Help finding a boost leak

[Ozynigma]

Hi All, I have a problem with my CX7 but I am putting it here because it is a general issue and I hope more people see it here and can help.

I have just installed a K04-882, JBR 3.5" WPI, ETS intercooler and a Stratified Base Map to suit.

My mod thread is here Seen the light and modding CX7 for more power

The car is running rich on WOT runs and Lex suspects a boost leak. I need to fix this before we continue with tuning.

I have attached the initial third gear log in Virtual Dyno format which shows AFR decreasing from around 11 to around 9.5.

After finding this I read up all the threads on finding boost leaks and made myself a boost leak detector. Here it is installed in the end of the WPI. I used a regulated air supply of around 30psi from my compressor and a tyre inflation gauge.

Untitled by ozynigma, on Flickr

The first thing I found was my Cobb XLE was dumping boost at about 10 to 12psi. I found this with the BPV return hose disconnected from the BPV and plugged. I put the OEM BPV back on and was able to get close to 20psi before leaking boost. I figure the Cobb needs a stronger spring. i don't know why the previous owner had it set so weak.

I also found the intercooler boost tubes needed tighter clamping. At this point I was pressuring the whole intake to around 20psi and the only leakage I could detect was around the PCV hose from the valve cover to the WPI. As this is not normally pressurised I didn't try and fix it.

I then did another uphill log run in 1st, 2nd and 3rd gear. The second gear portion of the log is attached. The AFR performance was better dropping from around 11.5 to 10.

I have also attached the VD boost and AFR graphs.

The car is a lot more drivable with the OEM BPV back on.

What else can I do to find and fix the rich running under WOT?

Have I done anything wrong in the boost leak testing? Even with the PCV plugged on the WPI I was still getting pressure leaking from the valve cover. I assume this is getting through the throttle body, manifold, valves and down past the rings?

[Ozynigma]

Edit. BTW the boost tubes are SURE. Sorry but they came for free with the ETS intercooler and I like the colour blue. If anyone seriously thinks these are the issue I will put the stock boost tubes back on.

[torquemaniac]

Did you check the PCV hose on the manifold for leak (its in between the manifold and block)?

Did you check ''commanded AFR'' versus ''actual AFR'' when logging...the CX-7 tuning is wicked...I found some very weird thing while going over all tables

[Ozynigma]

Originally Posted by torquemaniac Did you check the PCV hose on the manifold for leak (its in between the manifold and block)?

No I didn't. Should there be a check valve to stop manifold boost pressure bleeding back into the crankcase? I was wondering why I was getting positive pressure from the rocker cover breather, it seemed too much to be blowing by the rings.

Originally Posted by torquemaniac Did you check ''commanded AFR'' versus ''actual AFR'' when logging...the CX-7 tuning is wicked...I found some very weird thing while going over all tables

No I have not logged "commanded AFR" yet. I was assuming that the map is commanding appropriate values hence Lex concern at the running richer than expected. It is not an OTS map, it is specific to my mods. You make an interesting point though that the CX7 ECU might be handling the commanded values differently to a 3 or 6. After all Versatuner is the only company to have cracked our ECU's.

[torquemaniac]

Originally Posted by Ozynigma No I didn't. Should there be a check valve to stop manifold boost pressure bleeding back into the crankcase? I was wondering why I was getting positive pressure from the rocker cover breather, it seemed too much to be blowing by the rings. No I have not logged "commanded AFR" yet. I was assuming that the map is commanding appropriate values hence Lex concern at the running richer than expected. It is not an OTS map, it is specific to my mods. You make an interesting point though that the CX7 ECU might be handling the commanded values differently to a 3 or 6. After all Versatuner is the only company to have cracked our ECU's.

Im not saying the PCV is leaking in the crankcase, but rather that the hose may be leaking to the atmosphere (but the PCV may actually be leaking in the crankcase also at 30psi)

You absolutely need to log ''commanded AFR'' and ''actual AFR'' before suspecting a leak...if ''actual AFR'' is the same as ''commanded AFR'', you have no leak

FYI, Mazdaedit also has a software for the CX-7 epifanSoftware | mazdaEdit

I know a tuner that tune any Mazda with this software (Pro license...)

[rfinkle2]

Tell Lex that you did the leak test and that you feel everything is buttoned up. He may just need to touch up the fuel curve now.

[ALPINEST4RS]

Are your afr's high ten's @ wot? If so I'd have to think that's normal depending on gas type.

I've working with rob @ freek and my afr's were dipping into high 10's @ wot.

Switching gas stations helped. I'm right on target now.

What were your afr's like before you installed everything?

[Ozynigma]

Originally Posted by torquemaniac You absolutely need to log ''commanded AFR'' and ''actual AFR'' before suspecting a leak...if ''actual AFR'' is the same as ''commanded AFR'', you have no leak

Thanks, I will check everything again and do another test log to check that.

Originally Posted by rfinkle2 Tell Lex that you did the leak test and that you feel everything is buttoned up. He may just need to touch up the fuel curve now.

Thanks, I will do a bit more testing and send him the logs. Just trying to do my homework first to make sure I haven't missed anything obvious. The Cobb BPV performing worse than stock was a surprise.

Originally Posted by ALPINEST4RS Are your afr's high ten's @ wot? If so I'd have to think that's normal depending on gas type. I've working with rob @ freek and my afr's were dipping into high 10's @ wot. Switching gas stations helped. I'm right on target now. What were your afr's like before you installed everything?

The AFR's (and boost) were worse. See the attachment. Ignore the @crank power and torque estimates from using a correction factor of 1.6.

Edit: For clarity the worse was with the stock TMIC reinstalled with no testing of connections and with the Cobb BPV installed and the OEM tune. So it wasn't before all mods were done.

[Ozynigma]

I did another second gear log up hill today. I have attached the log file. I can provide another log file with more parameters if required. It starts a bit high in the rev range, my son was a bit slow to hit record.

The actual AFR is running about 0.5 to almost 1 below the "desired"/"commanded" AFR.

There are no obvious leaks between the turbo and the manifold.

The thing that is bothering me is that my crankcase seems to be pressurised with pressure escaping through the cam cover PCV.

Does this indicate my crank case PCV check valve has failed and is letting manifold boost pressure into the crankcase?

I was hoping not as the manual says I have to remove the intake manifold to remove and check the PCV check valve.

If it is the check valve should I replace with OEM or rig up something more effective on that line? Someone over here in Oz was making up stronger check valves to use with bigger turbos and oil catch cans.

[torquemaniac]

Originally Posted by Ozynigma The thing that is bothering me is that my crankcase seems to be pressurised with pressure escaping through the cam cover PCV. Does this indicate my crank case PCV check valve has failed and is letting manifold boost pressure into the crankcase? I was hoping not as the manual says I have to remove the intake manifold to remove and check the PCV check valve. If it is the check valve should I replace with OEM or rig up something more effective on that line? Someone over here in Oz was making up stronger check valves to use with bigger turbos and oil catch cans.

Check this out Boost leak testing

A catch can setup is always a good thing, especialy with more boost (more blowby) and if the EGR is still connected (blowby + EGR = gunk buildup)

[Ozynigma]

Originally Posted by torquemaniac Check this out Boost leak testing A catch can setup is always a good thing, especialy with more boost (more blowby) and if the EGR is still connected (blowby + EGR = gunk buildup)

Thanks I read that thread and a few others a few days ago.

I have just ordered a new PCV check valve for like $16 from Mazda.

I retested this morning and I have reduced leaking from the top PCV hose. I still can't get more than about 15 psi to hold as measures on my dodgy tyre inflation gauge. Why do they go to 150psi and have poor graduation and precision to 30psi?

I think I may have enough in the kitty to buy an OCC. There is a used CS one going cheap here on eBay but only the can and bracket no valves or hoses or clamps. I will have another look at the JBR and Matt Damond ones. I think one of them does CX7 specific hoses.

[jdmage_mx5]

The key is getting one with a check valve.

[torquemaniac]

I told you...to make it work with a tire inflation pump, you need ''da pink pyjamas''

AS for the catch can, the best ones have media to help catch oil (non-media CC have about a 30% efficiency)

[Ozynigma]

Originally Posted by jdmage_mx5 The key is getting one with a check valve.

Thanks for your input. The Matt Damond one has a check valve standard. The JBR has it as an added cost option. The MD is under $200.

Originally Posted by torquemaniac I told you...to make it work with a tire inflation pump, you need ''da pink pyjamas'' AS for the catch can, the best ones have media to help catch oil (non-media CC have about a 30% efficiency)

The cheap one on eBay is the older model Corksport. But at $120AU for the can only I don't think it is cheap enough.


Both MD and JBR do hose kits the right length for the CX7 with their catch cans. I almost ordered the MD OCC yesterday but ordered an Aussie made check valve from an OzMPS forum member instead. This should help me solve the immediate check valve problem this week as I will be waiting about 3 weeks for shipping after I order the catch can.


I think I have figured out my BPV problem. I don't think the Cobb has a problem. I used a scale to test it and it is taking about 8psi pressure differential to move the piston and 18psi to open it fully.

So the fact that I was generating the 8psi pressure differential at around 12psi boost pressure indicates I might have a leak from the vacuum/boost hose connection from the manifold to the BPV.

So the plan is to get the improved check valve tomorrow hopefully and pull the intake manifold and replace the check valve and check the BPV hose connection.