Dangerous cylinder deactivation trouble on 2018 CX-5 and Mazda6 models

[jjm86m]

For those who were asking, I was able to take a few pictures of what the rocker arm looks like when its fallen off. I stopped by the dealership to grab a few things from my car and talked with the technician for a bit. He was a lot more helpful than the service rep and showed me all the paper work he got and said the copy of the work showing that I had a P0304 previously never got to him. WTH! For those that are going ask how he didn't have the record, Hindsight is 20/20 I should've taken my car back to the same dealership that did the P0304 code work for me. The first dealership I took it too was the one I purchased the car from and I was having an oil change done that day as part of the year free service and 80 miles from my work/home. The second trip I ended up taking it to the dealership 10 miles from my place to avoid having to take off more time off from work than necessary.

Anyways, while I was there I asked to grab a few pics and was able to get the ones below.

I'm also grabbed a shot of the service bulletin with the instructions to check for the rocker arm from Mazda. The SA-015/19 posted earlier in this thread is for checking debris but SA-041/18 is the specifically for the rocker arm. Inspection step 3 is important for anyone owning a Mazda that may be affected. Make sure your oil change service records are verifiable, especially if you're doing your own oil changes. I wouldn't put it past a dealership to hassle you over something like that. They didn't for me but you never know.

 

[madar]

Looks like a lucky pocket to fall off into, not some place where it would have worked it's way into other moving parts.

 

[shadonoz]

For those who were asking, I was able to take a few pictures of what the rocker arm looks like when its fallen off.

Thanks for the great work following up with Mazda and sharing with us. Still hard to get my head around this design.

 

[madar]

Yes rocker arm is the lifter arm. I really feel this's a very strange experience for you as Mazda has officially sent a service alert to NHTSA, that means Mazda North American Operations have also sent the same alert to all of their dealers. It does instruct the dealer to open the valve cove and check the rocker arm at cylinder #4 if the engine fell into the limp mode when it revs at high rpm and got the misfire code on cylinder #4. I wonder why the Mazda dealer can't follow the service alert to do it and get paid by Mazda? The gift card thing for you to pay the cost of inspection of rocker arm is ridiculous to me.

Glad now you found the real issue and I wonder are you going to get a "new" engine, or the tech just put the rocker arm back to the position?


Although you're the first to report this issue, but nobody would know what's going on here at the time. Not until Go Hawks the OP came out and posted his research and experience on such weird issue, we realized Mazda made a big design mistake on cylinder deactivation and every SkyActiv-G 2.5L with CD would potentially have its rocker arm falling off on cylinders with CD. We should appreciate Go Hawks who started this thread, although some here was questioning he was spreading fake news.

Lastly, I don't believe a software change with higher oil pressure to Hydraulic Lash Adjuster or Hydraulic Lifter during the CD phase can truly resolve this problem. The HLA's for cylinder #1 and #4 need to be redesigned in such a way that even at the lowest oil pressure, it won't be too low and let the rocker arm fall off. Or some holder needs to be added to keep the rocker arms at the bay hence even the HLA's are at their lowest or shortest the rocker arms would stay put.

And according to the procedure it looks like engine replacement is their answer, gonna be an expensive fix!

 

[yrwei52]

For those who were asking, I was able to take a few pictures of what the rocker arm looks like when its fallen off. I stopped by the dealership to grab a few things from my car and talked with the technician for a bit. He was a lot more helpful than the service rep and showed me all the paper work he got and said the copy of the work showing that I had a P0304 previously never got to him. WTH! For those that are going ask how he didn't have the record, Hindsight is 20/20 I should've taken my car back to the same dealership that did the P0304 code work for me. The first dealership I took it too was the one I purchased the car from and I was having an oil change done that day as part of the year free service and 80 miles from my work/home. The second trip I ended up taking it to the dealership 10 miles from my place to avoid having to take off more time off from work than necessary.

Anyways, while I was there I asked to grab a few pics and was able to get the ones below.

I'm also grabbed a shot of the service bulletin with the instructions to check for the rocker arm from Mazda. The SA-015/19 posted earlier in this thread is for checking debris but SA-041/18 is the specifically for the rocker arm. Inspection step 3 is important for anyone owning a Mazda that may be affected. Make sure your oil change service records are verifiable, especially if you're doing your own oil changes. I wouldn't put it past a dealership to hassle you over something like that. They didn't for me but you never know.

Thanks to your extra effort to get more information out of this rocker arm falling off problem! Getting some pictures with another related service alert is much appreciated as a picture is worth a thousand words!

Are you going to get a new or factory-rebuilt engine? If Mazda can't tell you what fixes have been done on the new or rebuilt engine, how can they know the same problem won't happen again?

Service Alert SA-015/19 Mazda submitted to NHTSA is actually related Service Alert SA-041/18 you posted.

Subject: CHECK ENGINE LIGHT COMES ON WHEN DRIVING WITH P3400 STORED IN PCM

Service Alert No.: SA-015/19
Last Issued: 04/08/2019


APPLICABLE MODEL(S)/VINS

2018 Mazda6
2018-2019 CX-5
2019 Mazda3


DESCRIPTION

If you receive a customer concern with the check engine light coming on and DTC P3400 (cylinder deactivation system malfunction) stored, perform the inspection in the Repair Procedure.


REPAIR PROCEDURE

1. Remove the valve cover and inspect for any rocker arms that are broken or have fallen off, especially in the cylinder deactivation area (cylinders 1 and 4). Refer to SA-041/18 if found.

 

[Mulligan]

I don't understand how low oil pressure could be the cause of the rocker arm falling off. If it was, wouldn't the rocker arm be likely fall off each time the engine is started, since oil pressure is essentially zero for the first turn of the camshaft?

There's a drawing and explanation of what happens at the camshaft, valve, rocker arm, and hydraulic lash adjuster (HLA) during cylinder deactivation here: https://www.mazda.com.sg/car/mazda-cx-5/ (about halfway down the page). If something stops the spring inside the HLA from pushing the moving part of the HLA up fast enough to follow the cam action, a gap would occur between the cam and the roller on the rocker arm. That gap might be enough to allow the rocker to fall off.

I wonder if the "contamination or debris" that is referenced in Mazda Service Alert SA-015/19 could be what interfered with the movement of the spring inside the HLA. Perhaps blocked oil passages inside the HLA could lead to fluid locking the sliding inner portion of the HLA at the bottom of its stroke, creating a gap between cam, rocker end, and/or rocker roller.

The above is speculation, obviously. I'm just trying to understand how it's possible for the rocker to fall off. It's hard to believe Mazda would create a design where that's possible, even with a malfunctioning HLA, but apparently they did.

 

[yrwei52]

I don't understand how low oil pressure could be the cause of the rocker arm falling off. If it was, wouldn't the rocker arm be likely fall off each time the engine is started, since oil pressure is essentially zero for the first turn of the camshaft?

There's a drawing and explanation of what happens at the camshaft, valve, rocker arm, and hydraulic lash adjuster (HLA) during cylinder deactivation here: https://www.mazda.com.sg/car/mazda-cx-5/ (about halfway down the page). If something stops the spring inside the HLA from pushing the moving part of the HLA up fast enough to follow the cam action, a gap would occur between the cam and the roller on the rocker arm. That gap might be enough to allow the rocker to fall off.

I wonder if the "contamination or debris" that is referenced in Mazda Service Alert SA-015/19 could be what interfered with the movement of the spring inside the HLA. Perhaps blocked oil passages inside the HLA could lead to fluid locking the sliding inner portion of the HLA at the bottom of its stroke, creating a gap between cam, rocker end, and/or rocker roller.

The above is speculation, obviously. I'm just trying to understand how it's possible for the rocker to fall off. It's hard to believe Mazda would create a design where that's possible, even with a malfunctioning HLA, but apparently they did.

I have been saying this problem is a bad design on Hydraulic Lash Adjuster specifically for cylinder deactivation. The HLA can get too low or too short during CD phase, causing a gap between rocker arm and HLA which is big enough losing the support on rocker arm on HLA end and it falls off. The HLA for CD simply can't go that low physically creating a gap even if the oil pressure is too low, or a gate keeping the rocker arm at the bay should have added for CD cylinders.

The other thing I found is the hydraulic lifter is used to eliminate valve clearance adjustment. if this HLA designed only for cylinder deactivation using a lockpin to secure the non-CD phase movement, does this mean it also losing the purpose of hydraulic lifter eliminating the valve clearance adjustment?

 

[madar]

Sounds like they're trying to compensate by boosting oil pressure during this particular phase to keep the lifter collapse from happening via software.

 

[yugrus]

Same deal as with GM, as I suspected. But GM design ends up collapsed lifter as a huge racket in any mode, but nothing falls off. Then you now what's going on.
Mazda is more sophisticated. The impacted valve adjuster is not involved in day-to-day operations. It gets itself known only in high-speed, high-load situations, as additional air is needed. Jinba Ittai, sure.

 

[yrwei52]

Sounds like they're trying to compensate by boosting oil pressure during this particular phase to keep the lifter collapse from happening via software.

Like Mulligan said in the previous post, if low oil pressure would cause the rocker arm fall off, rocker arm would be likely fall off each time the engine is started, since oil pressure is essentially zero for the first turn of the camshaft. Preventing rocker arm falling off by increasing the oil pressure won't work as oil pressure varies a lot depending on engine rpm, oil temperature, and oil viscosity. If you have been driving older cars with oil pressure gauge, you'll know oil pressure is constantly changing. I really don't believe simply boosting oil pressure during cylinder deactivation phase would work.

 

[jjm86m]

I don't see any place with this design that Mazda could've pinned anything. The shifting fulcrum doesn't allow you to pin anything. It relies on 2 point contact at all times to hold everything in place.

The HLA can get too low or too short during CD phase, causing a gap between rocker arm and HLA which is big enough losing the support on rocker arm on HLA end and it falls off....... The HLA for CD simply can't go that low physically creating a gap even if the oil pressure is too low, or a gate keeping the rocker arm at the bay should have added for CD cylinders.

It looks like the oil pressure you're mentioning here is definitely part of the problem which is why Mazda addressed it with the PCM change/update. Like you mentioned perhaps the HLA dropped too low due to lack of oil pressure and loses contact with the cam. Poor design yes. I don't see why the PCM update wouldn't fix that. Increase pressure so the HLA doesn't drop as far.

The other theory could be a mistiming during the transition into CD. Say the valve is driven open (spring compressed) via the cam and all of a sudden the HLA is driven to the "unlocked" position by software before the valve springs back shut (spring nominal). That would create a gap large enough for the rocker to fall out. In this scenario the rocker arm would lose contact with the cam and both the HLA side and Valve side would be moving at the same time. The valve springing back up and the HLA dropping down with no cam contact in the middle.

Theoretically this shouldn't happen because the transition to CD happens while Cylinder #1 or #4 are not being fired. The reality is the software can't anticipate when someone is going to jam on the gas. Perhaps this happens right as the car is going to go into CD but the driver just mashes the gas. Just a theory though.

Following this theory out. This scenario only happens on transitions into CD because during the transition out of CD, the valve spring keeps the valve from opening up (ie dropping down). This would explain why on startup you don't see this happening. Startup is essentially a transition out of CD (HLA lifts up, springs keep valve from dropping, cam stays engaged)

Check SA-015/19 for another theory. If I'm reading and interpreting pictures correctly, a fouled oil pressure controller could lead to the scenario I mentioned above rather than mistiming.

 

[Mulligan]

It relies on 2 point contact at all times to hold everything in place.

Actually, it [the rocker] relies on 3 points of contact at all times: valve, cam, and HLA.

... perhaps the HLA dropped too low due to lack of oil pressure and loses contact with the cam.

I don't see how oil pressure figures into keeping contact between the rocker end (at the HLA) and rocker roller (at the cam). Look at the diagram at the link in my previous post. There's a substantial spring inside the HLA that pushes upward on the sliding inner part of the HLA (when the lock is disengaged), forcing it into the cam. It's that spring, not oil pressure, that maintains contact between the HLA and the rocker.

Say the valve is driven open (spring compressed) via the cam and all of a sudden the HLA is driven to the "unlocked" position by software before the valve springs back shut (spring nominal). That would create a gap large enough for the rocker to fall out.

I don't think it would. When the HLA is tall, that is, when the lock pin is locked, the spring inside the HLA is extended and in contact with the bottom of the movable inner part of the HLA. If the lock is disengaged at the wrong time, there's no gap, rather the reaction force that was previously being countered by the pin is now being resisted by the spring inside the HLA. This spring isn't as strong as the spring at the valve, so it will begin to compress as soon as the lock pin is disengaged, but there won't be any initial gap.

If we want to really get out on a speculative limb, I guess we could consider the possibility that a mistimed unlocking of the pin could drive the inner, movable part of the HLA downward so fast that its momentum overcompresses the spring inside the HLA, and that overcompression is what opens up a gap between the top of the HLA and the rocker.

However, I still think the gap that allows the rocker to fall out is likely caused when the compressed spring inside the unlocked HLA is, for some reason*, unable to extend quickly enough to follow the profile of the cam after the lobe passes.

*Above, I speculated that reason might be debris or a hydrolock-like force holding the movable inner portion of the HLA down, countering the force of the spring. A more parsimonious explanation would be simply a broken spring inside the HLA.

 

[jjm86m]

Good catch. 3 points.

Yeah I just looked at that diagram. The internal spring on the HLA is a part of the design but I believe there is a hydraulic aspect that that is pressure based. I'm not familiar enough with the HLA's operation to verify though.

 

[madar]

Good catch. 3 points.

Yeah I just looked at that diagram. The internal spring on the HLA is a part of the design but I believe there is a hydraulic aspect that that is pressure based. I'm not familiar enough with the HLA's operation to verify though.

Nobody here is familiar with it. It could boil down to something like a manufacturing tolerance issue. Mazda has had these in other areas.

 

[joemazdacx5]

well, I hope they put out a recall......
pisses me off , was going to buy a 2019 colorado , but they have that trans problem,
then decided to buy a honda crv , but they have that oil in gas problem, then finally
buy mazda cx 5 reading alot of professional reviews never talking about CD problem

have about 200miles on engine......... I'm sure within a month or two ,I'll have to give
it gas and see if rocker arm falls off or get error P0304

 

[Sig cx-5]

well, I hope they put out a recall......
pisses me off , was going to buy a 2019 colorado , but they have that trans problem,
then decided to buy a honda crv , but they have that oil in gas problem, then finally
buy mazda cx 5 reading alot of professional reviews never talking about CD problem

have about 200miles on engine......... I'm sure within a month or two ,I'll have to give
it gas and see if rocker arm falls off or get error P0304

There have been what, two reports of this issue on the forum? Plus, the guy that started this thread signed up just to post about the issue. I'm not diminishing the OP, or this thread, just pointing out that the issue seems pretty rare at this time.

 

[David Paul]

Mine has just over 10,000 miles on it.

Several times over the past couple of weeks I have tried to recreate this problem to no avail. No matter how much I try to abuse my CX-5 engine, no Limp Mode has occurred. All of this has been done on an Interstate in very light traffic situations starting at various speeds from 40 mph all the way to about 85 mph.

So far, so good.

 

[Avoidin Deer]

well, I hope they put out a recall......
pisses me off , was going to buy a 2019 colorado , but they have that trans problem,
then decided to buy a honda crv , but they have that oil in gas problem, then finally
buy mazda cx 5 reading alot of professional reviews never talking about CD problem

have about 200miles on engine......... I'm sure within a month or two ,I'll have to give
it gas and see if rocker arm falls off or get error P0304

I went through the same thing.

Researching cars is like reading the negative reviews on Amazon. You sometimes walk away feeling like your best option is to do nothing. Eventually you just hold your nose and jump in, hoping for the best (helluva way to spend ten$ of thousand$).

I dodged the Honda oil dilution issue because my brother (a mechanic) warned me against new engines, so I did research I would not have otherwise done (I was relying on the CR-V popularity.)

I dodged the Mazda CD issue only because life issues got in the way when car shopping in the fall of 2018. When I picked up again in February, the turbo had been released. But now there are folks talking about the turbo occasionally not working in certain weather conditions. Like you, I read a lot on these sites and independent reviews, and did not see any major negative issues (except the SCRM paint, which is easily avoided by making a different color choice).

This is not like buying a $50 drill that does not perform properly. You get much better customer service when spending way less money. And even if you have to eat it, it's only $50.

A car is not so easy to get out from under...and it ain't cheap. It's this lack of sufficient relief that spikes the angst.

 

[Go Hawks]

There have been what, two reports of this issue on the forum? Plus, the guy that started this thread signed up just to post about the issue. I'm not diminishing the OP, or this thread, just pointing out that the issue seems pretty rare at this time.

As I stated in a much earlier post, the number of people on this thread is a miniscule sampling of the total number of 2018 and possibly 2019 CX-5 on the road today. It is still to be determined how many actually have this issue due to the fact that it is very hard to determine if you have the problem. There may be many vehicles driving around with this but they haven*t discovered it yet. Or there may be some in the junkyard that ended up there due to this issue. And the reason I started this thread is because I went into my dealership to get an oil change a few weeks ago. When I walked into the service area the first thing the service manager said to me was they had heard nothing from Mazda concerning my issues that I had about this problem. It had been 6 months since my car was in the shop and not a thing to address this SAFETY ISSUE! I wanted to post something on a forum or facebook back when this first happened but I was certain that Mazda would do the right thing. So when I found out nothing had been done by the MOCO to get the word out, I searched and found this forum. This is how we have gotten here and so far we all know of one person who had the same issue and now is having it addressed. This info may have saved him from an accident. So I feel vindicated and relieved that this forum has helped someone. I hope this thread continues to grow because unfortunately, this could very well be the tip of the iceberg concerning this issue. I for one hope it isn*t, and that you are correct in your assumption of it being a VERY small number of affected vehicles. But all indications that I have scene show the number of these troubled vehicles is continuing to grow. And that scares me!

 

[Mulligan]

There have been what, two reports of this issue on the forum?

Some people might have missed where the OP mentioned in his initial post that Mazda had confirmed 26 similar occurrences.