requested brightness after the signal has been received.
"This bugs me. Claiming that OLED screens will be much faster than even CRTs and then claiming they must use active matrix is counterintuitive. The two cannot exist together. Active-matrix = SAH effect even with faster refresh rates and PWM or strobbing or whatever." The two cannot exist together? Is there something about transistor-gated power supply you perhaps are unclear about? There is nothing preventing active-matrix designs from using any switching scheme you can devise. The transistors under each cell can switch as fast as the transistors in the driver circuitry of a passive-matrix design.

You could build an active-matrix panel which used short-duration high-intensity activations, precisely emulating the (economically-driven) effect found in passive-matrix devices. Just for the fun of it you could activate alternate rows sequentially, for a neato psuedo-interlace visual effect. There would be no need to do so however, because the cheaper low-power drive circuit would allow you to address all the pixels simultaneously, unlike such passive-matrix designs. Thus the entire panel could be strobed in temporal sync, or each pixel could be strobed in checkerboard fashion, or whatever.

Hell, you could even drive it somewhat like a plasma, using full on-off PWM, but at much higher frequency (Khz.), thus avoiding the visual 'noise' of temporal dithering.

It seems such schemes are unlikely to be used, because they are perceived (correctly) by the industry to be unnecessary.

Following is a simple, easy-to-understand description of the basic differences between an active-matrix and a passive-matrix design.

Passive-matrix switches the cells in a driver circuit located on a PCB behind the glass. Power to twist the liquid crystals or glow the OLED emitter is sent through individual ITO traces to each pixel.

Active-matrix uses very low-power switching transistors (in an integrated circuit) on the PCB to send signal-level voltages through the ITO traces to the base pins of power transistors on the glass, which switch power from on-glass ITO power planes into the cell.

The net effect of active-matrix is we are only running the tiniest of currents through the individual pixel traces, instead of power to supply the cell. Also, the power transistors are photolith'd onto the glass instead of living within the driver circuit. The only time lost is we are now switching on two transistors instead of one, but that is incredibly tiny.

The advantages of active-matrix switching are so great that it is even making inroads into smaller devices. If you read the latest stories in this thread, you'll note that the industry (correctly) sees a dim future for passive-matrix devices.
post #183 of 9448
Iso, rather than get angry and argue I'd rather attempt to settle this with technical information. From the following quote from you I think I can see why we are not on the same page.

Quote:
Originally Posted by Isochroma View Post

Finally, my statement regarding OLED response time being better than CRT (and any other display device currently in existence, excepting ILED) was not speaking about motion smear or the sample-and-hold effect. I was speaking about response time in the normal sense of the word, which is the time taken for a pixel to reach requested brightness after the signal has been received.

I never once referred to response time which as we know is ultra-short in OLEDs and will not contribute to any motion video artifacts that AMOLED might have. What I was referring to is hold time. The SAH effect refers to how long each pixel is emitting light per frame period. The rating of technology I posted above refers to this "HOLD" time. It has been scientifically proven that "HOLD" time is the greatest contributor to motion blurring in display technology.

see link

Active Matrix and Passive Matrix: Do you agree that PM is an impulse display method with one very short pulse of light? This pulse length is determined by the addressing time since each line of pixels is addressed individually. Since there is so little time to emit light the current must be very high to get enough intensity out of the EL material.

Do you agree that AM is a hold type of display method with an extended duration of light emission? Since the amount of time for light emission is much longer the current can be much lower and our eyes integrate the light output over time to get a high brightness. The problem is we also see blurring due to retinal persistence.

Now, if you still disagree then just do a quick google search for AMOLED and you'll get plenty of articles telling you how AMOLEDs hold the image on the screen for the entire refresh cycle.

link 1
link 2
As a result, the AMOLED operates at all times (i.e., for the entire frame scan), avoiding the need for the very high currents required for passive matrix operation.

And if you want to read actual research papers that describe how AMOLEDs suffer from image blurring just like LCDs for the exact same reason as LCDs (SAH effect) and how they plan on fixing the problem (120Hz, BFI, interpolation) just like LCDs then you can read the following and also search the journals yourself
link 3
post #184 of 9448
Also, just to be thorough: The number one reason LCDs use AM is to give the liquid crystal enough time to transition to the desired state. PMLCDs activate each cell for such a short period ot time there was no way the LC could transition properly. This gave poor brightness and severe ghosting.

AM solved this by holding the voltage at each cell for the entire frame scan. At 60Hz this gave a full 16.7ms for the cell to transition properly. The result is high brightness and better ghosting performance. The problem is the SAH effect.

OLED on the other hand does not have a response time problem so the driving force for AM is due to "LIFETIME". Using AM significantly reduces power consumption and increases lifetime of the EL material. The problem is the SAH effect.
post #185 of 9448
Also from the same link

OLED displays are activated through a current driving method that relies on either a passive-matrix (PM) or an active-matrix (AM) scheme. In a PMOLED display, a matrix of electrically-conducting rows and columns forms a two-dimensional array of picture elements called pixels. Sandwiched between the orthogonal column and row lines, thin films of organic material are activated to emit light by applying electrical signals to designated row and column lines. The more current that is applied, the brighter the pixel becomes. For a full image, each row of the display must be charged for 1/N of the frame time needed to scan the entire display, where N is the number of rows in the display. For example, to achieve a 100-row display image with brightness of 100 nits, the pixels must be driven to the equivalent of an instantaneous brightness of 10,000 nits for 1/100 of the entire frame time.

So what happens when you have 1080 lines of resolution ?? Even with multiplexing you can see that you would have to literally blast the EL material with current and fry the lifetime and power consumption to get any usable brightness.

This is why AM is used as I've said about 10 times now. The only way to use PM in large TV OLEDs is to greatly increase the EL material luminous efficiency.
post #186 of 9448
Another paper from Science and Technical Research Laboratories in Japan:

link

First paragraph states as follows : Moving images displayed on active matrix displays such as AMLCD and AMOLED are essentially blurred because hold type displays use active matrix driving.

As I said above, AM is synonymous with SAH effect. 120Hz, BFI, interpolation reduce the SAH effect but do not eliminate it.
post #187 of 9448
It's well known that OLED should suffer from the sample-and-hold effect. What I'm wondering about is why there should be a change between passive and active? I mean isn't the OLED picture "on" all the time either way?
post #188 of 9448
Quote:
Originally Posted by madshi View Post

It's well known that OLED should suffer from the sample-and-hold effect. What I'm wondering about is why there should be a change between passive and active? I mean isn't the OLED picture "on" all the time either way?

In general Flat panel pixel based displays address the pixels in line by line fashion (multiplexing can be used but I'll keep this simple). Active matrix has a TFT and capacitor behind each pixel which allows the cell to be "on" even when not being addressed. So the pixels can actually be addressed line by line and then all turned on at once and held on for the full frame period.

With Passive Matrix there is no switch or capacitor behind each pixel. This means that when the panel addresses the pixels line by line that the pixels actually activate when addressed and de-activate when not. This is similar to a CRT and SED. So at any given moment only one row of pixels is actually on. And the time that the pixels are on is determined by the addressing speed.

Take 1080 lines as an example:

60Hz addressing = 16.7ms of time per frame to address all the 1080 lines
1/1080 * 16.7 = .015 milliseconds of time per line of pixels

This means using passive matrix each pixel can emit light for only .015ms. This is plenty of time for OLEDs but the problem is that to get a usable brightness in .015ms would require a massive amount of current during that .015ms. This would quickly burn out the OLED and kill the lifetime and power consumption.

This is why small screens can use Passive matrix as they have few pixels and less stringent brightness, lifetime requirements.

Large TVs with 6 million pixels just won't work with Passive matrix and current EL materials as the emission times are too short and power consumption to high along with extremely short lifetimes.
post #189 of 9448
Quote:
Originally Posted by xrox View Post

In general Flat panel pixel based displays address the pixels in line by line fashion (multiplexing can be used but I'll keep this simple). Active matrix has a TFT and capacitor behind each pixel which allows the cell to be "on" even when not being addressed. So the pixels can actually be addressed line by line and then all turned on at once and held on for the full frame period.

With Passive Matrix there is no switch or capacitor behind each pixel. This means that when the panel addresses the pixels line by line that the pixels actually activate when addressed and de-activate when not. This is similar to a CRT and SED. So at any given moment only one row of pixels is actually on. And the time that the pixels are on is determined by the addressing speed.

Take 1080 lines as an example:

60Hz addressing = 16.7ms of time per frame to address all the 1080 lines
1/1080 * 16.7 = .015 milliseconds of time per line of pixels

This means using passive matrix each pixel can emit light for only .015ms. This is plenty of time for OLEDs but the problem is that to get a usable brightness in .015ms would require a massive amount of current during that .015ms. This would quickly burn out the OLED and kill the lifetime and power consumption.

This is why small screens can use Passive matrix as they have few pixels and less stringent brightness, lifetime requirements.

Large TVs with 6 million pixels just won't work with Passive matrix and current EL materials as the emission times are too short and power consumption to high along with extremely short lifetimes.

Thank you - good post!!
post #190 of 9448
Thread Starter 
Capacitors are only used with LCD

And they are used because it is a constant potential which holds the liquid crystals in a particular position. The electrical leakage through the LC, while significant, is small enough that its voltage can be maintained by a parallel capacitor.

OLEDs are not at all the same: as energy-consuming devices, they must be supplied with current at all times (they are constant-current devices). There is no capacitor behind an OLED pixel. Thus there is no minimum 'hold time' as there is in an AM LCD design.
post #191 of 9448
Thread Starter 
To elaborate on my previous post, the liquid crystals in an LCD themselves act as a small capacitance (albiet with significant leakage).

The amount of twist that the liquid crystal experiences is proportional to the static electric field between the electrodes. Little current flows from one side to the other; the potential difference (voltage) between the two electrodes creates the electric field.

By coupling a capacitor in parallel, the particular potential last applied (by either the on-glass or on-PCB transistor) can be held (for a while), despite the small leakage current. However, to change the potential, the capacitor must be charged/discharged by the difference in potentials between the current state and the requested state. Larger transitions will be slower, adding a minimum 'hold time' which varies depending on the requested transition and the current state.

Plasma and OLED pixels are very different: they are energy-dissipating devices. Plasma cells are short-circuits of ionized gases, while OLED pixels are diodes with such minimal forward resistance that they are basically short-circuits also.

Such devices require a constant-current energy supply. They have minimal capacitance and there is no particular value in placing a parallel capacitor; it would be drained of all charge at the moment the transistor was deactivated.
post #192 of 9448
Iso, seriously, if you would read the links I provided you'd see that even though the AMOLED is constant current you need a capacitor to supply the current "AFTER" the addressing is complete. Otherwise there is no reason to use Active matrix.

Direct Quote "The TFT array continuously controls the current that flows to the pixels, signaling to each pixel how brightly to shine. Typically, this continuous current flow is controlled by at least two TFTs at each pixel, one to start and stop the charging of a storage capacitor and the second to provide a voltage source at the level needed to create a constant current to the pixel. As a result, the AMOLED operates at all times (i.e., for the entire frame scan), avoiding the need for the very high currents required for passive matrix operation."

Even better, go to google and type in AMOLED and capacitor.
post #193 of 9448
Check out the schematics for AMLCD and AMOLED. LCD is constant voltage and OLED is constant current. Both require a storage capacitor after the addressing TFT. OLED requires a second TFT to provide a voltage source to maintain contant current through the OLED.
LL
post #194 of 9448
Thread Starter 
The capacitor in an OLED circuit is not used to provide energy to the OLED. Rather, it is used to maintain base bias on the control transistor. There is a big difference.

Because the base of the transistor has virtually no leakage current, the base-capacitor's value (microfarads) is very tiny; it need only maintain a very small bias voltage to keep the power transistor open. Thus it takes virtually no time to drain its charge, unlike the capacitor which is directly parallel to a liquid crystal.

The roles of the two capacitors are very different, and so too is their performance. One is a high-value device which pisses charge into a leaky liquid-crystal solution, the other is a small-value device that keeps a pool of electrons swimming in the almost leak-proof base region of a FET or non-FET transistor.

The base-bias capacitor's value is no obstacle to switching the pixel at microsecond rates.

I was expecting you'd raise this point, but I knew to wait for you to further exposit your rationale so we could completely clear up this aspect of circuit design.
post #195 of 9448
Quote:
Originally Posted by Isochroma View Post

I was expecting you'd raise this point, but I knew to wait for you to further exposit your rationale so we could completely clear up this aspect of circuit design.

Nice First you say there is no SAH effect in OLEDs then you say there is no capacitor in OLEDS, then when you're totally proven wrong you belittle the person providing the correct info. I don't recall that I have questioned your intelligence or belittled you in any way. All I did was question your information and this is how you act. Please control your pride.

Quote:
The roles of the two capacitors are very different, and so is the performance. One is a high-value device which pisses charge into a leaky liquid-crystal solution, the other is a small-value device that keeps a pool of electrons swimming in the almost leak-proof base region of a FET or non-FET transistor.

Regarding SAH effect the roles of the two capacitors are identical. That is to hold the pixel "ON" well past the addressing period. Think about it, the data line charges the storage capacitor to the desired value. In LCDs this value "holds" the LC at a certain orientation to emit light. In OLEDs this value "holds" open the gate in the 2nd TFT allowing current to flow from Vdd to ground well "after" the addressing has passed by.

One thing I don't get is how they change the Cs value the next cycle. Maybe you can inform me on this as I am always willing to learn

Do you agree that AMOLED will have SAH?? If so that is enough for me to be satisfied. If not for the love of god please read the links I provided and comment on why they say that AMOLED suffers from SAH. Maybe the links are wrong but some of them are full university research projects.
post #196 of 9448
Thread Starter 

Perhaps you don't remember that the caps have values which differ by at least an order of magnitude. Think! The diagram omits the cap values because it's intended as a basic structural exposition; it's assumed that the viewer has an understanding of basic analog circuits. It is also assumed the viewer will be cognizant of the vastly different values of the two caps, considering their places in the circuit.

One only has to hold a charge on the base pin of a transistor, the other has to have sufficient charge to piss into a leaky LC for a frame. Do you have any idea how much energy-holding difference that is?

As for SAH, I've already told you that the driver circuit design is responsible for its presence. That is independent of OLED technology. It occurs whenever you keep pixels at the requested value for an entire frame.

What I'm saying is that OLED emitters and their active-matrix driver circuits don't have to do so. Unlike LCD active-matrix drivers, the hold caps do not place a reasonable limit on pixel transition rates, due to a different circuit design that removes the need for them to hold large quantities of charge by placing them at the base pin of the transistor rather than in parallel.

OLED active-matrix driver circuits can be built to run high-amplitude short-duration or long-duration low-amplitude mode, or anything in-between. Any display that uses long-duration low-amplitude driving mode will 'hold' the image, so generate SAH visual artifacts. There is no need to even discuss that. I've pointed out since my first post in our little discussion that the driving mode is determinate of hold-time.
For LCD, the minimum hold-time is the largest of the liquid-crystal rotation time and the hold cap charge/discharge time.

For OLED, the minimum hold-time is the time taken to suck the minute charge out of the switch-transistor's base-pin holding cap (whose value is very very tiny).
Depending on target emitter lifetime, OLEDs can be driven any which way you like. The studies you quote look at specific implementations of OLED, specific driver circuit implementations in particular.

How any particular manufacturer chooses to implement its driver circuit is not my business.

The implementation-independant characteristics of OLED are my business. It is those characteristics that differentiate each display technology, and it is there that the ultimate limits of each can be found.
post #197 of 9448
Thread Starter 
Even in an LCD, the limiting factor is the liquid crystal rotation time, not the hold cap's discharge rate. It is because of the slow LC that backlight strobing is required to achieve impulse-like performance.

There isn't even sufficient time in a frame for the LC to reach the requested rotation value, never mind to get there from full-block and back again. In contrast, the OLED emitter material is plenty fast enough to do so, providing the AM driver tells it to. And if the driver is told to, it has plenty of speed to make that happen, with room left over for fun

Thus, if you prefer the impulse-characteristic, then the correct approach is to lobby the soon-to-be manufacturers to design their driver circuits in that way. And if you're partial to the hold-type characteristic, then too you may be able to influence their decisions.

Remember though, that it is the flexibility of the emitter material which allows manufacturers the luxury of choosing their driving regime with such latitude. That speed consitutes one of the implementation-independent attributes of OLED which grant it a natural advantage over all other display types. Whether it is made use of will depend on many factors, which only the future will reveal in its own time.
post #198 of 9448
Thread Starter 

Let's take another look at this circuit

On the left, we'll note that the transistor supplies voltage to the combined capacitance Cs and Clc. To answer xrox's question about why Clc changes: when liquid crystals rotate, their capacitance changes. Now, the capacitance of Clc is pretty small in any state of rotation, and its leakage rate is fairly high.

That is why we need Cs. Cs is a capacitor of much higher value. It provides a source of charge at the voltage level TFT provided when it was on. Because Clc is leaky, charge drains out of Cs into Clc during the time TFT is off.

The problem with the specific circuit pictured above is that in order to keep the voltage across Clc reasonable during TFT's off-time, Cs has to be fairly large, due to Clc's high leakage.

Assuming that Clc is really fast (which it isn't, but let's just assume it is so we can look at the limitations of the other portion of the circuit), we can crank the voltage to maximum across Cs and Clc quickly by activating TFT fully.

The problem comes when we want to bring the voltage across CsClc down to zero quickly. This is because while TFT can provide charge, it cannot be used to remove charge. All we can do is wait for Cs to drain through Clc, which takes a while due to its large capacity.

The process can be sped up by inserting a new TFT, called here TFT2, across CsClc. This second transistor short-circuits CsClc when its base is activated, dissipating the charge in CsClc as heat.

Unfortunately, adding TFT2 means requiring a third connection, Data line 2, in order to activate its base. Even worse, each added transistor makes photolith more complex and prone to defects. Adding a third data line is unfeasible as well.

A third data line can be avoided but this requires making the circuit more complex by adding more components behind each pixel, increasing the failure rate and cost even more.

All of these procedures will speed up the non-LC portion of the circuit (TFT and Cs), but won't help with the slow rotation-rate of Clc's liquid crystal.

Turning to the right circuit (OLED), we see that as soon as TFT2 is turned off, the OLED extinguishes immediately. There is no need to wait for stored charge to deplete, or slow liquid crystals to rotate to their relaxed position. The OLED's excitation-de-excitation times are in the microsecond range.

The slowest component in the OLED circuit is Cs, whose value is low enough that it can be charged and discharged many thousands of times per second. We can make Cs small because TFT2, being a transistor, has extremely minimal leakage current between its base pin and source/drain pins; better electron retention means less reserve is needed to maintain pressure during TFT1's off phase.

Back to the LCD circuit: if we increase the leakage through Clc, or put a resistor in parallel with it, then Cs will be discharged quicker. Unfortunately, we will need a larger value of Cs to hold sufficient charge between TFT on-cycles. The larger value of Cs will precisely eliminate any gains to be had by its faster discharge.

To sum up, liquid crystal operation is a messy compromise. There are methods that may marginally improve performance of the purely electrical portion of its action, but such improvements are small and come at high material cost - not to mention being negated by LC's slow response. Thus, they are not implemented in commercial products.

Due to its different nature, the OLED emitter doesn't require such compromises, and thus naturally achieves higher performance. The one extra cost is the second transistor, needed because diodes require a constant-current source, rather than LC's constant-voltage.
post #199 of 9448
Quote:
Originally Posted by Isochroma View Post

Perhaps you don't remember that the caps have values which differ by at least an order of magnitude. Think! The diagram omits the cap values because it's intended as a basic structural exposition; it's assumed that the viewer has an understanding of basic analog circuits. It is also assumed the viewer will be cognizant of the vastly different values of the two caps, considering their places in the circuit.

Not only did I not say they did I also don't need to use such condescending attitudes You really need to stop that.

Quote:


As for SAH, I've already told you that the driver circuit design is responsible for its presence. That is independent of OLED technology. It occurs whenever you keep pixels at the requested value for an entire frame..........

So after saying there is no hold time issue in AMOLED you go and say that the driving scheme causes the hold time Of course it does, I've told you ten times. Active matrix is a driving scheme!! I've posted more than enough confirmed proof that AMOLED suffers from SAH and why.

Then you say there is no way a capacitor is in AMOLED and that is just my lack of knowledge. But alas when I point out using actual proof that there is a capacitor you say Oh! of course but tell me all about it's technical purpose (which I already told you) rather then saying you were wrong about that.

Quote:


Even in an LCD, the limiting factor is the liquid crystal rotation time, not the hold cap's discharge rate. It is because of the slow LC that backlight strobing is required to achieve impulse-like performance.

How can you say this when I've posted scientific evidence that directly contradicts this assumption of yours. Throughout your argument not once have you posted any sort of confirmed information. Just your thoughts which I have questioned. I on the other hand have posted many technical papers which totally contradict what you believe.

This is not my thoughts versus your thoughts. This is your thoughts versus what actually exists in scientific papers and journals. Did you read the paper that studied motion blurring on SAH displays and determined that even in current LCDs SAH is by far the major cause of motion blurring.

So answer me this: How can a large area TV OLED (AM or PM) using current materials achieve impulse like performance with high brightness and long lifetime. If it can't be done what must they do to fix the problem?

Better yet state simply why you think OLED needs to use AM? (number one reason)

What about LCD - why did they need to use AM? (number one reason), hint look a few post up
post #200 of 9448
Rather than continue with this I must just choose to ignore Iso's misinformation about certain things. Overall, this thread is a great resource thanks to Iso himself. I just cannot agree on this one topic as I actually have proof to the contrary. I don't know what else to do.

So in summary the actual facts are:

-PM has very short high current emission times that are not feasable in large area displays with many pixels and long lifetime requirements.

-AM has much longer lower current emission times that allow for use in large area displays with many pixels and long lifetime requirements.

This is due to the simple facts that OLEDs lifetime is inversly proportional to current through the device. AM is used to extend lifetime while maintaining high brightness.

If you won't take my word for it just google AMOLED versus PMOLED

http://www.onestopdisplays.net/FAQ/FAQ_AMvPM.pdf
http://www.universaldisplay.com/passive.htm
http://www.universaldisplay.com/active.htm

The one problem with AMOLED is that just like LCDs it will suffer from motion blurring from the sample and hold effect (not the same as response time).

120Hz refresh, Black Frame insertion, or pulse width modulation are all concepts to solve the problem.

link 3

An ideal solution would be to produce a high luminous efficiency OLED that requires very little power to achieve high brightness in very short pulses with long lifetimes. That is what research is still focusing on.
post #201 of 9448
Thread Starter 
You are confusing AM driving schemes with OLED. Neither OLED itself nor the various driving schemes require either long or short hold times, on a purely technical basis.

"So after saying there is no hold time issue in AMOLED you go and say that the driving scheme causes the hold time Of course it does, I've told you ten times. Active matrix is a driving scheme!! I've posted more than enough confirmed proof that AMOLED suffers from SAH and why."

It is not active matrix which 'causes' hold time. Both passive-matrix and active-matrix can run OLED emitters continuously or in pulsed operation. It is not economical to run passive-matrix in continuous operation, so that is not done. It is economical to run AM in either pulsed or continuous mode.

"This is not my thoughts versus your thoughts. This is your thoughts versus what actually exists in scientific papers and journals. Did you read the paper that studied motion blurring on SAH displays and determined that even in current LCDs SAH is by far the major cause of motion blurring."

And in my previous post I explained why LCDs always operate in hold mode. And SAH is of course the cause of motion blurring. And SAH itself is unavoidable with liquid crystals because of their limited rotation rate. Limited rotation rate creates both undesireable intermediate values and prevents pulsed operation. Barring any major breakthrough in liquid crystal formulation, it looks like we're stuck with only small improvements in that area.

Then you say there is no way a capacitor is in AMOLED and that is just my lack of knowledge. But alas when I point out using actual proof that there is a capacitor you say Oh! of course but tell me all about it's technical purpose (which I already told you) rather then saying you were wrong about that.

Context! My fault was to not specify where the cap is located, and its purpose (if it is used). You found one design which uses caps to hold the TFT base pin's voltage, so that the PCB drive circuitry's cost can be reduced by driving only one or a set of rows/columns at one time. That method is one implementation of active matrix, and is neither a performance constraining one nor characteristic of all implementations.

Regardless, in LCD passive- or active-matrix circuits, the cap is not the limiting factor preventing faster switching speed. In OLED circuits, due to the extremely fast response of the emitter, cap drain speed may very well be the limiting factor, but it is still orders of magnitude faster than any LCD, and well within the cycle rate necessary for visual strobing.

"The one problem with AMOLED is that just like LCDs it will suffer from motion blurring from the sample and hold effect (not the same as response time)."

The 'problem' is neither active matrix nor OLED, it is the driver design. Because you've found one company with a particular design which runs in continuous emission mode, doesn't mean that other AM designs won't run in pulsed mode. Active-matrix does not imply continuous emission any more than passive-matrix implies pulsed operation. In the case of passive-matrix, for economic feasibility pulsed-mode operation is used, but it is not required in a technical sense.

Active matrix means there's one or more transistors or other 'active' components under the pixel. These 'active' components allow voltage to be switched into the cell from ITO power planes via a low-current control signal. Active-matrix does not specifically require either constant-on or impulse operational modes. It can accomodate either one, depending on design - which is largely driven by economics. And when I say economics, I mean both the cost of on-PCB drive circuitry and the 'price' of burning out OLED emitters quicker using high peak driving schemes.

It is up to manufacturers to decide which mode they prefer. Whichever they use, it will be with active-matrix circuits for large displays.

There will be no end to this discussion until you understand how basic electronics works, at which point you'll be able to answer these questions yourself.
post #202 of 9448
Quote:
Originally Posted by Isochroma View Post

You are confusing AM driving schemes with OLED. Neither OLED itself nor the various driving schemes require either long or short hold times, on a purely technical basis.

LOL, I am not confusing anything Iso. I am just regurgitating what is in the posted literature. Believe what you will but the facts are straightforward and for all to see in the literature. You sure are stubborn and wrong, a bad combination. Stick to posting actual articles and scientific information as you have in the past. It is a very usefull thread.

Cheers
post #203 of 9448
Thread Starter 

You sure are stubborn and wrong, a bad combination. Stick to posting actual articles and scientific information as you have in the past. It is a very usefull thread.

I've run an electronics lab for over a decade. I've built analog and digital circuits, high and low voltage. If you have a better understanding of the electrical characteristics of driver circuits, you are encouraged to post specifics. As for my statements, I believe them to be correct and will continue to do so until you can refute them.

Having read "A new driving method introducing a display period for AMOLEDs", I'm not particularly suprised to find that their driving method is similar in some respects to plasma.

Though it solves the motion blur problem by lighting each pixel for only a portion of the frametime, it introduces some other problems, which heretofore have been characteristic of plasma. In particular, I'm disappointed to see the "Peak luminance" plan. It means that, just like plasma, bright scenes will have lower pixel luminance than dark ones. Whites will look 'dirty' just like with plasma.

The primary goal of their 'peak luminance' scheme is to reduce the cost of drive circuitry by making it too weak to power the entire display fully. Instead, total power will be constrained such that the product of the total number of pixels times their intensity will be a constant.

It is sad to see the potentials of such nice emitters being sabotaged before they can really shine. However, these plans may very well never come to fruition.

Strobing can be achieved without peak luminance limitation under a power * area curve, with only a minor increase in driver circuit cost.

Regarding this discussion, we can continue as long as you wish. Perhaps we may even provide this thread's many other observers with both education and entertainment, while we await substantive news from the manufacturers.

Since you're interested in strobing to overcome the SAH effect, and I've said that it can be accomplished with active-matrix circuitry, we both have good news to give one another. It's really good news because large displays both require active-matrix and also show SAH artifacting (with more pronounced visiblity than small displays).

So, I propose that between us, we develop a homegrown driving scheme for OLED pixels that allows both intraframe off-times and excludes the undesireable limitations of plasma drive circuitry, such as APL limiting. Thus we may grow sweet crops from the fertile soil of argumentation.
post #204 of 9448
Thread Starter 
Reading your link on active-matrix displays, I've pinpointed some areas which can be clarified. The following quote is from that article.

"As a result, the AMOLED operates at all times (i.e., for the entire

frame scan), avoiding the need for the very high currents required for passive matrix operation." It's important to understand that like the PM design, the AM design's on-PCB driver circuit operates in line-by-line or column-by-column scan mode. The main difference is that the AM design allows each pixel to hold its value continuously, rather than strobe at high intensity for a portion of the frametime.

This 'scan' mode is not necessitated by either AM or PM in particular, but is used by both designs to save money by using fewer components in the driver circuitry. 'Scan' mode is cheaper because you only need componentry to address one row/column at a time.

Both AM and PM can be run in 'scanless' or totally-addressed mode. This is where each pixel's addressing is handled separately from others. It costs more, but can be done. It is not likely to be commercially available due to economic reasons, regrettably.

We are thus left with active-matrix and scanning for large displays.

The active-matrix scanned mode can be used to achieve strobe action by increasing the scan speed such that there is time for two transitions per frame, instead of one. The first transition is from discharged to charged (with the charge voltage determined by requested brightness). The second transition is from charged to discharged (black).

The paper referenced in my previous post posits one of many methods to achieve such a device. There's plenty of room for new proposals.

The good news for you is that what you seem to want out of OLED - a reduction or total elimination of the SAH artifact, can be achieved at reasonable economic cost by using high-speed active-matrix scanning with two transitions per frametime, or equivalent methods using various mutations of the basic AM driver circuit.

Whether manufacturers will decide to use one of these methods in their products, is a question only the future will answer.

At this point I'm done with what needs to be said, and over the volume of material we've posted, I think we both have a good idea that what is wanted can be achieved, within technical and economic constraints. Personally, I hope that we can both find a resolution of our concerns in this positive outcome and return to our usual pursuits.

Waiting is hard, especially for those whose ideals are crushed by the current devices on the market. Expectations and projections can be tossed back and forth for eternities, but in the end only patience and time will deliver answers to those questions. In the meantime, my busy life will take me away from this subject once more, for periods of time unpredictable.

Before I go, I'd like to invite anyone observing this discussion, who is qualified by electronics experience to provide their input about the statements that have been made. It is always helpful to have third parties who can lend their voice to the exposition. In particular, I'm thinking about bringing some electronics professionals here to provide their advice on the subject.
post #205 of 9448
Quote:


Having read "A new driving method introducing a display period for AMOLEDs", I'm not particularly suprised to find that their driving method is similar in some respects to plasma.

I've posted for the last 24 hours that SAH is a problem with AMOLED and once you finally read this paper you say your not surprised ?

Anyway, at least you finally agree with the literature Did you happen to see what Conventional driving schemes are for AMOLED. This is what I've been saying from the very beginning, and several literature sources back me up.

Quote:


Though it solves the motion blur problem by lighting each pixel for only a portion of the frametime

It does not solve it at all. It only reduces it. Look at the ratings I posted a while back. In order to reach CRT levels you need emission times shorter than 2ms per frame period. SAH effect is purely persistence of vision. The shorter the emission time per frame the crisper the motion. And this driving scheme at best will match Plasma at an average 4-6ms.

Quote:


Since you're interested in strobing to overcome the SAH effect, and I've said that it can be accomplished with active-matrix circuitry, we both have good news to give one another. It's really good news because large displays both require active-matrix and also show SAH artifacting (with more pronounced visiblity than small displays).

So, I propose that between us, we develop a homegrown driving scheme for OLED pixels that allows both intraframe off-times and excludes the undesireable limitations of plasma drive circuitry, such as APL limiting. Thus we may grow sweet crops from the fertile soil of argumentation.

I personally think that to " fully" overcome SAH using AM or PM circuitry will require different EL materials at the least. Specifically higher luminous efficiency and lifetimes. This will give loads of lattitude when it comes to current driving and emission times.
post #206 of 9448
Quote:
Originally Posted by Isochroma View Post

Reading your link on active-matrix displays, I've pinpointed some areas which can be clarified. The following quote is from that article.

"As a result, the AMOLED operates at all times (i.e., for the entire

frame scan), avoiding the need for the very high currents required for passive matrix operation." It's important to understand that like the PM design, the AM design's on-PCB driver circuit operates in line-by-line or column-by-column scan mode. The main difference is that the AM design allows each pixel to hold its value continuously, rather than strobe at high intensity for a portion of the frametime.

Explained this 10 times already (look back)

Quote:


The active-matrix scanned mode can be used to achieve strobe action by increasing the scan speed such that there is time for two transitions per frame, instead of one. The first transition is from discharged to charged (with the charge voltage determined by requested brightness). The second transition is from charged to discharged (black).

The paper referenced in my previous post posits one of many methods to achieve such a device. There's plenty of room for new proposals.

Yes but all proposals are limited by the material lifetime issues. There is no way to get high enough brightness in microseconds without cranking up the current to lifetime killing levels, this is why we need higher efficiency materials with longer lifetimes.

Quote:


The good news for you is that what you seem to want out of OLED - a reduction or total elimination of the SAH artifact, can be achieved at reasonable economic cost by using high-speed active-matrix scanning with two transitions per frametime, or equivalent methods using various mutations of the basic AM driver circuit.

Whether manufacturers will decide to use one of these methods in their products, is a question only the future will answer.

At this point I'm done with what needs to be said, and over the volume of material we've posted, I think we both have a good idea that what is wanted can be achieved, within technical and economic constraints. Personally, I hope that we can both find a resolution of our concerns in this positive outcome and return to our usual pursuits.

I think OLEDs is a great technology and I think this thread is a great resource but I know that large AMOLEDs will not achieve CRT motion performance unless better materials are used.

Quote:


Before I go, I'd like to invite anyone observing this discussion, who is qualified by electronics experience to provide their input about the statements that have been made. It is always helpful to have third parties who can lend their voice to the exposition. In particular, I'm thinking about bringing some electronics professionals here to provide their advice on the subject.

Aside from this just do a search of SID or IDW display journals on AMOLEDs and how the SAH effect is being dealt with.
post #207 of 9448
Thread Starter 
Considering the high peak brightness and short on-times you quote, you are correct in that the primary problem to be solved if manufacturers decide that SAH warrants their attentions, is the emission material.

The OLED emission material would have to withstand significantly higher peak currents than it currently can to achieve such high-peak short emission times.

A large determinant of power-delivery schedule of the active-matrix driving circuitry in the first discrete OLED displays on the market, will be the lifetime vs. power density characteristic achieved by OLED forumulation at that time.

It may be the case that in order to save costs, manufacturers could use continuous-emission mode in their first models. This may be a horror to you, but a large fraction of the population is not disturbed by SAH artifacting; personally, I can barely notice it in LCDs fed 60 fps (interlaced) video.

Manufactureres may decide that other attributes of their OLED devices will be sufficient for buyers to pay its premium price: perfect black level, high color purity, high resolution, light weight, low power consumption, thinness.

SAH is but one attribute in a larger mix, in a market that is under very rapid change.
post #208 of 9448
Quote:
Originally Posted by Isochroma View Post

Considering the high peak brightness and short on-times you quote, you are correct in that the primary problem to be solved if manufacturers decide that SAH warrants their attentions, is the emission material.

The OLED emission material would have to withstand significantly higher peak currents than it currently can to achieve such high-peak short emission times.

A large determinant of power-delivery schedule of the active-matrix driving circuitry in the first discrete OLED displays on the market, will be the lifetime vs. power density characteristic achieved by OLED forumulation at that time.

It may be that, in order to save costs, manufacturers could use continuous-emission mode in their first models. This may be a horror to you, but a large fraction of the population is not disturbed by SAH artifacting; personally, I can barely notice it in LCDs fed 30 fps. video.

Manufactureres may decide that other attributes of their OLED devices will be sufficient for buyers to pay its premium price: perfect black level, high color purity, high resolution, light weight, low power consumption, thinness, etc.

The funny thing is, and you may laugh at me for this, that I have extremely fast retinal persistence and thus do not see much motion blurring even on LCDs. I don't even see those plasma rainbows people complain about but I know they exist. I am more interested in the science than anything else as that is what I do for a living.
post #209 of 9448
Thread Starter 
A final horror which I sometimes ponder: even with the lowest-cost circuitry and panel componentry, AMOLED may never displace LCD from the TV-size market, or it may take longer than the decade so optimistically forecast by certain analysts (see stories). The picture in my head is of us struggling to make a better chair arrangement on the Titanic's deck, as it slowly sinks below the waves.

It's not outside the bounds of probability that I may not be able to buy a 40"-class OLED TV inside my lifetime, though I certainly hope and believe that this won't be the case.

OLED is having a slow and complex birth process; right now the key to economic success is mass production on a big scale, and material cost reduction too. I'd be really happy to get any TV-sized OLED and, personally, wouldn't be too mad if it didn't have strobing. Others will have different priorities, I know.
post #210 of 9448
Thread Starter 
On the topic of SAH vs. impulse displays, some personal observations of myself and friends. I get headaches after a few hours of using my PC CRT at 88hz. I can't watch plasmas in the local store for more than a few minutes without eyestrain & the beginning of headaches.

LCD however, does not cause me such problems, even under extensive viewing. Two of my friends won't buy anything but LCD, despite my encouragement to get a good CRT. One of them gets terrible headaches from any kind of CRT.

Because of these problems, I will not purchase any display product that uses impulse operation. When my savings are sufficient, the CRT will be replaced with an LCD or OLED display, which uses the SAH driving method.

Both SAH and impulse modes have certain advantages, and certain disadvantages. Since we probably know well the advantages of impulse for motion rendering, maybe it's time to mention that a good portion of the population is sick of living so long with interlaced, flickering CRTs.

The problem of flicker (interlaced and non-interlaced {strobe}) and the health problems it causes in a significant fraction of the general population will drive them to pay the high cost of new display technologies only if they offer relief from the pain. A pain that gets worse as the display size increases and APL rises.

LCD has already given them the taste of a stable, flicker-free image. Some will notice and be displeased by motion smear on SAH displays, but at least as many and probably more will find the eased eyestrain and vanished headaches to be of greater value.

Considering the pressure to get OLED displays onto the market ASAP at a price and with a lifetime that have some hope of competing with LCD, manufacturers may decide that not only is impulse operation unworthy of inclusion due to panel lifetime degeneration, but that it is an active liability for their first products, where eyestrain considerations will be paramount: PC displays.

Now I'm not privy to these manufacturers' internal decision-making processes, but if I was in their board rooms casting a vote and giving a reason for it, I would make a logical argument against the impulsive driving method for at least first-generation products, for the reasons outlined above.

I would further add to that argument that if impulsive driving were to be implemented, it ought to be added to the top end of the product mix as a premium feature: for only those whose concern with motion rendering stands above all other priorities.

The rationale for this argument is that OLED's small (projected) market is already priced out of sight of such a large fraction of the population, that its production costs will be difficult enough to recoup as is, without further fragmenting or alienating its (prospective) toehold in the market by introducing a feature which decreases panel lifetime significantly (from the already limited OLED lifetime) and alienates potential customers who are affected by flicker sensitivity.

The more I think about it, the more convinced I become that it is suicide for an aspiring OLED display manufacturer to even attempt introducing their product with this 'feature', given that it is short lifetime that has kept OLED off the market for so long.

Their products will have significantly shorter lifetime than competitors' SAH products, and they will have to wait longer than their competitors for sufficient emitter material advancement to enable the release of a product with acceptable lifetime and color-shift (differential aging). The marginally higher price they (may) be able to get for it will not be sufficient compensation for delayed introduction and smaller market (flicker-sensitive segment).
New Posts  All Forums:Forum Nav:
  Return Home
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › OLED TVs: Technology Advancements Thread