The DME Relay explained

[Jazzbass]

Most of us Carrera and 944 owners learn the first lesson of Motronic when we get our cars fairly quickly: always carry a spare DME relay. When the DME relay dies, the car no longer works. I've already done a write up about how to fix a bad DME relay (http://dorkiphus.net/porsche/showthread.php?t=12684), but I wanted to post here to discuss what exactly this thing is and what it does.


DME Relay Explained

So what is the DME relay? Well, it's actually two relays in one - one of the relays switches power to the DME and the fuel injectors, and the other relay switches power to the fuel pump. That's right, SC guys, we don't have that magic red relay in the fuse box controlling the fuel pump. Instead, we have a crappy, failure prone two part square relay under the drivers seat (the red relay in a Carrera fuse box is typically the A/C relay). Since 944s, 964s and up have fuse boxes designed after the 1950s, they typically have their DME relay in there. Despite being in a better location than in the older Carreras, 944/964 DME relays are still crappy and failure prone.

Here's a picture of the actual internals of the relay, showing both relay coils:

11856

This is the schematic from the factory manuals showing the relay in circuit:

11855

The schematic shows clearly how both sides of the relay are wired up. The switched +12V coming in on pin 86 is the first step in turning everything on. When you turn the ignition switch to position II or start, +12V is switched to this wire. Pin 86 is connected to one side the the right relay coil, and the other side of the coil (pin 85) is connected directly to ground. So, switching the ignition switch on turns on the right relay.

Turning on the right relay allows +12V to come in on pin 30, go to the right side of the left relay coil, and then out of the DME relay on pin 87. The +12V coming out of the relay on pin 87 feeds power to the DME itself on pins 18 and 35 and to the positive side of the fuel injectors (the ground side of the injectors is controlled by the DME). As you might imagine, this is a pretty high current circuit. Power coming in on pin 30 is coming directly from the battery on a 4mm^2 (~11 ga) wire. Note also that this power is UNFUSED - this is why your DME can get destroyed if there is a short in it. That big 4mm^2 wire can supply as much power as a bad DME could possibly want to burn itself up inside.

So the right side of the relay controls the DME itself. What about the left side? That's the fuel pump relay. With the right side of the relay energized, the left side now has +12V on one side of it's coil. To turn on, it needs the other side of the coil (pin 85b) to be grounded. This is controlled by the DME itself. When the left side of the relay is turned on, current flows from pin 30 out of the relay on pin 87b, through the fuel pump fuse, and to the fuel pump itself.

Pin 85b of the DME relay connects to pin 20 of the DME. The internal circuit of the DME controlling pin 20 looks like this:

11865

Transistor T480 is a simple NPN darlington transistor being used as a switch to pull DME pin 20 to ground when activated. For the non-EE types, it's sort of like a relay, but it's an electrical switch instead of a mechanical one. T480 will turn on (and connect pin 20 to ground) when there is voltage at base of the transistor. This base voltage is supplied either from the DME microcontroller (uC) coming in at ADV7 or from the starter signal coming into the DME at pin 4. The point of this circuit?

Well, the DME doesn't want the fuel pump running unless the engine is running. The uC in the DME will only turn on T480 if it sees the engine turning at over a certain RPM (I've read 200). So, if your engine stalls, the first thing the DME uC does turn off the fuel pump. However, when you go to start the car, the starter doesn't turn the engine fast enough to allow the uC to enable the fuel pump in this manner, so a bypass is needed for starting. It's a pretty simple solution, too.

The +12V coming from the ignition switch going to the starter solenoid is also routed into the DME on pin 4. D752 and D481 form a "diode-OR" connection which turns on T480 when either input line has voltage. With positive voltage at the base of T480, the transistor switches on and connects pin 20 if the DME (and thus pin 85b of the relay) to ground, turning the fuel pump side of the relay on. Viola.


Troubleshooting tips

So, if you're troubleshooting start up/DME relay problems, the common issues are likely:

1. Bad DME relay. The solder joints inside crack from fatigue. See this thread for more info (http://dorkiphus.net/porsche/showthread.php?t=12684).

2. No power to the DME relay on pin 86. The factory alarm in Carreras disables the car by interrupting power line. If you are having alarm issues or have removed your alarm, make sure you run this this wire to a location in the fuse box connected to ignition switch bus 15. This last part is very important - bus 15 is hot in the "on" and "start" positions of the ignition switch. I've heard of people who remove their factory alarm and connect this line to another ignition switch bus that was hot in the "on" position, but not in "start". So the DME relay turned off every time they went to start the car. In my 85, the factory location for bus 15 is the unfused side of fuse 8.

3. Bad main power line. The big 11ga wire coming directly from the battery can corrode and not supply power to the DME. Without power on this line, you're SOL.

4. Bad speed sensor. This will cause the DME to not work for many reasons, but one of the side effects will be that without a good engine speed signal from the speed sensor, the DME uC will turn the fuel pump off.

[Jazzbass]

DME Relay Pinout

Here are the pinouts from the 84-89 Carreras and the 944s/964s/993s (which I think are all the same).

11869

Pin 85 [1] - Ground to the DME/injector relay coil
Pin 30 [2] - +12V directly from the battery
Pin 86 [3] - Switched +12V from ignition switch (bus 15)
Pin 87b [4] - +12V to the fuel pump, switched by the fuel pump relay
Pin 87 [8] - +12V to the DME and fuel injectors, switched by the DME/injector relay
Pin 85b [9] - 0V switched to the coil of the fuel pump relay from the DME, pin 20

Note that Porsche/Bosch changed the pin numbers of the relays in the documentation. In the older docs like I have for my 85, they're labeled with the 85/30/86 numbers. In later docs (and on the newer relays) they have the 1, 2, 3 numbers.

[Jazzbass]

Replacing a bad DME with jumpers to get home

So your POS DME relay went bad and you were too cheap to buy and carry a replacement? Well, you can remove the relay from it's socket and connect a wire between pins 30 and 87 to turn on the DME and between pins 30 and 87b to turn on the fuel pump. This is a hack, so do it only to get home, put a crowbar in your wallet and buy a new DME relay. Then fix your old one (http://dorkiphus.net/porsche/showthread.php?t=12684).

11871

[racer]

Am I to gather from all these posts that your car is being improved upon once again?

[}{arlequin]

hopefully he's running out of 'reasonable' improvements

[pookie]

That is awesome.

[tdatk]

That my friends is just how a technical write up should look like.

Thanks Chris

[Jazzbass]

Am I to gather from all these posts that your car is being improved upon once again?

Actually, it's mostly just out of my interest in this part of the car. This was the very stuff I loved doing before I realized there were better career opportunities in software. That said, I did just add 10 hp to my car for free. I'm working on a new post about how. ;)

[pookie]

Jazz I gotta question...

Testing for a no start on my car, and just testing the wiring from the 3.2 install as well. When I test my where the DME relay plugs in, or in my harness, my #30 has power at all times. It is at 12+ volts with the key on or off. My plug # 86 is hot with the key on, and my #87b is hot with the key on. My # 87b is also connected to the red/green wire in the front trunk. None of the other plugs are hot at all with the key on or off. Pin #87 is supposed to be hot with the DME relay in.

And the kicker is with the DME relay removed my fuel pump still runs. That's not supposed to happen right?

[Jazzbass]

my #30 has power at all times. It is at 12+ volts with the key on or off.
Correct. This comes directly from the battery.

My plug # 86 is hot with the key on
Correct. This comes from the ignition switch

and my #87b is hot with the key on. My # 87b is also connected to the red/green wire in the front trunk
This is a problem. This is the OUTPUT of the DME relay to the fuel pump. It is supposed to go to one side of the fuel pump fuse, and the other side of the fuse goes to the pump. My guess is instead of hooking this to the fuel pump, you connected it to a +12V source in the fuse box. Again, the red/green wire isn't a +12V supply FROM the fuse box TO something else, it's +12V FROM the DME TO the fuse box and then to the fuel pump

None of the other plugs are hot at all with the key on or off. Pin #87 is supposed to be hot with the DME relay in.
Yes. This is switched +12V to the ECU itself.

And the kicker is with the DME relay removed my fuel pump still runs. That's not supposed to happen right?
Not at all. Start at the fuel pump and work backwards and find where you cross wired the fuel pump to a +12V switched source.

[Dr K]

Man this is a great board. Jazz, you're a resource, and not the only one on this board, either.

[73and88]

Jazzbass,

I just read your explanation of the DME Relay. The engine in my 1986 911 was starting intermittently. It would run for 4 or 5 minutes in idle, sometimes a little rough, then abruptly shut off. It would restart sometimes. I removed the DME relay. I noticed a white film of corrosion on pin 30. I removed the black plastic relay cover to expose the circuit board & two relays as you pictured.

Upon close scrutiny with a 7 power jewler's eyeloupe I was able to spot what may have been cracked solder connections on some of the circuit board. I reheated and added solder. Still no start after re-installing relay.

I noticed in your photo the copper braid on the relay furthest from the pins was discolored, probably from overheating. I see a similar discoloration on the braid of the same relay on my relay board. I'll call this relay (B). I also noticed a darkening of the coil winding wires on this relay. This usually indicates overheating of the wires.

There is a small resistor directly under the relay (B). It appears to be electrically across the coil of relay (B) yet this resistor is not shown in the diagram of the DME Relay you posted. It appears to have the color code of Blue, Green, Brown [ 650 ohms ]. I cannot make out the color of the tolerance band. When I measured across the coil of relay (B) with an ohmmeter I read 676 ohms which is in accord with the color bands and tolerance. The other relay nearest to the pins, I'll call relay (A) shows about 83 ohms across it's coil. This resistance is more in line with a 12 V relay of this size.

When applying 12 vdc across the coil of relay (A), relay (A) did indeed actuate. Applying 12 vdc across the coil of relay (B) had no effect. It did not activate. If the coil of relay (B) was indeed still good, I should have seen about 83 ohms or slightly less when measuring it's resistance. It seems that the coil of relay (B) is open.

I ordered 2 new relays that appear from the photo in the vendor's web site to be made in Hungary. My DME relay says made in W. Germany and it must be the original. It has 1986 date code.

That corrosion I spotted on pin #30 may have contributed to the early stages of poor starting. I plan to apply a thin film of silicone grease to all the pins of the new DME relay before installation in the socket. Over time I have found this reduces the risk of corrosion on electrical connectors.

Jazzbass, your posting was very helpful to me while trouble shooting. Photos excellent. Thank you for your concise info.

[73and88]

Here are some photos of the DME relay from a 1986 911. This appears to be the original unit. Note the area of corrosion on Pin #30.

Also see the areas of overheating on the copper braid of the relay on right and the discolored coil wiring. Compare colors of both coils.

The discolored coil appears to be "OPEN" when measured with ohmmeter. It reads the resistance of the 650 ohm resistor that is across the coil. Reading should be around 83 ohms.

[73and88]

Just received two DME Relays from Automotion/aka Eckler's. Their web site advertised it as a genuine Porsche part. The photo on the site shows the relay with the Porsche factory numbers and white lettering showing "MADE IN HUNGARY".

http://www.automotion.com/dme-relay-for-911-carrera-porsche-1984-1989.html

The two relays I received from Eckler's have no factory numbers. It had a label marked uroparts. I found the company that imports them. APA in Simi Valley,CA.
I called them. Asked where this DME relay is made. They replied, "CHINA". In addition, there was no indication of the Country of Origin on the part or the package it was boxed in. This is possibly a violation of U.S. Customs regulations.

I refuse to knowingly use Chinese products in my car.

When I called Automotion/ aka Eckler's and complained about misrepresentation, the customer service rep claimed we don't go by the photos!

Moral of this story is: Be careful of where a part is manufactured before installing it in your car and don't assume the photo shown on a web site is what you will get. Especially if dealing with Automotion / aka Eckler's.

[BlackTalon]

This is pretty common these days. Not unique to Porsches, or to Automotion.

It's a chore to photo/ catalog a couple thousand parts; can't say I fault them for not having up-to-date pics of each and every part.

But the moral of the story is if you are intolerant of things from China, you may want to call and order your parts, and ask them to check each one.

[Lupin..the..3rd]

When I called Automotion/ aka Eckler's and complained about misrepresentation, the customer service rep claimed we don't go by the photos!
I'd be pissed too. Seems like false advertising. If you or I listed a part for sale on craigslist or ebay, and shipped something that wasn't the same as the photo, we'd be labeled as scammers.

But even if we ignore the photo, it is advertised as a "Genuine Porsche" part. What you received doesn't have a Porsche part number on it. And URO Parts isn't a supplier to Porsche. What you received is not a Genuine Porsche part. I'd complain and return it for a refund.

Also, URO parts has a poor reputation, at least in the MB world.

[Trak Ratt]

^ bet Porsche doesn't make their own relays or many other parts either. How about all the gen Porsche parts with VW and Audi emblems on them? Gen. Porsche parts can mean many things to many people. Global economy at it's best.

[Jazzbass]

Gen. Porsche parts can mean many things to many people. Global economy at it's best.
Regardless of who the OEM is, "Genuine Porsche" does mean something - it means the part has passed Porsche's QC. Like Lupin said, it should be in a Porsche bag with a Porsche part number on it. What you see with a lot of places like Pelican, Automotion, etc, is that they buy parts directly from the OEM and not from Porsche. Sometimes this leads to the issue that the parts they get from the OEM are not quite up to the standards that the parts they sell to Porsche.

I learned this lesson the hard way with some OEM Lemforder ball joints I bought shortly after I got my car. Supposedly the exact same part that was on my car. The car at the time was 20 yrs old and while the ball joints were worn the boots were still intact. Within a year of installing, the boots on the replacements were cracked and torn.

If someone sold you a "Genuine Porsche" part and it wasn't in Porsche's packaging, send it back. Pelican is usually pretty good about letting you know the difference on their website.

[73and88]

I carefully examined the DME relay received from Eckler's. It is marked Uroparts. I could see some differences in the construction of the board and layout of components as compared with the original genuine Porsche DME relay board. More on this Uroparts DME Relay later.

I removed the defective relay from the original genuine Porsche DME Relay P.C. board. The coil was indeed "open". No connection. The other relay on the board is still operational. Looking at the white plastic molded parts of this defective relay I could see the plastic had darkened from what I think was overheating.

I have seen many postings in a few forums talking about DME Relay failures with some suggesting we all carry a spare in our glove box. There must be a reason for all these failures.

I had a few similar relays around that are made in Germany. Same coil resistance, physical size and shape and ratings as the original, now defective relay. For simplicity I'll call this relay (C). I applied 13.0 VDC to this relay's coil for 10 minutes. Each minute I checked the temperature of the relay coil (C) with an infrared temperature sensor. After 10 minutes elapsed the temperature reading was 173 degrees F. Impossible to touch for more than a split second. I then lowered the voltage to 12 VDC and after 10 minutes the temperature of the relay coil held at 152 degrees. Then I dropped voltage to 11 VDC and after 10 minutes the relay coil temperature dropped to about 125 degrees F. Bear in mind, this test was done in a room with 70 degree ambient temperature.

Based on the resistance of the coil, I calculated a current draw at 13.4 VDC of 161 milliamps. That equates to 2.16 watts. Since there are two of these relays in the black plastic box that house the "DME Relay" we are looking at a total wattage of 4.32 watts. That is a lot of heat that is being dissipated into a tiny box , 1"x 1"x 1.7". With no place for the heat to go those relay coils are probably reaching temperatures well over 173 degrees. It is no wonder the soldering connections on some failed DME Relay boards have been found to be cracked. It is also no surprise that the fine coil wires can eventually open and fail from this excessive heat. I'm certain the discoloration or darkening of the coil wiring and the surrounding components is from this extremely high heat being trapped inside that black 'dust' cover.

Possibly Porsche engineering realized this heat problem and tried to reduce the voltage going to one of the relay coils by putting a diode in series with the coil. Or they may have used the diode to reduce inductive spikes on the 12V line. In any case, the diode lowers the voltage around 7/10 of a volt. If the battery supply voltage getting to the DME Relay box is 13.4 VDC, which is a typical voltage it could see when the engine is running, the actual voltage to the relay with the diode in series with the coil would be 12.7 VDC. 12.7 VDC is only 3/10 of a volt less than the 13.0 VDC I used in my temperature test of the relay.

What can be done to lower the voltage to a point where the temperature of the coils would be at least 25 degrees or more cooler. And how can we get rid of some of that heat being generated by the both relay coils?

One method of cooling those coils would be by inserting 2 more diodes in series with the relay coil. That will give a voltage drop of 2.1 volts. With a supply voltage of 13.4, a 2.1 volt drop would leave us 11.3 volts. This would reduce operating temperatures by at least 25 degrees.

These relays will activate from 7 volts and up. Running them on 11 to 11.3 volts should not be cause for concern.

I recently replaced the bad relay in my DME Relay housing with one of those German Relays. I plan to install 2 additional diodes or use a Zener diode if I can find one that would be suitable for the proper voltage drop and current.

I have also drilled six 3/16" diameter holes in the upper part of the black DME Relay dust cover. Three on the top and 3 on the bottom. The top being the surface that faces up when the relay box is installed under the seat. This series of holes should allow some of that heat to flow out of the black housing which should reduce the internal temperatures.

I'll post more about the use of additional diodes after I install them and then try the complete assembly in my 1986 911.

[Dr K]

73and88,
That's an awful lot of posting, detail, testing, and labor on your relays, but I really want to know - did the car run with the new (Chinese) relays???

[smdubovsky]

... After 10 minutes elapsed the temperature reading was 173 degrees F.
...

I'll post more about the use of additional diodes after I install them and then try the complete assembly in my 1986 911.

F? That equates to 78C in proper engineering units;) Magnet wire and solder aren't even *remotely* phased at 80C. Most common magnet wire is good to at least 150C. Lower is avail but so is insulation up past 250C. Old lead based solder is good to ~180C (the newer RoHS stuff even higher.) You've got a long way to go before you're at those temps. The pcbs and plastics are usually the limiting factors.

Normal running voltage should never hurt the relays. Vibration is the typ killer. Yours being discolored does indeed indicate it was overheated. But why one and not the other? It probably wasn't due to battery voltage. Its from the current load and resistive contacts. A dirty contact or circuit overload will send the temps skyrocketing. We have controls that come back from the field where the large 30A+ terminals have literally desoldered themselves if they weren't tightened properly. A couple watts in the coil is no match for the I^2*R heating of the load contacts.

FWIW, If you add more diodes you may not be able to pull in the relay during cranking on a low battery.

IMO, DME relays aren't severely under designed. Most make it 100k+++ miles and several decades. They are mechanical parts. Mechanical parts fail from time to time.

did the car run with the new (Chinese) relays???
x2. Just because something is made in China doesn't automatically mean its crap (though I do share your general sentiment.) Stuff under designed and built to be as cheap as possible is complete shit no matter where its made. But you can build very good parts anywhere, China included, if thats what you spec (and pay just a little more for.) Many very high quality relays are made in china now. As an example I have work experience w/ Zettler (good stuff) and all made in China afaik.

[73and88]

73and88,
That's an awful lot of posting, detail, testing, and labor on your relays, but I really want to know - did the car run with the new (Chinese) relays???

Dr K...yes it ran with the Chinese relays and it ran with the original DME relay once I rebuilt it. I would not use the Chinese made relays as their construction is not as good as the original Porsche part.

It appears there is a design fault in all these DME relays regardless of who makes them. That is why they often fail and why people suggest always having a spare in the glove box.

I will not install any DME relay in my 911, Chinese, O.E.M. or otherwise unless I am completely satisfied with my new modification to eliminate the design fault.

[73and88]

I agree that the high temps inside the DME relay are not enough to melt solder. However, according to one Relay manufacturer, anything over 120C can be detrimental to the insulation used on the coil wires with insulation failure and wire shorting a possibility. From what I can see with my tests, I would not be
surprised if the interior of the relay housing reached 120C and above.

In my own case, the contacts of the failed relay looked rather clean. No signs of burns or heavy pitting. The coil is open and the coil wires certainly look discolored from heat.

I'll try the extra diode and the cooling holes in the dust cover and see how the combination works out.

[Jazzbass]

Why would the interior of the relay reach 120C?

[Parenn911]

hopefully he's running out of 'reasonable' improvements

now thats a stretch..

Great write up!

[73and88]

Further tests running the DME Relay with the plastic dust cover on for 10 minutes at 13.8VDC shows very high relay coil temps....as high as 120 degrees C.

The two internal relay coils dissipate around 4.5 watts of heat total. That heat is trapped in the closed dust cover. It just keeps getting hotter & hotter until it reaches thermal equilibrium. Take my word for it, with the dust cover off, touch the relay coil or the metal relay frame and you'll quickly burn your skin!

The very high operating temps along with vibration are probably a major cause of fractures in the solder joints on the printed circuit boards. The high temps are probably also responsible for some degradation of the coil wire insulation. This will result in shorted wiring in the coils, even higher current draw and eventual failure.

Long live the DME relay....and if your in car heater fails on the coldest day of winter, simply put your hands under your seat and grab that DME for a nice toasty feeling!!:lol:

[Elfer]

Actually, it's mostly just out of my interest in this part of the car. This was the very stuff I loved doing before I realized there were better career opportunities in software. That said, I did just add 10 hp to my car for free. I'm working on a new post about how. ;)

Did you post this yet?

[Jazzbass]

Did you post this yet?
You know, I don't know if I ever did. I must have been referring to upgrading the early DME ROM to a later one. Since your car is a later Carrera it wouldn't be much help for you - you already have that ROM in your DME. Really applied to 84-86 guys.

[Elfer]

You know, I don't know if I ever did. I must have been referring to upgrading the early DME ROM to a later one. Since your car is a later Carrera it wouldn't be much help for you - you already have that ROM in your DME. Really applied to 84-86 guys.

roger that.. thanks...

[Bird watcher]

Jassbass has skills. I hope he don't mind me printing a page as an addendum to my Bentley
manual. I have searched hard and long to find this. Great work..

[frankM_]

Pin 85b of the DME relay connects to pin 20 of the DME. The internal circuit of the DME controlling pin 20 looks like this:

11865

Transistor T480 is a simple NPN darlington transistor being used as a switch to pull DME pin 20 to ground when activated. For the non-EE types, it's sort of like a ...

Can I ask you where I can find a full schema of a Motronic 1.0 box ?

[ClayPerrine]

Just an add on to this thread. The DME relay evolved from the "Double Relay" used on L-Jet injected 914s. But it evolved badly. In 35 years of running an L-Jet 914, I have only replaced the "Double Relay" once.
And I have had lots of experience with DME relays. I used to work on Porsches for a living. I got lots of 944s, 911 Carreras, and BMWs towed to the shop with no start issues. First thing I did was to replace the DME relay with a test one. About 90% of the time that fixed it.

So I made a replacement. I took a DME relay and stripped the base of everything but the pins. Then I got a pair of sockets for a standard Bosch square relay. With a little wiring, I was able to build a plug in replacement for the DME relay that used standard Bosch square relays.

When I put the 964 motor into my 914, I just rewired the harness to eliminate the factory DME relay, and put in a pair of Bosch solid state relays. I should never have a DME relay issue.

If it will help everyone, I can workup a wiring diagram that will show you how to redo an old DME relay to use a pair of standard Bosch relays.

Clay Perrine

[rothaus]

If it will help everyone, I can workup a wiring diagram that will show you how to redo an old DME relay to use a pair of standard Bosch relays.

Clay Perrine

That would be great.

Cheers
Engelbert