I don’t have my car here. Does anyone know the model number for the Zytek ecu in an XJ-220?
Does anyone know if the XJ-R used the same Zytec ecu?
1 Like
Hi Steven don’t think there is a ‘model number’ … and NO the XJR-S and the Xj220 systems are not interchangeable…
Steve I think you best bet LOOOONG SHOT is to try to call GARY BARTLET.
If your in the states…If you don’t have his number pm me……
Godfrey is correct, I passed on a local xj220 which was NOT up to par.
Any simple situation will be a BIG hassle.
gtjoey1314
Thanks to both of you. I’m working on a possible solution with a gentleman at Zytek who started at TWR as the xj220 was being produced and has a colleague who is the last unretired person to work on the Zytek units. He may have a solution.
I did find out that GC performance in Tennessee, where Gary Bartlett’s car was serviced by Gail (the dad), is still working on xj220’s under his son. Waiting for a callback, see if they have the diagnostic kit.
You are correct. 2 different ECUs for the 220 and the XJR-S.
xj220: JaguarSport p/n P220-P-81-001
Zytek p/n 20/015/0001
XJS-R: Jaguarsport: SPD 1122
Zytek: 20/007/0006
While all this detective work is interesting, I do have businesses to run, so if it just doesn’t work out I’ll swap to a Motec or Link or similar system, and get an anti-knock sensor, air/fuel ratio, egt, adjustable boost, multiple maps,and a few other things the Zytek doesn’t have, so I don’t blow up the motor with a bad tank of gas.
Cheers guys!
2 Likes
Found the 2nd Zytek interface for the xj220. It’s at a shop in the US, and they are happy to service xj220s, but they have a 5 month waiting list to get in (they do all Jaguars). They have been servicing xj220s for many years. They are running the Zytek software on a modern laptop in emulation mode and it’s working fine, so no need to hunt through garage sales for a 1992 DOS/Win3.1 laptop.
There may be another interface for sale soon. Original purchase price back in the late 90’s was $75,000 so I have no idea what it will go for, if it ever comes to market.
If you need the shop details PM me.
maybe a MegaSquirt system , they do all that the 220 would need and great followup tuning/and upgrade over the older Zytek!
and about 75% cheaper also parts/sensors etc, easily obtained ,low costs!
ron just an opinion!
I thought a few representative pics would be nice for future reference.
XJ220 ECM Markings.
XJRS Markings:
Note this one is SPD1002. and is labeled Zytek Type R1-0-12
(Oddly I can’t find an reference to SPD1122 or Zytek: 20/007/0006) Dr. Solly, can you say more about that PN?
Edit. Nevermind I see it on this Facelift XJRS ECM’s label:
I can’t make out the SPD part for 100% certain though.
I haven’t found the SPD1122 PN in the XJRS parts list either making me think the image is just too blurry to properly read it.
R1-0-12 is apparently the same thing as part SPD1162 identified as existing in a 1992 XJRS. (Looking a the board markings -on another forum - it’s clear the board numbers inside are the same between these two units regardless if there are PROM updates or not. The SPD1162 ECM is listed as a component of the JS1 Equipment Kit that included all the diag stuff.
<10/13/19 Edit:>
Jaguar EPC list these XJRS ECM PNs.
Under Client Code 24 (whatever that means)
SPD1161 for all countries (USA not listed)
For another variant, this time some shipping to the USA:
SPD1162 (USA)
SPD1140 (Europe)
SPD1132 (Europe from VIN 179737)
This unit uses a Hitachi Motorola Clone CPU. AN HD6303X - pretty common processor that’s well documented and still available for projects these days. I should note that the totally hackable re-programmable Lucas 14CUX (software available) uses a variant of this same chip.
Now interestingly, we know that some DB7’s that used Zytek can use the same diagnostic box, but it’s quite possible? Maybe? that Rolls Royce/Bentley could use it too? (Given the Zytek part number similarity).
1997 Bentley Turbo part:
Zytek PN 20/021/0007A
Aston Martin DB7 i6 ECM:
Zytek ECM SPD5333 (various issued versions?) Maybe SPD5489
Zytek PN 20/020/0003E / XYRATEX PN 06H 2995
MAF based so no MAP port.
Also, Janspeed 6cyl Turbo XJR’s used a modified XJRS V12 ECM (part number unknown)
For further interest, the guts of the XJRS ECM look like this:
Note the 12 obvious MOSFETS for the Injectors, and the 13’th & 14’th odd man out for some other purpose.
I like Megasquirt as much as the next guy, and although it’s easier than this ECM for sure, I think this ECM STILL has it beat for sophistication given the number of VE & advance tables it has + diag code output. Plus it has enough MOSFETS for the Sequential EFI if that’s considered important.
A Closer Look:
And one brave XJRS owner got creative, de-soldered his PROMS and ECM and did a data dump to file for analysis (he didn’t get far, but he discovered some tables. If you speak Japanese, that blog is here:
It suggest that with the right BIN files & format files, the data from the ECM could be read live in a standard Tuner program built for talking to different ECMs like Tuner Pro RT or other software package (free).
One Last Look:
Eventually maybe enough documentation will come together so that owners of cars w Zytek’s made in the late 80s to late 90’s can actually do diagnostics on their cars. It’s a tall wish but maybe eventually.
BTW, does ANYONE know what else a PM4 connector is called or who made it? What catalog it might be in? I’d like to find the pinout (what position is pin 1-9 by drawing). That would help w/ replicating the data connector cable at least. Interestingly the PM4 connector wasn’t JUST used as the diagnostic connector on the XJRS, it was used on the XJS from at least 1992, and the XJ40 as the diagnostic connector. One would think it would be available.
Dr. Solly, you uploaded these on another forum that I haven’t had a chance to reply to yet:
Diag Box Side:
Car Side:
Edit: I answered my own question.
What a friggin PITA! I had to work this backwards from a similar 3 pin connector, figure out it’s product group, and then it’s alternative name. Eventually found it on TE connectors, and I found out that this OBSOLETE series of connectors is called: “Positive Mate Mark Iv Connectors And Contacts”
I got lucky with an RFP place that listed a green 9 pin assembly and it’s PN. Well that PN didn’t show up at TE’s website NOR did it show up when doing a PN search, but that 5Pin connector of the same series did. Looking at THAT connector’s URL for the datasheet PDF I figured I’d try to hack it by changing the PN in the URL. I got damned lucky
Note car black connector mating PN is 343066 (no pdf for that one)
And THERE IT IS. The pinout we need:
Still a little unsure about orientation. I presume the black connector above is for the 220, not the XJRS, so maybe a little different pinout than given in the XJRS S80 datasheet.
So from the green connector we know that #5 is used. Connector 5 in XJRS datasheet leads to an ECM pin described as a 5V input (despite the fact that the illustration shows it connecting to the grounded shield wires on 7 & 9.
That means the pinout (which should be confirmable by wire colors going in should be:
So
1 = +12V (Red/White line on an XJRS black connector)
5 = 5V Signal/Ground (confused on this one - but on a red wire - XJRS)
7 = Ground (black wire XJRS)
9 = TX/RX 0.670V (green wire XJRS)
That leaves the only other direct to ECM possible communication wire #6 (5V input also but unshielded) left with nothing talking to it. But that’s probably ok. It may be used w/ a different connector for reprogramming or ? Don’t know…
So really (as expected) only pin 9 and maybe pin 5 would need to be watched by oscilloscope during communication to reverse engineer the protocol and box.
(Borrowing your pic from the JEC forums: Zytek ECU Tuning/Diagnostics - www.jec.org.uk)
240° 5 pin din connectors on side of box:
Next it would be nice to know for sure what the pin to pin layout looks like for the 240° 5 pin din put to 9 pin RS232 port looks like (and if there are any resistors or crossovers in that). As well as the 240° 5 pin din to PM4 plug pinout
Din plug numbering convention assuming Zytek used it:
Serial port pinout
~Paul Kobres
I will never work on one of these–not out of revulsion; just that I am no longer all that interested in cars–but, this is some excellent documentation!
Thanks for doing it.
3 Likes
Great job Paul. But this is so far out of my league I don’t understand a lot of it. Biochemistry was easier.
Taking this from your above post:
1 = +12V (Red/White line on an XJRS black connector)
5 = 5V Signal/Ground (confused on this one - but on a red wire - XJRS)
7 = Ground (black wire XJRS)
9 = TX/RX 0.670V (green wire XJRS)
I always thought the whole point of serial connectors were that they were full duplex, send and receive at same time. You have pin 9 both transmitting and receiving; wouldn’t that be half duplex? We know there is one ground (#7), and at least one power wire (#1). Could #5 and #7 be one for TX and one for RX?
It would be a good idea to centralize this discussion. I know we are talking about it on the UK forum, and it looks like on 2 forums here. Pick any one and copy all the info there?
I kind of wish I’d done biochem. I did developmental biology as a grad student and made it through Organic chem III. This electronic stuff is just stuff I’ve either had to learn or learned to put up with a hobby, and I’m no expert or even proficient either, but I get a few things, and I think I’m decent w/ Jag wiring diagrams. A LOT of it is over my head functionally too… I get just enough of it to go Ah! I think I’ve seen similar on vids/doc sites researching my own stuff… Ok… I think I see how this can be done.
I wish I had the XJ220 EFI diagram to compare. That would tell us a bit more about the pinouts.
Yes #5 could be a signal as well. The documentation vs illustration is confusing in XJRS pub S80. I thought #6 (slate/orange) was going to be one of the RX wires. I can’t eek enough logic out of what I know/have read to discern any further at this time.
Two thoughts on the full/half duplex thing. You may be right about #5 and that’s how RX/TX is achieved. Only an oscilloscope would tell us for sure. For an XJRS computer mostly I would expect it to be transmitting (so that inputs can be viewed and adjusted). Although RX would be needed for code clears.
As far as I can see, at least some GM OBD1 units appear to be half duplex combining the TX/RX wire. Here’s one of the familiar-ish ALDL to serial adapter diagrams. Straightforward. ALDL signal (RX/TX) on the left— separate serial RX & TX lines on the right.
Huh. and this indicates that at least as of 1986 the GM protocol was half duplex: https://en.wikipedia.org/wiki/On-board_diagnostics
I know a bit about the GM protocol from owing an old GM truck & looking into the E&C bus for some hobby upgrades & having to do a bit of diagnosis now and again. Interestingly some Lotus owners benefit from this and created their own Lotus tuning software for the cars that used GM EFI.
Whether the Zytek expects to communicate like that… don’t know. It may be a communication mode. Is it possible that the advanced version of the software (the one that tweaks fuel maps) requires another wire utilized? It’s also possible that the proms themselves are the UV Erase/Exterior programming type and have to be physically changed out to adjust. The microcontroller should be electronically rewritable though. I really need to figure out how to create an account on the website that host Jack’s Japanese XJRS blog and see if he’ll send the data dump files (probably DAT or BIN files) of the proms & microcontroller to me. I probably cant’ make it much farther than he did w/ a disassembler, but they’ll be on hand for others to take a crack at.
I agree central conversation would be good, but which forum? This one host images more easily as far as I can see. It’s also the grand-daddy forum - but maybe not as widely used in the UK? Or by official / officially connected Jag people?. But then again folks look for info on half a dozen very good Jag forums. Jag-Lovers (the granddad), JaguarForums, JaguarForum, JEC forum, Facebook (terrible to search)., and maybe specialty XJRS and XJ220 forums. Minimally crosslinking helps so that folks can flip back and forth and piece conversations together. I’ll try to duplicate as best I can, and I’d probably pick this forum myself (it’s where I started back before 2000 when was an email forum) if I were forced to choose - plus it has better picture handling-. A wide audience makes for the best input though. Minimally I think it’s critical to make sure any future audience has all the relevant forum links to follow, and to try to duplicate on the forums any critical information found on blogs/commercial websites that’s likely to disappear over time.
i.e. the Jec forum discussion: https://jec.org.uk/forums/viewtopic.php?f=40&t=9616&sid=dbe5ff63fac1e9962dbd5388ee303b8e
Hopefully at some future point, the bottom line (the protocol and a DIY circuit board + USB interface) can be documented on all relevant forums!
1 Like
All this is clever… and interesting BUT without the technical service manuals to decipher the information provided by the Zytek system… you are no further ahead.
No further ahead? Sort of. To someone looking for the complete info we don’t have it yet.
I’m going to assume we might never get our hands on any internal Zytek stuff about the protocols those ECMs used. That doesn’t mean we’re lost, or not laying ground work to get ahead even if it’s several years in the future (there’s great info on these forums going back over a decade).
Dr. Solly will relate to this. It’s like studying a biological system. You may be looking at a certain phenomenon and not know how it works mechanistically. So you start poking around the edges. Observe behavior. Take samples. Analyze those samples. (maybe a protein which you work to isolate). You sequence part of the protein and work that back into a possible cDNA sequence you can check against RNA produced in a particular part of the organism. You use your partial sequence to amplify the RNA (protein instructions set).out of your tissue sample, and get it’s full sequence (which you can do) which tells you the full protein sequence. You can then take that and compare it against other proteins in a database and infer possible function, test that… etc etc etc.
Where we’re at now is getting an idea of how the cables are constructed so that can be replicated (maybe at some point into a future USB dongle). As far as communication goes, the computer-side serial protocol is known. We know from instructions/screenshots in the manual what the Zytek software displays and asks for. For example, if one sends the clear codes signal out of the software, that ought to be a pretty clear consistent instruction that can be seen on an Oscilloscope. Someone versed in ECUs tuning and repair & protocols (they’re out there) might get one of these cars at some point, or have a buddy with one) and might be willing to start poking if they could get their hands on the interpretation device. The interpretation device is likely setting voltage levels, maybe inverting signals coming from the serial port. Some modern computerized oscilloscopes are even able to recognize and auto-decode well known protocols.
There’s a dump file of the contents of some of the PROMS on an XJRS out there assuming the guy who pulled them off the chips still has a copy). It’d be nice to archive that. It might be possible to work that backwards into a “Mask” file. What I mean by that is in programs like Tuner Pro RT (just getting into it myself) talking to different ECMs requires a translation mask so that the program knows what to do with and how to display live data. Tremendous work has been done for a huge variety of GM ECMs and other mfgrs ECMs. That means the expertise to do such a thing exist. Maybe it’s a matter of time before that expertise get applied to a Zytek ECU.
Moving ahead sometimes happens like the turtle, but a lot of times it happens like the rabbit… lots of work that takes time to gel before it suddenly comes together.
We need to check with Jaguar Heritage Trust her in the USA as well. George C seems to believe that they’ve got a lot of Zytek documentation… but he may be thinking of the WDS system. I hope he’s right though and that they have most of the Zytek diag floppies and not just Genrad WDS stuff.
~Paul K
2 Likes
Rimmer Brothers in the UK offers the complete set of xj220 service, diagnostic and information manuals on a USB drive. Just ordered one, will take 6-8 weeks to get here but then I’ll have the codes. Also saw a screenshot of a diagnostic unit being used, output was in plain English. Different than an OBD1 scan tool which only gives you codes. Engine RPM, temperatures at all the sensors, vacuum, battery voltage, etc. I was able to diagnose his fault just from reading the screenshots.
1 Like
Yep… that sounds like what a tuner software woud display, whether the Zytek specific stuff, the Tuner Studio whatever for the Megasquirt and other DIY ECUs, or like what I have, Tuner Pro RT (probably what the megasquirt version is based off of) that can talk to just about ANY ecu given enough parameters are set for it to interpret the communication. The Zytek software could be ditched for some standard software that everyone knows if that were known, and even be able to do so wirelessly via a bluetooth to RS232 adapter.
Who would have thought 20 years ago that someone would be able to display OBD1 GM computer parameters LIVE wirelessly!? There’s hope.
~Paul K
1 Like
I fully believe that your and others’ work will solve this apparently intractable issue: were I a betting man, I’d wager you, and others on the case, will successfully reverse engineer this issue.
Well Done! I wasn’t aware of these… G
This was a Zytek unit in an XJRS that the owner managed to assemble a kit for, and somehow got superuser software. It’s on the UK forum at https://jec.org.uk/forums/viewtopic.php?t=9616
The pictures were very clear a week ago. Now they have Photobucket stamped all over, but you’ll get the gist.
BTW, I was one of the dumb ones at Columbia, so I became a surgeon. The really smart people went into research:) Don’t know if you’ve heard the old saying: Internists know everything and do nothing. Psychiatrists know nothing and do nothing. Surgeons know nothing and do everything. Pathologists know everything and do everything, a day too late.
2 Likes
LOL! I hadn’t heard that one. IDK…depending on your grad school mentor that Ph.D. to be a researcher can seem a LONG way off even if you have the hours. There’s a little bit of luck involved in that.
I hate Photobucket too. What a bait and betray that was. Glad I never used it but it was so integral to forums that trusted it that many many informative images are now unviewable. Now forums host their own images (it’s become cheaper) and other services have popped up. It’s going to be too late for photobucket after taking everyone’s pictures hostage though…
Re. the comp w/ the superuser software… THAT is the situation in which one would want to clone the hard drive to an image file or a usable VHD (virtual hard drive) since one probably isn’t getting access to the install floppies again. Yeah that’d be a nice image to play around with. Probably EVERYTHING you ever want to know about the ECM’s datatables are encoded in that software somewhere.
Oh, if this problem gets licked, and folks DO turn to something like Tuner Pro RT, then something like this becomes possible to use on the XJRS & XJ220
Just drivin’round… in the 220… payin’ too much attention to the phone for all the right reasons…
~Paul K.
1 Like
Paul- Do you know if the early 90’s Zytek ecus used ROM or EPROM? If it’s an EPROM chip (which by 1992 I suspect it is, EEPROM would be too much to ask for) can’t it be removed from the ecu and read by other software like your tuning software? Then buy a pack of EPROMs and a burner, burn either the original map or a tuned version with a remap. This would avoid the whole communication problem via the interface. Just a thought, don’t know if it would work.
My other thought is a piggyback ecu, but again if that interface is necessary for communication with the Zytek ecu the piggyback wouldn’t be able to “talk” to the Zytek ecu without a permanently mounted interface box. We really need to find out what is in that box. Since they were all the same there should be a number of them in the UK, and some of the interfaces may not work any more and we could get one of those and dissect it. I’m going to start asking around and see if I can find one.
2 Likes
I don’t know yet. The Facelift JagSport cars’s ECM had at least 3 types of memory:
-
There’s processor memory on the Hitachi HD6303X CPU (datasheet available) , - where the core program resides - and presumptively where the “DTC” codes get written given the Japanese XJRS owner/ECU hacker did not see them written to prom.
-
At lest 2 PROMS (I haven’t IDed yet) EPROMS or - as described on the Japanese Blog, EEPROMS which probably contain data tables / ve tables / spark correction tables / temp value lookup tables etc.
The markings on those chips are: DE2864-350 9026 - with a white sticker noting their positions as IC13, and IC 14 but possibly obscuring other information. ) -
And a Hitachi HM62256 CMOS SRAM (static ram - closely related datasheet available-) chip for low power storage fo rewritable memory. --also possibly where codes are written.
I have written to the Japanese XJRS’s owner via that blog interface asking if it’s possible he still has the dump files from the PROM and ECM and if he’d be willing to share. Don’t’ know if I’ll ever hear anything as it looks like the blog was abandoned in 2016.
One thing that has become clear is that the TWR XJRS cars 88-89 during their last years experimenting w/ using the Zytek ECU used a different computer than the JagSport cars did. A wiring diagram for that has just been published on the Jaguar XJRS owners Facebook Forums and there are some clear differences including the diagnostic connector (which might just need a different dongle to work). But pin 7 is ground on one and signal on another:
I don’t’ know if they’ll get around to it, but I’ve asked a couple of the XJRS owners if the need to send their CPUs off to Zytek for a fix involving the fuel pump, if they’d mind getting some high resolution shots of the guts so that the chips inside can be identified.
Getting a dead interface box with which to play 19th’ century doctor would be great!
EDIT: Just found this.
Some time back Peter Fritz “MauiHoss” noted he was able to ACQUIRE/PAY FOR the diagnostic equip from Zytek directly, along with getting his ECM checked & gives contact info:
End Edit
Just so it’s here, this is the best I’ve got At The Moment, for the older TWR Zytek wiring:
EDIT:
The chips marked DE2864-350 9026 are Timer E2 64K Electrically Erasable PROMs
I found the closely related PDFs here for the series: DE2864 datasheet & application notes - Datasheet Archive
The base number is “DE2864” “350” represents the minimal nanoseconds for “read cycle time,” I.e these weren’t the most expensive chips in the series. I’m not the same board slot could accommodate the DE28C256 larger memory chips even though their both 28pin DIPs. The pinout is close but… might not work.
This was a hard to find datasheet. I’m not sure how hard it would be to obtain a supply of these chips… OR… maybe… a way to slap a modern EEPROM in such that the pinout still works.
To get an idea of possible variations of the chip these are series part numbers:
DE2817A150
DE2817A200
DE2817A250
DE2817A300 Timer E2 16K Electrically Erasable PROMs
DE2817A350
DE2818C1072AS1R0N
DE2864-200
DE2864-250
DE2864300 Timer E2 64K Electrically Erasable PROMs
DE2864450
DE2864450
DE2864A150
DE2864A250
DE2864H250
DE28C256200
DE28C256250 Timer E2 256K Electrically Erasable PROM
DE28C25625D
DE28C256300
DE28C256350
DE28C256A200
DE28C256A250
DE28C256A350
1 Like