Fixing an Early XM Clarion PU9358A Radio - Or Trying To!

[CitroJim]

(Mods, I trust this is in an appropriate section, if not, please move it for me)

I have two of these devices I'm looking to fix for the owner of a certain white S1 Citroen XM...

Today, finally, my head has been in the right place and time has been on my side to get deeply stuck in...

These are quite clever wirelesses and sophisticated for their time. At their heart is a 4-bit microprocessor/microcontroller which controls all aspects of the operation of the unit, including the cassette deck.

Both units initially presented with the same fault. Totally dead. No lights, no 'Code' light flashing, no display...

I'm concentrating on just the one unit for now after determining the second unit does the same but appears to have a power amplifier fault too...

My first check was for power to the microcontroller. That was good and on the 'scope I could see the microcontroller clock was running.

Looking at an old and long thread on the Club-XM forum detailing how to defeat the codes and reset these wirelesses, I tried holding down the BND (Band) button to do a hard reset and after a minute or so it fired up and appeared to work. FM radio good and all four audio channels good. And it never even asked for a code! Happy days

Own Work

It was too good to last... After a couple of minutes it died and not even a hard reset would work until the power supply was disconnected completely and reconnected. A further hard reset would allow the thing to work for a further two minutes or so.

I'd noted this had been mentioned in the Club-XM radio code thread with co conclusion reached as to why.

During the two minute working period, all was good. It would turn on and off normally and retain its memory.

The way it works is quite bizarre. The microcontroller has no program memory as such and all instructions to be executed are synthesized with a combination of front panel button presses, signals from the cassette deck and some 'fuzzy' logic built around a number of transistors and diodes...

What appears to be happening is the controller seems to crash after a couple of minutes of operation and corrupt its volatile memory... This happens even if the thing is turned off but is being maintained by the battery. Switch it off and on within two minutes and no problem. Switch off and then try to switch on after two minutes and it's crashed and dead...

Oddly, it's never asked for a radio code and I've never seen the code LED flash... I even unsoldered all the code bridges to try and force the code to 1112 but it made no difference. Still no code request...

I shall carry on my fight and do all I can to determine what's happening to crash it. I have some thoughts and tests to perform...

I hope I can fix it, it's a very high quality unit and performs well...

Watch this space

[GiveMeABreak]

I have the service manual (if you can call it that) and wiring diagram (not brilliant) if any help Jim.

[xantia_v6]

The processor will have on-chip ROM containing the software code.
Does the processor chip get warm? That would be a bad sign for a CMOS processor. You can try gently cooling the processor chip (e.g. ice in a plastic bag) and see if it runs longer before crashing.

[CitroJim]

I have the service manual (if you can call it that) and wiring diagram (not brilliant) if any help Jim.

Already got it thanks Marc... Indeed yes, it is a bit light to be called a service manual but it's better than nothing. Data sheets for the various ICs help too...

Does the processor chip get warm? That would be a bad sign for a CMOS processor. You can try gently cooling the processor chip (e.g. ice in a plastic bag) and see if it runs longer before crashing.

One of my next checks Mike To my finger it does feel a little warm. What I really need is a can of freezer spray and one will be on order as it does appear to be prima-facie heat related...

[mickthemaverick]

I still have a brand new pc cooling fan in the drawer from when I modified the satnav in the mk1 C5 if its any use to you Jim

[CitroJim]

Happily, nothing to do with temperature Mick

From a post on the Club-XM forum...

In my experience they don't ask for a code, just the display is feintly illuminated, but blank.
I assume you have tried entering the code when the display is blank?

That's cracked it Will Excellent. How bizarre. Yes, a blank display. Entering the 1112 code using the numerical buttons followed by 6 (after clearing all the code bridge links) and started normally without needing a hard reset...

The first set has now been running for an hour without fault... I'll now leave it off with 'ignition' off but battery connected and see if it properly wakes and has retained its memory..

Thanks so much! Now to diagnose why the cassette deck only works on one channel and to try the second set and see how that responds...

Progress

So is this two minute timeout after a hard reset without a code a deliberate piece of programming to give it a sort of 'demo mode' in the absence of a valid code... If so, very neat

I'll post how to clear the code bridge links to reset the code to 1112 in a subsequent post...

[CitroJim]

Sorry or the lack of updates recently... Competing in a Middle Distance Duathlon and parkrun got in the way of any radio work this weekend...

Today has been a rest day for me so a great opportunity to continue.

Of the two on my bench. Both have been successfully de-coded and one works brilliantly well on radio but is very quiet on tape and only one channel works irrespective of direction. This rules out the heads... Swapping decks between the units makes no significant difference...

The other is totally silent on both tape and radio although it can be seen the radio section is working and receiving signals...

Scoping through the circuit, the audio from the radio is present at the input to IC401 - see diagram - but there is no output from IC401. IC401 is not just an Integrated circuit - it's a small module and acts as the tape head amplifiers and an audio preamplifier after tape/radio switching selection logic ahead of the audio processing IC and its onward path to the power amplifier modules. IC401 is particularly inaccessible, being a daughter board on top of a daughter board! It's the item circled in yellow in the picture... The IC on this daughter board appears to be a special as Google knows nothing of it...

A corner of the circuit showing IC401

IC401 circled in yellow

I suspect this component is the root cause of the tape issue in the first unit too... Maybe the Achille heel of these radios? I'm suspecting duff electrolytic capacitors as the little daughter board is full of them and of a different style to the ones in the rest of the radio...

This component being suspect may paradoxically assist dumping the tape deck and converting to Bluetooth... With the owner's permission I may, in slow time, look into this and perhaps do a bit of modification in this area to work around a faulty IC401...

It's been an interesting time and I've got to understand these fiendishly complex radios quite well. Due to the component density they are not at all easy to repair or event fault-find on.

More news as and when!

[CitroJim]

I'm now looking to try an experiment bypassing IC401 completely as a precursor to a potential conversion to Bluetooth in place of tape.

IC401 is, inter alia, the Dolby processor and prima-facie, not required if tape is discarded. In any case, trying to source a replacement may be trickier than tracking down some rocking horse poo... It's a thick film module rather than a normal IC and I suspect, originally supplied by Dolby Laboratories...

To de-code these radios, remove any solder 'bridges' in the area highlighted in the photo below... The photo shows a number of 'bridges' in place. Solder wick followed by a good cleanup with Isopropyl and a careful check to ensure the 'bridges' are all clear is the best way to do this. Don't try to de-solder the big blob on the corner of microcontroller... In retrospect I should have made the yellow circle a bit smaller to encompass only the bridgeable pads...

Radio Code 'Bridges' within highlighted area

Once all solder bridges are clear, the code will be set to 1112.

Stay tuned for the next instalment of this adventure One I must admit I'm quite enjoying...

[CitroJim]

Much more work done today...

To see the latest news and how I'm developing a Bluetooth conversion follow my thread over on Club-XM

https://club-xm.co.uk/Forum/viewtopic.p ... 06#p126306