JP86SS

iTrader: (1)

I've always had an issue with fouling plugs whever "Limp" mode comes into play. When I changed to the MSD ignition the problem got better but still exists. I'm sure allot of guys have the same issue running high fuel pressure and larger injectors. Car runs great on the code but is painfull when limp mode takes over.
If we knew which pin has the fuel control we could adjust this and possibly have a good running limp mode with modified motors if it was ever needed.

If anybody has a memcal laying around they could measure the resistance on and send the info to me, I'll make up one big spreadsheet showing the values.

I made a sheet for putting values in so the comparisons will be easier to be consistent. There are spots to measure across and pins next to each other.

Please indicate:
The application (year, model, Auto/manual, MAF/SD, etc)
Broadcast Code (AXXC, ...)

Download the XLS file and fill in the values,
Send back to Johnshome ---AT--- WoWway.com
Throw up a post here so everyone can see which memcals have already been added so people don't waste time duplicating efforts.

I'm not saying we will figure this out, but at least som more information will be available to try.
TIA.
Jp

scuzz

I can't seem to open the zip file.
What pins on memcal do I try to read. I've got a couple different memcals I'll read for you.
Ron

JP86SS

iTrader: (1)

Sumthin' wierd with the file.
If I open it, nothing there.
If I save it and then open it, it works ??

I was going across from side to side, then from each pin to the one next to it.
The part on top with the OEM prom is not needed, just the resistor sections on the 14 pin and 16 pin chips.
thanks,

If the file keeps messing up shoot me an email and I'll send it over.

jwscab

I can't seem to open the file. I have the lotus document with what I beleive is a prototype memcal, using discrete components, I think i pulled it from DIY-EFI awhile ago. I can send it to you, just PM me.

I've looked into trying to replicate the r-packs in order to get around having to buy the now ridiculously expensive stock memcals, and figuring out something else for knock detection. I will tell you this. They are an extremely complicated network to try and solved based on measured resistances.

I tried to find the inner workings of the redundant fuel device chip so that you can figure out what each node does. I also have a pretty neat document about that, but it's on a different PC. I did not have any luck getting any chip info however, so I think i left it at that. too many simlutaneous equations for my simple brain.

the resistors are part of a bunch of op-amp circuits, both inverted and non-inverted, so it's tough to tell what is a divider circuit vs. what is part of a feedback network for an op-amp.

I have some of the info you asked for, again on a different PC. if i find it in a reasonable time, I'll forward it to you.

jwscab

here is a link with some pertinent info:

http://www.iroczone.com/resnet165.htm

JP86SS

iTrader: (1)

I've made a new zip to see if it works for you guys.

s24a

Still doesn't work for me! Says the ZIP file is corrupt.
What ZIP program are you using (incl. Version) to ZIP this file?
Is the file inside an EXCEL file?

I am using WinZIP to try and extract, as well as Total Commander, and both report that the archive is corrupt!

jwscab

Yeah, sorry, I am getting the same thing. Still says corrupt. If you want me to try, PM me and email it to me, and I'll try and stick it on there.

RBob

iTrader: (1)

Need to use FireFox to download the zip file. For some reason IE can't be used on TGO for this anymore.

Note that pin 56 on the MEMCAL is the cylinder select. Ground for an 8 cylinder. It is a voltage level output. Which sets up the injector firing rate.

One thing that can be done to help decipher the pins into the RFD: ECM on test bench, scope on injector output for PW, remove EPROM. Place various resistor values between ground and an RFD pin, then between VCC and an RFD pin. See what it does to the PW.

Also vary the CTS , MAP & TPS during this. Some inputs will change the position of the knee in the PW curve.

RBob.

JP86SS

iTrader: (1)

The spreadsheet was just indicating measuring across and next to each other.
The diagram on the 165 page looks more like what I should be doing.
It indicated that measurements were taken from every pin and referencing all the others. Probably will take all those measurements to complete the puzzle.
My sheet was allot less.
Here's a pic of it.

I will eventually do as you suggest RBob, Just wanted to try and narrow down the items if possible.
If things only change in certain areas then its best to concentrate on those first.

On a side note, does anyone have any info on the RFD?
(Redundant Fuel device)
If we know where power and ground was, it would be a starting point.

vernw

iTrader: (2)

The links are downloading just fine for me using a Dell 600 laptop, IE 6 and XP SP2. Did a download and save to hard drive, then opened it with winzip 8.1 R1

RBob

iTrader: (1)

That is good to know. Maybe it is IE7 that has an issue. I know that one of them does. I too am using winzip 8.1 r1, seems to have been the last good release of it.

RBob.

jwscab

well i decided to scrap one of my many non-v8 memcals for the greater good. I'm in the process of pulling the netres packs off the memcal, and managed to pretty cleanly snap the epoxy off of the top of the packages. They are thick film deposited resistors that are laser trimmed, so I may ruin the actual values, but I will be able to get the schematic for what is going on internally. The ceramic substrate appears to have the resistors on top and bottom, so I have that challenge ahead.

this is a 1987 memcal, ASMZ BCC, which should be a 2.8L auto v-6 car.

let you know what i find......

jwscab

I've attached what the resnet's look like without the epoxy molding. From this, you can figure out which pins to measure to get the value of each resistor. Now you can build a spreadsheet to have people measure specific pins for values.

As Rbob said, you can play with the values to see how the PW changes on a test bench from here.

jwscab

Oh, some clarification. There are 2 pinouts per package, one on the top side, and one on the bottom. If you mentally overlay the two views, this is what the network looks like physically. Each pin of the package connects to the top and bottom layers.

s24a

Which side is "top" and which is "bottom"?
Reason for asking is that the way you have the pins labeled, they are both "top" views. Look at your "1" and "14" as an example.

If I flip the package mentally, then "14" would be at the top left....

Please clarify.

Thanks!

JP86SS

iTrader: (1)

Thanks Joe,
That will help this immensely! (sp?)
OK, allot anyway

jwscab

No prob, fun project to try and get the things out of the packages....worked easier than I had hoped.


s24a, not flipped, overlayed. they are all top views. I have pin 1 and pin 14 labeled. the reason it was done that way was that I flipped the second view of each chip once I traced it out so I wouldn't confuse myself while tracing it out.

s24a

OK...So not flipped.

SO...is the view on the LEFT looking down at the top, and the view on the RIGHT looking THROUGH the IC with xray vision to the resistors on the bottom side?

jwscab

yes, exactly.

although technically not an IC, it's a ceramic substrate with printing on both sides. Left view, top of substrate. Right view 'x-ray' of bottom of substrate.

As mentioned previously, each pin ties the top to the bottom of the assembly, so pin 1 on left view is the same physical pin 1 on the right view.

JP86SS

iTrader: (1)

If I understand correctly, they should look like this
They are in the memcal with the 14 pin chip at pin 33 of the memcal.
Chip select memcal pin#56 is at pin 13 on the 16 pin chip.

I've got some measuring to do in the morning.

Here's some info I figured would be good to link into this thread as well.
https://www.thirdgen.org/forums/diy-...es-values.html


One more thing, Do you have the part numbers off the 14 and 16 pin chips still intact?
That way if we find different patterns we can follow what chips have similar patterns.

jwscab

nice job on the new pics. They are exactly correct.

I have the part numbers, but I have them at work. I'll post them on monday.

JP86SS

iTrader: (1)

I went and measured my butt off today.
I checked every pin to every pin on the AXCR and ANJF memcals that I have.
Both contain the same part number chips but have different resistance values in them. The pattern is not the same as the diagrams posted above (or so it appeared when I was measuring)
Here's a pic of the differences I found.

The Spreadsheet is also attached with all the values obtained.
Rename the PDF to XLS extension.
Figured since guys are having trouble with zip, not to use it.
Jp

jwscab

so are these differences in ohms? meaning, the most some of these are different are less than a hundred ohms?

if so, its probably tolerance on the laser trim. Probably less than 10%.

JP86SS

iTrader: (1)

All the readings are in K ohms.
Most of the readings I took had a 0.1 K (100 ohms error maximum) from one reading to the next on the same pins.
I'm not sure how much tolerance there would be on the actual parts.
I used a conditional format of 0.20K ohms on the color indication so that only the large magnitude changes were highlighted.

jwscab

OK, here is the part number and all the detail I have for this memcal.

prom ID ASMZ

Rpack 14 pin 16060954 93 26
Rpack 16 pin 16060956 93 25

knock circuit 1A3176

1987 2.8L Auto

I took at quick look at your chart, and the one thing that jumped out was the 'short' measurement from Pin's 7 and 9 of the 16 pin netres. It looks like in that cal, they took the TPS signal rather than a MAP signal on CAL61.

the thing to remember is that these resistors values can vary widely, so even if the pinout has a resistor in what I drew, and you measure an open, or short, its probably becuase they are not using a specific function.

there is a link within the link you posted that has a very detailed description of the redundant fuel device. I believe post #16 from Haulna$$. check that out it will make some things much clearer.

I'll take a look at some of the memcals I have stashed, and let you know how they compare.

JP86SS

iTrader: (1)

I've been through that thread several times looking at that link.
"To determine resistor values refer to TSO CT MANUAL $RFDCAL procedure"
Thats the one I'd like to find as well a a schematic to indicate which resistor the author is talking about. They are all numbered but according to what diagram?

jwscab

yeah, that sure is the bugger of it. wish we had the last piece of the puzzle.

I'm going through the memcals I have now.....I went through various ecms I have, including a '165, a '730, a '427(late tbi), 16159280(late tbi, caddy) they all have the 54995 IC which is the redundant fuel chip, so in all cases, the net res packs should all hook up the device exactly the same way. that means they should have the same basic structure as the ones I stripped.

I've thrown the pics into a spreadsheet with a listing of pins to measure each resistor singly. It appears various cals changed the input like I mentioned above, (MAP or TPS), depending on the vehicle.

I beleive some of the cals also change whether they use some of the other functions, like cold start enrichment and etc, from that document, where one cal might have a resistor on the unused pins(from above) that you might have measured on other net res packs. I should be able to pin down some of that stuff in the next week or so, as time permits. I have about 20 memcals to go through, TBI, TPI, MPFI V-6, a nice cross section. eh...better make it 2 weeks....

I'm attaching the .xls sheet, let me know if you can get it.

jwscab

I managed to collect the resistor network values for a handful of different cals that I have. I will post them hopefully tongiht, if I don't get home too late. These include a 4cyl TBI, several V8 cals, and a couple v6 cals (I believe).

I haven't really studied them at all, too busy. there is a schematic on ludis' site of an ecm that has the same RFD device 54995, that uses discrete resistors on its main board for the fueling calibration. I have to dig up the ecm number.......OK. 1228321 is the one.

jwscab

OK...sorry for the delay....check out sheet 2 of the spreadsheet. Sheet one is the schematic and first memcal measured.

jwscab

I decided I'm going to do a layout for a memcal replacement, as a test board, so the resistor packs will become 0805 or 0603 discrete parts, so the values can be tweaked much more easily. I'm going to lay it out so that it would directly replace a memcal for a '165 type ecm, or with the addition of the knock module, a '730 ecm. I plan to add provisions so that you can run an external knock module (jumper resistor) so it will work in either ecm (probably others as well).

I'll let you know when I get it all wrapped up.

jwscab

ok, board is layed out. I'll probably get 9 made. I'll post a snapshot of the layout soon....

JP86SS

iTrader: (1)

Well now, that was quick

jwscab

Yeah, it's sort of what I do as a full time job, so I'm familiar with layout, you could say.....It's not a super cad program, but i really like how they have the PCB prototype service, and the price is nice. I don't like the fact that the trace function doesn't snap to specific angles on corners, so they aren't perfect....

I've attached the panel layout (saves money if you can build more than one board in the area the proto service gives you). This pic is 3 per panel.

it's using 0603 components, you can install a regular prom and knock module if you prefer, or you can use an external knock module by adding a zero ohm resistor and add the external harness and wire the output to the 'knock in' on the ecm harness.

I'll probably order them after new year, since this holiday is so quickly coming up this year, then I'll do a bench check out. I'll have some available, if you want, I just want to recoup the cost of the boards, so it would probably be about 10 bucks each, give or take. Once I get 'em and check 'em out, I'll let you know for sure what they cost me. I'll leave it up to you to add your own resistors!.....

s24a

If its not too much of an issue, if you make a layout for surface mount chip resistors, I think I could get a place to to make those as maybe a better package. Can be double sided, too!
Would still want to keep the external knock as a through-hole jumper......

jwscab

I'm not sure I follow what you are suggesting. The resistors ARE surface mount, 0603 devices. Mounted on top side for ease of access. The board is double sided, I just chose to place very few traces on the bottom(the green lines). The knock option is a zero ohm resistor added or deleted. It's not a modification that I would make as a jumper, since it's either/or, and not really both on the fly. The thing to remember is that this is a pretty tough enviroment, so any chance of failure I would avoid if possible. A jumper could fall off from vibration.

If you are referring to manufacturing, certainly, if you purchase bulk boards, the unit cost goes down, but you are paying more overall. This CAD program offers a pretty cheap proto service, which is why I use them. If I found I needed to make hundreds of these, I would probably switch to some professional software, generate gerbers, and quote a few PCB manufacteres to find the best deal. At that point, they would get silkscreened and soldermasked. These are just a quick turn, non masked proto.

that said, I definitely appreciate the offer to help. If people actually want these, I'd like to get more info, later on....

s24a

OK...

Blew the picture up to larger size and see what you mean.
I was looking at the through-holes that exist in nthat, and did not see the green underside , but now see the 603 chip resistors.....

Anyway, I might still be able to get some made.
I know I definitely would like to get around 10 myself, as I have quite a few MEMCALs with various NteRez DIPs that I could convert (with or without the knock Thick-Film)....

jwscab

Hey, just a quick update......It took longer than I expected to order the boards, with the holidays, and a new lil' baby who was a few weeks early, so he has been occupying most of my time.

the boards are ordered and should be here Monday. I expect to order up some resistors (if I don't have the values at work to borrow), and the rest of the parts (socket, header).

I'll let you know how they look and when i can get to testing them.

jwscab

well, I got the boards in, and they look pretty good. I'm not extremely happy with them becuase of an oversight on my part, but its workable (small interference with a socketed prom and the knock circuit, since the original prom sits below the knock assembly, but all you need to do is just slightly push the knock module over, bending the pins to allow clearance.

ordered the headers and prom sockets, so I'm going to build one up and check for fitment, and hopefully in early feb. fire one up on the bench. I'll post some pictures.

jwscab

well, in case anyone is still paying attention to this thread....I had to end up ordering just about all of the 20+ resistor values because I had none in stock. I got them today, so I'm going to build one up this week. I'll snap a couple photos and attach later....it's annoying because they are like....less than a penny each, but you have to order min quantities and multiple values to make it even worthwhile to ship.

the values I'm going to start with are the default AUJP values I measured.

s24a

I am still watching your progress!
Let us know if your board comes through as transparent to the actual NetRez ICs....Hope so, as I am interested in getting some boards when you feel comfortable with them....

jwscab

here are a couple of pics:

jwscab

and one more, if it lets me post it:

JP86SS

iTrader: (1)

Congrats on the baby, hope all is going well.
Boards look good for first go round. The chip spacing with the KS stuff is just one of those things. Workable so no biggie.
Are you planning to do some PW measuring on the bench to see if you can get some modification to it?
That was my original plan but time and work are keeping me from the bench lately.

jwscab

Thanks for the congrats, Noah and mom are doing awesome. It's really a life changing experience.

Yep, the 'plan' is to build one up and compare it to the stock aujp I have. depends on the time I have, obviously. I'll start with a pin to pin comp, then move to firing up the bench and comparing signals.

otherwise, in the meantime, I'll probably tweak the layout to improve a couple things like the knock module spacing. (easier to watch the lil guy with a laptop, than with a scope and soldering iron!)

I also noticed one thing last night. Cal54, (or pin 54) of the memcal actually has a value tied to ground inside the one IC, however, it doesn't appear to be used in the '7730 schematic. Very interesting, I'm going to try and take a look at other ecms to see if that pin is actually used somewhere.

roarin_mouse

Does anyone have a schematic and/or information on the knock module (for an AUJP or ANHT)? It looks like once this ResNet is tackled, all that remains is the knock module. If someone wanted to use an external one, how would it be done?

jwscab

I don't have a schematic, and I haven't found one. The knock module is a thick film circuit, so you could trace out some of it, but there is an IC on it as well, would need to be studied under a microscope.

I placed a jumper resistor on the board so that you can feed the knock signal like a '165 or '747 ecm, that is, using the external knock module, the black box, about 4" square usually on the firewall.

there is another option, but its a long shot becuase someone would have to design it. there is a sweet programmable IC that you could use along with a small micro or pic to make your own knock module circuit. there are some issues, since the module that we have can stand alone-ie, see knock, generate a pulse. the newer IC's are tied to the firing sequence, and are microprocessor driven, so you would need a pic or micro to simulate the older module. I dunno if that made sense, but it gets technical from there.

884+3

Rbob and another member posted how to use an external knock board with the 7730 in this thread

https://www.thirdgen.org/forums/dfi-...sensor+730+165

I did the conversion several months ago on my iroc but never got the car on the road since. I have done some datalogging while running the car in the garage and everything seems to be as it was with the AUJP memcal and knock board. No ses light with the 43 code either so it seems to pass the knock test fine at key on.

roarin_mouse

Okay, I'm learning how the ESC works but have a question or two. I found data sheets for TI and Intersil's knock sensor processing chip but I think that the method used in the TPIs (all of them) is much simpler. In the newer ICs, the knock signal is detected and the amplitude of the detected signal determines the rate of change in the ESCs integrator. These ICs use the ECM to tell it when to integrate and when to hold. I assume that the integration period starts when an ignition pulse is sent to the coil.

The ESC used in the TPIs don't have an input from the ECM. It looks like (and please correct me if I am mistaken) there is nothing more than a detector and an integrator. If this is true, than this could easily be implemented on an aftermarket MEMCAL board. This, in conjunction with the resistor networks, would essentially give one everything needed to create MEMCAL board asemblies that work like factory.

Any thoughts? Anyone?

jwscab

yep, the newer IC's are fully programmable, and are processor driven to coincide with spark events. I referenced this above, it would need to be built up with a small micro that would run the integrator much faster then the highest rpm the engine would run, and send out a pulse for a knock event to mimic the simpler circuit. You could build a simpler circuit, but you would need to know the dynamics of each knock sensor, gain, bandwidth, etc etc etc.....the nice thing about the IC is all that is already done, AND it's temperature stable, as designed.