Fix for Benq FP737s 17" (Q7T3)

[davmax]

Yesterday my monitor momentarily flashed on and then only the power LED functioned as normal ie green with signal and amber without. No display. It was quickly identified that the backlighing had failed by shining a bright light on the screen to faintly identify a video image.

Dismantling the unit and testing revealed a short circuit bi polar 2SC5707 in one of the two backlight inverters. Why did it fail? Checking around the circuit no other problem could be found, no shorts and no failed caps.

Tracing the circuit of the inverter revealed a simple push-pull free running switch inverter design. Each inverter switches two transformers. The attached diagram is simplification intended to reveal a strange omission. The diodes D1 & D2 are usually included in this type of design to ensure that the feedback winding connected to the bases of the two transistors has a complete circuit and they also protect the transistor emitter base from excess reverse voltage. However in the monitor inverter these are missing!!!! Meaning that the feedback /drive current can only flow through reversed bias junctions.

I have double checked the circuit and cannot believe that this is the case. Flowing drive current through the reverse voltage zener of th emitter base junction is definitely not a good feature, not recommened by manufacturers/designers. This would reduce the transistor life with higher EB junction temp.

I presume this is done to save two diodes worth very little ie 1N4148 at a few cents.

I would be interested in any comments.

Also it can be difficult to source the 2SC5707 transistor. Currently they are available cheaply from eBay.

I will replace the pair of transistors in the failed inverter and speculate whether to insert diodes D1 & D2, this would increase drive current, perhaps a single resistor should be added to limit base current. I will post further.

[starfury1]

Humm this sort of rings bells there was a post...I think here on inverters and transistors and diodes...I think Benq was the brand....now I am wondering if you may have hit the crux of the matter Davmax.

Transistors were dying and people were whacking in big ones....I think TIP 41C's were mentioned

I'll see if I can find the post,

ok found it...

now it runs over 8 pages...not sure if its in relation to your post cause I didn't re-read it all again (late shouldn't be here)
And it was DELL made by Benq..

Dell E172Fpb

there is part schematic on page 6. (thanks to UMD_EE)
benq inverter pdf (very similar)

Dell info pdf

another post
only really pics of possible interest

Like I said dont know how relevant it is ,
it just rang a bell when you mentioned it...

HTH some how or someone

cheers Davmax , All

[davmax]

Thanks for links. The closest ref to my circuit is the lengthy explanation by electrical engineering student. There are circuit references and no accompanying diagram on the PDF.

I am going to use a bench power supply to get one output inverter operating to check op conditions. Unfortunately the Benq PCB mounting makes monitoring difficult when installed.

Will post results later.

I see that others have been substituting TIP41C transistors. There is no way that these can replace the extra ordinary 2SC5707, it is a very good high speed, high gain switch, the TIP41C is NOT. It is important to use the 2SC5707.

[starfury1]

yeah I was bit Eh? about that when I read it, RE the Tip 41C...its seems to me to be a band aid fix and not a good one at all if the original are used cause of the High Speed spec...I hadn't looked really at the issue just read thought the post awhile back and thats was what twigged me on those transistors when you posted, cause it seemed to be a common failing...and they fail again
(I had to quickly check the transistor number in the thread but it was what I thought the same.)

I remember there was something about checking a diode and if a condition is not met chances are they will go again...like I said Id have to fully re read it...take it in and digest it

Anyway not the guru, so I'll be interested and what you find Davmax
BTW how did you go with Acronis? if you have had as yet a chance to get back to it.

Cheers

[davmax]

Re Acronis. Still trying to get the support guy to read and understand my report. His last reply was still off the point so I edited his reply pointing out the misunderstandings, hopefully that will work.

[davmax]

Well the investigation of the Benq lamp inverter continues.


I have attached a diagram taken from the Benq service manual marked up to match actual circuit.


First it must be stated that to measure the inverter under normal operation is nigh impossible. It is hard to get at and video signal must be present. To test the lamp inverter it is simply a matter of connecting the lamp load and the positive lead of a bench power supply to the cathode of D751, the negative being connected to the ground plane. It seems that 10volts DC is plenty. See the recorded test results.


Block Diagram

The attached block diagram illustrates the control method used in this monitor. There are a number of different methods in other monitors. This inverter is supplied by a variable input voltage (Pulse width modulated (PWM)) to change light intensity, very simple principle.


The Inverter


The inverter uses some very special bipolar power transistors, they have a gain so high that they approach a darlington gain device (min about 200 @ 1amp). The Ft (gain x frequency) is also high at 330Mhz typical. It is very hard to find an equivalent. A TIP41 will burn up trying to do the same job. Use the correct transistor for replacement.


The inverter is a simple push-pull free running oscillator with current limit protection.
The base drive current is relatively low and limited by a 1.5K resistors. This takes advantage of the high gain transistors. See diagram mark up.

• The static limiter of lamp current is a 27pf 3KV cap in series with the lamp.
• A positive DC feedback voltage across a 2.2uF cap monitors lamp current, this modifies the PWM at the input to the oscillator.

The base feedback winding produces a voltage low enough to ensure that the reverse emitter base voltage is not exceeded. (A higher supply voltage would cause a failure here).


Testing before assembly


It is a good idea to test the inverter before re-assembling the monitor. This requires a 5 -10V DC supply capable of 1.5 A at 10V and 1 amp at 5V. A 5 volt supply can be taken from a hard drive connector on a PC.


Why did the 2SC5707 fail? No reason has been found yet. Perhaps a transformer shorted turn, yet all the working transformers test the same under static low voltage conditions. Where would you find a replacement transformer???


The feedback circuitry tests OK. Perhaps some transistors do fail after a lengthy period!!!


This will only be finally tested with new transistors.

[davmax]

The Circuit diagram is attached. Ran into size and time limits so it got seperated.

[davmax]

To avoid a misunderstanding the monitor has two CCFL inverters. The operational inverter was tested, the other awaits replacement transistors.

[starfury1]

Humm well it seems that this transistor failure is common and I really think its a symptom of a more hidden fault not the root cause
(if this be by some over sight in design or more likely a failing component)

I did read a post but unfortunately I don't think it was here and stuffed if I know were the link is.

It touched very much on the fact that transformers were going shorted turns.

Any repairs would be doomed to failure because of this, was one persons opinion...other were saying change caps transistors, which did needed to be done, I think he was "ok thats fine but you need to change the transformers as well or it will just fail again further down the track.

There was a source listed but I think its now gone, they were on ebay USA
I cant recall what brand model of inverter it was about either
I'll see if I can find the post.

Not sure, but maybe Bobs "shorted turns LOPT tester" may be of use here.
(it does mention "switching power supply transformers" how close or far that is from an inverter transformer I don't know but suspect you could)

Anyway just my thoughts on it...and not the guru.

HTH cheers

[Harvey]

If its the same as the Dell E172Fpb then the underlying cause is heat, as the Dell's give off enough heat to fry an egg on even without the back fitted.

I've never been keen on TFTs having both PSU and Inverter in the case for just that reason - the heat just builds up with nowhere to go.

[starfury1]

Thanks Harvey and yeah I see what you mean both on the one PCB...I do remember mention of distortion and something...I think it was in relation to the Dells about being riddled with dry joints too

[davmax]

I have not been aware of heat on inverter transformers. I would think that failure in the high voltage output winding is due to poor insulation quality. These devices should have vacuum impregnated insulation treatment, the manufacturer dos not appear to have used this process.

It is hard to find transformer replacements but after much searching here is the place to get them. http://www.lcdinverter.co.uk/coils.htm

[starfury1]

not an answer davmax's orginal post in anyway
but since I was searching thought id throw these links up

there a couple of pdf's that might be worth a read to those that want to know a bit more on CCFL inverters

This is a pdf design ic specific, does offer an insight into the inverter

http://www.microsemi.com/micnotes/1401.pdf

Another app note pdf

http://power.tdk.com/dcac/applicatio...on_Note_97.PDF


http://www.maxim-ic.com/appnotes.cfm/an_pk/3375

some listing at manufactures

http://www.lishin.com/english/01_pro...=1&ID=1#result

http://www.delta.com.tw/product/cp/i...erter_main.asp

http://www.delta.com.tw/product/cp/i...?pcid=9&ptid=1

http://www.topmagnetics.com/index7.htm

Anyway hope its of help to someone out there

Cheers

[davmax]

Well the Monitor is working again. Before replacing the transistors I thoroughly checked for a cause of failure. None could be found. I even removed the inverter transformers and made up a very simple tester that delivered a 20uS pulse to the transformer primary with pulse voltage up to 15 volts to place voltage stress on the high voltage secondary. All transformers were the same in terms of the magnitude and duration of the result ringing waveform. So not a transformer problem. Satisfied with all checks the transistors were replaced and all works. Must have been a weak transistor. I will post the tester circuit later, it uses just one 555 integrated circuit.
I also discovered that inverter transformers are in stock in Australia. Detail: Thomas Electronics 1300 303751
Price $40.70 for two including post. P/N 19.26038.001

A blown transistor also takes out the tiny 3 Amp PCB fuse. Replacement is available from RS Components.

[starfury1]

glad to hear Davmax and even better to know there is a local source
So thank you for your research feedback on it.

Wish I could have been of more help to you on this.

Cheers

[davmax]

Really good links Starfury1. The second relates very closely to the Benq monitor circuitry.

I have included the diagram of my little transformer tester. It is only usable by those who have a bench power supply and an oscilloscope. The test voltage is applied to one half of the transformer primary winding.

Close examination of the Benq circuit borad has left me with the impression that it was designed by an inexperienced student. The board was made for many unnecessary components, most of which were not installed, however some pointless components remained.
1. The 27pf capacitor in series with each CCFL lamp is a high impedance that creates very effective current limiting. At max brilliance lamp current being about 8mA and a short circuit current of 14 mA max.
2. From 1. above there seems no reason to have lamp current control, only voltage control is required. The PCB implements current control on one lamp in each pair of lamps. This is sort of OK.
3. The circuit diagram details short circuit protection for each lamp. This is unnecessary because of the 27pf cap. Inspection of the circuit design reveals it could not possibly work. Looking at the PCB only two lamps are supposedly protected by this circuitry. What about the other two lamps even if the circuit could work!!!!
4. The switching transitors can blow because there is no protection at this point. If there were protection it would protect the transistors against a shorted turn transformer or a lamp short.

Not impressive. Why would the bean counters care? The product must go out the door!!! (It works does'nt it???)

[gonzo0815]

I think this happens from time to time, when the new engineers must be productive real soon.
With some supervision from an experienced engineer this would certainly not happen.


But for the current control of the tubes, i think it is necessary, as those ccfl tubes are IMHO considered as an short after they ignited.
I am not quite sure, if the cap current could sufficiently be regulated by the voltage alone, a variable frequency could at least do this.

[davmax]

The cap alone even restricts short circuit current so it is a good current limit/control. Varying the supply voltage to the inverter is enough and improved by changing th effective voltage and current with PWM.

[davmax]

The final fix. After testing components and replacing two inverter transistors I set about implementing an overload protection circuit for the inverter transistors seeing it was non existent. The attached schematic shows 9 added components that replace 31 components to the right of the double crosses. Each of the twin inverters is current sensed with a 0.11ohm resistor and the emitter/base junction of a BC640. When the current exceeds about 5 amps the BC640 charges up the 47uF cap at the SCP (short circuit protection) input of the controller and the time taken to trip the controller OFF is about 100uS. This trip current and speed will keep the transistors within their safe operating area. Once tripped the monitor or the video signal must be removed to reset.

The result means that the inverter transistors are protected against any over current eg transformer shorted turn. Each transistor can maintain 8 Amps DC when switched ON and a pulse of 11 amps.

The original Benq design was intended to detect only lamp shorts and it could not work anyhow. For fewer components a better protection has been implemented.

[davmax]

With further testing the current sense resistor can be rduced to 0.1 ohm and collector resistor of the BC640 to 27 ohms. This decreases the time to trip to about 50uS and increases the safety margin of the current sense under all conditions. The margin is close with a 0.11 resistor and at full screen brilliance.

Certainly my last post was a good indication of the original Benq design with excessive components and no real short circuit protection.