1993 Civic Hard Starting Problem

the back emf arcing is pretty much dealt with by putting a diode in
parallel with the coil - it conducts the breakdown current rather than
allowing the points to arc - and that's what these relays have built in.
so that reduces the rate of contact point wear, but they still wear,
and when they wear, they run hot. that's why the relay is of massive
copper construction - to conduct some of the heat away from the
contacts. the part of the relay that's running the fuel pump is also
responsible for switching a big reactive load, which doesn't help. you
can argue that the relay is not big enough for the job, but otoh, a 10
year relay's not bad in the grand scheme of things.

reject rates on machine soldered work are /way/ less than human soldered.

 

Jim Yanik wrote:

The problem with the failing main relays is not related to poor
soldering from the factory. The problem is that solder has very poor
mechanical properties. If you subject it to stress, solder will fatigue
quickly. What holds the relay in place mechanically in the main relay
module is the solder joints only. The relays are heavy, and after a
decade of shaking, bouncing and vibrating, these solder joints will
fatigue and fail.

This is also why, contrary to popular belief, it is a bad idea to tin
the leads before puting on crimp connectors, screw terminals etc.

Honda is not the only make having this problem. Check out
http://www.bmwe34.net/E34main/Maintenance/Electrical/LKM.htm for a very
similar problem with BMW's "Lights Control Module".

 

Jim Yanik wrote:

The problem with the failing main relays is not related to poor
soldering from the factory. The problem is that solder has very poor
mechanical properties. If you subject it to stress, solder will fatigue
quickly. What holds the relay in place mechanically in the main relay
module is the solder joints only. The relays are heavy, and after a
decade of shaking, bouncing and vibrating, these solder joints will
fatigue and fail.

This is also why, contrary to popular belief, it is a bad idea to tin
the leads before puting on crimp connectors, screw terminals etc.

Honda is not the only make having this problem. Check out
http://www.bmwe34.net/E34main/Maintenance/Electrical/LKM.htm for a very
similar problem with BMW's "Lights Control Module".

 

Jim Yanik wrote:

The problem with the failing main relays is not related to poor
soldering from the factory. The problem is that solder has very poor
mechanical properties. If you subject it to stress, solder will fatigue
quickly. What holds the relay in place mechanically in the main relay
module is the solder joints only. The relays are heavy, and after a
decade of shaking, bouncing and vibrating, these solder joints will
fatigue and fail.

This is also why, contrary to popular belief, it is a bad idea to tin
the leads before puting on crimp connectors, screw terminals etc.

Honda is not the only make having this problem. Check out
http://www.bmwe34.net/E34main/Maintenance/Electrical/LKM.htm for a very
similar problem with BMW's "Lights Control Module".

 

Howard,

Well, I certainly hope I'm not in your car when your relay fails in the
middle of the night on a dark road. Unless you carry around a solder iron
with you. And I hope you got a few more spares, like some pistons, a starter
motor, brake pads, maybe CV joint. Or maybe you can fix all of them with a
solder iron? I'm not gonna buy stock in Napa till you throw away that iron!
Good luck.
Howard

 

Howard,

Well, I certainly hope I'm not in your car when your relay fails in the
middle of the night on a dark road. Unless you carry around a solder iron
with you. And I hope you got a few more spares, like some pistons, a starter
motor, brake pads, maybe CV joint. Or maybe you can fix all of them with a
solder iron? I'm not gonna buy stock in Napa till you throw away that iron!
Good luck.
Howard

 

I bet the early coil/breaker ignitions did NOT have any diode to proect
from back EMF.

Also,I suspect there haven't been many -worn out- main relays that were due
to contact wear,if any. Just cracked solder joints due to aging of the
solder alloy itself,or thermal cycling unrelated to current flow thru the
contacts.The current drawn by the main relay is not enough to heat up those
contacts,especially with the low make/break duty cycle.

 

I bet the early coil/breaker ignitions did NOT have any diode to proect
from back EMF.

Also,I suspect there haven't been many -worn out- main relays that were due
to contact wear,if any. Just cracked solder joints due to aging of the
solder alloy itself,or thermal cycling unrelated to current flow thru the
contacts.The current drawn by the main relay is not enough to heat up those
contacts,especially with the low make/break duty cycle.

 

I bet the early coil/breaker ignitions did NOT have any diode to proect
from back EMF.

Also,I suspect there haven't been many -worn out- main relays that were due
to contact wear,if any. Just cracked solder joints due to aging of the
solder alloy itself,or thermal cycling unrelated to current flow thru the
contacts.The current drawn by the main relay is not enough to heat up those
contacts,especially with the low make/break duty cycle.

 

no, but they had a condenser, which helps. you can't completely kill
back emf on a contact breaker circuit because that's what creates the
spark on the ht side of the coil.

the contacts don't heat much when new, but after a while in operation,
even with spark protection, the contact points get eroded. it may not
look much to the naked eye, but reality is that the actual surface
contact area decreases - you can see all the little hills & valleys that
have been eaten away under a microscope. as the current per unit area
increases [same current divided by less area], operating temperature
increases.

tin/lead solders don't "age", but as stated by randolph, solder has poor
mechanical properties. thermal cycling & mechanical stress can easily
fatigue it. soldered joints not subject to these conditions can last
virtually forever.

 

no, but they had a condenser, which helps. you can't completely kill
back emf on a contact breaker circuit because that's what creates the
spark on the ht side of the coil.

the contacts don't heat much when new, but after a while in operation,
even with spark protection, the contact points get eroded. it may not
look much to the naked eye, but reality is that the actual surface
contact area decreases - you can see all the little hills & valleys that
have been eaten away under a microscope. as the current per unit area
increases [same current divided by less area], operating temperature
increases.

tin/lead solders don't "age", but as stated by randolph, solder has poor
mechanical properties. thermal cycling & mechanical stress can easily
fatigue it. soldered joints not subject to these conditions can last
virtually forever.

 

no, but they had a condenser, which helps. you can't completely kill
back emf on a contact breaker circuit because that's what creates the
spark on the ht side of the coil.

the contacts don't heat much when new, but after a while in operation,
even with spark protection, the contact points get eroded. it may not
look much to the naked eye, but reality is that the actual surface
contact area decreases - you can see all the little hills & valleys that
have been eaten away under a microscope. as the current per unit area
increases [same current divided by less area], operating temperature
increases.

tin/lead solders don't "age", but as stated by randolph, solder has poor
mechanical properties. thermal cycling & mechanical stress can easily
fatigue it. soldered joints not subject to these conditions can last
virtually forever.

 

Well,now we have diodes that can take the back EMF,back then we didn't,had
to use big stacks of selenium wafers for rectifiers,and they were SLOW(in
turn-off time,critical for flyback operation).Your PC power supply works
just fine using diode snubbers in the flyback mode supply.They even make
the switching transistors with on-substrate diodes for this purpose.

Yes,and for the -main relay- we're discussing,how often does the make/break
happen? Not anything like a automotive ignition doing it thousands of times
per second,nor is there a high current.

 

Well,now we have diodes that can take the back EMF,back then we didn't,had
to use big stacks of selenium wafers for rectifiers,and they were SLOW(in
turn-off time,critical for flyback operation).Your PC power supply works
just fine using diode snubbers in the flyback mode supply.They even make
the switching transistors with on-substrate diodes for this purpose.

Yes,and for the -main relay- we're discussing,how often does the make/break
happen? Not anything like a automotive ignition doing it thousands of times
per second,nor is there a high current.

 

Well,now we have diodes that can take the back EMF,back then we didn't,had
to use big stacks of selenium wafers for rectifiers,and they were SLOW(in
turn-off time,critical for flyback operation).Your PC power supply works
just fine using diode snubbers in the flyback mode supply.They even make
the switching transistors with on-substrate diodes for this purpose.

Yes,and for the -main relay- we're discussing,how often does the make/break
happen? Not anything like a automotive ignition doing it thousands of times
per second,nor is there a high current.

 

jim, sorry if i'm coming across as unclear.

diodes on relays aren't used in normal forward conduct mode. think
about it: if they conducted the current used to energise the coil, the
coil would never activate. so they're oriented the opposite way.
where's the benefit? well, when the magnetic field energy of the
solenoid breaks down after switch-off, it changes rapidly. a rapidly
changing magnetic field induces an electric current in the coil that
just happens to be wrapped around it. depending on the number of turns
& rate of decay, that current can have many thousands of volts. diodes
break down and conduct reverse voltage at some level, and this level of
breakdown can be carefully designed. so, the diodes used to protect
relay circuits do this at some appropriate level - whatever - and so the
voltage spike never goes beyond this, hence protection.

the devices used in switch mode power supplies are a very different
creature. unlike simple full wave rectifiers that conduct the whole
waveform, switching devices need to be able to switch a forward current
in the middle of a wave. this means a very large momentary power
dissipation within the device, so to do this successfully,
[semiconductors are not great power dissipators] the switch needs to be
/fast/.

the two operations are very different and there is no simple analogy
between them.

well, you're right that the switch count is not high, but take one of
these relays apart, particularly if you get a chance to put it under a
microscope, and you'll see the contact erosion. also, it's switching
the fuel pump which has a fairly large coil in it....

 

jim, sorry if i'm coming across as unclear.

diodes on relays aren't used in normal forward conduct mode. think
about it: if they conducted the current used to energise the coil, the
coil would never activate. so they're oriented the opposite way.
where's the benefit? well, when the magnetic field energy of the
solenoid breaks down after switch-off, it changes rapidly. a rapidly
changing magnetic field induces an electric current in the coil that
just happens to be wrapped around it. depending on the number of turns
& rate of decay, that current can have many thousands of volts. diodes
break down and conduct reverse voltage at some level, and this level of
breakdown can be carefully designed. so, the diodes used to protect
relay circuits do this at some appropriate level - whatever - and so the
voltage spike never goes beyond this, hence protection.

the devices used in switch mode power supplies are a very different
creature. unlike simple full wave rectifiers that conduct the whole
waveform, switching devices need to be able to switch a forward current
in the middle of a wave. this means a very large momentary power
dissipation within the device, so to do this successfully,
[semiconductors are not great power dissipators] the switch needs to be
/fast/.

the two operations are very different and there is no simple analogy
between them.

well, you're right that the switch count is not high, but take one of
these relays apart, particularly if you get a chance to put it under a
microscope, and you'll see the contact erosion. also, it's switching
the fuel pump which has a fairly large coil in it....

 

jim, sorry if i'm coming across as unclear.

diodes on relays aren't used in normal forward conduct mode. think
about it: if they conducted the current used to energise the coil, the
coil would never activate. so they're oriented the opposite way.
where's the benefit? well, when the magnetic field energy of the
solenoid breaks down after switch-off, it changes rapidly. a rapidly
changing magnetic field induces an electric current in the coil that
just happens to be wrapped around it. depending on the number of turns
& rate of decay, that current can have many thousands of volts. diodes
break down and conduct reverse voltage at some level, and this level of
breakdown can be carefully designed. so, the diodes used to protect
relay circuits do this at some appropriate level - whatever - and so the
voltage spike never goes beyond this, hence protection.

the devices used in switch mode power supplies are a very different
creature. unlike simple full wave rectifiers that conduct the whole
waveform, switching devices need to be able to switch a forward current
in the middle of a wave. this means a very large momentary power
dissipation within the device, so to do this successfully,
[semiconductors are not great power dissipators] the switch needs to be
/fast/.

the two operations are very different and there is no simple analogy
between them.

well, you're right that the switch count is not high, but take one of
these relays apart, particularly if you get a chance to put it under a
microscope, and you'll see the contact erosion. also, it's switching
the fuel pump which has a fairly large coil in it....

 

But once the switching transistor turns OFF,there's still the flyback EMF
that has to be dealt with (on the primary side),otherwise the switching
xstr breaks down and fails due to those high voltages.That's where the fast
recovery diodes enter the picture.(along with other snubbing circuitry) The
fast diode allows less of the back EMF to reach the switcher xstr.

BTW,the switcher transistor only dissipates power when it is ON,when it's
off,it passes no current,thus no dissipation.That's why they use low ON-
resistance devices,to reduce dissipation inthe switcher xstr.

And the inductance of the long wiring between relay and fuel pump windings
helps limit that.

 

But once the switching transistor turns OFF,there's still the flyback EMF
that has to be dealt with (on the primary side),otherwise the switching
xstr breaks down and fails due to those high voltages.That's where the fast
recovery diodes enter the picture.(along with other snubbing circuitry) The
fast diode allows less of the back EMF to reach the switcher xstr.

BTW,the switcher transistor only dissipates power when it is ON,when it's
off,it passes no current,thus no dissipation.That's why they use low ON-
resistance devices,to reduce dissipation inthe switcher xstr.

And the inductance of the long wiring between relay and fuel pump windings
helps limit that.