How to integrate the Guardian Angel with a WMI Failsafe
 

We got this question a couple of times this week alone, so here's a diagram showing how to connect the GA to an AEM Failsafe Gauge. It is that easy!

Hi Lex,

I just ordered the GA, after the bent valves incident, please can you assist with a fail safe diagram for DVC 30 controller from devils own,

Regards
Ricardo

PS : hope to be up and running soon !!

This is the exact setup I did. If you look carefully, you can see the Guardian Angel LED indicator between the gauges.

http://images60.fotki.com/v367/photo...GA_LED3-vi.jpg

The cool thing is that the AEM failsafe has a USB cable and Windows based program that allows you to trigger the failsafe condition through the gauge. This means you can test the whole failsafe end to end after install and not just in the heat of battle.

Ricardo, thanks for the GA order. The DVC-30 does not actually detect a methanol flow problem. What you need is an AEM failsafe system that measures WMI flow and triggers a failsafe - similar to what dale_gribble just below has setup.

The AEM failsafe will work with your DVC-30 pump controller. If you'd like an AEM failsafe, send me a message, we can get one sent your way.

@dale_gribble; are the ends of the flow meter sensor changeable if one wants to run a complete AN/SS line system vice push in fittings?

The ends are changeable, yes. I think you can even buy the kit with threaded fittings.

Having said that, my newly installed kit isn't working. Meth is spraying fine but flow guage isn't reporting anything.

I double and triple checked all connections and fittings, flow sensor direction, 0 install issues.

UPDATE: THE AEM flow gauge is a total piece of shit. It isn't rated for temps over 100F and I have already broken 2 of the flow sensors due to heat. AEM recommends that the flow sensor be in the passenger compartment or trunk (LOL) or basically completely away from any heat.

I don't know where on a car you can guarantee ambient below 100F. For the reason, I cannot recommend AEM flow or failsafe gauges because their flow sensor is garbage.

Well shit.

Actually now I'm not sure if it's heat or pump pulses. I'll provide more later.

I don't get how 100F would be the max temp. We see >100F ambient temps here in Florida. Bah!

I will be upfront about my experience with AEM electronics and this is mostly limited to their widebands. I would run them in my standalone builds and they would overheat and melt the connectors off the back of the gauge. The analog readings would be about half point off the digital readings on the gauge. I moved to Innovate and didn't have these issues.

I've seen a lot of builds using AEM ECUs and with their popularity I thought they had a turned a corner reliability wise. Dale, let me know your findings - 100F seems unreasonable for the sensor ... is the sensor a propeller type?

In 4-6 months we should have these woes resolved with a much more resilient failsafe but until then it's hard to find something to reliably recommend so I hope this can be resolved for you and we have something that works with the AEM kit.

Ok guys here is the AEM flow sensor. It's the thing in the lower left hand sensor.

http://www.maperformance.com/images/T/aem_30-2030.jpg

It only took me a little while searching to find their flow sensor supplier, it's gemsensors.com and the part number is FT-210.

http://www.gemssensors.com/~/media/I...hx?h=270&w=270

Website: FT-210 Series TurboFlow; Turbine Electronic Flow Sensors

Now, what first comes to mind is they have an 'Instruction Bulletin' it's here. Of note is the following: "The system needs to be filtered to 50 microns prior to the sensor, and pulses/water hammer effects should be minimized to prevent unit damage."

http://www.gemssensors.com/~/media/G...ns/FT-210.ashx

Well shit, of course AEM doesn't include a filter (which isn't a bad idea anyways), but the part that gets me is the 'pulses/hammer effect'. My DVC-30 pulses the pump. The higher duty cycle (more pulses), the higher the flow and vice versa. If you've ever had a meth pump pumping into a water bucket, you can see the pulsing/hammer effect on full display. I assume that all controllers pulse the pump. So, I suspect the pulsing meth pump is what is damaging the flow sensors instead of temperature.

So, AEM is basically is selling a kit with a piece of shit flow sensor.

BTW, my AEM EGT sensor is also bouncing around when temps go over 1400F unless I accelerate very slow. Sometimes it doesn't do this but 90% of time it does. I have been over things 10 times, and if you know my work I am pretty careful with things.

I am in full AEM hate mode right now.

EDIT: Keywords for anyone ever searching MSF or even google for this POS kit so they find this: AEM Water/Methanol Failsafe Gauge, 30-3020

AEM's filter is only good for 150psi and 50/50 mix meth. I'm pretty sure snow and DO use the same 5800 series pump.

Not sure if snow's safeinjection / gauge combo is any better.

It really comes down to the robustness of the flow sensor. If SNOW sells the kit alongside their meth systems, I would be more inclined to believe they selected a better flow sensor to work, whereas AEM probably just picked whatever flow sensor they could find. I could be wrong.

I emailed AEM my findings. I asked them to source me a flow sensor that won't crap out that is compatible with their system.

The snow flow sensor is not any better from my research. It uses the same turbine style sensor.

There are two other current solutions on the market that are different. Aquamist. Their DSS3 was just meant to be the failsafe but I don't think they sell it as a standalone anymore. I've worked with their sensor and it is not a turbine style sensor and 100% meth compatible.

The Coolingmist failsafe. I am about to install one of these systems - it doesn't use a direct flow sensor but rather monitors pump current draw for deviations. This is an excellent idea as long as the rest of the system is calibrated ... in other words you might need to use their system.

Ugh, really sorry to see this. I'm surprised the guys at MotoIQ recommend this system ...

Lets see what AEM comes back with. It might be a bad batch of sensors or maybe something they aren't aware of. Unfortunately I have a kit arriving tomorrow. Maybe I'll have better luck.

Looks like we may have mis-read. -4C - 100C. So its really good until 212F. I doubt it could see that heat operating with methanol/water flowing through it.

I'm going to set my system up with filter -> solenoid -> flow sensor -> nozzle. From what I've read, all the pumps pulse, even the pump with the AEM kit. When I tested my system you could hear the pump pulsing(or maybe just pumping?), but the flow looked stable. It looked like a hose flowing water and with the nozzle attached it looked like a cloud of mist going bananas.

If there are pulses, maybe the inline filter would stabilize them a bit.

The filter and solenoid might help mitigate and smoothen pump pulses. My flow sensor is the first thing after the pump.

The other thing I noticed is that the sensor input voltage is 5-24vdc. When I checked, I am reading just under 5v at the sensor. I reckon the aem gauge is outputting 5v and there is voltage drop over the 24 - 28awg 6' cable length.

It could be the 5v is also putting the flow sensor in a borderline state. From my experience, devices do not like operating at the lower end of their voltage input range. It's usually the least efficient there.

All these pumps do pulse. Basically the top of the pump has several one way valves on a platform and it oscillates filling the space with fluid on one end and pressurizing it out the other. The less voltage you provide the slower the oscillation and the more it "pulses" All these stage 2 kits vary the voltage from min to max at the pump.

Most of the issues I've had with aem are cheap connectors and components so this is something to watch for.

As a general rule, these flow sensors are designed for water and they can swell and seize if too much methanol is present so they are not a good 100% meth solution. 50/50 should be fine. Chemically resistant sensors are much more expensive.

With filters, check valves and sensors in line it creates pressure drops. You don't want to drop pressure too much as it will cause an on delay for he system and poor atomization especially at lower pump voltages. My suggestion is to run the system as a test outside the charge pipe with just water to make sure everything looks good and sensor is reading.

AEM told me 100 degrees over the phone but I reckon they don't know the difference between 100f and 100c.

Another note. You can siphon fluid through these pumps. So a check valve or solenoid is a must. Last thing you want is to gravity feed your charge piping with fluid and hydrolock.

Solenoid all the things!

I've been on the do check valve for 1.5 years without issue.

I ran a hose with a nozzle through the two AEM flow sensors I have. One flow sensor's turbine seemed to work; I could hear it whirring and changing frequency as I passed water through it. The other didn't. So, I put the good one back in.

On the way to work, one time randomly on the highway it briefly registered flow when I accelerated and then of course the gauge hasn't moved since even though I can see meth spraying by watching BATs.

Apparently inline filters are supposed to go on the low pressure side of the pump.. so between the tank and the pump. There goes my idea.

Any word from AEM? Even with a product that has spotty reliability I can't imagine they don't work right out of the box or they wouldn't sell them.

Haven't heard yet. I keep reading about 'WMI' flow sensors, and it is pretty apparent they are problematic as a general statement. I have read a number of cases where coolingmist's sensor is also unpredictable and read other things where flow sensors from WMI are just crap and not designed for this application (high heat, pulsed).

The only one I can't find anything negative about is aquamist's sensor. It looks like that one is a ground up design made for meth.

I don't know how to get a read on this. Of course they must work to some extent, but there is plenty about griping about WMI flow sensors out there. I found a BMW forum post where people were particularly down on WMI flow sensors. I suspect a lot of people piss about it or just deal with it and don't say any more about it. Who wants to say they got duped into spending $250 on a total piece of crap that doesn't work?

Lols

AEM FAILSAFE REGISTRY: AEM Failsafe Fail Registry - Subaru Impreza WRX STI Forums: IWSTI.com

Cobalt owner reports not working out of box:AEM failsfe meth gauge - Cobalt SS Network

Mad subi guy: Having issues with you AEM Methonal Failsafe failing? Check here - NASIOC

Another one: http://forums.nasioc.com/forums/show....php?t=2639502

Bad reviews:

http://www.summitracing.com/parts/avm-30-3020

http://www.rallysportdirect.com/AEM-...Monitor-System

There's quite a bit more. Quite a bit on coolingmist too. There is some for the Aquamist failsafe and temp sensitivity also although I get the impression that the aquamist may be a better part. Interestingly, not a lot of bad feedback on the Snow Performance failsafe.

Here's the aquamist flow sensor. The aquamist appears to be a custom engineered design for/by them and improved since 2002.

http://i85.photobucket.com/albums/k4...flowsensor.jpg

Here's the GEMS sensor, off the shelf, not designed for meth/automotive.

http://i85.photobucket.com/albums/k4...lowsensor2.jpg

Some decent reading here besides the pissing contest.

http://www.e90post.com/forums/showthread.php?t=566589

My takeaway is that the flow sensor setup is marginal for WMI and Aquamist appears to have the best one; snow's may be decent too. Lots of issues with CM/AEM if you do searches like I did. It's just really spread out and not a lot of people run them. It looks like the Aquamist isn't compatible with AEM failsafe (analog vs. pwm). It looks like AEM is only PWM based input; the rest are all 0-5v analog.

Honestly, I can't find a lot bad about the Snow Performance safeinjection setup and that thing's been out since 08. The Snow Safeinjection setup and gauge can be had for about $75 more than the AEM garbage. You can at least tell that snow attempted to make the flow sensor more rugged/robust. With the size of the box, I reckon they are using a beefier sensor.

http://i8.photobucket.com/albums/a50...snow_30020.jpg

That could just be another bum flow sensor in a pretty box. However for the good of this discussion I will acquire said pretty box and open it up and post my findings.

I found this. This appears to be their old design. I sure as shit hope their new design doesn't include gem flow sensors. Their old design included usage of a gem FS-4 flow switch.

https://www.google.com/patents/US758...ed=0CDQQ6AEwAw

The fs-4 flow switch indicates flow/no flow and nothing in between when the switch contact is open/closed.

That patent mentions an implementation with a flow sensor as well which is what they would need to meter flow.

So who wants to call snow and dig?

I emailed this this morning asking for specifics of the flow sensor. Hopefully they'll provide some feedback.

The chemical resistant flow sensors exist but are expensive.

There's another method and that is to use a pressure sensor. It has to be a sealed unit so their prices for the sensor alone is 100-150. They do offer a 0-5v output so they may actually work with the gauge you already have. Let me see about finding a few part numbers for compatible sensors.

Snow lists their failsafe as 100% methanol compatible. We shall see! ...well that is if it is at least 50% methanol compatible

I have one more thing going on here that might be a manufacturing error on the AEM system.

I happened to order an AEM WMI flow kit before I decided I wanted the failsafe. So I had that on my desk here. it is essentially the same thing- same board, same flow sensor, same cable, etc. I started probing the brand new flow gauge and measured the following on the flow sensor pins (when the unit was powered on)-

Pin 1 - 5v
Pin 2 - GND
Pin 3 - 4.681v
Resistance pin 1 to 3 (supposed to have pullup resistor): 30kohm
Calculated I (load on 5v for pullup)= 170mA

Now, here's the same readings on the failsafe

Pin 1 - 4.95v
Pin 2 - GND
Pin 3 - 4.95v
Resistance pin 1 to 3 (supposed to have pullup resistor): 10kohm
Calculated I (load on 5v for pullup) = 500mA

Now it's starting to make some sense. I think the resistor in the failsafe is actually too high for the flow meter NPN collector to sink to 0v. The flow sensor guide states that the unit requires max 20mA current to sink the output to 0v. The pullup resistor needs to be just low enough so the collector can sink to 0v when sunk to ground but high enough to pull right back up. The 500mA of the failsafe is definitely too high for the flow sensor to overcome.

EDIT: Update. When I put the flow sensor on the new flow gauge, the supply voltage remains at 5V. When I put the flow sensor on the failsafe gauge, the supply voltage drops to a volt. Makes perfect sense. There's too much load on the power supply. I am going to swap the 10k resistor to a 30k resistor on the failsafe and see what happens.

My response from Snow.

Hi Neal,

Sorry for the delayed response.

I do not know the specific temperature that the SI is good for but it has
been built to withstand under hood temperatures and conditions. The engine
bay is typically where customers mount the SI. Our sensor is also compatible
with 100% methanol. All of our system use PWM to command the pump and we
have never had issues with the SI's. Our SI units do have a maximum flow
limit of 1500 ml/min.

Sincerely,

Joshua Guritz
Snow Performance Inc.
jguritz@snowperformance.net
#719-633-3811 Ext. 104

This doesn't seem to be going anywhere. I swapped gauges and still 0 reading with either flow sensor. I am back to the flow sensor as the culprit. It seems like the pullup resistor may be matched to a flow sensor based on variation of the flow sensor- kind of like a calibration. It's just a hunch though. The 30kohm resistor is allowing the flow sensor output to sink to 0 volts but I can't make the flow sensor move to see if it pulses on the output. I don't really have a way to test flow on the bench here.

Would blowing air through it be enough to make it spin a bit?

Naaa, tried. I am officially closing book on AEM failsafe and moving on. I am going to mail flow sensors + gauges back to AEM, get refurbished/repaired and sell on fleabay.

I called Snow and talked with Josh. Unfortunately I did really want to speak with a designer about it and see what flow sensor they used. Josh did state the system was designed for 100% meth and was designed for pulsed systems as @neganox; said.

I too think I will go the safeinjection route. The Snow safeinjection has been out since 06 and I can't find any real dissatisfaction with it whereas the AEM has been out for less time and I can find a lot of folks griping about it. I did like how AEM mapped the progressive flow, but if shit doesn't work, shit doesn't work. The safeinjection relies on a minimum flow setting (based on nozzle) + time delay for flow to build up once flow starts.

There are parts I like of each system but no systems have nailed it.