Cracked turbo manifold and studs - what could cause this?
 

Hello,
I have had my 1991 Miata turbocharged for about 4 years now. This week I noticed that the turbo manifold has started to crack. It had sounded like a mild exhaust leak for perhaps a couple of weeks now, but nothing out of the ordinary that made me go looking. In the past I have struggled with nuts on the turbo constantly coming loose, and 2 studs breaking (at different times). The studs were drilled out and tapped larger, then replaced with a helicoil and another M8 stud. Recently I added new nord-lock washers, and the nuts have not since loosened. The exhaust was custom made by a shop nearby and went on the car at the same point that the manifold did. The manifold was bought from ebay, though it is not a cheap china model, but something a local fabricator made.

I'm posting here (and registered) because I worry that there is some underlying cause to the manifold breaking other than it just being cheaply made. I don't consider 4 years to be a particularly horrible lifetime as compared to other manifolds, e.g. china ones that might last a year. I was however surprised that it would break so soon as it seemed quality. The car is driven almost daily, and has not seen the track since being turbocharged. My downpipe on my exhaust does not have a flexi pipe, and the turbo/manifold is not braced in any way. The exhaust hangs by rubber hangers in a number of places, but none of these are on the downpipe, the closest one being about the middle of the car from the manifold. Motor mounts are poly.

I guess what I am questioning is if the loosening nuts and broken studs point to an issue - e.g. the exhaust moving or not supported properly. Could this also have caused the cracking of the turbo manifold, perhaps in combination with the less play in the nuts now from the nord-lock washers? Currently the manifold is still in one piece, though I expect it will crack through soon. What would be the best course of action? Buy another of the same manifold and hope for another 4 years with the same exhaust setup (coincidence that it broke now and another would last similarly)? Or make some modification to better support the turbo?

https://cimg8.ibsrv.net/gimg/www.mia...e51be334c8.jpg
https://cimg9.ibsrv.net/gimg/www.mia...c02d385e2c.jpg
https://cimg1.ibsrv.net/gimg/www.mia...7f470af32d.jpg
https://cimg7.ibsrv.net/gimg/www.mia...5f0c9f8fac.jpg
https://cimg0.ibsrv.net/gimg/www.mia...4f2a5b7713.jpg

Welcome to turbocharged Miatas. Our manifolds and hardware must be made of heavier materials and engineered stronger than most other brands. Our engines are very rough and agricultural from a harmonics and vibration standpoint compared to silky Hondas and Nissans.

If you don't have a flex pipe you are compounding the issue. The engine trying to move and the exhaust fighting against it puts pressure on the components. Plus you're trying to use the exhaust manifold and turbo mount bolts to hold up the weight of the front of the downpipe. This occurs even with rubber mounts.

Higher exhaust temperatures at higher boost levels can exacerbate the issue. So can overly retarded timing in boost, which can be necessary to avoid detonation.

I mean those pictures make it look exactly like the cheap eBay manifolds that are known to crack.

Perhaps a similar design, yes. I've never had one of the cheap China ones so wouldn't know if this is any better. It feels heavy and well made but I have nothing to compare it against.

My worry with changing to a different manifold style is the downtime with having to modify the exhaust (due to different turbo position).

https://cimg0.ibsrv.net/gimg/www.mia...f25eb4aea0.jpg

The welds have some undercutting at the toe, and not a lot of filler material was used. One if the runners is also cracked beyond where I think is the HAZ, probably too thin on the pipe material. Inconel studs (Trackspeed) will fix the breakage problems- dont over torque them, follow the directions. Another manifold will most likely still fit your exhaust, but I would defiantly add a flex pipe section before the cat.

How long do your EGO sensors last? It is commonly found that mounting the sensor so close to the turbine outlet shortens its life; unless a heat sink is used.

Will definitely look into inconel studs for the next manifold thanks! I guess the issue with another manifold fitting my exhaust is that the rotation of the T28 exhaust flange has now been dictated by the angle of the T28 intake flange e.g. at ~45 degrees from the horizontal vs something like the flyin miata manifold being dead vertical.

Would the ebay cast manifolds perhaps be an idea at improving longevity? They seem to have the flange at a similar angle, so could work for the downpipe. Only issue being that the mount location is a bit further forward (flange not centred between the two middle runners), but if I'm already getting the exhaust cut to put a flex pipe in then that probably doesn't matter.

This is the style I mean:
https://cimg1.ibsrv.net/gimg/www.mia...0d31e81c6f.jpg

Cheap tubular manifold. Known quantity that they do not last on Miatas.

If you're wondering what exactly causes manifolds to crack, it's a combination of thermal cycling and vibration.

Welds are very hard, and harder = more brittle, which is why cracks tend to form at the welds. The welding process also raises the temperature of the base metal quite high, which can anneal it (remove the heat treatment), which weakens the metal. The vibration of the engine basically flexes the metal over and over again and eventually fatigues it and causes it to crack. It's similar to how you can break a wire hangar by bending it back and forth over and over. Also, a poorly designed component may have a resonant frequency within the operating range of the engine, which will cause it to amplify the vibration from the engine, and crack sooner.

In addition to vibration, the alternate, repeated heating and cooling of the component also weakens the material, which is why exhaust components are much more prone to cracking than, say the intake manifold. It might be plenty strong enough when the part is new, but over time, the thermal cycling degrades the strength, and eventually it will crack.

These are the more or less invisible differences between a high quality manifold and a cheap one; harder welds, large heat affected zones, poor resonant frequency characteristics, lower quality material. Because of this, you can have a component that appears to be solid and well made, but will still fail eventually.

Those cast manifolds work ok on the street. They are known to also crack on the track.

All this is very true. I very much believe that all welded together manifolds for a miata will eventually crack after hard usage. A higher quality Artech/Absurdflow will last a hell of a lot longer then a MKTurbo manifold, but given enough time and use will also crack. On the street the amount of time to crack a manifolds is probably in years and tens of thousand of miles. Once you start tracking the car that time until it cracks probably goes down significantly.

Im not entirely convinced.. my latest one from ArTech did not impress me at all, minimal penetration IMO. I left it as is, for experimental purposes, and then put a $200 coating over it...

^ Agreed. *edit* That "agreed" was originally for Lar's post...I took too long to write.

OP, honestly, getting 4 years of daily use out of a cheap manifold sounds pretty good to me. Getting a cheap cast manifold vs a cheap fabricated one won't necessarily make it last much longer. Castings can be prone to cracking as well, albeit for different reasons.

For example, in the picture you posted, see the relatively sharp transition between the flange and the runner? That corner will cause a stress concentration around each runner, and will probably crack there eventually. Also, sharp corners are especially bad casting design, because the molten metal has a hard time flowing around the corner, which can cause porosity and seriously weaken the component. A higher quality casting will have smooth transitions in thickness and larger fillet radii when possible. Also, the flanges for each cylinder are all cast together, which will cause stress in the runners due to thermal expansion. That's why both FM and Kraken cut their flanges after casting.

Most importantly, the material of a casting will have a huge effect. If the right alloy isn't used, the material strength will degrade with heat, and the manifold will fail relatively quickly.

Anyway, if you want something that will last forever, you'll need to shell out for a high quality casting. I agree with Lars that even a high quality fabrication will fail probably eventually; welds just don't deal with thermal cycling very well.

use a flex pipe around the area that is the bellhousing and the subframe. they last longer there (no road debris or bottoming out) and it will help.
The manifold you pictured doesn't look like a pipe manifold and there was no support. Thin walled manifolds, especially in which have no exhaust flex and no downpipe/turbine housing brace will crack. Thick walled manifolds aren't exempt, but they have more probability of lasting longer.


Last thing no one talks about with manifold durability is your tune. most metals don't like the high heat that pressurized turbo manifolds put into them and when you get a bad tune that's pumping lots of egt into the manifold (or you have anti-lag enabled achieving the same thing) you drastically reduce the life cycle of your exhaust system.

That said, Im building a replacement for another 1.6 with the same manifold design right now - the one you have looks no different than a standard ebay manifold. FWIW a schedule 10 pipe manifold like you would see in my photo thread is roughly 17lbs.

Thanks to all those that have commented, it's given me a lot to think about.

Engi-ninja - Thanks for the explanation of what causes the stress in a manifold. It makes more sense now why the manifold would crack after all these years. In saying that, 4 years is a good lifetime as you said. Probably not worth looking at cheap similar style cast ones if they aren't significantly better in terms of lifetime I guess. Your comments make me think that the best solutions are either the current style or an FM/Begi manifold.

DNMakinson - So far the ego sensor has lasted the full 4 years the car has been turbocharged with this setup. I did read about the issues placing it there but did not realise the exhaust shop would do so till I picked the car up!

Shlammed - Will take on your advise and get a flex pipe added regardless of where I go with the manifold next. Would you have any suggestion specifically for bracing in addition with a similar style manifold? The seller has told me that the wall thickness of the tubing used was 3mm and that the manifold weighed "over 3kg". Not sure if this is good or bad, seems to be schedule 10 from googling? Of course if the welds/design are bad I guess this doesn't save it. The car doesn't have antilag, but I will look into how the tune is.

I'm thinking now that perhaps the best course of action is to get another of the same manifold and hope for another few years of life, and get the car up and running quickly for now. I will then drive it to an exhaust shop and get a flex pipe added in the downpipe as suggested, hopefully this helps things a little. Is there any further bracing that could be of use either on the turbo/manifold itself to the engine or the exhaust to the gearbox etc that might raise chances of the manifold lasting as long as before? After some research I've been thinking of something along the lines of the FM exhaust brace (to the bellhousing) before the flexi joint.

Later when I am better prepared (time-wise and financially), I will go with something like the FM cast manifold and get the exhaust remade for this change. After the next couple of years my priorities with the car may likely have changed, e.g. bigger power goals (1.8 engine?), and getting it on the track. For the immediate future I won't be changing the engine setup or tracking the car so perhaps this manifold with bracing/flex pipe will be adequate and the best choice at the present in terms of getting the car running again and financially?

3kg is pretty light. the weight at 17lbs for a schedule 10 pipe manifold is closer to double that (7.7kg). Granted the design I have does use a bit more tube length for a more standard long collector, I still think something is up with that manifold being 3kg.

As mentioned before, the quick angle right at the turbo flange does lead to localized hot-spots. A longer collector like the one pictured below that I just welded will put more of the heat into the turbo instead of into the material and improve spool up and improve service life.
Attachment 229837

As far as bracing, there are a few ways to go about it. Most OEM's have a threaded port on the turbine housing that bolt to the engine itself (think of the intake manifold brace on the Miata engines, very similar but engine to turbine housing) I know that people don't like welding to turbos (I personally don't mind doing it) so the next most common is welding a large threaded bung to the downpipe or using a pipe clamp that connects a bell housing bolt to the downpipe. I believe that FM does/used to offer a bellhousing to pipe clamp option for their kits, a good place to get inspiration from.
Some people have discussed turbo cranes on this website, but after having been more involved in turbo system design and inspected more and more OE/performance systems I believe that hard bracing to the turbine housing is the best method.

That is a fair difference in weight with the manifolds then, and perhaps not the best design as you've pointed out. Sounds like it can't be both SCH10 and 3kg.

I was just looking at the old FM brace that you mentioned now in fact. I think I could come up with something similar to brace my exhaust to the bellhousing. Presumably I would want this before a flex pipe to keep the downpipe section and turbo all solid mounted, and then the exhaust isolated from the hard mounted section (on the other side of the flex pipe)?

Thanks for the other suggestions, I'll look into them too now, and see how I might apply them also to my setup.