How to integrate the Guardian Angel with a WMI Failsafe

[neganox]

I have more faith in snow as they make ONE type of product and seem to do it well. My only concern is that I can't put this flow sensor inbetween my solenoid and nozzle. My nozzle holder has a built in check valve so maybe having both is overkill anyways.

[kbz0129]

Glad to see you guys putting in the time to research. I've been running the Snow stage 2 kit for a couple months with 100% meth and so far so good. With my new build I was looking into a fail safe that was compatible with the GA as well as reliable. My first instinct was to stick with Snow because they haven't done me wrong yet and it looks like that was a good idea, you guys just confirmed it.

[Lex]

The snow kit is compatible with the GA as well. This weekend I will have a coolingmist kit installed and will write about its failsafe performance as well.

[jdmage_mx5]

I'm just waiting for a Guardian Angel that works with push & pull bpv's. @Lex; where's it at man?!

[Lex]

Originally Posted by jdmage_mx5 I'm just waiting for a Guardian Angel that works with push & pull bpv's. @Lex; where's it at man?!

Due to the nature of the pull type valves it's not possible to force them open unless you have a large vacuum canister of sorts. On top of this boost seals them - the more boost the more they are sealed so you need a good amount of vacuum to pull them open. This is why the HKS valves flutter.

[jdmage_mx5]

I still don't see why it's not available yet. Just kidding.

[KCi]

Is the GA can work with AEM meth kit (not failsafe)? If not mistaken, the AEM boost safe signal is ground triggered.

[England15j]

The aem kit has a ground and power signal wires

[Lex]

Originally Posted by England15j The aem kit has a ground and power signal wires

This is correct the AEM is fully compatible. If a kit only has a ground trigger, it can still work with a GA and a relay is needed. Very simple setup and I and I can provide the diagrams.

[dale_gribble]

Ok here's the snow safeinjection guts. It is the good old gems ft-210 sensor although it and the guts are all well potted and it appears snow opted for the better material housing than aem.

[Yatta]

Originally Posted by Lex Knew it. I believe it's the wheel/seals that swells so I am not sure why they are claiming 100% meth compatibility. The potting protects the electronics but it's not what fails as water will do the same to electronics.

Unless I am reading this wrong the sensor comes in a form that is meth/eth compat?

[Yatta]

Originally Posted by Lex Due to the nature of the pull type valves it's not possible to force them open unless you have a large vacuum canister of sorts. On top of this boost seals them - the more boost the more they are sealed so you need a good amount of vacuum to pull them open. This is why the HKS valves flutter.

So I thought I knew the answer to this previously, but also will not work with a synapse BPV in push? (allowing it open while under vacuum?)

[Lex]

Originally Posted by Yatta Unless I am reading this wrong the sensor comes in a form that is meth/eth compat?

You're right, looks to be the case. Wonder if AEM simply doesn't use the methanol compatible version.

Originally Posted by Yatta So I thought I knew the answer to this previously, but also will not work with a synapse BPV in push? (allowing it open while under vacuum?)

The GA will work with any valve in push mode.

[dale_gribble]

I can confirm that the aem definitely has a shiny body that differs from the snow.

[ItsNox]

Do you mean snow is shinier than the aem?

[dale_gribble]

Aem shiny. Snow dull.

[Yatta]

Originally Posted by Lex You're right, looks to be the case. Wonder if AEM simply doesn't use the methanol compatible version. The GA will work with any valve in push mode.

Thanks, will it prevent it from staying open off boost or only have an effect when it triggers?

[Lex]

Originally Posted by Yatta Thanks, will it prevent it from staying open off boost or only have an effect when it triggers?

If the VTA feature is toggled on it will prevent the bpv from being open at idle and allow you to safely run it in vta mode.

[dale_gribble]

Hey @Lex; I could use your help here.

So, in the Devil's Own System, the pump gets a constant 12V and the DVC-30 closes the ground to the pump to energize the pump.

The Snow SafeInjection system input trigger requires a +12V input.

I used a relay that should use the DVC-30 ground switch to close the relay coil and provide the 12v signal to the SafeInjection.

Problem is, the relay latches as soon as I connect power to the car even though the power should be isolated. It seems like the relay is somehow getting energized by the 12V constant through the pump wire?

Any ideas? this is a diagram. I can't figure it out.

[Lex]

Do you have a link to the SafeInjection instructions? The link on their site doesn't work. I want to see what the 12V is about.

[dale_gribble]

http://www.snowperformance.net/produ...e-file-119.pdf

[dale_gribble]

Hey Lex, don't sweat it. I found guys from the Evo forums with the exact same issue, and they cleverly (and simply) solved it with a pressure switch.

It's very simple, I'll hook up a vacuum line to it and +12v line, and when pressure gets to 15psi, it closes the switch and sends 12v to the SafeInjection trigger input. I used a Honeywell 76052. It's adjustable between 14 psi, 15 psi, 18 psi, 20 psi, 22 psi, & 23 psi (factory set at 15psi).



In a way I kind of like this better also, because now it makes the SafeInjection completely standalone and independent from the meth controller.

[Lex]

@dale_gribble; This is a good solution.

Just so everyone knows why he had to use a pressure switch:

1. The SafeInjection from Snow uses a 12V signal to tell it that the pump is spraying in order to diagnose a failure. Otherwise it wouldn't know how to diagnose a low flow condition. It would always say "low flow" when the pump is off. So we have to tell it when the pump is spraying.

2. The Snow controller varies the voltage on the positive line going to the pump. At full spray the positive wire will read battery voltage and at lower spray volumes it will read less than this. This is how the Snow controller varies the amount of WMI delivered on stage2 kits. The Snow controller uses what is called a high side driver.

3. The DO controller varies the ground signal to the pump instead and keeps the 12V constant at battery voltage all the time. They use a low side driver.

So the DO controller can't tell the Snow SafeInjection when the pump is spraying. In order to address this problem, a pressure/Hobbs switch is used. Once this switch sees X psi (set it close to full spray pressure for the controller) it switches 12V to the SafeInjection and now the SafeInjection can detect the low flow condition only when the pump is working.

[dale_gribble]

Ok guys, good news, the pressure switch worked PERFECTLY and the Devil's Own DVC-30 and the Snow Safe Injection are playing together. I just hooked the switch up, and the SafeInjection kicked into gear and the flow sensor and gauge are kicking (reading 250-280ml/min = 4-4.5gph which is pretty much dead on with how I have DVC-30 set). I tested the failsafe and it did trigger the Guardian Angel as expected.

I am running the SafeInjection on the output of the check valve, and I am trying to decide if I should have the failsafe before or after the check valve.

I left the pressure switch factory set at 15psi, because that's a good place for it to be. Now I just need to dial in the SafeInjection delay/min flow pots.

The pressure switch is Honeywell PN MFG Part No:76052-00000150-01

Mouser, Digikey and others have them in stock.

76052-00000150-01 Honeywell / Hobbs | Mouser

[Lex]

I would leave the failsafe as close to the nozzle as possible with the least number of connections in between.

I can see it having no problem detecting a 0 flow condition but can it detect a nozzle disconnected or a leak in the tubing with accuracy?

[dale_gribble]

This guy is only good for low/no flow. It has a setting for min flow and it has to hit that or triggers fault.

There aren't any other hooks than that.

[dale_gribble]

I tested the failsafe function quite a bit more and the Snow Failsafe is playing very well with the GA. This is in STARK CONTRAST to the AEM setup, which, even when I got to play with the GA once for like 5 minutes, the AEM caused the GA to freak out a bit. That AEM flow unit/failsafe/software is great in theory but a complete and absolute mess in practice. I would go so far as to recommend Lex remove that AEM option from the OP until someone much smarter than me gets it working right because I couldn't.

So, the SNOW SafeInjection has a flow delay once the 12V pump trigger signal goes high. Well, the pressure switch I installed to interface with DVC-30 triggers at 15psi. I can create a no flow condition by setting the DVC-30 to start spraying at 18psi and that creates a delay fault (the time the pressure switch triggers 12V to the failsafe to the time the DVC-30 actually starts spraying). It triggers the GA, the GA failsafe LED comes on and the BPV opens as expected. As soon as the fail condition clears (i.e. I let of gas) the GA returns operation.

Next, I set the DVC-30 where it would begin flowing at 15psi to match the pressure switch and then I upped the failsafe minimum flow rate past what the DVC-30 would flow at 15-20psi. Again, the Failsafe triggered a fault and the GA responded appropriately.

So, I simulated a low flow and a no flow situation and the failsafe + GA operated as expected automatically in real time.

The optional SNOW Digital Flow Gauge is acting a little odd, but I think it's a matter of dialing the meth curve in on the DVC-30 controller when we actually start tuning meth again. The flow gauge is just kind of fun, it's not like I can actually read it during a pull or when passing on the highway anyways. The way the gauge looks, there is a high flow inrush and then it stabilizes. Since the flow sensor is on the 'checked' side of the check valve this makes sense to me.

I am starting to get confidence in this failsafe setup.

[dale_gribble]

Couple days in, flow sensor and setup are acting a little erratically. The behavior is repeatable, the flow sensor logs 200-300ml/min briefly and then drops to 0 (triggering the failsafe) even though I can monitor BATs and clearly see the temps are still down and meth is still flowing. This is pretty repeatable. There also seems to be times where the flow sensor doesn't respond at all and seems 'stuck' and then it will start responding again.

I called SNOW. They have never seen this before (so he says).

I called Chance @ Devil's Own and he said he couldn't build a failsafe because he said the low end GEMS sensors are garbage and the one he would need to use is the $100 sensor and that would push the price point out of acceptable (well beyond $200).

Chance recommended triggering a failsafe off of the BATs (i.e. when under boost >15psi and BATs >100F). I thought that was a pretty smart idea. Otherwise he recommended going with the high end GEMS sensor.

I might send my SAFEINJECTION kit to SNOW under RMA and see if it passes on their test bench to decide whether its them or whether the setup just isn't going to work for me.

[SparkySparks]

Lame, I wonder what it would cost to make a Fail safe the would require you to source your own sensor and then provide a list of sensors that would work or maybe just sensor output requirements.

[dale_gribble]

If you have run methanol for a while, you know that even in the hottest days BATs aren't going to get above 100F with meth @ wot under boost.

Here's the product design. Read in BATs from the MAP sensor to a standalone unit. Monitor BATs. Read in boost from a boost sensor. When boost goes >15psi (or adjustable via pot), track BATs; BATs must drop a certain level within a certain time with adjustable delay (from when boost goes >15psi to where unit samples BATs). Something along those lines would probably be pretty cool.

[dale_gribble]

Ok, guys. This is the hard part about forums, facebook, social media, etc. You always read about the successes, but often the fails are not equally on display. It is with great humility that I post this.

I decided to see how the meth was flowing out of the nozzle last night to see if the flow through the SNOW unit would hold steady when I applied boost pressure to the controller/failsafe lines (with air compressor). What if I told you the SNOW SAFEINJECTION actually CAUGHT a partially clogged nozzle AND that the changes in flow were due to what is likely increased backpressure from the partially clogged nozzle? I am probably lucky that I didn't pop a line or fitting.

I am face palming myself. It appears that some sealant goop from the nozzle threads got onto the filter screen. It was maybe 25% blocked, which apparently is a big deal to these nozzles.

I cleaned the nozzle screen and nozzle in solvent and reinstalled. I ran the boost/flow test with the car turned off in the garage, and sure enough, the nozzle flowed evenly and the gauge/flow sensor held perfectly even for 10-15 seconds at a time. I put the car back together and took it for a spin and the flow sensor held perfectly even.

I created a total hell for myself when the culprit was likely the simplest variable. I was so confident that it wasn't the nozzle after a recent cleaning that I didn't think to even look at it. This brings into question the operation of the AEM setups which I suspect would work if the nozzle was cleaned. I think one of the AEM setups would still work.

In short, I am a retard and I was a huge detriment to this discussion. I feel pretty shitty honestly because the last thing I want to do is misinform/redirect someone on false/incorrect info so I want to apologize to all of you for that.

[Lex]

Thanks for the heads up on this. Do you still have any of the AEM kits? Just to throw in line in the garage (without installing) and see if indeed it works.

[dale_gribble]

I am going to rma aem kits and let them sort them out. When they come back I am going to test.

[MazdaspeedKills]

@dale_gribble

I just finished my meth install using the snow progressive controller, the AEM failsafe, and the GA. I had problems with the AEM gauge. Broken on receipt. Sent to AEM for repair and since it appears to be working. I haven't encountered any issues with the flow gauge at this time. I preferred the AEM since it does have both high and low flow alarms. But the fact that it was broken when i received it didnt give me a warm and fuzzy. it was a diode in the gauge that failed not the flow sensor.

@Lex I took a look at your diagram. This will cover you for one fail scenario but not both. I ended up utilizing a relay to revert the low flow ground output to a power source and linked both the blue (low flow ground) and yellow (high flow power) wires to the GA. So i'm now covered on both ends. I also added a audible alarm in case for some reason I'm not watching the gauge.

If you guys have questions I might be able to help.

[MazdaspeedKills]

@Lex Upon further testing of my setup, I would like to retract my previous post. Your diagram is correct. AEM's manual is not. Reading the manual, AEM suggests that the blue wire be used for a low flow alarm scenario where ground is required and the yellow wire for a high flow alarm scenario where power is required. I found that if the sensor picks up either alarm scenario it will provide ground and power via those two wires at the same time. Therefore in the event your sensor triggers an alarm use EITHER the yellow or the blue wire depending on what you want to occur for your setup. For the AEM WMI failsafe and Guardian Angel combination the yellow wire is the only one you will need.

[Lex]

Originally Posted by MazdaspeedKills @Lex Upon further testing of my setup, I would like to retract my previous post. Your diagram is correct. AEM's manual is not. Reading the manual, AEM suggests that the blue wire be used for a low flow alarm scenario where ground is required and the yellow wire for a high flow alarm scenario where power is required. I found that if the sensor picks up either alarm scenario it will provide ground and power via those two wires at the same time. Therefore in the event your sensor triggers an alarm use EITHER the yellow or the blue wire depending on that you what to occur for your setup. For the AEM WMI failsafe and Guardian Angel combination the yellow wire is the only one you will need.

AEM just wants to make sure their system can trigger a device either with a high signal or a low signal.

[dale_gribble]

Ok update on my end, the snow failsafe started working intermittently and now won't read at all even though bats drop as expected and meth is definitely still flowing. All meth/vacuum/electric connections are fine and the snow failsafe is kicking in since it isn't measuring flow.

I am RMA 'ing the 2 aem setups and the snow setup tomorrow. We'll see what comes back.

FML...

I'm thinking about just running the DO status leds to big dummy lights on the dash and calling it good.

[neganox]

I'm starting to think running a failsafe is more trouble than its worth.

[dale_gribble]

Right there with you. My faith is very low atm.

[dale_gribble]

So, I already disconnected the SafeInjection failsafe from GA because I got SO SICK of the SI triggering the GA when it didn't read flow. So, I just watched it today driving in, and sometimes it would read a number for a second but most of the time not at all even though BATs were diving into the 50s-60s through an entire 3rd/4th pull.

If I hook it up when the car is off and I trigger a steady flow, it reads fine. When I use it on the car in real world, it sucks. I have done everything possible to mitigate any setup issues. Everything is working like it should but the flow sensor. It just can't seem to get a decent consistent read at all and it reads inconsistenly just like the AEM sensor (when it read at all). If I had to guess, I think the flow sensors can deal with flow ok, but they don't like transients (i.e. rapid ramp up/ramp down)

So, even when it worked good, it didn't work for that long so it's back to sucktown.