Question for fuel delivery gurus

[Groesbeck Hurricane]

Does anyone have a theory on why a fuel relay would go bad? I had one go about 3,000 miles ago and yesterday morning the replacement one went bad.

The first one was made in Europe and had been in the car more than a decade with possibly 100,000 miles on it. The second one was made in china and, like I said, had about 3,000 miles on it.

Could there be something in the car causing the units to go bad?

I traced wires yesterday but could not find any issues.

Any thoughts are appreciated.

[docrodg]

Possible that the coil on your 3000 mile relay was poorly made and the varnish on the coil wire came off and it shorted out. These relays do consume quite a few watts and that can cause shorts and other issues if the relay cannot take the heat produced. Not sure of this specific relay but another potential failure mode is when the power to the relay is removed you get a back-pulse of energy from the collapsing magnetic field of the coil, this is a very large pulse and can melt the coil out. Most relays are protected by a diode across the coil or in the circuit.

[psalt]

How old is the fuel pump and how many amps does it draw while running ?

[Groesbeck Hurricane]

Yep, five pin one on the boot, I believe number seven slot.

Hmmm, old one was plastic and had been there a VERY long time.

Paul,

Fuel pump was replaced about four years ago/16,000 miles or so. I have no idea how many amps it pulls as I have never tested. What would be a safe way for me to test? Safely raise the tyres on one side, and double support. Crawl under and remove the aluminum panel. Start car and make sure it is not going anywhere (blocking drive wheels, keeping in neutral), then how do you suggest I test? Thanks!!!!

[Grady Clay]

Looking at the ’82 Current Flow Diagram page 0.1-3/91, the Fuel Pump Relay has an internal diode across the coil. This is to protect the coil or some other device (RPM Limiter Switch?) from over-voltage due to ‘inductive kick-back’. It is notable that there isn’t a diode or some other ‘protection’ at the fuel pump.


I agree that the relay should be sized to have a ‘safety factor’ above normal operation. This includes temperature, current flow and voltage transients.

One problem we have is no one seems to have diagnosed the relay failures. Is it from over-temperature? Is it from diode failure? Is it from coil failure? Is it from contact failure? If contact failure, is it from over current or reverse voltage when opening?

Given that there is a powerful motor (fuel pump) in the circuit, I suspect one cause of relay failures is the ‘inductive kick-back’ when the relay cuts off the motor. It may be the diode in the relay is undersized. It may be the motor needs added ‘protection’ that the engineers didn’t foresee. That ‘protection’ isn’t for the motor but to protect all the other circuitry from the ‘inductive kick-back’.

What to do?

A reverse diode across the fuel pump will give the ‘inductive kick-back’ current a preferential path, reducing the voltage at the opening relay contacts. This diode only conducts when the motor is being turned off, not during operation. This needs to be a very substantial (high currant and heat dissipation) diode.

Another is a voltage protection device across the relay contacts and motor. This can be a ‘ZNR’ (zinc-oxide non-linear resistor), zener diodes and resisters and others. These are devices commonly used in your computer ‘surge protectors’ only with voltage limiting characteristics appropriate for 12-14 V automotive conditions.


This relay failure seems to be sufficiently common that PCNA and Porsche AG should have addresses it. Is there a Porsche recommendation?

If Porsche hasn’t addressed this issue, well-protected automotive relays are ubiquitous. I’m sure we can find a suitable replacement from Bosch, Ford or Toyota.

Best,
Grady

[T77911S]

you can pull the FP fuse, connect an amp meter where the fuse was, make sure it is set to amps and not volts, start the car and measrure the current. i think mine is around 7 amps.

[T77911S]

grady,
the inductive kick back is not in the circuit that feeds the FP. well. i guess in a way it is since all voltage is from the battery, or alt.
the diode is across the coil, not the contacts. the power for the FP comes from the batt, for simplicity, through the contacts to the FP. when power is removed from the coil, the collapse of the field would be felt back to the source, the batt, or maybe even other electrical deives, no? but, its the relays that are going bad, not the FP's so i dont think any kickbacks are effecting the relay.

i am not sure putting something on the motor itself would do any good since agian, the motor is not in the circuit with the coil of the relay.


it would be interesting to test some of these bad relays. is it the contacts, which see the brunt of the high current, or is it the coil it self. do the red relay coils draw any more current than the black ones? if not, then it would be the FP current making them hot. are the diodes shorting bypassing current from the coil to energize it?

when ever i have worked on someones home AC unit, usually the first thing i look for are bad contactors due to the contacts being burned from the current. perhaps the current for from the FP is too much for the relays and/or the relays are just not as good as they use to be.

[docrodg]

I think we are all on the same track as far as this specific relay is concerned, although a post-mortem on others would be interesting. This relay failed, and I would question the quality based on the fact that the original worked for 100,000 miles and this one failed at 3K. Also, it is a china part and may be of lower-grade components. low-grade contacts may heat due to resistance and burn out, the inductive kickback diode could be bad due to improper handling at manufacture or high temp stress on the silicon, or the coil burnt out from heat.

I should think that you could take the relay out, purchase a high-quality relay from an electrical supply that will fit in the case and wire it in there for use. Given proper specs should not be a problem.

[Gunter]

The prices of these relays tell the story:

OEM (Bosch?) = about $ 34.-
Junk (URU) = about $14.50

Check it out, search for 5-pin round relay, black or red.

[lindy 911]

I think the relay failure has to do with the socket it's plugged into. The pins over time become looser in the fit to the socket and heat up due to a bad contact there. If the pin heats and the connection is bad the coil suffers and becomes overloaded. If the coil is not holding the contacts tight enough the contacts overheat and fail.

Lindy

[Groesbeck Hurricane]

I will search and see if I saved the really old relay, but I doubt it is still available. If I can put my hands on both of them would someone like to take a look into them for gee whiz sake?

Just for fun, I put the failed 3,000 mile relay on the car and it started right up, got hot, died.

I know very little about this, but I am thinking the circuits in the relay are poorly constructed and heat is causing the issue.

[Grady Clay]

Originally Posted by T77911S

grady,
the inductive kick back is not in the circuit that feeds the FP. well. i guess in a way it is since all voltage is from the battery, or alt.
the diode is across the coil, not the contacts. the power for the FP comes from the batt, for simplicity, through the contacts to the FP. when power is removed from the coil, the collapse of the field would be felt back to the source, the batt, or maybe even other electrical deives, no? but, its the relays that are going bad, not the FP's so i dont think any kickbacks are effecting the relay.

i am not sure putting something on the motor itself would do any good since agian, the motor is not in the circuit with the coil of the relay.


it would be interesting to test some of these bad relays. is it the contacts, which see the brunt of the high current, or is it the coil it self. do the red relay coils draw any more current than the black ones? if not, then it would be the FP current making them hot. are the diodes shorting bypassing current from the coil to energize it?

when ever i have worked on someones home AC unit, usually the first thing i look for are bad contactors due to the contacts being burned from the current. perhaps the current for from the FP is too much for the relays and/or the relays are just not as good as they use to be.

The issue here is the ‘inductive kick-back’ causes an arc at the contact interrupting the current to the motor in the fuel pump. That arc can damage the contacts. Many relays can handle a resistive load (like a headlight) will fail quickly with an inductive load (like a DC motor).

This happens because the ‘inductive kick-back’ produces voltage that tries to keep the current flowing. That voltage appears across the relay contacts and causes an arc. With a 12 V system and only ,say, 7 A flowing, there can be 500 V momentarily appear across the contacts. In this situation, the coil in the motor is producing the voltage with the negative (ground) end of the motor positive and the power supply end of the motor is negative. The motor momentarily acts as a very powerful ‘battery’ trying to maintain the current flow. The faster you try and interrupt the current, the higher the voltage.


Some features of relays designed for inductive loads are: Fast opening; large gap when open; magnetic field across the contact, perpendicular to arc current flow. In serious situations they break under insulating oil or in high-pressure inert gas.

The other way to address the situation is to provide an alternative path for the ‘inductive kick-back’ current. This is where a diode or ‘surge suppresser’ across the motor can be useful.

The disadvantage of providing an alternative current path is that the motor will tend to run a bit longer after the relay opens. I don’t know if this could be an issue since the relay is also controlled by the RPM Limiter Switch and acts as the ‘rev-limiter’ for the engine.



We don’t know the failure mode. As I noted above, it can be many things. I didn’t note socket/pin contact failure. All we know is the relays quit working.



So … do we need a relay with temperature ratings so the coil insulation doesn’t fail? Does the diode fail? Do the contacts fail? Are the pins and socket undersized or being corroded? Is the relay improperly designed for an inductive load? Something else?

Best,
Grady

[mca]

Originally Posted by don gilbert

David, if you are talking about your lambda control relay on the 83sc, you have found out what many have already. The 02 relay on 81-83 run very warm, sometimes hot, why? I don't know, bad design I guess, You must buy the original Bosch metal relay, (Pelican has them for 20.00)the plastic 5 pin uni's will not last very long at all.

Little off topic ... but I bought mine from Pelican and it was plastic.

The pic shows metal ... but it ain't metal. And it only costs $10.75.

Pelican Parts - Product Information: 821-951-253-M14

[Gunter]

I think Grady is referring to the 5-pin red FP relay.

These stock relays, black or red, are Wehrle and no rating is shown on them other than part numbers.

They last a long, long time provided the current-draw is not excessive.
If the pump, or electric window motors have to labor, the current draw will actually melt the socket and/or the relay itself.

I recommend for anyone to occasionally remove the relays, inspect and smell the sockets and relays.

We cannot blame the relays for these failures. The cause of the excessive draw has to be remedied because no amount of changing relays is going to fix the problem.

Unfortunately, the round 5-pins available at Pelican are either cheap URU junk for $14.50, or the good stuff for $33.-

For the relay controlling the Lambda brain, it's a 5-pin Bosch 30A? and it also lasts very long. (My plastic Bosch 5-pin was in situ for more than 10 years before I removed the Lambda)
Heat is not a problem, IMHO, but rust is because the relay and brain is mostly neglegted under the seat.

[dshepp806]

The diodes are there to prevent relay contact "chatter"....as explained, inductively, above.

Usually the relay failures will be contact-related, moreso than the coil itself.

What wattage ratings have you guys found on these diodes across the relays' coil?

Best,

Doyle

[86 911 Targa]

Originally Posted by Gunter

The prices of these relays tell the story:

OEM (Bosch?) = about $ 34.-
Junk (URU) = about $14.50

Check it out, search for 5-pin round relay, black or red.

+1 One the ÜRO.

Those things are junk!