Several Weeks ago I bought an Ensoniq DP4+ Effects Processor. It is a nice and flexible machine, that has 4
independent Ensoniq Effect Chips on board and you can route the audio signals in many ways.
The first version of this processor (DP4) was sold in 1992 and was built as a through-hole board, the
successor (DP4+) has some minor additions and is mostly built with SMD Parts. That version came out in
1995. If some of you remember, that was the time where many electrolytic capacitors were built with low
quality chemicals (capacitor plague). And as I will describe later that is the root problem with these
machines.
So I have written this report to tell my story and help others that have similar problems with these devices.
The machine was advertised with "distortion on some channels". As naive as I was I thought that could be
scratchy input/output pots in the signal path and some cleaning should be enough. So I contacted the seller
and bought the machine.
In my opinion the machine has some positives: You can get the full schematics of the "Plus" version, and on
most of the boards they used mostly the same components (e.g. many opamps are NJM4565, with some TL072, and
there are only 3 types of SMD Caps on all boards). So if you order enough parts, you can use them in many
places.
After I got the device, I opened it and checked the inside. The first thing I noticed was that some of the
previous owners inserted a jack on the backside to remote switch the "Mute"-Button on the front.
So I connected my phone to the input and switched the device on. On my headphones the audio sounded OK, as
expected only the pots were scratchy. I cleaned the pots, removed the mod, and started to play with it. The
processor can be driven by up to four sources (it has 4 input jacks, that is 4 mono signals). My first test
used just one source (mono, using just on jack or stereo using two). As soon as I switched it to more than one
source, the output got distorted. Ouch.
Another thing I noticed was that the signal level led for channel 3 was red (too high). So I took another look
at the audio board (the machine has 4 boards: audio, digital, frontend and psu) - this time with my
microscope. And there I could see leaked electrolytics and the damage it has caused.
So I ordered new caps, opamps and some 0805 caps and resistors, and began to clean and replace the
boards. After replacing the input-opamp on channel 3 the signal level was OK, but now I was greeted with the
error message "Bad DC Offset(s)". More searching, more following the signal path. I inserted a sine wave
signal from my function generator and checked several points with my scope. The last opamp right before the
ADC looked suspicious. After replacing the 4565 the error was gone. But I still got distortion.
What was weird was that the same signal was fine until I switched to more than one source - but with the
signal still on the same input jacks. More thinking, more head scratching, more staring on the schematics...
Then I started looking at the output stage, right after the Phillips DAC (the famous TDA1541A). There was only
one opamp for both channels (L and R), each channel carrying two signals (later switched to 2 channels). I
replaced the opamp and some caps and resistors in the feedback loop, without succes.
The mystery was that the 1-source configuration worked. My assumption was that the signal on channel 1 comes
out of the L-Channel of the DAC, because that channel looked connected to channel 1 on the output. So what
puzzled me was that the same signal got distored when changing the config... Where did that come from? But
after some more checks I realized that that was not true: in this configuration the signal comes out auf the
R-Channel of the DAC and is later routed to the channel 1 jack on the back side. After this insight I was sure
that the problem was in the left output channel of the DAC.
The next assumption was that the problem could be in the digital signals that the DAC got from the digital
board. Between the connector and the DAC is only one 74HC86D XOR-Gate, and one 74HC04AD inverter nearby. Both
chips are located near one of the leaking caps. So I decided to replace both of them - but no luck.
The signal on the defective channel looked kind of funny: My signal was stretched out near the zero
crossing. One of the algorithms in the device can generate a sine wave signal, so in order to concentrate on
the output stage I used this signal instead of my own on all 4 outputs. And there the problem was better to
see (sorry - I forgot to take pictures): The top of the wave was ok, but then there was a shift of about 0.5V
downward near the zero crossing...
I started to fear that there could be a fault on the digital board, because nothing on the analog side
helped. It looked like the digital data was transmitted correct to the DAC, so where else could the error
be. I checked the Firmware EEPROMs, but the checksum was OK...
But before starting on the digital side I looked up more information on the DAC. And then I found the crucial
information on some page: there are 7 bypass caps on each side of the chip, and each of those is used when
some bit is set on the DAC. When one of those is shorted, that could explain the shift in voltage....
I replaced the whole row of caps on this side of the chip - and it worked. Distortion gone!!
The bypass-caps on this side of the DAC were placed near the voltage regulators on the board, with some of the
leaking smd-caps between. So it looked like heat and electrolytics in combination damaged one of those caps
and caused the shift.
Mission accomplished. Maybe this machine can now work another 20+ years...
Home
Help
Search
Login
Register
