HKC Monitor Type 2615

 My 26 inch monitor which has been working for many years without a problem suddenly failed. I wondered when I'd bought it so rummaged around and discovered it was in October 2014 so it's obviously out of warranty. It started life as my main dsplay, then after many years, I replaced it with a new 4K monitor and relegated it to show the picture from my camera in a dual-display set-up.

 A few days ago in September 2024 the picture had suddenly gone so dark it was difficult to see anything. In other words, the picture was actually present as a vague shadow but unlit. The HKC 2615 uses LEDs to illuminate the screen and these had obviously extinguished. I unscrewed the monitor from the dual-screen mount and investigated the problem by first removing the rear cover and scrutinising the parts fitted to the metal panel on the back of the screen. There was a pair of interconnected circuit boards plus a cable to a control panel carrying a green LED and a two others plus a thin cable to the screen LEDs and a cable to the top of the flat-screen panel.

Nothing looked amiss so I detached the board carrying the power supply. I'd discounted the second board responsible for driving the PC as faulty beause the picture was actually present. I also reckoned the power for that board must be present so I reckoned that the LED part of the monitor had failed.

The fault could be one of several things.. first the power supply to the LED circuitry, second the LED driver chip or its components, thirdly the boost circuit which provides the enhanced voltage to drive the LEDs, and lastly the LEDs themselves.

Below the blameless video board

 

 

And now the suspect board showing the key parts for the LED supply voltage... 

 I checked the various capacitors and diodes etc in the area of the power supply responsible for generating the 12 volts supplied to the LED area and everything checked out OK. Surprisingly all the capacitors measured close to their marked values with very low ESR values even though some had been heated from adjacent diodes. The rectifier diodes for the LED circuitry had been very hot as you can see from the scorch marks on the plastic cover above. The diodes looked like 3amp types and wired in pairs (not particularly good practice unless they're well matched).

Powering up the system from a mains supply proved the main LED circuit supply voltage was sitting at 12 volts and the circuitry appeared to be in good order. The green on/off LED was flashing but plugging in a PC caused the LED to stay on. As this is the first time I've tackled this type of repair it was a case of learning about the design used by the HKC people. Previously I've only repaired older modern flat screen monitors using fluorescent tubes for backlighting.

The method used in this HKC 2615 monitor is to drive the LEDs from a special chip coupled to a voltage boost circuit which provides just the right amount of power to set the correct screen brightness.

The next step was to identify the LED driver chip and the MOSFET used for boosting the LED voltage and figure out why the LEDs weren't lit. As usual chip identification was not straightforward because their markings are truncated.

After a lot of investigation I found the LED driver was an OB3362 and the MOSFET (mounted on the underside of the board) an AOD4454. Neither of these devices is particularly common but I tried a vaguely similar MOSFET to no avail. This meant that the LED driver or its associated components could be bad. In theory I suppose there's a slim chance the LED circuit itself is bad. Because, for example, if all are out, it may be an open common feed. As checking this would have meant more disassembly, and anyway I'd discounted this as most unlikely, so had decided to follow a simpler course. It might be possible of course that the LED driver is so clever it shuts down all its four outputs if one of these sees a bad circuit. Certainly possible but I decided to first swap the LED driver after measuring all the parts to which it's connected.

Click the picture below to read the full spec

 

 Studying the OB3362 documentation showed that the HKC engineers had basically followed the circuit in the spec (shown above) so I was able to confirm the parts were most likely in good order. With everything powered up I confirmed the voltage at the on/off pin on the chip (PIN 1) reflected the state of the green on/off LED and the voltage at the brightness pin (PIN 2) suggested the dark LEDs should have been lit. Looking at the MOSFET gate pin with a voltmeter suggested it was dead as my meter showed zero volts AC or DC. The boost circuit wasn't operating and the implication was a bad OB3362 so I ordered a replacement together with a new AOD4454 in case excessive heat in the removal process had degraded it.

I should mention that I'd carried out a simple test on the LEDs by connecting a bench power supply to their connector. Connecting a 12 volt supply had resulted a zero current to all four LED circuits although I hadn't at that time wanted to try a higher voltage Maybe I should have done this, but I'd already ordered the new chips and tidied up by the time I thought of doing this.

 As I'd imagined the new chips failed to fix the lack of backlights so I opened up the screen and inspected the LEDs. These were all fitted to a thin strip glued to the top of the metal screen surround. I counted a total of 88 LEDs and after checking for continuity found these were connected in four chains of 22. My bench supply of 30 volts max failed to turn on a chain but, after setting the PSU to about 3.5 volts, all 88 proved to be good and also each chain connected to the six pins on the connector.

What next? Well I'd not actually tested each of the capacitors out of circuit having relied on in-circuit results. Of course all tested OK although a little low in values, so time to do some more detailed testing. I found a better version of the spec for the OB3362 and read up about the operation. There are several features explained which were not apparent before. In particular there are some fault detection features that will disable the current to the LEDs. Also its possible that either a resistor or capacitor in the circuit has failed or drifted away from the allowed limits imposed by the chip.

I noted that a bad LED chain will not switch off the other three if it goes short-circuit or open circuit.

The turn-off feature will activate if the OVP voltage is less than 75mV; the ISET voltage at Pin 5 will determine the LED current; a voltage of greater than 5 volts at and LED drive pin will turn off that chain. I turned on the mains power to the boards and monitored the boost voltage. This started at 12 volts and for a few moments rose slightly then tripped which would seem to indicate a turn-off condition. It's also possible that one of the surface-mount resistors or capacitors is bad. All the resistors measured around their marked values but the capacitors will need to be removed to confirm whether their values are correct for the OB3362. Below is the suggested circuit taken from the chip suppliers spec. In the mode used by my monitor the BF_DEMOD pin is grounded. Resistor R6 is actually a 10K plus a 4.7K in series with the latter decoupled to ground and the voltage is 175mV which should be OK.

 

 I decided to dispense with the mains supply and instead power the LED circuit from an external 12 volt power supply. I then fitted a few resistors to set the enable and brightness pins to voltages that should turn on the boost circuit. I also plugged in the LED chains. Turning on the 12 volt supply proved there was no difference with the fault still present. The four LED drive voltages are without boost and all measured about 12 volts. I tested the original MOSFET and it was OK so that particular component wasn't the problem.

 

 

 This is a picture showing a section of the 88 LEDs mounted on a thin circuit board glued to the metal frame.

 The HKC LED currents are defined by a pair of 120k resistors in parallel (=60k) rather than the 20k in the above circuit. R3 is 430k as shown. I suppose a critical part I haven't checked is the choke L1 but surely that wouldn't prevent the boost circuit from trying to pump up the voltage... The choke was of course OK so I pondered over the AliExpess parts that had arrived the day before. Could the chips be faulty? I couldn't think of anything else to swap so replaced the MOSFET with the original which had tested OK on my meter. Still no change in the boost voltage so I swapped the new OB chip with another (there were ten in the packet) and switched on the external PSU 12 volt supply and this time the boost voltage rose from 12 volts to 26 volts (an improvement). The LEDs remained out so I swapped to the mains supply and plugged in the video board and switch panel. I turned on the mains supply and the boost voltage rose (again) to 26 volts (with no LEDs lit) before dropping back to 12 volts (ie. no video input detected). Clearly the circuit has a problem because 26 volts is too low for the LED chains. Could the second OB chip also have a fault? I reckon a 16-pin DIL socket might be the best bet as I can try yet another OB chip. I remember many years ago when "red spot" transistors were sold... rumour had it that these were "floor sweepings" or parts that hadn't met their spec but too good to chuck away so sold to "experimenters". A clue might be the oscillation frequency which measured about 100KHz.

What level should the boost voltage be?

Assume the LEDs need a typical forward voltage of say 3 volts then 11 x 3 volts = 33 volts.

I removed the second OB chip and fitted a 16-pin DIL socket so I can try other chips. I did this and tried two more OB chips with the same result... the boost voltage rose to a little under 26 volts before turning off. Admittedly I didn't have a video input, but I'd have expected the monitor to come on with "no input" or similar before turning off. Could the MOSFETs be faulty.. I'm using the original after trying one of the new ones?

At this point I decided to test the LED strips. My bench supply was limited to 30 volts so I connected two in series, connected the common to plus and the first bank to the negative supply and gradually increased the voltage with a 40mA limit. The LEDs started to illuminate at 55 volts but only at 61 volts did they all come on brightly and drew 10mA. All four banks lit perfectly. So each LED requires 61 divided by 11 or 5.5 volts for decent brightness. Clearly the boost voltage is way short at 26 volts.

I decided to fit another MOSFET but after checking both the original and its replacement I discovered a parameter I hadn't noticed. The input capacitance for the old MOSFET was 800pF on my test meter but that of the new one was close to 2nF. Could the extra capacitance be damping the pulses from the OB chip? But because the old MOSFET was in place and the boost voltage was only 26 volts that idea was discarded (also because none of my junk box MOSFETs had a low enough gate capacitance to match that of the AOD4454).

But I'd already removed the original ready for a new one so decided to just refit it and continue with my experiments so I first cleaned the area of old flux residue and excess solder. It was then I noticed a short length of shiny copper close to R10. In my years of repairing stuff I'm aware that if a copper track fuses it will intially burn off its solder resist leaving a tell-tale shiny surface, so I peered at the area and noticed the tiniest sign that there was a break in the copper track. The break had been masked by a length of channel ground into the surface of the board presumably to prevent leakage of solder from under the surface mounted MOSFET. You can almost distinguish this in the picture below. The break had occurred at the weakest (ie. hottest) point on the track and was completely covered by the gate pin and was clearly weaker than that zero ohm resistor (R7).

 

 

  I fitted the original MOSFET after testing it (yet) again and added a short length of wire between R7 and the top of the set of paralleled current monitoring resistors. I fitted the board, connected it up, placed my meter probes at the cathode of the boost rectifier and ground then plugged in the mains supply. The boost voltage went from 12.5 volts to 70 volts and all the LEDs lit up.

 

 

 

Now I have to reassemble the various screen elements (removed to see the LEDs) and put the monitor back together. The original fault will have to remain a bit of a mystery. Possibly Pin 12 of the OB chip developed an internal short to a positive voltage and the current fused the thin track. Once the connection to Pin 12 had opened, the new replacement OB chip would have seen a hard "rest" voltage (maybe circa 10 volts) instead of a zero to 270mV from the source pin of the MOSFET and shut down the boost supply (or not allowed it to get beyond 26 volts).

 
 
 

 Reassembling the monitor's screen was a bit tricky but I realised eventually the panels needed to be fitted into the metal surround in a specific order otherwise the flexi-circuit board would get jammed. Once the panels had been fitted I replaced all the screws. Some are machine screws with a thread and others self-tapping. The latter are used for securing into plastic and its important to fit the short screws in the correct tapped holes to prevent damage.

 

 Above.. The securing clips for the push-button panel are very weak and one broke off. Fortunately a design change had left two mounting posts so the panel can be held in place as shown.

 

This picture shows the connector for driving the monitor screen via a very long flexi-circuit.

 

Having assembled everything I added the monitor to my PCs dual monitor configuration and booted up. The correct sequence of display information appeared and I was able to see a second desktop, however just as I was dragging the picture from my camera across to the HKC screen it went black.

What had happened? There was no red LED showing so maybe a fuse had blown? I tested the mains lead and 245 volts was present so this left me no alternative but to detach the monitor from the stand and investigate the power supply board for the nth time.

I removed the rear cover, the metal shield, a couple of cables and unscrewed the PSU.

A quick check revealed a puzzle. A resistance of about 80 ohms was present across one of the diode pairs, in fact those responsible for providing the 12 volts to the LED drive circuitry. Now, that resistance could be due to a bad component, or maybe it was present to provide a minimum load for the 12 volt supply? There are lots of suspect parts and if I'd got a thermal camera (*** see below) it would be a straightforward way to see the culprit as 12 volts and 80 ohms represents a dissipation of nearly 2 watts. Rather than remove each suspect component in turn I decided to connect my bench PSU across the 12 volt supply. The crowbar kicked in but the MOSFET got very hot so that's the problem. I removed it and found about 180 ohms across gate and source and that 80 ohms puzzle resistance across drain to source.

Why had the MOSFET failed? I'd unsoldered it a few times and refitted it so lots of heating was involved, but it had actually worked for about 5 minutes. That fused track might be the clue. Some semiconductor devices use a parallel arrangement of smaller chips to provide a high power performance and it's possible the track fusing not only cut off the MOSFET operation via the OB chip, but open-circuited a section of the MOSFET output transistor array (if indeed it has such a thing). Once this had effectively been detached the MOSFET rating would be much reduced. Running it as a 20Amp device with a 94milli-ohm Source/Drain rating when it may have had say a 10Amp 200milli-ohm rating may have pushed its chip temperature too high for the pretty crude heatsink fitted to the board?

Anyway, with a fair degree of confidence I checked another from my bag of ten from AliExpress and, seeing it was good, soldered it in place. I should point out here that, although these MOSFETs carry the correct code, their gate-source capacitance is double that of the true device spec at something between 1.6 and 2nF instead of circa 800pF (0.8nF).

I didn't bother testing further but just refitted the PSU and screwed on the rear cover. It's now working fine in my dual set-up, running at 1920x1080.

Cross fingers it's not going to let me down again...
 

 

*** I ordered a thermal camera a few weeks ago from AliExpress but it seems to have gone missing shortly after it left UK Customs and probably placed in the care of Royal Mail. Was it stolen or just "lost in transit"? It seems I'll be in Limbo for 40 days, after which I can get a replacement.

 Return to Reception