Fix for Benq FP737s 17" (Q7T3)

[davmax]

Re: Benq 17" Q7T3 fix

I have been waiting for parts to complete the short circuit protection of this monitor by causing it to trip in 10uS or less.

Whilst waiting I happened to notice another circuit flaw (see previous posts for first flaw). Again reinforcing the previous view that a student designed the electronics.

Benq Q7T3 (FP737s) LCD monitor design flaw.



MOSFET drive circuit.



Refer to the attached diagram.



Inspection of the circuit shows that there is no current control through transistors Q753 & Q752. They are directly connected between the +15 volt rail and ground there is nothing to limit the current between the supply rails (only the ill defined current gain of Q753). With Q752 saturated ON drive to Q753 is about 4 mA, a current gain of 50 makes about 200mA Ic (gain of 100 = 400mA) and the voltage across this transistor is about 13 volts (between 260 and 520mW x duty cycle).



The same current flows through Q752 that has an absolute max current of 150mA.

Additionally the absolute max current for the Ic drive is 21mA.



Clearly these components have survived but it is a poor design without a current limiting resistor between supply rails.



This problem has been seen before where there is a focus on switching speed more than other parameters. Not a good practice for reliable design.



The design should guarantee absolute maximums are not exceeded.



A 100 ohm resistor inserted at the emitter of Q752 will ensure that 150mA is not exceeded and likewise the 21mA IC load is not exceeded.



How does this mod affect performance? Increasing the MOSFET gate source resistance will increase the turn ON time by about 150nS when switching 7.2amps (based on data sheet). However the circuit max operating current is 1 amp so the increase in turn on speed and switching loss is much less significant. The switch OFF speed will not be affected because the 100 ohm is not part of the switch off circuit.



To conclude the insertion of 100 ohm will not impact performance significantly but will make sure current through devices and across the supply rail is controlled.

[davmax]

Re: Benq 17" Q7T3 fix

The last post is not correct. The 100 ohm resistor is in the wrong place causing a prospect of emitter/base reverse voltage being exceeded on Q753. I am still concerned about uncontrolled current paths through Q752. I will come up with something better shortly.

[davmax]

Re: Benq 17" Q7T3 fix

Well I worked carefully through the MOSFET drive circuit, documenting as I went to make sure that I did not slip up. The result is that I have discovered some further interest things about this circuit. Not the least that the actual PCB has one important resistor value changed from that given on the schematic. I have implemented two simple changes added one resistor and changed the value of another. The changes bring certainty into the design performance.

If any body is interested in the documented detail it is attached.

[davmax]

Re: Benq 17" Q7T3 fix

I have now completed setting up and testing the current trip protection. Somewhat optimised now. Circuit components and values have been changed so that the trip time may be <10uS ie within one inverter half cycle, a recommended design timing.

If anybody is interested the attached document covers the concept, the design example in steps and the modified schematic.

[davmax]

Re: Benq 17" Q7T3 fix

There was a fair bit of exploration and testing to reach final solutions. It seems helpful to now create a reference to the significant information in this thread.
1. 09/11 Source for 2SC5707 transistos (eBay)
2. 09/12 Warning against use of TIP41 transistors
3. 09/16 An inverter transformer source.
4. 09/19 Aussie transformer source and a transformer tester
5. 09/28 Detail covering the fix of a MOSFET driver flaw
6. 10/01 Final design detail of a fast over current protection to replace the original ineffective circuit.

[dan8139]

Re: Benq 17" Q7T3 fix

Good man, although I don't have one of these monitors and don't know anyone who does, you really do deserve credit for your analysis & thanks! Too much stuff these days is designed to be shite!

Dan

[davmax]

Re: Benq 17" Q7T3 fix

Discovered I had left one component out of the modification diagram ie a 2.7K resistor. Revised diagram attached

[davmax]

Re: Benq 17" Q7T3 fix

Some have asked for PCB photos. Well I have just fixed and modified another monitor and taken photos.

This monitor was received with a damaged ribbon connector. This is usually because the release mechanism is not understood. To release the ribbon cable lift /rotate the black locking bar outward from the centre of the top edge. Do not force.

[davmax]

Re: Benq 17" Q7T3 fix

Having now repaired and modified two Q7T3 monitors I can say with reasonable certainty that the failure of the 2SC5707 transistors is due to a weakness in the transistors. Testing of inverter transformers has found no shorted turns and no other cause of failure.

[davmax]

Re: Benq 17" Q7T3 fix

An interesting observation. Under certain boot conditions the monitor is caused to pull excess current in the inverter circuits and my new protection shutsdown the backlighting. I found that by waiting for initial boot to complete then switching the monitor off and then on again all will reset to normal. Great proof that the new current sense trip is protecting tightly.

[davmax]

Re: Benq 17" Q7T3 fix

To clarify my last post. The only time the monitor draws the excess current is when I restart the system and just before re-boot I switch on the power to my external SATAII drive. This seems to throw the bios into some momentary confusion that affects the monitor. I can only assume that the timings are such as to mess up the slow start of the backlight converters and excessive currents occur. This may be a clue as to how the transistors have previously become stressed and eventually fail. Hopefully they will not fail now due to the very fast protection.

This problem does not occur if the system is shutdown,the drive power switch set to ON and then the system started. Only in the restart procedure.

[gonzo0815]

Re: Benq 17" Q7T3 fix

Very interesting find, may be there are other scenarios, where this could happen.
May be another design fault in this circuit, or the OCP was altered to fix the original fault (assuming that a customer would RMA a new TFT who would go into protection mode from time to time).

[davmax]

Re: Benq 17" Q7T3 fix

Yes Gonzo.

The backlight inverters are switched ON & OFF by the BLT_ON signal that is ON when video is present. The BLT inverters have a soft start that is simply controlled by a charging cap. Now if the video turns OFF and then comes back very quickly the soft start capacitor has not reset and therefore soft start is corrupted. I suspect this corruption causes a transformer saturation and therefore excess current. This I believe is what is happening when the bios gets upset as previously described. There could be other possibilities.

I suspect that other monitors have the same problem and that is why the 2SC5707 transistors pop off without previous explanation.

I am going to insert a circuit with a one second blocking action that ensures the BLT inverters cannot be restarted in that time. If the problem goes away it will prove my assessment. Tomorrow I will install it.

I am very thankful that my high speed overcurrent trip circuit allows me to protect the transistors and additionally enable detection of a problem that may explain the transistor failure mystery.

[gonzo0815]

Re: Benq 17" Q7T3 fix

I think the simplest fix would be to put a bleeding resistor parallel with the soft start cap, if the circuit design allows a steady leakage current by a pull down resistor. An simple transistor which discharges the cap at PS_on =low should do the same.

An yes, i think this kind of root cause analysis you are performing is not seen very often and i appreciate this kind of information, even if i haven´t got one of this LCD`s on my desk.

[davmax]

Re: Benq 17" Q7T3 fix

Thanks for your inputs Gonzo. I really appreciate your many inputs and comments.
Yes I have considered resistor and transistor at the soft start point. However there are two soft start circuits, one for each inverter. It is much better to block the BKL_ON signal at one point for one second. This is also much more precise, operating in nS thus ensuring no possible weak point. Also it is very simple to adjust the time independently of the soft start.

I bench tested the delay circuit at 5 volts it worked fine. I put it into the monitor (15V supply) and the 4047 chip fused. Investigation found that I had been supplied with a HEF4047 that works up to 6 volts when I expected to get a CD4047 up to 18 volts. The supplier is stocking the wrong version. So will try again. Another trap!!!!

The 4047 chip drives a transistor that clamps down the BKL_ON signal. There is no changing components on the Benq board. I use a very small vero board with four leads. +15V, ground, input taken from a board transistor output and output that clamps the signal. As the board transistor turns OFF the input to the same transistor is clamped to ground for 1 sec, this achieves the required result.

[Krankshaft]

Re: Benq 17" Q7T3 fix

Clearly these components have survived but it is a poor design without a current limiting resistor between supply rails.

It could also be that the bean counters removed the limiting resistor to cut costs.

The way manufacurers see it if the circuit runs without it eliminate it.

Just look at the cheap china-mart power supplies without EMI supressors the supplies run fine without them so scrap em. They may emit high frequency noise but who cares.

If it bites the dust in a few years good time for another monitor.

[Krankshaft]

Re: Benq 17" Q7T3 fix

Davmax:

Ironically enough shortly after reading your post I recieved an HP 19" monitor manufacured by Benq.

It had the exact same problem you did it powered up with a green light and shining a flashlight revealed backlight failure.

So once I open it up and see Benq on all the boards I immediately check out the transistors near the inverter transformers (C5707).

And low and behold one of them was shorted no voltage drop with the diode testing function of my DMM 0.00 volts. Ironically enough the other 3 of the 4 transistors tested just fine.

I will replace them all but you say that the cause of their failure is due to an internal flaw. In that case would you be able to recommend another transistor to substitute it?

Also another thing one of the pico fuses was cracked it still had continuity but it was cracked. Its 5 amps buy what voltage should I get?

[davmax]

Re: Benq 17" Q7T3 fix

From my experience the failure is only the 2SC5707 transistor. See my most recent posts in this thread ie I have not found any other component problem , only two design problems.
1. The transistors are not protected from excessive current.
2. Excessive current can occur or rare occasions. See my posts from 6th Nov. There is a time window that can allow corruption of the soft start circuits.

Re fuse. In 17" monitor the fuse is a 3A fast PCB fuse. I use a TE5 fuse (125V) available from RS Components.

Fixing problem No.1 will protect the transistor circuits. It requires some skill. Details including diagram and board photos are posted in this thread. It will protect against problem 2 which I believe is the hidden cause of these transistor failures.

Davmax:

Ironically enough shortly after reading your post I recieved an HP 19" monitor manufacured by Benq.

It had the exact same problem you did it powered up with a green light and shining a flashlight revealed backlight failure.

So once I open it up and see Benq on all the boards I immediately check out the transistors near the inverter transformers (C5707).

And low and behold one of them was shorted no voltage drop with the diode testing function of my DMM 0.00 volts. Ironically enough the other 3 of the 4 transistors tested just fine.

I will replace them all but you say that the cause of their failure is due to an internal flaw. In that case would you be able to recommend another transistor to substitute it?

Also another thing one of the pico fuses was cracked it still had continuity but it was cracked. Its 5 amps buy what voltage should I get?

[davmax]

Re: Benq 17" Q7T3 fix

Krankshaft

See my reply to your other post re transistor and fuse.
Actual the design specifies 30+ components for a current protection that does not work. It is replaced with one that does work having only 9 components. So no bean counter saving just bad design.

I do not know if your 19" monitor has the same circuitry. If you can download a service manual and compare the circuitry with the 17" that I have posted. That may help if you can understand these things.

[davmax]

Re: Benq 17" Q7T3 fix

Well I have completed the modification that overcomes the likely hidden cause of the transistor failures.

It works a treat. There is no evidence of slowstart corruption causing excessive transistor current.

To summarise the monitor problem of blowing transistors.
1. Only transistor failure has been found, no shorted turn transformers. Why?
2. It has been discovered that if the video signal to the monitor switches OFF and then rapidly ON again the two softstart circuits are corrupted and excessive transistor current occurs. This can typically occur at post time (computer start) in fact I can recall my monitor failing right at that point of time.

The best fix is to protect the transistors from excessive current with a 10uS current trip. See design detail in this thread. This protects against all fault conditions and it was this protection that enabled the detection of the cause described in 2. above.

For my better understanding a 1 sec clamping circuit was introduced to stop the video signal returning rapidly and disrupting the softstart circuits ie allowing them time to reset.