Mac Valve Boost Solenoid Backfeeding Issues

[loca5790]

Let me start with saying whenever the vehicle power is shut down (Key Off) and the boost solenoid is connected to Aux 4 the injectors start machining gunning.  See wiring diagram below for current configuration.

I have tried on other Aux channels with no success.

 

Checking Aux 4 hovers around 10-11V assuming because of the flyback diodes internally and on key off it grounds.

 

All.... bear with me for a second as this is a kind of long winded discussion as to what I have done and where I am at.

I have tried multiple relays with quenching diodes, built in resistors and normal relays... no avail.

I have tried adding and ensuring grounds are adequate no change.

I relay isolated the link instead of running it on the key on signal (that signal is already relayed by the factory BCM).

I ran diodes inline with the mac valve to try and stop it. 

I added essentially a pull-up resistor on Aux4 to try and not float it low on the link.

I ran the mac valve ground (aux4) line relayed from pin 87 to 30 with the trigger being the key on power same source as everything else.

I tried isolating the solenoid on it's own relay.

 

All of this did nothing.  I believe the simplest thing here is to allow the driver be high side (cannot).  Anyone else have any ideas or ran into similar issues?

 

I have transistors on the way as another effort.  Originally this was completely piggybacked into my factory car with factory fuse box.  The whole thing is now standalone simply because of this issue and nothing has fixed it.  Testing the Test PWM the boost solenoid does work.   Injectors pulse until the boost solenoid (single wire into the ecu) is unplugged.  This issue does not show up with a noid light... only when the load of the injectors are plugged in. 

 

Here is the diode isolating of the relay.

 

 

[Adamw]

The only way the solenoid can click is if there is still 12V on the feed side of it when the ecu is off.  So for whatever reason either that Ethanol sensor relay isnt shutting off or the wiring isnt like you draw it.  

Did you try with the flex sensor unplugged?

Another quick test would be if you snip/disconnect the 85 wire off the ethanol relay when it is back feeding does it then all shut down?

 

[loca5790]

Hey Adam,

 

Glad I could catch you here, hope all is well. 

 

100% is wired per the above schematic.  When I disconnect the FFS no changes.  If I pull the relay the back feed stops.  If I disconnect only Aux 4 to the ecu the back feeding stops.  Meaning if it is in the middle of back-feeding remove aux 4 pin stops it.  I can test this by having the aux 4 pin ran through a single inline duetsch connector I was using for diagnostic purposes.  After the connection was broken I could check and there was no voltage at that pin.

 

I have a work around and have implemented it.  Relatively clean fix just don't believe it should have been necessary.

 

The fix entails removing the boost solenoid 12v power from the flex fuel and moving it to it's own relay.  I have this relay trigger signal coming from the Link setup on GP Output to enable above 1500 rpm and greater than 50% TPS.  The logic here is that the boost control is not required under these conditions and ignition should not be turned off above 1500 rpm.  This was a clean fix and no longer has any backfeeding issues.

I'm guessing what is happening is on shutdown the mac valve is de-energized resulting in a large voltage spike caused by the collapse of the magnetic field that the flyback diode is not capable of removing... not an expert in electronics as to why this may occur.  It is somehow self-sustaining as the injectors also create voltage spikes as they backfeed.  This is logical because as you unplug injectors the frequency changes, it decreases as the number of injectors plugged in drops from 8 to eventually 0.  If this is the case eventually the energy would be low enough to stop the cycle.  I could put my o-scope on it to watch what's occurring at the mac valve/injectors during this condition.  I'm assuming there is no snubber diode across the mac valve.  This may have also fixed it from high to low at the valve, but it is extremely difficult to get to and do that in my situation.  I'd rather not continue to try and determine what's occurring and use the method I am.

The mac valve current draw during operation is extremely low and could be driven by a high side driver on the ECU and this would also address the issue ensuring the mac valve does not have power during key off.

[Adamw]

No, you are making it more complicated than it actually is.  It is simple - the absolute only way Aux 4 can backfeed is if there is some voltage on the other side of the boost solenoid.  According to your diagram there is only two ways this could happen - either there is 12V coming out of the flexfuel sensor (via the DI pull-up) or that flex fuel relay is staying engaged.  If it still happens with flex sensor unplugged then that rules that path out.  The only 12V source left is from the flex relay.  So this is why I say to unplug or snip pin 85 on that relay.

 

1 hour ago, loca5790 said:

If I pull the relay the back feed stops.

So this proves the voltage is coming from that relay.  If that relay really was already "off"/open then unplugging it would do nothing.  

[loca5790]

Hey Adam,

 

This is the third relay its been on.  Factory relay first, relay with built in resistor second, normal relay (current) and then I diode isolated that relay per the above diagram on the coil side of it.  If I unplug aux4 pin and watch voltage on it.  It immediately goes to 0 on key off.  That is with the FFS installed.  All of my last round of diagnostics before doing what I outlined above was with the FFS unplugged to completely eliminate it from the equation.

 

I 100% get where you are coming from and I would have agreed with you if I didn't just spend 2 weeks diagnosing this and rewiring the car 15 different ways.  

 

I changed relays from factory fuse box first because the factory always has constant power on pin 30 and 86 and ground controls 85.  I was told this was the issue so I rewired that.

 

The second relay I was told if it as not a normal relay and had a diode/resistor in it it could cause these issues so I removed it.  

 

The last relay was void of both but still did it so I tried diode isolating it to no avail. 

 

When the relay potentially "sticks" perhaps microwelds on three separate relays now, there is no voltage at pin 85 as seen by a wire probing the lead from the back side.

 

I also tried previously using the link Ign relay feature to power all relay coils on key up that also did not work.

[Adamw]

So did you try disconnecting the relay pin 85 when it is back feeding? This simple test would confirm right away if the relay is sticking (very unlikely) or if it is staying energised due to your wiring on the pin 85 supply. 

The fact remains the solenoid can only click if it has near 12V across it.  The aux output is effectively a ground when the ecu is off.  So that means there is 12V on the relay side of the solenoid, that 12V can only be coming from the relay or something else connected to that same relay pin such as the flex sensor.  

[Matt Dunn]

I have had similar issues before with this, and struggled to find someone who could explain,

I had been told ot was not possible but what I believe happens is,

When the key is on, there is power from relay output 87 to the valve, through the valve to the ECU aux.

What happens when you switch the ignition off, is that that power goes through the boost solenoid, into the ECU, comes back out of the ECU on the link power in wire, and goes back to pin 85 on the relay keeping the relay engergised to keep the loop going.

Whenever I wire and ECU now I always have the ECU on a separate relay,

or wire the power feed to the ecu to a relay output wire, not and relay input wire.

 

To test, you can change the feed to the Link to pin 87 on the relay, rather than looped off 85.

[loca5790]

Yea, I did try that as well Matt.  No success I just gave up and wired the relay to be switched by an rpm enable and/or a 2step enable.  Haven't gotten a timer setup yet for that but plan on it.

 

100% without a doubt was wired per the manual.

 

Adam, the issue is no with pin 85.  I tried about 50 different things over 2 weeks before I broke down per a friend's request who had said he saw the same issues that isn't isolated to one ECU.