Digiplexes --- several on Ebay currently

You don't see these come up available very often, so I thought I would pass this along in case anyone is looking to replace / rebuild their units.

Interesting variety ---- US, Euro, NOS, used, rebuildable...

Ferrari 308 Magneti Marelli Digiplex | eBay

Magneti Marelli Digiplex Ignition Module Med 801 A Ferrari 308 Mondial | eBay

Ferrari Mondial 8 308 Gtbi Gtsi Digiplex Orig 115977 | eBay

Magneti Marelli Digiplex Ignition Module Med 801 A Ferrari 308 Mondial | eBay

Ferrari 308 Euro Digiplex Ignition MED803 118451 New | eBay

Magneti Marelli Digiplex Ignition Module Med 801 A Ferrari 308 Mondial | eBay

 

3 of those auctions are for non functional units.

Of the several failed Digiplex's I have seen in the last few years none have been repairable.


New ones have not been hard to get.

 

There is this outfit in Florida that claims to be able to rebuild / repair them --- but, I don't know anything first hand about their capabilities.

Ferrari 308 QV ECU Magneti Marelli Digiplex MED 805 A - Specialized ECU Repair

 

I have used them. The component that most commonly fails is propreitary and cannot be sourced.

 

They have failed to repair any Motronic units I have sent to them FWIW. Always try to upsell another unit.

 

They have one of mine as we speak.

 

Brian,

Does that statement include the original ones (801A IIRC) used in the 80 81 and maybe 82 308 cars?

 

I almost never work on those so I have not looked for one in twenty years but if AW is getting them they are out there.

I know one US dealer that is sitting on like 8 QV units but wants full on retail for them.

For the prices these thing are getting I would go aftermarket anyway. Ferrari left a lot of HP on the table with the advance curves so a custom tuned aftermarket piece would be a real upgrade.

 

One box is working but the other is presently with The ECU Doctors right now. They are having trouble with it but say that they are replacing a "part" this week.....who knows.

If they cannot fix it, what do you suggest I do?? I did locate a few new ones, but they are going for the rip-off price of $1,695 +

What you you suggest?

Frank

 

Thats a great price.


What makes you call it a ripoff?



How much do you charge for what you do?

 

Frank ----

Which model is your unit ? ---- should be an 801 if your car is a US-spec MY'80...

One of those Ebay listings is for an 801 ---- used, but good working order, asking $495. Not cheap, but a helluva lot cheaper than any aftermarket ignition setup currently out there --- not to mention easier to install.

Seller is an F-Chatter, "Colorado Tiger", I believe. He recently did a 3.5L / ignition upgrade, so is getting rid of his stock system ---- seems like a good, stand-up guy from all his posts.

Also, AW has a "new" 801 unit listed for a BIN price of $1,265 / offer......but, at that price point, some of the aftermarket systems probably have to appear in the trade space .

 

Brian,

Thanks so much for your helpful response.

 

Thanks for the heads-up!

I do have the 801A unit. I guess that I will wait and see what the ECU tech's say about mu unit before I make any decisions.

I appreciate your input.

 

Just curious, which component is the one in question, the MCU or APS sensor perhaps?

Thanks,
Adrian

 

I don't know what Brian's experiences have been, but I can share mine.....

Many years ago, I started "collecting" failed units because I had the idea I would play around with them, and see if I could repair them. Alas, that "project" evaporated among the constraints of not enough free time, not enough sustained interest, and too many good alternative solutions becoming available on the aftermarket. So, I ended up with a box of partially dissected as well as intact (but fried) units .

++++++++++++++++++++++++++++++++++++++++++

What I did learn from playing with them was the following, as far as most common failure mechanisms ---

1) Rarely (if ever) does the most important component (the CPU / EPROM [MCU, if you prefer]) fail. And when it does, it typically burns along with several other discrete components and ICs as a result of some power overload ---- either from improper attempts to jump start the car (a big "no no" with these cars), or some improper manipulation of the car's electrical system that spikes / shunts power to the ECU harness (particularly vulnerable if the ECU ground loop is compromised --- which happens often). Obviously, when these ICs are fried, it is catastrophic and cannot be repaired (unless you have the ability to spec, and source, all the components and burn a new EPROM with M-M's data / code ).

2) The APS sensor (if you can call it that as it is more of a generic pressure transducer) is a very crude design LVDT-type device actuated by intake manifold vacuum. Internally it has a conductive rod / plunger (moving within 2 coil windings) attached to a rubber diaphragm that reacts to changes in vacuum pressure. The diaphragm often ruptures --- it is un-repairable and un-replaceable. If a healthy one were properly characterized mechanically and electrically, an alternative device could likely be found and substituted, I suppose, if one were so inclined and properly motivated to go to the effort.

3) However, of all the dozen or so units I took apart, by far the most prevalent failures were within the structure of the PCB. These PCBs are poorly constructed and are mounted / sealed in such a way within the case that they are subjected to excessive heat and significant moisture. As a result, from corrosion and repeated thermal cycling, the connections (traces) themselves fail (open circuit). This is a simple failure to isolate and the connections can be easily repaired ordinarily, however and unfortunately, removing the boards or attempting to expose them in order to locate the failures (and have access to repair them) is extremely difficult to accomplish without doing a lot of collateral damage to the PCB --- they are simply too delicate and it takes too much "aggression" to remove the potting material. The result is that you typically destroy (effectively) the board while you are trying to access / repair it ---- rendering it into shards of hopeless.

++++++++++++++++++++++++++++++++++++

Based on my experience, and IMHO, the best way to "repair" a bad board would be to actually build a new one from scratch, and to pull off and re-use the key ICs and main CPU from the old one. For someone who has a good knowledge of PCB design and construction, this would not be too difficult ---- as the board circuitry is reasonably simple (and only single layer).

FWIW

 

Time constraints are always an issue, interest is high though.

1) Common problem also with other types i've worked with, but not too difficult to add a TVS/inductor based voltage supply protection filter to older units though, but won't be of help if the ground connection is bad as you already mentioned.

Nothing is really catastrophic, it is just more work The MCUs are typically OEM specced types with internal part numbers, so no datasheets. Marelli often used Motorola 68K family derivates with more io lines but it shouldn't be that difficult to design a pin compatible adapter board with an Atmel Automotive grade MCU to replace the existing MCU (need to adapt the TTL to 3.3V level though). The MCU is just doing table lookups based on the input states, so with a working Digiplex, a good digital scope and a logic analyzer you can derive all of the operation states on the test bench.

2) Interesting, that is the same approach used by the Mareli SAE predecessor of the Digiplex I series. Wired to the LVDT is most probably an op amp supplying a linear DC voltage based on the map pressure, so one could replace it with a Freescale MPX pressure sensor which directly supplies a linear voltage. The MPX sensors are quite small so they should fit in the existing enclosure i think.

3) Hmm, that sounds more like a complete redesign. The original PCB is quite simple, shouldn't take too long to design a similar one but then you need to do full test cycles including a climate chamber (which i don't have) and EMV tests, that is the really difficult part. From what i see on photos the Digiplex 1 units is covered with white silicone coating, is that correct?

Thanks, very interesting input,
Adrian

 

I really do not remember but I suspect it was the APS. The advance control was always a source of problems with them. The Ferrari diagnostic unit used on those called the Jofatron had a set of go/nogo lights to indicate if the advance was within a rather broad range through the rpm range and failing to stay in that spec was a common problem even when the cars were still in production.

I do little work these days on cars of that vintage but it seems to me offering a plug and play kit based on a modern aftermarket unit with programmable advance is the way to go rather than repairing the old units. Ferrari left a lot of performance on the table, at least with the US units in the advance curve.

 

Brian,
thanks for the input! I'll have a look at the PCB, if i really have to do a full redesign with a new MCU then adding user programmable flash memory for the map should be relatively easy.

Adrian

 

Actually, MM used the old-school RCA 1800-series, gold / ceramic, 40-pin, processor for these applications. Although you can still find these things available (NOS) laying around in vintage electronics inventories, if someone were to do a rebuild, it would certainly be a good idea to replace it with a better, more modern, processor .

You would not even have to go through the effort to decode and extract the data from the original CPU or even fully characterize how it is operating. Something very "handy" for someone looking to create one from scratch is that Ferrari supplied charts (in the 308 / Mondial workshop manuals) that have all the ignition curve data (degrees vs. rpm vs. vacuum load) for all the various iterations --- 801, 802, 803 & 805 MED units --- and, they are simple, perfectly linear curves, divided into just a few discrete on / off / next segments (a no-brainer for programming) .

Precisely --- that is the scheme. IIRC (and don't trust my memory too much..... I'm getting old ), the LVDT is not instrumented directly as it would be in a proper test application --- whereby it would supply an output voltage. Instead, low current (again IIRC, it is rectified to DC prior to input) is passed through the coils and it (output current) is increased / decreased as the magnetic plunger position changes and varies the equivalent inductance of the coils. The output current then feeds into a Current-Controlled Voltage Source OP Amp (off-the-shelf item) which generates a variable voltage signal to the CPU --- this gives the CPU operating points for engine "load" to select the appropriate advance curves from the function generator element.

Correct --- the earlier units ('80 to early '81') had the PCBs potted in a white silicone, but from then on, for some reason, they switched to a black silicone. The white is easier to remove, btw .

 

Interesting thread guys, I'll be keeping a close eye on it.

I'm surprised though of the comments about the units being well sealed with silicone. That last ebay listing (Magneti Marelli Digiplex Ignition Module Med 801 A Ferrari 308 Mondial | eBay) in the first post shows a malfunctioning unit that has it's underside removed and is completely free of any type of silicone packing... Is this unique to the item being sold, or are they all open like this inside (once a cover is removed)? Please forgive my ignorance, I've only been playing with carburetors and mechanical distributors and am not familiar with Digiplex (not yet at least, but one day with any future purchase of an '80s-era Ferrari! ).

 

Peter ---

Actually, that unit is fully sealed / potted ---- in the "white" silicone. The PCB is entirely hidden, but some of the tops of the surface mount components protrude partially through the coating The CPU (that is the earlier version 801) just happens to have a white-colored package casing, by coincidence.

 

Gotcha, thanks...

 

That might make it a tad easier, 40 pin single chip sounds like 1802 family. The photo on the ebay offer Peter referred to shows a 24 pin chip though, looks like 89711? Can you read the chip number on the boards you have?

Ah, i don't trust workshop manuals I think you get more precise data when profiling the actual units, the manual shows how it should be but the unit will reflect how it actually has been implemented. Good info though, i'll have a look at the Mondial workshop manual to get acquainted with the curves.

Yes, that sounds familiar. I don't except the RCA having an ADC so there must be another op amp somewhere, probably doing a ramp compare with the APS output. A replacement APS could be directly adapted to that op amp if the original LVDT APS is broken, might not be too difficult to do once you created a pressure to voltage chart of the original LVDT.

 

Oops....my mistake....thinking of something else! Told you my memory is fading It IS the 89711 chip, 24-pin.

Correct ---- there are a couple more discrete amps on the board

For reference on the CPUs used, I dug up these old close-up pics I took ---- first one is the earlier version CPU (and entire unit), second one is the later version. In both, you can more clearly see the potting scheme --- white vs. black.
Image Unavailable, Please Login
Image Unavailable, Please Login

 

Very interesting. When I spoke to Ted Wentz about my QV engine and more HP he said don't bother unless you rework the entire motor as Ferrari pretty much did everything they could at the time. Why would they leave HP on the table?