Marantz DD82 RDMUX TP


Would anyone in the forum have their own image of what to expect on the RDMUX Testpoint of the RW pcb?

I am working on a faulty board due to the leaky capacitors corroding the board. All signals appear correct on all other test points apart from the RDMUX which doesn’t appear as shown in the service manual.

The PLL has been checked and set to the exact frequency as specified in the manual ± 0.1MHz

All other timing signals appear correct.

Hoping one day someone will be able to produce a clone PCB replacement.

1 Like

You can expect 9 pulses (high or low) for the 9 tracks of the tape and one empty spot with a clock of 3.072 MHz. So it is different depending on what you are playing.

From the DCC900 service manual:

From an unreleased document I got this year:

I plan to use these three signals for dumps, especially to conserve unique tapes for the museum.

What did you measure?


RDMUX is an analog signal that represents the level of magnetization of each track. On an analog oscilloscope the signals will change so fast that you’ll just see a blur, represented in the picture in Max’s post as a sequence of blocks. On a digital scope you will see the levels of the 9 heads, and if you put it on hold and look at multiple blocks, you will see that each head changes between high and low at least every few blocks. If one of the heads is defective or dirty, you will see a gap in the blocks. If the entire head is dirty, or there’s not enough head to tape contact, or if the amplification needs to be adjusted, the blocks will look flatter.

At recording time, the tracks are magnetized to full saturation in either direction by sending current through the recording heads in either forward or backwards direction, where the polarity can change up to 96000 times per second per track (except the auxiliary track which runs at 1/12th of that frequency). Because of a system called “eight-to-ten modulation”, there can only be a few (I don’t recall how many) consecutive 0-bits or 1-bits per track. The tracks are recorded in a staggered pattern: no two heads change polarity at the same time; first track 0 changes polarity (if necessary), then track 1, then track 2 etc.

At playback time, the system has to recover the bits from the tape of course, but it doesn’t know where each bit begins. So it reads the tracks repeatedly at a fast enough pace to get a “slice” of each track (in the form of the RDMUX signal) at a high sample rate (not 96000 Hz). Because the system knows that the number of transitions per second per head is between 96000 and 96000/n (where n is the maximum number of consecutive unchanged bits according to the right-to-ten modulation algorithm), it can recover the clock of the original signal (96000 x 8 = 768000 bits per second). It uses this to adjust the capstan speed. And it uses it to sample the bits.

The bit stream of 768000 bits per second includes synchronization patterns, SYSINFO data, error detection/correction codes and music data. The music data is exactly half of the bandwidth: 384000 bits per second.

A very clever system! For more info about how it works, see the DCC system description on

=== Jac


3 posts were merged into an existing topic: DCC System description discussion

Hi Jac

Thanks for your reply what a great description !

This is exactly what I’m seeing just rapid random pattern on the oscilloscope as it is indeed analogue. However it does not match the pattern in the service manual (which I assume is also an analogue scope)

If I stop the tape transport manually there appears to still be noise apparent.

I’m possibly looking at another fault/leak damage somewhere on this pcb.




I was seeing as Jac mentioned below just random fast waveform but it does not match what’s shown in the service manual

There must be another fault somewhere on the pcb…

Can you try and capture a photo or video and share your setup? Maybe your oscilloscope bandwith is not high enough to properly represent it? Are you sure that the head and the rest of the recorder is okay?

Yes I have checked all 9 coils in the record and playback head cable all are around 45ohms and the record from memory lower resistance but all equal.

The DA board wasn’t too badly damaged and has had all its surface cleaned and recapped.

I will send you a screenshot tonight of what I’m seeing

I am waiting on a couple of spare RW pcb to arrive worst case I’m hoping to make a good one out of three.

Another interesting thing I notice during analog cassette playback, (Which I have working ok now, after removing Q153 DCC/CC switch, and cleaning the backside of the component and pcb with IPA) if I ground the ground lug of the pcb to the deck assembly (the screw on the corner) the left channel analogue signal becomes distorted. There is a Voltage difference of about 2.4V between this screw point on the board and the chassis. I’m thinking there may be more residual electrolytic somewhere under one of the components on the board…

The whole pcb was soaked in IPA for a few hours and dried. But it still appears there may be electrolytic trapped under some of the smt resistors/capacitors… most of the semiconductor smt have been removed cleaned and remounted.

The digital RdMux looks like digital RF I will shoot through an image tonight


Please see the attached video and images, Hopefully this help to give an idea of whats going on…

RD CLOCK (Pin 9) (Above)

25HZ Tape Speed Reference Signal

PLL Frequency Test Point (Above)

RD SYNC Waveform(Pin 11)

This is the waveform on RDMUX TP on the output of the RW PCB

Note my probe is set to X10 for all these readings

1 Like

Come to think of it, this is to be expected. I don’t have personal experience (I’ve never tried to mess with a DCC mechanism like that) but I expect that the read amplifiers have a built in auto-gain control that maximizes the signal from each head. So if the mechanism is stopped (held) or the player is playing virgin tape, you still get a full signal.

So my previous statement that you would see a blank in the middle of the signal trace, is probably wrong.

=== Jac

1 Like

In order to show the RDMUX signal, you should connect a second probe to the RDSYNC signal, and use RDSYNC to trigger your scope.

On a digital scope like yours, you will see a different waveform than the service manual because the scope is fast enough to follow the signal: instead of the “blocks” you will see something that looks like a wooden fence where the boards appear to go up and down.

My logic analyzer has a digital scope too and I took a picture of the RDSYNC signal once, but I can’t find it right now, sorry…

=== Jac


Thanks I’ll give it another go as per your suggestion using the second trace