How to integrate the Guardian Angel with a WMI Failsafe

[Lex]

I have the Coolingmist up and running with their failsafe. They don't use a direct sensor but rely on pump current variations to determine if something happened. Ran out of fluid on purpose and it triggered right away. I will write more as I put more miles on it.

[neganox]

Factoring in the price of the snow kit + their failsafe + gauge + headache = price of an aquamist kit. I regret everything!

[jack_hammer]

Originally Posted by Lex I have the Coolingmist up and running with their failsafe. They don't use a direct sensor but rely on pump current variations to determine if something happened. Ran out of fluid on purpose and it triggered right away. I will write more as I put more miles on it.

This kit interests me since it is cheaper than aquamist but still has a failsafe that "sounds" promising based on their claims

[radbrad]

Originally Posted by Lex I have the Coolingmist up and running with their failsafe. They don't use a direct sensor but rely on pump current variations to determine if something happened. Ran out of fluid on purpose and it triggered right away. I will write more as I put more miles on it.

This sounds good. If I do end up adding fuel via meth injection. It will not be without a fail safe. Please let us know if it works as if It should.

[Lex]

Originally Posted by jack_hammer This kit interests me since it is cheaper than aquamist but still has a failsafe that "sounds" promising based on their claims

The science is sound but the implementation I need to make sure supports this. So far it has caught the tank running out which is a common failure. Clogged nozzles will be a little harder to catch as will small leaks.

False positives are very frustrating so this is something I am keeping a close eye on.

[Lex]

Put together 3 videos showing the install of the Coolingmist system in our Focus ST. So far no false positives and it caught the system losing prime and running out. Although this is a different platform, the ST is close to the MazdaSpeed and a lot of the same principles apply so I'm posting this video to show how the kit looks.

The pump feed through the top of the washer reservoir is a nice idea especially for gen1 cars where pulling through the OEM WMI pump is a bigger hassle. This is in Part 2 of the video.

Part 1: Under the car
https://www.youtube.com/watch?v=ZmGY...ZrMSXjBCfhIg3g

Part 2: Under the hood
https://www.youtube.com/watch?v=Kr1t...ZrMSXjBCfhIg3g

Part 3: Inside the car
https://www.youtube.com/watch?v=TehV...ZrMSXjBCfhIg3g

[dale_gribble]

I'm glad to hear the cooling mist system is doing ok. I have a call directly into GEMS sensors to discuss if they have a more robust/HD version that is functionally compatible with the FT-210 (even if cost is higher or form factor is larger). I want to get to the bottom of what isn't working here.

[Dust]

Originally Posted by neganox Factoring in the price of the snow kit + their failsafe + gauge + headache = price of an aquamist kit. I regret everything!

Picked up a used HFS-6 for $100 last week.

[jack_hammer]

Originally Posted by Dust Picked up a used HFS-6 for $100 last week.

You suck

[neganox]

Originally Posted by Dust Picked up a used HFS-6 for $100 last week.

lucky duck!

[Dust]

Originally Posted by jack_hammer You suck

Originally Posted by neganox lucky duck!

Picked up one last month for $350. Don't know what I need 7 meth kits for, but one day I will.

[neganox]

Clearly to form some sort of devastator but made with said kits.

[Dust]

neganox, is it bad that I saw a Transformer and your screen name became meganfox?

[neganox]

Maybe.

[kbz0129]

Originally Posted by Lex 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 can provide the diagrams.

@Lex Can you hook a brotha up with those diagrams? I was looking into the AEM wideband failsafe and it looks to only have a ground trigger. But, I may be wrong...

[HawkeyeGeoff]

@maisonvi; this could be useful

[Lex]

Originally Posted by kbz0129 @Lex Can you hook a brotha up with those diagrams? I was looking into the AEM wideband failsafe and it looks to only have a ground trigger. But, I may be wrong...

Here it is how to hook up a simple relay. The square is the relay.

This is not going to be an issue with the GA V2 as it triggers on both high and low input signals.

[dale_gribble]

I just got an update from AEM on my RMA of the two flow sensors/gauges I sent them. Both flow sensors tested bad.

I spoke with SNOW on the SAFEINJECTION and they swear up and down the intermittent behavior from the flow sensor is bad wiring or a loose connector. They have had reports of intermittent connectors causing this behavior. Given that they use a non rugged computer style commercial 5 pin connector, maybe they are on to something. They sent me a new and tested connector to try. I don't get why they POT the entire body of the unit and then use a crappy loose connector.

So, results to post soon. I am going to test the AEMs when they come in.

[MazdaspeedKills]

@dale_gribble

I've been following this thread closely as you and I seem to be on the same upgrade path with the same car.

I've encountered a variety of issues with my meth install. The latest I blamed on the AEM failsafe. Very similar symptoms. First issue: the gauge read wildiy 700cc on a 441cc nozzle. Then it dropped to 250cc and then nothing. Through troubleshooting I found that the high flow read was due to air remaining in the system ( System was not fully primed). The low flow issues I encountered were the result of blockage to the sensor impeller. Initially, I was not running a tank tap filter or in line filter. When i looked in my tank, I had a bunch of shit in here (small bugs, particles etc.). Funnel i used sat around in the garage awhile so I blamed that. Filtered the meth mix with a coffee filter, flushed the tank and installed both a tank tap and in line filter between the tank and the pump. I removed the flow sensor and blew into it just using my mouth and was able to get the impeller spinning again. Saw the same high flow results as before but that disappeared once the air from the system abated. The gauge is now reading plausible numbers however now I'm seeing a lower than expected flow rate ~390cc vs maximum (nozzle limiting) flow rate of 441 cc.

Current setup is as follows:

Snow Performance progressive controller running in boost mode/AEM failsafe/Guardian Angel

trunk installed tank -> Devil's Own filtered tank tap ->AEM inline filter->Cooling Mist pump->AEM flow sensor->Devil's Own solenoid->Devil's Own DO7 nozzle (using a snow performance CV integrated holder)

My current troubleshooting objective is to determine whether I'm seeing normal, parasitic loss due to the filters, flow sensor, and CV i have installed in-line or a partial blockage in one of the components in the system. A 12% flow loss may be reasonable but i want to make sure. Plan to pull everything apart inspect and clean. I also called AEM and got an RMA #. I'm having them bench test the gauge and flow sensor since I already encountered one issue with the gauge already and then had this flow issue.

Frankly, after my initial issue with a bad gauge, then the erratic flow, I don't trust the failsafe either. If this thing is going to be blocked for any reason, I cant see trusting it for high or low flow scenarios. But right now I'm still optimistic that i'll be able to get it to work.

Will keep you posted. My plan for meth going forward is to run it untuned during the winter to gain a little comfort with the reliability and then tune this coming spring if all goes well.

Originally Posted by Lex Here it is how to hook up a simple relay. The square is the relay. This is not going to be an issue with the GA V2 as it triggers on both high and low input signals.

This is exactly what I did for the low flow ground on the AEM failsafe output until I realized both power and ground are supplied when the unit triggers an alarm scenario.

[Dust]

If anyone needs/wants a simple filter, the plastic Purolator F20011 will fit outside the 1/4" push to connect hose. You could use it to clean up a bad tank of injectant, or, like me, you could use it full time(with 20-50% meth). I will say that Julio from AlkyControl did not like the idea. I used it for almost a year until I went to the AEM inline filter and 3/8" pre-pump hose due to reduced pressure from a dying pump.

Mazdaspeedkills, unless you are running two nozzles I don't understand why you are running a check valve and a solenoid. I didn't even know snow had a check valve. You could test flow by removing the solenoid and check valve separately. The check valve will reduce pressure and therefore flow.

[MazdaspeedKills]

Originally Posted by Dust Mazdaspeedkills, unless you are running two nozzles I don't understand why you are running a check valve and a solenoid. I didn't even know snow had a check valve. You could test flow by removing the solenoid and check valve separately. The check valve will reduce pressure and therefore flow.

The only reason I'm running both is to cover the volume of fluid left in the line between the nozzle and my solenoid that could induced into the engine. It also ensures this run of line remains primed. But we're talking 10 inches of line so its not much liquid. It's also my last line of defense for the whole system if the solenoid were to fail open (which again would be a very very low risk). It's really overkill but I'm risk adverse.

Snow used to provide a standard check valve. They changed this recently. They're now providing new nozzle holders that include a check valve in them. Pretty slick IMO.

I like your idea of eliminating the CV & solenoid just to rule them out. But my $10 says there's something up with that flow gauge especially after reading Dale's issues with his flow gauges. My sensor already got stuck once...

[Dust]

This check valve
Snow Performance : 82204 Nozzle Holder-check valve

Looks alot like this

Metal Nozzle Holder Fitting

I agree with your assessment, but you know that any CSR will try to troubleshoot/blame other things first.


Another thought. The coolingmist CMGS v1.07 has a "delay" for their flow sensor that offsets the displayed flow based on the distance of the flow sensor from the nozzle. Does the AEM have a delay? It might not be the reason for the reduced flow, but might help...something...

The solenoid failing is very rare, and could only be detected by the flow sensor or wideband, and I don't even know if you would know about it on the flow sensor unless you saw flow when the kit was off.

[MazdaspeedKills]

Originally Posted by Dust This check valve Snow Performance : 82204 Nozzle Holder-check valve Looks alot like this Metal Nozzle Holder Fitting

You're absolutely right. DO and Snow Housings look identical but the Snow holder has additional internals where the DO is simply empty on the inside. When initially looking at nozzle holders I came to the same conclusion you did, figured Snow's description was screwed up and bought the Devil's Own holder because it looked similar to Snow's but cost less. I also had a separate (by used) check valve I could install. I tested my check valve; it was stuck closed so that went in the trash. I called Snow to confirm their holder did have an internal CV, the rep confirmed it did, and I bought it. When I got it in, a visual inspection confirmed the internal CV.

Highly recommend the Snow Performance holder over the DO if you want to run a CV.

[Lex]

Originally Posted by Dust This check valve Snow Performance : 82204 Nozzle Holder-check valve Looks alot like this Metal Nozzle Holder Fitting I agree with your assessment, but you know that any CSR will try to troubleshoot/blame other things first. Another thought. The coolingmist CMGS has a "delay" for their flow sensor that offsets the displayed flow based on the distance of the flow sensor from the nozzle. Does the AEM have a delay? It might not be the reason for the reduced flow, but might help...something... The solenoid failing is very rare, and could only be detected by the flow sensor or wideband, and I don't even know if you would know about it on the flow sensor unless you saw flow when the kit was off.

The CMGS does not measure flow directly. It measures the current drawn by the pump and infers if there is a problem. This is why I went with this system. It is actually an idea I had about a failsafe and was testing in on the side when I realized someone already made one. So I bought the kit.

[Dust]

Originally Posted by Lex The CMGS does not measure flow directly. It measures the current drawn by the pump and infers if there is a problem. This is why I went with this system. It is actually an idea I had about a failsafe and was testing in on the side when I realized someone already made one. So I bought the kit.

I edited the post you quoted for the version I am using. The older CMGS, with 1.07 firmware uses a flow meter instead of their FSB. I went used 1.07 instead of the new 1.10 because the older firmware could use the "failsafe" output to control a second stage independently.

MazdaspeedKills, If I hadn't just ordered 2 DO check valves I would probably pick one up. $27 shipped on ebay. Darn it.

[Lex]

Just had another failure picked up by the CMGS. This was a leak at the check valve. Not a fully blown hose, just a leak, and the Coolingmist failsafe picked it up. So far no false positives.

[dale_gribble]

Here's an update on my failsafe saga-

SNOW failsafe. It appears as though the SNOW failsafe is working better with the replacement harness they sent me. It does occasionally not show a reading and trigger a false positive. I suspect this may be used to the pressure switch trigger I am using though; there might be some hysteresis in the pressure switch trigger pressure and sometimes it doesn't trigger the SNOW. When the SNOW reads now, it doesn't bounce around and it holds a steady 500ml/min reading under full spray. I am thinking about seeing if I can get a +12V trigger using one of the DO controller LEDs (i.e. flow LED). Maybe I'll get lucky and they don't do any DC-DC stepdown for their LEDs. I still need to build confidence in this system, but it seems to be ok for now.

I did get the AEM units back. I decided to try the failsafe again and see how it does. On the plus side, the first AEM unit/sensor I tried triggers flow every single time. Every single time. On the downside, it doesn't measure flow right. I saw it briefly measure correct flow last night for about 30 seconds (fluid movement to peak ~500cc/min), otherwise it jumps around between 700-750ml/min very abruptly. So, on one side, I can use it for an indicator and failsafe as it is, but it doesn't meter for shit so I really couldn't set the upper/lower failsafe limits very well. I might swap in the other spare AEM parts I have (flow sensor and/or gauge) and try to isolate where the shitty metering is coming from.

I guess it's a step in the right direction. The SNOW appears to be working save some intermittent issue with the trigger and the AEM is detecting flow although the metering is shit.

[Dust]

Dale, just for clarification, did the Snow and AEM sensors get installed in the same location? Are either running a filter before the flow sensor?

[dale_gribble]

Ok, big update here.

I decided to take a flow sensor and AEM failsafe and test them on a bench to get to the bottom of this. First, I plotted the curve of the FT-210 flow gauge (which is linear). This is pulse rate (hz) vs. flow from Gem's website.

I used an o-scope to measure the pulse rate and plotted on the flow sensor's curve. You can see the red line plotted on the blue line. The reported flow from the sensor matches the failsafe gauge reading. I repeated this with two flow sensors and results were consistent.







Because the results were consistent across both flow sensors and the flow gauge matched the sensor readings, I think I feel better about the AEM metering flow right. So, why then am I seeing ~700cc/min when the D07 is rated to ~440cc/min? It's elementary as I found out (and feel stupid about).

Devil's Own (SNOW too) rate their nozzles at somewhere between 60-100psi (I read snow is 60psi and Devil's Own is 100psi). All of their pumps are factory set at 200psi. What these means is, the nozzles flow a significant amount more at 200psi than they do at 100psi. So, when you read that a D07 flows 7 gallons/hour (441cc/min), that's at ~100psi. In reality, the D07 flows much more at 200psi pump pressure (650-700cc/min).

This cool guy on a vw forum flow checked a bunch of nozzles at various pump pressures. Here you can see he measured the D07 at ~655cc/min @ 200psi.

Has anyone ever done a WM nozzle flow benchmark? - Page 4 - VW GTI MKVI Forum / VW Golf R Forum / VW Golf MKVI Forum / VW GTI Forum - Golfmk6.com



What does this mean? I think the repaired AEM failsafe has been metering right all along.

Believe it or not, the repaired AEM setup has now been running consistently for about a week and hasn't missed a single trigger or had a weird flow event. What I suspected was a flow metering issue was in fact shitty marketing on the part of the nozzle manufacturer.

I did find a slight issue with the AEM failsafe in its design. The spec for sink current of the open collector is 20mA max. I measured the sink current of the collector and found it to be around 30mA with a 10kohm resistor AEM used. This is out of spec. I put a 13kohm resistor in place of the 10kohm AEM used and the sink current is in spec now (20mA). The gauge doesn't seem as jumpy now.

[Dust]

Hey, those are my nozzles. I only thought he had posted on the vortex. The DO numbers are for their new nozzles, not the older micro-droplet nozzles. The Aquamist nozzles were new to several years old.

[Lex]

I believe all manufacturers rate their nozzles at 100psi.
Also AFAIK most pumps have a settable internal pressure bypass and this sets the maximum pressure they run at.

Good work Dale!

[neganox]

How do you have all of these toys? I'm jealous.

[dale_gribble]

I'm an engineer with a well supplied lab. :-)

[ItsNox]

Droppin' knowledge bombs like panties after prom. Good shit dale. Thanks for keeping us updated

Sent via Tapafail

[Lex]

Originally Posted by dale_gribble I'm an engineer with a well supplied lab. :-)

All these engineers and their attention to details ... sheesh!

[dale_gribble]

Look who finally got his shit together! After months, I was able to finally get the whole AEM setup going and actually build a failsafe flow curve! I am not going to hook up the failsafe just yet, but it's almost done!

Here's the flow plot after about 3 4th gear pulls. Not super tight, but a curve nevertheless. I clicked the automatic curve function to build the initial curve.



aaaaand, the curve is officially built.

[Lex]

It's a big range for the error (50% of total flow). How does it end up reading a flow of 400cc and 800cc at the same Injection % (is that pump duty?). Does it start sampling the flow a little early perhaps when you first get into boost and the system hasn't built up pressure? Does increasing a delay (perhaps alarm delay) help tighten the curve at all?

[Addikted To Boost]

So this trigger could also be used with a low level sensor on my washer tank right?

[dale_gribble]

Originally Posted by Lex It's a big range for the error (50% of total flow). How does it end up reading a flow of 400cc and 800cc at the same Injection % (is that pump duty?). Does it start sampling the flow a little early perhaps when you first get into boost and the system hasn't built up pressure? Does increasing a delay (perhaps alarm delay) help tighten the curve at all?

This is AEMs curve on their website. I don't think it is super tight. yes, that is pump duty cycle. The failsafe triggers on the controller pump signal line. I suspect the variation has to do with 2 things:

1. The pulsing behavior of the pump itself. If the pump was continuous flow, I am sure the curve would be much tighter as the sensor is spec'd to hold +/-3%.

2. It would definitely benefit from some averaging as well. I don't think the AEM does any averaging/exponential averaging, so it samples and reports, samples and reports, etc in whatever duration it's controller samples the incoming pulses from the flow sensor.

[dale_gribble]

Guys I am really happy to report that the AEM failsafe has been working flawlessly now for a solid week. Everything is working well consistently and I am feeling confident in hooking the AEM up to the Guardian Angel. Finally!

For anyone going through this, don't hook up the failsafe to your GA until you have vetted the failsafe by itself. It's one thing to fight a failsafe, it's that much worse when it is constantly triggering the GA and popping the BPV. A lesson learned for me, hook up the failsafe, run it for a while, get confidence in it and then hook it to GA.