Lcd Troubleshooting - 2

Dear coachcarter,
I can no longer find the original thread - LCD Troubleshooting (Inverter?) and your post is also gone. Since I do not have 50 messages, yet I have to start a new thread. There also seems to be no way to contact the admin about this. A google search will still return results in the "cached" link.

The prescription was given by "Littlest Hobo". <- Thanks again!
Here is part of his post:

I have isolated the cause of this issue on my NEC 1700V to what may be a design deficiency. I have also noticed basic electronics design deficiencies with 2 of my 6 LCD monitors.

The NEC MultiSync 1700V (17") had a malfunctioning fluorescent backlight inverter board; indicated by the screen going blank after 1-2 seconds after startup.

I traced the issue back to the PWM IC (IC1, TL1451A) output driver not switching (however, the controller ICs seldom blow, so the IC's output must have been disabled due to an external control signal as I have a clock signal on pin 1. I confirmed that the PWM IC's internal short circuit protection is the cause of the inverter board shutdown, by grounding pin 15 / SCP on TL1451A; the inverter comes back up and the monitor works after toggling the ON/OFF switch with pin 15 grounded (shorted to pin 8).

So what is causing the short circuit condition? Seems to be a design deficiency.. I populated a 39K resistor at site C10, [in shunt with C12 which is 10uF] and monitor works. The TL1451A datasheet / application notes recommend a 50K potentiometer in shunt with C12, my suspicion is that its for fine-tuning the Short Circuit current threshold. This component was essentially missing, which may have led to the inverter board going into short-circuit protection mode unnecessisarily after external component parameters have drifted over time).
;------------------------------------------

Back to your original question: TL1451 is a 16 pin, surface mount, Texas instruments device. The board it resides on is the long circuit board attached with two screws on the side of the panel shield. The board has a plastic insulator wrap around it, glued in two spots on the bottom. This board is called the inverter - it drives the fluorescent tube.

Disconnect and remove the inverter (the board described above), take the plastic insulator off. The longest chip on that board is the TL1451, it is almost right in the middle of the board. There is quiet a few surface mount caps and resistors around it. Within 1/2" of one corner - near pin 15 you will find an empty spot - that is where I too, soldered a 39k resistor. There is some info about this chip on the ti website www.ti.com. Type the part number and search for literature... See SLVS024E.pdf
It would be nice if you can find the inverter schematic somewhere, I would like to see it.

 

If a design deficiency, problems would be happening with large samples of that model monitor and we are not hearing of that. Therefore, it is likely a manufacturing deficiency, either in the production of one of the assemblies, or more likely, an individual component failure or defect causing an out-of-tolerance condition. Even if the maker is using poor quality components and/or lousy quality assurance to weed out defective samples, that would not be a "design deficiency" - more likely a marketing decision to use cheaper components from supplier with less than stellar reputations.

Without a schematic and parts breakdown, or an identical monitor for comparison, anything more would be a guess.

My point is, don't blame the engineers - they rarely hold the purse strings.

 

Digerati, I really wish to be on your side here but in light of the facts I will take Littlest Hobo's side.
The facts are: - No circuit diagram available
- same problem on many monitors
If a part needs to be of a certain quality than it will be marked that way in the parts list and on the schematic. An LCD does not require fancy components like say a CRT does. Apart from the backlight it should almost be everlasting.
Design deficiency is very likely the cause here. That is why I would like to see the schematic. Just last month I had to fix an acer LCD - 19" that would not turn on. The notorious problem there is the under-voltage reset circuit. It is active high and drives the micro. The mistake? The reset circuit is connected to a 5V supply, but drives 3V logic. Not a visible problem on most monitors but...

 

Oh? Same monitors? Got model numbers? Links? Same exact problem caused by the same exact faulty device in the exact same circuit??? If you can say yes to that, and show (with links) that this exact failure is widespread, then and only then will I believe it is a design problem.

Whoa! Sorry, that's how it works in theory and on paper, but that's not how the real world works. There's no such thing as "fail-proof quality"!!! Until man can consistently create perfection, there will always be less than perfect to poor quality raw materials lurking about. Even with the cleanest factories using the purest raw materials and most advanced manufacturing techniques, there will always be the premature failures. These parts are made by the millions so even if with a quality percentage of 99.999% success, that's still 1000 out of every million that fails. And one Acer a month ago hardly indicates a pattern.

"Design" is based entirely on the Laws of Physics. Taking the design from paper to application involves human interaction - and a budget. To suggest otherwise is the same as saying any no-name generic 600Watt PSU is equal in quality to a 600W supply from a major PSU reputable maker. Or that every 22 inch LCD monitor performs the same. For nearly anything, are cheap "entry level" models that use cheap parts, poorly trained workers, and inferior manufacturing techniques, and there are "top-quality" models that use quality parts, highly trained people, and the best techniques.

lol What??? Ummm, tubes are hardly fancy, or high tech. I would say the same for analog, which all CRTs are.

No doubt. But that is rarely possible these days.

I don't know about that. Inverter replacement is pretty common.

Then without a diagram, there's no way to determine what the designers said was required. So I contend it is cheap parts (a marketing decision) and not a design deficiency.

 

I did not say tubes were fancy, but that they require fancy electronics -
such as the 35kV power supply where a transistor must be capable of withstanding 800V on Uce - developed in the late 70s. Compared to that, a fluorescent tube driver board (aka inverter) is a much simpler beast. No new components had to be developed, and max voltages are at least an order of magnitude lower. (100 times lower in practice).
Since the 17" NEC monitor is already obsolete, posting a schematic would not reveal any "trade secrets". Most of the time when big firms are hiding something it is for one reason alone - the design was a quickie that had to be done yesterday... So it works, but that is usually the only good thing you can say about it.

Getting back to the Acer - driving 3V logic with a 5V signal on one of the inputs (reset) is not considered a good design practice... You would do such a thing only in special circumstances - entering some other mode of operation, programming internal ROM and so on but then you probably do not want to use a reset pin for that...

Sure, but is it not interesting that the most references you can find on the Web about an LCD monitor that has a problem happen to be describing the same problem you are currently trying to fix?

 

This is not about assigning blame to engineers, its about products lacking provisions for repair and maintenance. I referred to this as a design deficiciency because there is not even a placeholder on the PCB for a potentiometer, so the board design does not consider the backlight current draw increasing after years of operation.

I have been using this monitor since adding the original post and it still works fine with the 39K resistor. Anyone else would have added this into a landfill by now. Believe it or not, I paid $1300 after tax for this monitor several years ago, I consider it an investment. I am not happy that a 50 cent component was omitted.

Wojski... it is commen that 5V signalling can drive 3V inputs provided the inputs are 5V tolerant, which many logic families are. You might want to check that chip for 5V tolerance.

 

When I mention backlight current draw increasing, I don't know if it is caused by the tube or the transformers on the inverter board... My guess it that the enamel coating on the magnet wire of the transformers may be breaking down over time and causing the transformers to become less efficient.

It's not something that can be detected at the driver level....

 

Well, that's a whole different issue - and as a technician myself, dear to my heart. The desire for companies to make notebooks thinner and lighter than the competition has created a nightmare when it comes to repairs. There are no Form Factor standards proprietary parts are used everywhere. Simple required tasks, like cleaning, are nearly impossible for the common user, and a chore for pro.

We can't blame the engineers for that because the choice to leave out [real estate consuming] maintenance friendly capabilities is a marketing decision, not a design decision.

I don't see that as an example of any design or even marketing deficiency. Voltage regulator circuits are there to compensate for the affects of aging too, so the circuit "theory" does not support the need for major adjustments, as would be provided by a pot. The problem is theory and real world rarely mesh perfectly, 100% of the time, for all of time - and won't until man can create perfection - as in perfectly pure raw materials and perfect zero tolerance manufacturing and assembly techniques. It is more likely something failed prematurely, or more drastically than "normal" and it is only after the fact, as in after the products have been on the market for some length of time and historical data on common failures is obtained could the engineers revisit redesign the board (if paid to do so - for a now obsolete, out of warranty, product) for a pot to compensate for voltages varying outside theoretical and historical expectations.

But even if the engineers foresaw a need to vary the circuit in the monitor's later years, it is not engineers who don't design in placeholders, it is the bean counters. It costs money to design in a placeholder for something that may or not be needed. It costs money to manufacture in a placeholder on the board. And it costs money to document, verify and publish (maintain on-line) the documents.

It is also the bean counters, and reality in many cases, that say simply replacing the circuit board is cheaper than repairing it (hence the landfill/environmental concerns).

I sympathize with you on your expensive monitor. At the time, I was happy I was getting a good deal by spending nearly $800 for this Samsung 17" LCD. It still has a gorgeous display, but it is only a 4:3 17". Who wants that?

My point is, I guess, is design deficiency suggests designers screwed up. I [would like to] think it is more likely their hands were tied by the bean counters and marketing weenies and it was a marketing decision, rather than a design deficiency.

 

I have never seen accountants get involved at the board design level, i could see them possibly sticking their noses into and scrutinizing the bill of materials out of boredom, but something such as a missing component placeholder would more likely be the decision or oversight of the board designer. Potentiometers can also introduce drift and intermittency, so you wouldn't populate for mass production unless its really needed, but this design does not even provide a proper voltage divider (ie. 2 resistor placeholders) to create the required circuit as per TI's app note in the datasheet.

There is a point at which one must speak out and assign blame. I personally feel its improper for companies to experiment with OPM, while meticulously counting their own beans and quite frankly I'm fed up with it.

 

Have you actually seen the schematic?
What is OPM ?

 

I have been working in power electronics (specifically with switching power supplies) for more than a decade now and really need to put my $0.02 into this topic.

As the backlight lamps age, they require more current from the inverter to fire and to stay illuminated. This is a simple consequence of the physics of the lamp itself.

A previous employer of mine had a product to fire and light a large lamp, and the unit had an OCP trimpot to set the current limit higher (within certain limits) to allow older lamps to be used.

As for the monitor: yes, if the backlight lamps are old it is entirely conceivable that the OCP could kick in and prevent the inverter from running. Does this constitute a "mistake" on the designer's part? Hardly.

The lack of a 50k pot does not in any way mean that the OCP threshold has not been carefully considered.

What must be made clear is that there is risk in arbitrarily adjusting the OCP limit higher (which adding the 39k resistor does).

The risk is that you are (1) potentially allowing the inverter to run at a power level higher than it was designed to safely do (depending on the lamps) and, (2) potentially risking destroying the inverter or starting a fire if there is an actual fault in the monitor that relies on the OCP for safety (i.e. an actual short)

If you are inclined to undertake this sort of repair, be aware of the risks. I would not arbitrarily go with 39k - I would fit the highest resistor value that allows the inverter to run (minimizing the OCP level change). I would also never leave the repaired monitor on while unattended just in case of a failure.

I would consider this fix a temporary measure, to give you something to use until you get a new screen - it's not a safe permanent fix in my estimation.

(end rant)

 

Note you added your 2 cents to a thread that has been dead for almost 5 months so not sure anyone cares anymore. But to you rant, I don't think there was anything "arbitrary" about the decision to use a 39K resistor. I note he said the data sheet recommend the use of a 50K pot, but the pot was missing. A fixed 39K resistor is a fair compromise to a pot, IMO. Though he suspects a design deficiency, I suspect it was more a marketing decision to cut costs.