Hello again...

I was sitting through an otherwise uninteresting electronics lecture yesterday, when I thought "Why don't I apply some of my knowledge to this PMD problem?"

I have read somewhere on this board that others, too, have considered making their own PMD's.

I think it was QM who suggested getting the schematic of the original device, and reverse engineering it.

I am sure Stanadyne has this information under lock and key, however each and every one of us has access to the final product.

I am wondering if there is any way to seperate the components from the potting compound?

Perhaps using some sort of thinner to disolve the resin, and free the PCB from its casing. (although that would most likely destroy the markings on resistors, diodes, and caps)

Is there a way to distribute the load on the transistors over a larger circuit?

I am going to take a trip down to the local diesel injection pump shop to see if they have inoperative PMD's I can disect.

If anyone else has dead PMD they are willing to disect and post pictures of, I am sure it would help the process.

Maybe we could all perform our own "autopsies" and compare our results for analytical purposes.

I find it very difficult to believe that with the advances in electronics since 1995, we are still fighting this problem by replacing a broken unit with an equally inferior unit.

However that does seem to be the case.

The 6.5 overall has a lot of issues to be dealt with, but I think they can be resolved on an end user level, by the individual.

It does make sense, from an economical standpoint, that industry has made no significant attempt to remedy the problem.

But it is my belief that with the resources this board has, both intellectual and physical, a solution can be derived.

Ok, enough motivational talk... who has some ideas?

I would like to look at a distribution of the load over a large4r circuit.

Anyone with any thoughts on this is encouraged to give their opinion.

www.oplnk.net/~vittelusa/pmd.html (http://www.oplnk.net/~vittelusa/pmd.html) Has 1.3 megapixel pics of a PMD opened up.


mdhorban has also posted about mapping the PMD curcuit and making a better version. Search for his FSD/PMD posts or PM him.

I have opened ( destroyed 2 PMD's). One I used a torch on and melted the epoxy, and the other I drilled. In both cases, the results left the box unindentifiable. I have scoped the inputs and outputs to the PMD at idle and found the FS signal (pin B) has a 12v pulse of approximately 5 ms, followed by a 7ms 0v then a -80v spike. My next project is to measure the current flow from the PMD/FSD to the FS and see if 500watt transistors are really required. I agree we need to find a better FSD and I am willing to help.


Ray

By the way, the PMD is actually 2 PC boards back to back with the power transistor leads going through both. This will make decipering the schematic even more difficult.

Where did you get your info, hrj?

The PMD is not just an amplifier, it also plays a role in calculating Closure Time.

http://www.oplnk.net/~vittelusa/pmd.html

Here are some pics that QM informed me of. It looks to me like quite a few of the circuits have been destroyed during the "stripping" process. I am looking at it right now trying to make a sensible diagram of it. It may take a while though. Anyone and everyone is encouraged to interpret this on their own. This should give us good grounds for comparisson. Well, I am off to get started on this project, wish me luck.

gmctd, by info I guess you mean the signals to and from the FSD? I have an oscilloscope I attached to the FSD while the truck was running. Only way to find out what is going on for sure.

Wondering about the two back-to-back pcb's in the FSD, hrj

http://www.dieselplace.com/forum/uploads/5Z3_Imag0005.jpgFailed Edited by: hrjack99

I tried to add a picture of the two boards, but I guess I do not know how to post a jpg. If you know how, I will post a picture showing the upper and lower board seperated by a small amount of space.

Scroll back up to a post by quantum mechanic, click on PM to send him a private mail - he knows how to show and tell.


Me - I'm computer illiterate, don't know how.http://www.dieselplace.com/forum/smileys/Confused.gif


That picture would be very illuminating, to say the least......Edited by: gmctd

gmctd, let me know if that pic is good enough.

Good shot, hrj - show and tell rules!


The single pcb is a high-pressure bonded two layer assembly, and what you've somehow managed to do is separate the two layers.


The lower mask peeled away intact on mine.


Check mdhorban's post\thread and the Houston website in quantum mech's post for further pics and infoEdited by: gmctd

I really do not think we will ever be able to duplicate the FSD as is. Maybe we should just try to duplicate the output signal, given the input signals.

I personally do not believe the FSD is a bad design - I've been running a used 'repair' on my truck since late summer '01, mounted on the injection pump, over 40k miles.


Empirical data points toward some quality control issues, but not bad design.


Only posts are from the failures- no need to post about the successes.


If the FSD was as bad as people think, we'd have to 'take a number' to post a complaint on any Diesel forum.

Well i'm finally able to post. (Damn PC.) The PMD board (Only one that I've found) Is a double sided board. The components are all discrete. By this I mean there is no programmable micro's or flash memory. The 5 major IC's (Yah only 5!) are as follows:


HC132A High-Speed Schmitt trigger http://www.toshiba.com/taec/components/Datasheet/TC74HC132DS.pdf


CD4013B or 14013B Flip-Flop http://rocky.digikey.com/WebLib/On-Semi/Web%20Data/MC14013B-D.pdf (http://rocky.digikey.com/WebLib/On-Semi/Web%20Data/MC14013B-D.pdf)


LM285 Micropower voltage ref. http://rocky.digikey.com/WebLib/Texas%20Instruments/Web%20data/LM285-2.5_385-2.5_385B-2.5.pdf


372 (I beleive is a LM372 OP Amp) http://rocky.digikey.com/WebLib/Texas%20Instruments/Web%20data/TLV2370,71,72,73,74,75.pdf


SAS230 or 72282 Unknown at this time. May be able to guess when I finish circuit layout or may need to sacrifice another pmd.


There are other components as well (Transistors, caps,resistors, VR, etc..) But these are the only major IC's.


As TDG has stated the PMD sends a signal back to the ECM for closure time. This is what has my attention at the moment.


hrjack99 has been in PM with me and has helped a lot with scoping out the signals as my truck would not run long enough at the time to do so. hrjack99 has notated that a -80v pulse is on the line to the solenoid. This I am absolutely sure is a "flyback" from the solenoid closing. Now the question that I am working on is if they are using this for reference for closure time or do they use a current calc. Just guessing at the moment as I am working on this from both sides. (Circuit and signal) I have almost completed my layout of the PCB and will then try to exstrapulate what is exactly going on. Even though in electronic terms the circuit is simple it is still a very time consuming process. Well gotta get back to work. If I didn't have so many other things going I probably would have this damn thing done by now!

Here is a shot of the layout so far. No values as most everything I have is hand written. I had it layed out more but had a drive failure and had to start again from a previous PCB. Red is top traces and blue is the bottom. It is not exact as dimensions (Width, Size) but traces arehttp://www.dieselplace.com/forum/uploads/mdhorban/2004-10-05_171531_PMD.gif Edited by: mdhorban

P.S. After looking at the pics of hrjack99. I think what you are seeing is just the imprint of the one board on the epoxy. This happened to me but if you dissolve it entirely you will see there is only one board. At least with the 2 that I have de-capsulated. I'll get some pics up.

Are you interpreting the emitter resistors wattage from physical size, or were your markings intact?


Excellent job, btw.....

WOW


Nice Job

Are you interpreting the emitter resistors wattage from physical size, or were your markings intact?


Excellent job, btw.....





Yes. Those are marked correct. There is a few things missing in this layout but I will have them back in and marked with values.


DALE LVR-5 .1 ohm 1% 5 watt resistors http://www.vishay.com/docs/30206/lvr.pdf


Thanks! I've been busy and haven't been able to post lately but I am working on it!

P.S. If anyone would like the actual PCB file I will post that also.(When complete) Everything will be posted when I'm done.(Even my version) I used Eagle Layout editor for this. I usually use Protel DXP but it is a $6000 program and Eagle works just as well on small stuff for a whole lot cheaper. You can download a light version. (Board size limited but PMD is within this) Here: http://www.cadsoft.de/freeware.htm

Got your Emitter\Base pads reversed..........


Also, the -80v pulse is back emf, when solenoid drive is removed, transistors off.Edited by: gmctd

http://www.dieselplace.com/forum/smileys/Confused.gif


As my eyes glaze over........all I know is this electronics stuff works on 2 princpals. 1. the PFM principal 2. The smoke princpal.


Hope you fellows have some success none the less. Good luck!

Got your Emitter\Base pads reversed..........


Also, the -80v pulse is back emf, when solenoid drive is removed, transistors off.





They should be correct. It is hard to see from this Gif as it is showing both sides from one plane but this is copied directly by hand from the PMD. (Pain)


The -80v you are saying is still present without the solenoid in the circuit? This is usually typical of a solenoid flyback (Field collapse) There is a schotky diode that is not in this pic that I will have in soon and it is only capable of a -60v peak max. I do not know the PIV (Peak inverse voltage ) of the MJ15004's but -80v is not good. I'll post some more soon.


Edited by: mdhorban

Emitters tied to the emitter resistors to the large land left center, to the schottky diode MBRF1635 cathode right pad, left pad anode to +12v thru module connector pin D.


Syntax error on my part - pads are correct, labels wrong


15004's are 140v piv - always in circuit, collectors tied thru module pin B to solenoid coil pin A, coil return is from coil pin B thru parallel wires to module pin F and to ground on inj pump body.


Keep up the good work......Edited by: gmctd

Now that I look at my pic I see that the E,B & C are labeled. You are right they are not correct. In eagle I used a generic foot print and still have not gone over everything. I used the same foot print on all the caps and resistors also. MBR1635. You must have disected a PMD to have seen this bugger as it is well hidden in the epoxy!

MBR1635 Datasheet: http://www.chipcatalog.com/ONSemi/Datasheet/MBR1635.htm (http://www.chipcatalog.com/ONSemi/Datasheet/MBR1635.htm)


I'll update the pic to show this and other stuff (Traces) continually when I can.

Hope you don't take it as criticism, dude, just offering assistance.


I did the second version, first major release from '94, back in ought one.


Two revisions have been released since mine - yours appears to be the latest version


Am interested in what the changes encompassed, OBD-I to OBD-II, and etc.


My stuff is scratched out on quadrille pad, so again, an appreciative 'Excellent work' goes to you.

Hope you don't take it as criticism, dude, just offering assistance.


I did the second version, first major release from '94, back in ought one.


Two revisions have been released since mine - yours appears to be the latest version


Am interested in what the changes encompassed, OBD-I to OBD-II, and etc.


My stuff is scratched out on quadrille pad, so again, an appreciative 'Excellent work' goes to you.





None taken! I did not know they had a few revisions but I noticed a few differences between the 2 I have. No trace difference just IC changes. (Brand) Wish I would have wrote down the numbers on them as to know the revision. I imagine there is not much difference as they all have to drive the same thing.

http://www.dieselplace.com/forum/uploads/B23_Imag0002.jpg

http://www.dieselplace.com/forum/uploads/D6C_Imag0001.jpg

http://www.dieselplace.com/forum/uploads/2AF_Imag0003.jpg

First pic is back side of top board, second pic is front side of top board, and third pic is the front of the second board.

I don't know that this will help at all, but I was wondering what value there would be to putting the diagram on a solid modeler?

I will try it and see what it lloks like.

It may be easier to discern between the layers and what not.

My appologies for having started this thread and subsequently vanishing... I've been a wee bit sick.

I am starting the 3d circuit now though.

FYI---This is great stuff. I knew there was a wealth of knowledge to be found here, I just did not realize it would be so in depth and accurate. Thank you.

That's the same revision level (later production) as mine, and same thing that happened to mine, hrj.


Check it out and see if this isn't correct -


The first two pics are heatsink side, then circuit side of the 1994 pcb.


The third pic is where the solder mask peeled off with the potting compound between the pcb and the heatsink.


The Emitter resistors (Dale) are still buried in the potting compound, as are the darlington pre-driver, and schottky diode, and you can see the General Instruments markings on the diode, with date code of 9520.


The TO-225 transistor collector and TO-247 diode cathode are butted up against the heatsink for max thermal transfer, so are oriented face-up against the pcb.


The large Emitter resistors solder to the lazy "L" circuit lands, upper left and lower right in the first picture, which exhibits no green solder mask on the raw copper circuitry.


The blotch of solder mask on the copper land, upper center, is where the diode covered the pcb, excluding potting material from that area.


N'est ci pa?


Btw, good work on the 'scope, dude.....Edited by: gmctd

Just started work on the 3D version of the PMD's schematic

Here is a "really simple" portion of it.

http://www.dieselplace.com/forum/uploads/B3A_part2.jpg

I know it is not to the dimensions that mdhorban's is.

I am still trying to figure out how to put an image in the background, so I can simply transpose md's drawing into 3D.

It should be a pretty handy to be able to view the circuit from all angles though.

I will post more as I "accomplish" more.

Thank's again mdhorban for all your work.

Noticed the on-going upgrade to the pic - might need to re-etch the direct anode to cathode short on the 1635 diode.


Also, the diode and transistor are mounted under the pcb, but face up toward the pcb, so the component labels and pinouts should be reversed.


As is, the diode and transistor are reverse-biased, and our trucks won't starthttp://www.dieselplace.com/forum/smileys/smiley19.gifEdited by: gmctd

I measured the current on the FS ( not sure if this is a good one), not mounted on truck, draws 11.78 amps. This does not include any resistence which may be in the FSD. I tested it several times to be sure. This means 11.78 amps x 14 volts = 165 watts. The power transistors are rated at 250 Watts, so forward current should not be a problem.

Think that one is probably defective.


Consider this -


MJ15004's are paralleled pnp open-collector drivers, with Fuel Solenoid to ground as the load.


The hi-current path from ground thru the coil to the collectors would be open-ended off the pump.


The series schottky diode from Emitter to +12v power is 16a\35v rated.


All internal componentry is Cmos, but for the bi-polars, incl 2N5195 pre-driver, so I think that 12 amps is suspect.


But, don't pitch it - got other tests, if you're interested.....Edited by: gmctd

tahoe2dr, Nice 3D model. If I used my Protel it will give me a 3D render but not as nice as that. I can separate the layers on this PCB to show the circuit better. Some of the footprints are generic and I have not yet checked the labeling but I will continue to update. Some things may look like they are on the top layer also but they are not. I will try to update more tonight as I am at work right now. Will also be leaving tommorow for a few days so I hope to get some done before. Gotta go boss is here! Edited by: mdhorban

mdhorban, Are we going to pool some money ( for boards and components) so we can get say, 10 boards made up, because making one board is almost as expensive as 10 boards?

I would like to contribute as well.

hey put me in for a contribution as well. How much and how ?

I will contribute also.

mdhorban, Are we going to pool some money ( for boards and components) so we can get say, 10 boards made up, because making one board is almost as expensive as 10 boards?


Wow! Thanks! I still have a ways till that as I'm not sure of the values of all components to do a exact copy. May have my own done before? I actually am set up at home to do a small amount of boards myself and just 25miles from my house there is a PCB manufacture that I have had work done through before. I am posting from work today and am leaving till Sunday from here. I will get back with you guy's as soon as I get back. When the time comes I will post how to either make the boards yourself (Pretty easy) or what it would cost me to have them done. (Typically a minimum amount of dough and amount of boards but cheap per board) I'll even post the gerber files and all. (get them done anywhere!)


Most of my other projects are winding down so I should soon be able to get my butt in gear on this.

Not an exact copy. There's potential for slight modification here. I'd like to run an improvement and keep a spare in the glove box, just in case.


gmctd said there are ss studs and nuts on the transistors, that should be brass or some thing conductive, a slight change.

Maybe a bit bigger so as not to concentrate all the heat in one spot?

Still working on getting a 3D done... This is taking longer than I had expected. And now it is even harder since one of my monitors crashed. Us design engineer types have this need for huge amounts of desktop space.

I am doing a prototype pmd cooler for my metal casting lab.

I am considering two designs, both of which wil mount in front of the radiator.

I am trying to get a feel for what they final size of this homemade PMD is going to be.

Another question is would we encapsulate these in some sort of "potting compound", or just leave them open to atmosphere?

Whenever this project is realized I would be happy to make investment cast "Coolers/Housings" for all those that contributed to the cause.

The cost would only be shipping... I'm not paying for materials... besides I feel like I need to contribute something of value here...

So, what would the majority like to see... an open backed cooler design, or something more like a hermetically sealed case design?

I am really impressed with everyone's contribution to this project... there is a wealth of knowledge here that seems unsurpassed.

I hope to get some more 3D work done this weekend... should help to visualize the device.

I would prefer a case that can be opened, so we can repair to the component level.

Yes but something weather proof as much as possible would always be a plus

So, something that can be openned, yet weather resistant?

I can make the casting so that it could be machined.

I can leave a little excess material so that the surface could be fitted with a rubber gasket, and maybe you could drill a piece of plexi and tap the casting so you would have a weatherproof enclosure.

I would think that you would be able to do the drilling and tapping with a hand drill and tap handle.

The surface could be "machined" with a piece of sandpaper and a flat piece of wood.

Is that something everyone would be comfortable doing?

I work in a machine shop, so I am able to do small things for myself, but I am unsure if doing a bunch would fly... I'll ask.

Sounds good to me!

I have access to a cnc machine without strings attached. I'll volunteer to make anything we need.

Just got back from my crab&camp weekend! I see this thread is still alive and kicking. All the issues with the PMD as far as design flaws will be issued. (Fingers crossed) The on going map-out is a exact duplicate of the PMD that I have. (Stanadyne PT210004S REV B). Once I have this done and take the time to make sure exactly what is going on I will finish a prototype of my own! One thing that I would like to do is go with Mosfets. Much less heat generated.


gmctd actually I just learned was working on this many moons ago! He also was looking into using mosfets. I will be picking his brain on this to see what he has as this may be a simple circuit electronically (So far)but it is still a time consumer.http://www.dieselplace.com/forum/smileys/smiley12.gif


Here is some pics of the board in each layer. Red being the top traces, Blue being the bottom. The components are still hard to decifer what side they are on though. I figure this might help tahoe2dr in his cool 3D model. gmctd and hrjack99, as I know you 2 are well familiar with the PMD, if you can "proof read" my traces you might see something I missed. Also if you could get me a list of the major components (Ic's) in your earlier revision PMD's I'm actually curious what steps the Stanadyne techs were doing to eliminate their problems. I will keep adding component values also.


http://www.dieselplace.com/forum/uploads/mdhorban/2004-10-04_171742_pmdpads&vias.gif





http://www.dieselplace.com/forum/uploads/mdhorban/2004-10-04_171723_pmdbottomtrace.gif


http://www.dieselplace.com/forum/uploads/mdhorban/2004-10-05_170508_pmdtoptrace.gifEdited by: mdhorban

P.S. The Blue (Bottom traces) Is looking through the board so if you looked at it from the bottom it would be mirrored of this. Here it is looking from the bottom.(Directly at it) The scale and proximity of traces is not exact on all of this as is was hand done but the actual connections should be.http://www.dieselplace.com/forum/uploads/mdhorban/2004-10-04_172712_pmdbottomtrace2.gif

Great stuff MD.

I will resume on the 3D modeling ASAP.

School has put me at a somewhat slower pace than usual.

I also have a lot of other "stuff" going on, as I am sure most of you do.

I will make a concerted effort to finish up my model though, as it seems to be a very worthwhile cause.

Imagine a 6.5 that has no issues with the PMD... it sounds too good to be true...

Like I said ASAP... I am really curious as to what kind of design flaws everyone will see.

Everyone has been amazingly helpful in all of this.

I only hope I am able to contribute something of use here.

Anything model or analysis wise I can handle... this electrical stuff may have been a bit over my head.

Always a learning experience, always made easier by those willing to share their knowledge.

Again Thank You all for your contributions. http://www.dieselplace.com/forum/smileys/smiley32.gif

I can not find anything wrong.

Component-side Anode-Cathode short at MBRF1635 diode

Here is Stanadynes attitude on our collective "problem" I borrowed it from the Diesel page forum





Dear Sir,

Our experience shows that in most instances that when people think their Pump Mounted Driver has failed, that it actually has not. There are many other things in the GM 6.5L system that may lead people to think the PMD has failed. One that comes to mind is the vehicle PCM itself which occasionally can send incorrect signals to the PMD causing engine stalling and leading people to think the PMD has failed. The best advice I have is to find someone who is very familiar with the GM 6.5L application and have proper diagnostics performed. Some of our Authorized Distributors and Dealers can do this. A complete listing of them can be found on our website at www.stanadyne.com (http://www.stanadyne.com/) in the Dealer Locator feature. As far as actual PMD failuress, we have also found that remote mounting PMD's can cause them to fail and if a new harness is not installed with a replacement PMD that this can lead to arcing in the connector and subsequent PMD failure.

Sincerely,
Mark Dionne
Manager, Product Support
Stanadyne Corporation


Ain't no wonder this thing is still broke after all these years, I'm on FSD 3 myself actually would be 6 if you throw in the 3 times I recovered FSD #2 with a retorque of the transistor lock nuts, fix EVERY TIME was FSD replacement.


Maybe we ought to try to find Mr Dionne's address and educate him via email and letter mail what a PITA this problem really is.


Won't do much good probably but will maybe give him some of the anxiety ( http://forum.thedieselpage.com/ubb/redface.gif another PMD/FSD fail letter-will they ever stop!!! http://forum.thedieselpage.com/ubb/redface.gif ) we have been feeling wondering if we can make it round the block or not.

Its ridiculious for me to have to carry a $300 spare part in the glove box so I can get home; and nothing better than the feeling of pulling a 12K load and having the engine go dead in traffic.



Edited by: Turbine Doc

Component-side Anode-Cathode short at MBRF1635 diode


Thank you! Updated. May need verification on input labeling as i'm not positive.

Did anyone ever come up with a new design????