Ensoniq DP/4+ (cap leakage, corrosion)

Help Support GroupDIY:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
More caps --almost all with fixed traces around them. In order of dodginess.

Around trimmers -- plenty of space here
1627184709616.jpeg


Getting tighter
1627185060228.jpeg


Traces running underneath ICs -- luckily it all connects
1627184765963.jpeg


Tight around caps on right -- traces around almost all caps needed some attention -- styro awaiting reinsertion
1627185211680.jpeg


And the winner is...
1627186831975.jpeg
...traces running underneath and between pins of SMD IC. Hmm. Me not like much !! But can't always avoid in layout, I guess. ( Note to self: avoid wherever possible, not least to make troubleshooting easier. Better keep everything visible at a glance. )
Connectivity here is OK. So I placed white marker near IC pin and on trace as hints, should this area freak later. Alternative is to remove the IC for inspection underneath -- kept in mind but want to avoid for now.
 

Attachments

  • 1627186162813.jpeg
    1627186162813.jpeg
    105.2 KB · Views: 2

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
-- Quite some progress despite heavy workload at the moment -- will try to update bit by bit --

Trimmer holes as rreconstructed before
1628401458658.jpeg


Trimmer underside, wired for stability, one eye "gone" (doesn't need to connect on underside)
1628401597390.jpeg


Leads needed soldering on board underside and one (on left) underneath trimmer to connect -- tight there
1628401670671.jpeg
Doesn't look too nice -- but who cares


Trimmer reinstalled (looks good enough for me)
1628401729885.jpeg
 

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
Regulator underside. That trace on the left needed to be wired to create wire loop to connect to regulator lead. That trace struck me as very thin for the periphery of a power regulator (5V). Might touch up with nail polish.

1628402334476.jpeg
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
GROUND ZERO

--Cleaning--
I had first gently scratched off leakage as much as possible -- that is before throwing any chemicals at it. Isopropyl is a bit hard to get hold of in Japan (at the moment). Hand sanitizers are too low in concentration. Also I didn't want to use acetone here or even break-cleaner (experimental). Ended up using a product (sold as a flux cleaner here) containing something that translates from Japanese as methylene chloride or dichloromethane. Banned in Europe AFAIK, not sure about the US.

--Tracing--
After tinning all pads and traces in the worst affected area, I traced the entire area there (four input channels) following schematics. Also measured all components. Two diodes were off in value and not connecting, revealing that one refilled "eye" still didn't connect. So I fixed the "eye" with a wire through it (white circle), which then is soldered onto the traces on the top und underside of the board.

1628403987782.png

All ttraces around here were badly affected . Also notice how bad the adjacent pad looks (the one next to + sign). A lot of electrolyte must have drained through this "eye" (top left on next image), damaging the adjacent trace.

1628404346873.jpeg
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
REPAIRING LIFTED SMD PADS

Right from the start there were a few pads that had (partially) lifted in this area. Lots of solder paste and solder wick is the way to clean them first. However, some were too compromised and just lifted off the board -- no matter how gentle the wicking process. Electrolyte had seeped into the pad beds and underneath them.

I used different "techniques" for fixing these SMD pads -- also cos I wanted to try out what works and what doesn't, and how long it takes (!). -- SPOILER : just as I had expected from the start, some didn't work out satisfactorily,, but I wanted to try them anyway.

------
Here's a pad with discoloration around it. It had lifted completely but still connected to the trace.

1628407350894.jpeg

I folded it up sideways, cleaned underneath and superglued it back into place (very quick process). Then I used the blunt side of an exactor knife to scratch off ecxess solder (from tinning -- solder is malleable, but this takes long) -- hoping that it would later withstand the soldering process.

I did the same with one of the partially lifted pads (top left in image below). Notice the fold line there. Pad was in a very bad state from the beginning. While for the pad below "C479" I used off-the-shelf epoxy as glue plus scratching off excess solder to form a plateau. -- I am aware that SMD pad kits are available with heat-resistant epoxy exactly for this, but I am cheap and didn't want to wait.

1628408589576.jpeg

The pad in the middle (above C479) is experimental. It sits flush with the board but works like a cap. There is a wire soldered onto or rather next to the trace and extending to form a tinned wire "inlay" underneath the pad. The wire inlay sits fully recessed into the board. The pad is then soldered onto the tinned wire inlay from above to create firm connection. Then lifted up together to epoxy-glue it firmly back into place. It sits flush indeed.

The idea here was to make the fix invisible. Overall this is pretty crazy -- no, it's bonkers and just time-consuming. Also, it lifted off later, although not during soldering the component but during cleaning the board. -- Overall, the only good thing about this fix is that I have now come to deeply appreciate the work and art of dentistry, haha : ) :)

--More feasible fixes to come.--
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
An overview image of the entire channel input area.

All traces confirmed for connectivity and pads fixed (as described above). This is before insalling caps, obviously, but also before final touch-up with lots of solder paste -- meaning all eyes/holes in that area (in all affected areas across the entire board actually) got resoldered, as did diodes and some IC pins. -- Why ? -- I sure wanted more practice with SMD, but also cos some solder connections looked dull even after cleaning and others were quite dark due to corrosion. Some needed scratching off of the top surfece for solder to take on at all, even with lots of SMD flux paste.

Reheating some of the joints produced a somewhat evil acidy smell. Not sure why but I guess electrolyte residue, despite 2 x cleaning.

1628422406011.jpeg
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
FIXING SMD PADS -- WHAT WORKS ?

Using superglue and off-the-shelf epoxy for reattaching lifted pads didn't make much of a difference. Both did not withstand the soldering heat later. One superglued pad did hold though, much to my surprise.

The caps all connected alright but were wobbly, hanging on lifted pads now and, one even worse, on a partially lifted trace... So time to resort to the tried-and-tested techniques that I knew of before starting with the ones described above.


1. Wire onto component lead (clamping down)

Wire is soldered onto trace (from CR32 and CR33), then bent over cap lead to hold component down (like a half-clamp) before applying solder. Comparatively fast and solid result. Also, the component sits flush on the board.

1628441649515.jpeg

As for pads that had come of, supergluing them quickly back into place seems absolutely fine in my case. This is DIY. Professional is a different game. Anyway, glued into place, such pad is not for establishing the primary connection, but for getting both a flush board surface and a larger surface for solder to sit on and around the wire and component lead.


Same technique as above, but a tad trickier.
1628441862280.jpeg
Two times wire soldered onto short trace between caps. Firm result. For a more solid connection, wire could be soldered into a nearby "eye" along the trace (if there is one). Wire is then soldered onto the tinned traces and finally over the cap lead. [Note: Still some fibre here from first rough cleaning.]


2 Wire as pad underneath component

A bit more work (maybe, but not really) is same as above but instead of half-clamping over component lead, the wire forms a loop that sits on top of the glued pad.
1628442103862.jpeg
[A lot of flux paste from touch-up -- and out of focus, man !]


OK, different angle
1628442198976.png
[Now, there is some fiber on the lens !]


Here after some cleaning. Wire through ‘eye’ and loop on pad invisible. Component solidly held into place. The cap sits slightly higher than others, by exactly one wire diameter (see left corner of component).
1628442448225.jpeg

This is the technique and result that I ended up liking most by far . It is not slower than half-clamping component leads with wire, and it sure is also more reliable than soldering pads directly onto the thinnest of traces.

Ironically, this is also the one technique that I had envisioned (but dismissed – snarked at) when first looking at the damaged board. --- Humans -- a funny creature. With us, some things seem to just need a bit of time, as first ideas that come to mind, especially when we instantly dismiss them entirely, might just need looking into a little more to make sense after all and maybe even make us change our mind.
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
Another technique deserving mention. It's probably the fastest.

3 Lead-onto-trace

If only one pad has come off, an alternative technique is to solder one component lead onto the remaining pad and the other lead directly onto a tinned trace. This makes the component sit at an angle. -- Who cares about aesthetics of a PCB once the lid is on ? -- It sure is not me. Still I refrained from it here, cos traces in this unit are very thin = too little trace surface for firm solder connection.

-----------------

Caps in profile. Spot the two outliers. A bit out of line cos of thicker trace tinning and bulging eyes. Note to self : Keep in mind for next unit ; )
1628443531360.jpeg


Top view of entire input section (cap platoon in place)
1628443595794.png
Next up is cleaning this area a bit more, to be followed by some more green nail work in the beauty joint.


-- Still remaining are a few minor construction sites. --
 
Last edited:

Whoops

Well-known member
Joined
May 9, 2008
Messages
5,688
Location
Portugal
I’m following this thread with much interest,
such an amazing and hard work you re doing.

I really hope it all works in the end, since this is a great effects processor

Keep up the good work
 

hobiesound

Well-known member
Joined
Jan 29, 2006
Messages
390
Location
holland
What works great as heat resistant epoxy is the uber cheap Mechanic UV curable solder mask you can get on aliexpress.
 

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
Thank you very much .

That looks like a product I could enjoy and it comes in a variety of colours too. Will likely order cos now I feel almost tempted to rip out a cap or two again just to see how it works ; )

Honing my non-existent SMD skills is what I am in for here.
 
Last edited:

j.colt

Member
Joined
Aug 6, 2020
Messages
11
Location
Washington DC
Thanks for this incredible thread... my DP4+ was declared dead many years ago with the usual corroded via and traces. I hope you will succeed and I also hope this thread will deflate the insane DP4+ prices!
 

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
Not sure whether whatever I do or anyone does with their unit would have any effect on the overpricedness of these 'rare vintage "cult" units'.

And they are a cult indeed I have found out meanwhile -- justified or not. Well, I tend to think that for today's standards it's rather the latter. It's a 24 bit processor but with only 16 bit input and output -- that plus component choice, I guess, make for that classy 'grainy' sound -- but what do I know ; )

Also, I think repairing units that are in very bad shape like this one is not a feasible business model. lt's not the cost of components, but the sheer cost of work time. Leaves DIY or the junk heap for good.

Refurbishing with new caps seems due in these units -- even if still fully functional with original components. Those SMD caps in there are real time bombs and after 20/25 years easily the cause of extensive and nasty damage when dilly-dallying

The unit here was officially declared dead and destined for the junk. Personally, I'm in for learning how to handle SMD 0805 -- could easily have chosen a different unit -- either way, for me, this is just some kinda 'carrot' ; )

Should I not be able to fix it, it is all still satisfying -- but should I succeed, I will audition it and might keep it, but I can also easily see myself giving it to a good friend who sure can make better use of it. I'd quite like that...
 

Whoops

Well-known member
Joined
May 9, 2008
Messages
5,688
Location
Portugal
Those SMD caps in there are real time bombs and after 20/25 years easily the cause of extensive and nasty damage when dilly-dallying

I really have to open my DP4+ unit to check is there's any leakage from the capacitors.
Last time I used it was 1 year ago, and it worked and sounded great but now Im afraid of what might go inside
 

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
As for my unit, the power supply worked unloaded. II's all back together provisionally. Easy coseverythingl wired via connectors. Plugged in power and... lights up... but doesn't load greeting screen... Some error message instead. Hoped it was just a brain fart, so tried 'reinitialize' software per front buttons. Didn't work. Then tried reinitialize per power button as described in service manual. Nothing. Still have to try hard reinitialize. According to manual, that error message could mean anything really -- including that there's still something wrong on the boards, which is very likely.

Anyway, the regulators on the audio board are running quite hot. Two of them too hot to touch for long actually . but I guess that's what they generally do -- maybe not. Well, quick test showed that voltages on IC pins are fine, but there's 15V on pin 1 of one of the dual op amps in the input section. So obviously something wrong there : input protection diode or IC itself, or something down the line.

But all good, cos now the work can start ; )
 

j.colt

Member
Joined
Aug 6, 2020
Messages
11
Location
Washington DC
As for my unit, the power supply worked unloaded. II's all back together provisionally. Easy coseverythingl wired via connectors. Plugged in power and... lights up... but doesn't load greeting screen... Some error message instead. Hoped it was just a brain fart, so tried 'reinitialize' software per front buttons. Didn't work. Then tried reinitialize per power button as described in service manual. Nothing. Still have to try hard reinitialize. According to manual, that error message could mean anything really -- including that there's still something wrong on the boards, which is very likely.

Anyway, the regulators on the audio board are running quite hot. Two of them too hot to touch for long actually . but I guess that's what they generally do -- maybe not. Well, quick test showed that voltages on IC pins are fine, but there's 15V on pin 1 of one of the dual op amps in the input section. So obviously something wrong there : input protection diode or IC itself, or something down the line.

But all good, cos now the work can start ; )

I remember the regulators to be super hot.. In fact, one channel crapped out (overload led all the time) and the unit was then beyond repair after I forgot to turn it off for a couple of days....
 

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
Thanks for confirming

Seems to be a pattern. Red overload LED permanently on was reported on this one too (caps already leaking, I assume) -- until kaboom. Two fuses blown.

Hot regulators. Yeah, those heatsinks struck me as small, considering what the PSU spits out. SMD caps right between those heatsinks plus X5R Z5U 0.1uf bypass caps, which are rated for 85°C, sitting right next to them.
 
Last edited:

Script

Well-known member
Joined
Nov 26, 2008
Messages
1,446
Location
Japan
U1 (4565) dual op amp on input CH 1&2 was definitely fried. Replaced it and voltages on all pins look better.

U101 (4565) dual op amp on input CH 3&4 seems OK, but fried diode CR34 in front of it. Have to order and wait.
 
Last edited:

Latest posts

Top