Mitsubishi swollen caps

[lcaillo]

Many discussions can be found on the Internet regarding the swollen caps that are sometimes found in the Mitsubishi DM modules. The typical failure is that the set will not complete the boot cycle and the timer light just flashes continuously. This is an indication that the DM microprocessor and the main micro on the Signal board are not communicating to load the data to the micros from the eeproms.

I have fixed a few of these and thought I would share a few things that I have learned. First, it seems clear that Mitsubishi bought a bad batch of caps from Jamicon. All of the swollen caps have been from this manufacturer. All of the DM supply caps have been 1000uF in either 10v or 16v ratings. I have found the same caps with the same failure mode in other modules and in 330uF caps, all Jamicon made.

On the last set that I had in the shop, we decided to change the caps in the DM even though they were not swollen. The light box was in the shop for another problem (bad front panel switches likely due to windex syndrome) and the client is some distance away. We wanted to go through it and catch any potential problems before they developed because of the distance, and on general principle. When we took the caps out I decided to test them. We use a Sencore unit that tests four parameters for caps, capacitance, ESR, dielectric absorption, and d.c. leakage. All tests were well within acceptable ranges except DA. This is probably one of the least understood parameters for capacitor failure and few meters test it. What we have noted is that when we see caps with high DA they tend to fail in very strange symptoms, and sometimes intermittently. Eventually they break down and go leaky in many cases. The Sencore limit for a good reading is 15%, but I have found that caps with readings for DA as low as 10% have created problems. 4 of the 7 caps in this set were between 10% and 15%. This suggests that a preliminary test prior to failure of these swelling caps might be DA. The typical readings for good caps are usually just a few percent. None of the other parameters were anywhere near concern. Previous testing on swollen caps from other sets yielded the same findings. The ones that were not leaky or open, but were swollen all had high DA readings with other parameters OK.

The bottom line is this. Don't count on testing these caps to find those that might be failing. Most testers don't measure DA, and you have to take them out to get a good reading anyway. These double sided boards with large ground surfaces are difficult to work with we think the problem is a bad batch of capacitors anyway. Caps are cheap, just change them. Just don't use Jamicon caps.

[bkalbfeld]

Could one symptom of bad caps be that the TV input levels become too sensitive? I'm running an Panasonic BD30 Blu-Ray player into an Onkyo 605 AV receiver into the Mits. All 1080i component. When there's a very bright flash in the movie, the TV looks overloaded -- picture shrinks or goes away momentarily. It recovers on its own.

I tried putting a 1db pad on Y, Pb and Pr, but then I just lose the picture altogether. Step the output resolution of the Blu-Ray down to 480p, movie plays fine. Run the same rig into a newer TV (Aquos -- still running component), movie plays fine. And no problem running 1080i OTA signal through the Onkyo into the Mits.

Any thoughts on this problem much appreciated.

Thanks!

[lcaillo]

Sounds like either a sync problem or a power supply problem. Have you tried running the signal directly into the set? Model?

[bkalbfeld]

It's a Mits WS55807. When the BD30 is run directly into the Mits, it works fine. So I figured I'd bypass the Onkyo and use a simple component switcher. Picked up a Psyclone Source Selector (4 in 1 out component/S-video switcher) and hooked it up. That made it worse: even the OTA signal (thru the Onkyo) caused the Mits to wig out!

[lcaillo]

The switchers are likely altering the sync signal slightly. This series of Mits were very touchy about distorted sync. Hook it up directly. I will search to see if there are any updates or mods for the set, but I do not recall any at this time.

[bkalbfeld]

So what's the chance that a different AVR would have tighter sync tolerances? How would I tell (I still have a few days to return the Onkyo if need be)? Plugging and unplugging cables to switch from OTA video to a Blu-Ray disc is less than ideal.

[lcaillo]

Well, since you have tried two units, I would not think the probability would be very good.

[Krankshaft]

I regularly post on a forum dedicated to bad caps and Jamicons are marginal caps at best.

In some applications they will fare well and in others they will go belly up.

The big companies that produce good quality caps (in no particular order) are:

Panasonic (Matsushita) (These will have a M on them)
Nippon Chemi-Con (these will have a crown logo on them)
Nichicon (stay away from HN and HM series)
Rubycon
Samxon (stay away from GF series)

Nichicons failures for the HN and HM series were a defect in manufacturing which has been rectified.

Samxon is a good brand however I have removed TONS of bulged GF series caps from LCD inverters that have gone belly up. Just stay away from the GF series.

If you use one of these brands for your capacitor replacement and stay away from the series' mentioned you won't be sorry.

[lcaillo]

Over the years we have seen batches of bad caps from most of the cap vendors at times. The one that I would say is an exception is Panasonic. Nichicon, for instance was notorious for leaking in the V11 and V12 chassis of the Mits sets in the early 1990s. We would have to change dozens in each set.

[Mike P.]

.........bad front panel switches likely due to windex syndrome.........

Could you explain what you mean by "windex syndrome"? Thanks.

[lcaillo]

We often see sets that have erratic controls or are locked up such that the microprocessor is not responding. The cause can be corroded front panel control switches. When this is caused by over-zealous cleaning in which something corrosive like windex is sprayed upon the screen and drips down onto the tactile switches in the control panel, we call it "windex syndrome." This would probably be a good entry for the glossary.

Basically what happens is that one of the switches passes enough current that the system thinks it (or thinks some other switch) is depressed. When one control is activated, the bus is typically tied up so that others cannot. Sometimes just pressing the buttons repeatedly can clear it, but usually only temporarily. Once damaged, it is usually necessary to replace the switches.