Hello
I have recently bought an XV-5080 on ebay.
after some tests I noticed some crackling noises in the sound.
some search on the internet confirm a sound issue known as "static noise"
here is the problem in video


it sound like a clock sync problem so I've checked the master clock, PLL and all was fine.
after checking the power supply of the DSP something was strange. there was a big ripple on the 3.3V
here are some pictures of the tests
where the scope is connected with short ground link (ripple noise method)


and here the results

420mV of ripple noise!
the frequency seems to be near 378KHz

and now with the cap connected with the same scale as before


the current drawn from the regulator is measured at 350mA max


at the beginning I was thinkibg about a faulty LDO regulator and I have searched for a new one with the closest specs
the output noise must be near 55µV and rejection ratio at least 64db with line regulation near 15mV and load regulation 25mV.


so I've chosen the KF33DBT


replacing the regulator has solved the problem
a video before:


and after replacing the regulator


but the ripple measured was strangely high so I've removed the 47uF cap at the output of the regulator and measured it.
it was fine under my multimeter but bad under the LCR meter!
measured at only 1uF instead of 47uF!
replacing the regulator has solved the problem because the new one is stable with a lower bypass capacitance.
to completely solve the problem I've chosen to totally recap the unit.
I've used the Nichicon CD serie 105°C - 2000Hrs (the 5000Hrs were not available in low ESR/high ripple and same size)

10uF 16V : UCD1C100MCL1GS
22uF 16V : UCD1C220MCL1GS
47uF 6.3V : UCD0J470MCL1GS
47uF 16V : UCD1C470MCL1GS
47uF 25V : UCD1E470MCL1GS
100uF 16V : UCD1C101MCL1GS
220uF 6.3V : UCD0J221MCL1GS

I think pictures are better than words

I've replaced the regulator and added a 3.3V transil diode as protection. (I do this on each unit I work on. it's not expensive and can save the main electronic life in case of power supply failure)

yes! nanofarads! for a 47uF capacitor!


I've measured each old cap value and made a table


so I highly recommend to the XV-5080 users (perhaps 2080 and so) to make a recap of the unit because all these capacitors seems to fail. And they are the source of the "static noise" or "sparkling noises" issue.
The ripple caused by the lack of bypass capacitor create DSP master clock sync failure and I2S DAC conversion problems.
The original LDO regulator was not the culprit but was unstable with very low bypass capacitance. But I've chosen to keep my new KF33 LDO instead of the old one because it has more stability with less bypass capacitance, lower output noise and higher power supply rejection.

Now I'm replacing the caps on the power supply and I've trimmed down the output voltage from 5.2v to 5V . the XV-5080 seems to works hot, the power supply PCB is brown!
This is why I'm replacing caps, resistors and diodes because I have some doubts of the components reliability after working under high temperature stress and these components drives the main side transistor.
so a failure could be catastrophic.

here is my original topic: http://forums.rolandclan.com/viewtop...298986#p298986

Great report, a pleasure to read Thanks for sharing

you're welcome

yes great work! keeping the roland rompler dream alive!

I understand you did an extensive work to trace the source of the problem and that it would be difficult to list all the parts you replaced. I too have a XV5080 with the same crackling noise and I will follow this guide to make it sound nice again. I aready bought the PSU and I would really appreciate a guide to the connector pinout if you wouldn't mind sharing it. Thanks.

here is the link to my facebook album with all the infos
I've just made an album to explain the modification
check if you can access to this album.
XV5080 PSU REPLACEMENT | Facebook

I have been performing this repair and for anybody worried about crimping the molex connectors for the PSU, you don't need to go all out for a pressmaster if you want a terminal crimp tool.

I have found the Engineer PA-20 Universal Wire Crimping tool does a really good job. You have to crimp onto the bare wire with one size die and move to the next largest one and crimp again to make the crimp onto the insulation.

But as long as you are careful, the crimp is every bit as good as you can get with the pressmaster. I used to have the pressmaster at work, and it is easy to mess up your crimps with that tool too!

The Engineer it about 1/6 the price, and well worth it if you are just a hobby user.

Thanks to Genesis92 for this extensive repair guide

During my re-cap of the main board, I see

C138 and C154 on the main board in the XV5080 main board circuit diagram are shown to be non-polarized caps listed as Roland part: 01455845, 16CV22NP

These look to be bi-polar 16V 22uF caps from Sanyo

https://www.datasheetarchive.com/pdf...2520capacitors

BUT my main board has 25V 22uF polarized electrolytic Al caps from the factory.

I have replaced them with UCD1E220MCL1GS Nichicon 25V22uF, without problem.

This question has been discussed on the RolandClan post about the recapping, but the reply just referred to the service manual, which shows the bi-polar caps.

Looks like there has been a revision to the bill of materials since the 1st edition of the service manual

Ozz

as explained in the PM these caps can be polarized too.
they are used in the word clock and SPDIF out coaxial circuit so the signal is AC and the best is to use NP caps but like in the audio path, polarized caps can be used to if the right orientation is used.
(I've not seen your picture before my PM and I've tell you the same polarity as the picture looking to the schematic drawing so it's ok to replace them by polarized caps with the same orientation (C154 positive side on the input cinch and C138 negative side connected to FL1)