Old Chemi-Con KMF went out with a bang

Chemi-Con KMF went poof

Poof.

This capacitor was situated in a power supply under my bed, it was a rough awakening to say the least!

Advertisements

HP L1906 power supply failure

HP L1906 PCB

HP L1906, manufactured in February, 2007, model no. HSTND-2L09

Symptoms unknown; screen was not powered on before disassembly.

Electrolytic capacitors C822, C823 and C824 (1000 µF/10 V, bottom center of image) had failed. They are situated on a 5 V rail and can be substituted by 6,3 V units. They were successfully replaced by various used low-ESR capacitors. C824 and C822 are connected in parallel and can be replaced by one 2200 µF component.  The 35 V capacitors on the inverter power rail were in good condition and as such were not replaced.

IBM L170 power supply failure

IBM L170 power supply

IBM L170, manufactured in May 2004, model no. 6734-AB1

The screen would turn on fine, but when a signal was applied, the backlight would turn off.

Capacitors C201, C223, C922, C924, C925, C926 had failed. Capacitors C924, C925 and C926 are situated on the 5V rail and were successfully replaced by used 6,3V Chemi-Con KZG and Nichicon HM parts of equivalent capacitance. C201, C223 and C922 are situated on a well-regulated 12V rail. C201 and C223 were collectively replaced by one used 1500µF/16V Chemi-Con KZG (they’re directly in parallel with about 1cm of trace in between). C922 was replaced by another such.

IBM L170 power supply after repairs.

I should end with noting that these are some of my favourite consumer LCD screens ever made. Aside from the the power supply, the build quality is staggering. They weigh six kilograms and ship with a proper, extremely high-quality 100×100 VESA stand that you can bolt onto any monitor you wish, for an instant improvement in stability. They’re screwed together rather than snap-locked, and the serviceability is just staggeringly good all-around.

Acer V193 power supply failure

Acer V193 power supply

Acer V193, manufactured in March 2009, model no. ET.CV3RE.A08

The screen would flash its power LED once and then give up any signs of life until the power cord is pulled and re-inserted.

The 1000µF/10V capacitor C820 had failed. It was replaced by a used Chemi-Con KZG 1500µF/16V. Raising the voltage rating of the capacitor is not necessary, I just used the part I had in stock. All other capacitors measured OK, and were not replaced.

Acer AL1717 power supply failure

Acer AL1717 power supply

Acer AL1717, manufactured in November 2005, model no. ET.1717P.014

The 17″ Acer list on this blog is nearing completion, with the AL1711, AL1716 and now AL1717 down!

The monitor would squeal and flash its power LED, occasionally flashing the backlight for a second or two.

All the larger electrolytic capacitors on the secondary side of the power supply tested bad. The big 2200µF/10V capacitor sits on a well-regulated 5V rail, and was replaced by a used 2200µF/6,3V Rubycon MBZ. The 1000µF/25V capacitors were replaced by various 1000µF/16V units. However, since the 12V rail they’re situated on seems to be poorly regulated and lying at almost 15V unloaded, replacing them with 25V caps is advisable. If the 220µF capacitor above the heatsink is not replaced, the screen will run noisily. Replacing it by a 470µF or 1000µF part will considerably lower the noise emitted by the screen still.

Samsung Syncmaster 206BW power supply failure

Samsung Syncmaster 206BW power supply

The screen would squeal and flash its power LED, showing no other signs of life. The two 1000µF/25V capacitors (C263 and its mate) had failed. They sit on the 13V rail. They were successfully replaced by two used 1500µF/16V Nichicon KZG. All other capacitors tested OK, and were not replaced.

The screen also suffers from a loud buzzing noise when the brightness is turned down. I have not been able to identify the fault, but I believe it to be a fault in the lower CCFL of the panel itself.

Worth noting is that the screen had only logged a run-time of 3000 hours in the service menu before the power supply failed . . .

Fixing the LG PDP42V5 “coloured dots” problem

While unrelated to this problem, replacing C18 and C77 on the Z-SUS board is good practise. They measure 0,5-1,5 Ohm of ESR on all my boards. Of course, that won't remedy an exploded IPM . . . 😉

The short version:

  • Replace all 33 µF SMD electrolytic capacitors on the Y-SUS board. Start with C72.
  • Replace C21, C24 and C72 with bigger capacitors, 100-330 µF.
  • Adjust potentiometer VR1, labelled Setup, until the dots disappear.
  • Make sure to fine-tune VR1, as over-compensating will cause dark dots to appear instead.
  • Adjust the VS voltage on the power supply. Lower VS reduces the number of dots but increases overall picture noise. 180V seems to work well.
  • Adjust brightness and contrast in the menu to compensate for any changes in panel behaviour.
  • VR2, Setdn, seems to make very little difference unless turned extremely, in which case it causes serious image artefacts. It could be worth adjusting if adjusting VR1 fails to improve on the problem.
  • The through-hole electrolytics are of poor quality, check and replace if necessary. In particular C71 and C77.
  • Component locations here.

The long version:

Everybody who has dealt with televisions based upon the LG PDP42V5 and PDP42V6 plasma TV platforms will probably know about the problem of random coloured dots appearing in dark areas of the screen when the picture is moving. It is an extremely common problem that seems to affect most (if not all) 42V5 and 42V6-based plasma screens. Despite it being such a common issue, little information regarding it is available on the Internet.

There is an LG service/assembly manual floating around on the net (link). It is fairly useful in diagnosing these screens, but it’s written in shoddy English and it addresses the “coloured dots” problem (plasma mal-discharge) by replacing the Y-SUS and Z-SUS boards in complete. I felt this solution was unsatisfactory, as new boards often are hard to come by, expensive, or both. The manual does, however, hint at adjusting the “Setup” potentiometer to adjust plasma discharge, an adjustment that can make all the difference. More on that further down.

Before touching the boards past re-soldering and replacing random electrolytic capacitors, I decided to check the Internet for hints. I stumbled upon this page, which addresses the issues on both 42V5 and 42V6 screens, but fails to acknowledge the similarities between the two; setting the VS voltage down also works for 42V5-based systems, for instance.

During the experiments, I used parts from three different televisions, all PDP42V5-based.

Televisions used:

One Andersson X1, which had the “coloured dots” problem very, very bad before the panel failed due to hitting the floor at speed.

One LG MZ-42PZ44, which I received with a failed Z-SUS board. Its Y-SUS board proved to have the “coloured dots” problem upon replacing the Z-SUS board, although not as bad as the Andersson’s Y-SUS board. The original Y-SUS board in the LG is of a newer revision than the ones found in the other televisions.

One Grundig Tharus 110 PW 110-9410 TOP, which I received with a failed power supply. It did not have any faults beyond the power supply failure. It did, however, immediately develop the “coloured dots” problem once I installed the Andersson X1’s Y-SUS board into it. Upon re-installing the original board, the problem disappeared.

The test rig:

I used the confirmed bad Y-SUS board from the Andersson, installed in the LG television. Upon making the adjustments, the problem was almost completely remedied. I have not made any changes whatsoever to the original Y-SUS board from the LG.

I’ll start by posting some before and after pictures of the TV that I experimented on.

TV off

TV on with original Y-SUS board. Results were similar with the un-adjusted Andersson board.

TV on with the adjusted Andersson Y-SUS board. Aside from the problem disappearing, the contrast ratio improved immensely.

Showing a picture of wood with the original board.

Showing the same picture with the adjusted board.

So, those are the results. Onto the fix.

Adjustments made to the Y-SUS board

The Andersson Y-SUS board. Red arrows mark 33µF capacitors. Blue arrows mark the Setup/Setdn potentiometers.

As you can see, all of the 33 µF/25 V, surface-mounted capacitors are arrowed. When I measured the capacitance/ESR of them, not a single one was above 25µF. I incrementally by-passed them with 10 µF and 22 µF electrolytics, and every time I added a capacitor, the problem became less severe. However, one made a considerably bigger difference than the others, and that was C72. It’s the one below the transformer, to the left.

After replacing all the 22 µF capacitors, the problem was not completely gone, however. While considerably better than before (the TV went from “unwatchable” to “tolerable but poor”), I tried adjusting VR1. Turning it clockwise would produce a darker black level and less coloured dots. Counter-clockwise did the opposite and introduced black dots.

After setting VR1 to a position where some (very few) coloured dots were present, I set VS on the power supply down to a lower value. The problem can more or less be remedied by only adjusting VS to a very low value (170-175V), but the results are poor in comparison to also repairing the Y-SUS board; the lower VS is, the noisier the picture is. Contrast ratio seems to also be negatively affected by setting VS down. Chances are also that the problem will return when C72 deteriorates further and the panel gets more worn.

I found that a VS of about 180-183V to be optimal in this TV, but it probably varies from case to case.

Conclusion/notes/misc:

Upon by-passing the capacitors as in the picture, adjusting VR1, and setting the VS voltage down to around 180V, the MZ-42PZ44 shows no signs of ever being broken. The adjustments did, however, make the image somewhat darker, and I had to turn the brightness up a tad in the menu to compensate for it. It is a very minor complaint, however; the picture is still very bright.

I have tested the TV for a few hours, and have not spotted any signs of the issues returning, nor have I spotted any other abnormal behaviour.

I also found that the big 680 µF/250 V capacitors on one of the Y-SUS boards had very high ESR compared to the others. They may be worth checking. Note that they’re connected in parallel.

To the left is a measurement of a fresh cap pair, to the right, a failed pair.

I have applied the VR1/C72/VS-adjust fix to a PDP42V6-based TV as well, with good results. Replacing capacitors past C72 did not make any significant improvements to it. I might make a follow-up article covering PDP42V6-based televisions soon. The procedure is identical on them.