When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
C6 Corvette ZR1 & Z06General info about GM’s Corvette Supercar, LS9 Corvette Technical Info, Performance Upgrades, Suspension Setup for Street or Track
Like I said, unless we know what happened we don't know if it is a QC issue or design flaw. Could be either. If it drawing says concentricity between 8-12 and Joe programs it, has a bunch between 8-10 and the ones at 8 are failing, that is design. If he gets a bunch between 6-10 and the ones in the 6-7 range are failing, that is a QC issue. If someone has the actual released drawing with all the specs for the head, and takes a brand new one out of the box and does a full inspection and it has dimensions out of the drawings specs, that is QC, but it could be on the low or high end of a range that is in spec. We will never know unless someone from GM releases the drawings.
And it could get more complicated with process capability.
If the spec is 8-12, and the machine is only able to hold a +/- 5, then that is a design issue too because your process isn't capable of hitting the spec. It would be an engineers wet dream to call out perfect specs with no tolerance but we can't do that. You can't say head flatness across the deck of .0000000001 inch and then complain when manufacturing can't do that and say it is a QC issue because they can only do +/-.0001
We need to look at what we have available to make judgements on whether we are dealing with a design flaw or a Quality issue.
We do know with some degree of certainty that many vehicles had performed a test of guide tolerances and found to be in spec. Some with high mileage , some with moderate mileage, and some with low mileage.
The fact that we have vehicles designed with specified tolerances and found to be in spec after use is a valid indication that the design is appropriate and robust.
The question one needs to ask is , why are some ok and some not ok.
What random factor(s) accounts for this? The processes that occur post design is the likely source, I suggest.
The vendors that are the most familiar with the heads in tear down and re-work , I do not believe have indicated the problem to be a design flaw?
If the tolerance on a print says +- .000000001 and the machine and or the machinist
cant hold that tolerance that is the machine shops fault not the design. Take it to another machine with better equipment and machinists and they can hold .00000001
and the part is good has nothing to do with design flaws its a machine/machinist flaw
then it also got past quality control.
No it isn't. Having a few in spec doesn't mean design is robust. Like I said, you have tolerances. Just making numbers up, but if you have a range from 8-12, and you find several at 10-12 that are fine and work, but some come in at 8-9 which is still in spec, according to the design, taking a sample and finding a few in the 10-12 range doesn't mean the design is valid and robust. Not even close.
Like I said, you don't know it is a design flaw or not without seeing the design. If heads are outside of the allowed specs, then it is a QC issue. If heads are within the allowed specs and failing, it is a design issue.
No part has zero tolerance. Everything has a range. You can say head has to weigh 35lbs, but without a tolerance that is impossible. So you put 33-37 or whatever the design requires, and test at the limits to ensure the design is valid.
Having a few in spec means those are probably on the highend or low end of the process, whichever leads to less wear. Plus it probably isn't one spec, but a stack up that leads to failure.
Think of if you have a basket that can support 10lbs of load. So you have 5 ***** that go in to the basket. The weight is spec'd to be 1.9lbs a ball with a +/-.2lbs. So most of the time if you have a normal process and the spec of the ***** is centered around the center of the spec, you will have 5 ***** total weighing less than 10lbs. But if you get 5 on the high side, you can exceed the weight limit. Sometimes you have ***** weighing over 2lbs, some times less. Now if the vendor that makes the ***** finds it is cheaper to machine off less material and the mean weight of the ***** is now 2.05lbs, meets spec, all good, they just measure each ball, QC passes them all, no issues. But you start getting 4-5 ***** over the 2lb weight in the assembly, and you now have failures. Not all of them pass, not all fail. That is a design issue because you stack up allows the failure.
All passing, all failing, some passing, some failing doesn't give you any idea if it is a design flaw or not. You need a root cause analysis and see if the failure is within the specified design. Once again, since we don't know what the specified design is (besides certain published specs like <.037 for valve to guide clearance) we don't know if this is a design issue. I'm sure the actual spec on the drawing isn't <.037" for guide to spec clearance. It could be .005-.010", and you get parts in that range. The parts with tighter clearances may work fine and last a long time while the ones sent out on the loose end fail.
If the tolerance on a print says +- .000000001 and the machine and or the machinist
cant hold that tolerance that is the machine shops fault not the design. Take it to another machine with better equipment and machinists and they can hold .00000001
and the part is good has nothing to do with design flaws its a machine/machinist flaw
then it also got past quality control.
You can't design something that isn't manufacturable then blame the machine shop. DFM (design for manufacturing) means you look at what the vendors can do. If they can only hold .0001" and the best shop in the world can only hold .00001" then you either make the design more robust with the looser spec so the one shop can do it, or you pay extra for the high end shop to do it. Either way, you own the design and have to work with your vendors on if they can hit the spec.
You can't say you want a 2000hp emission legal 150mpg 1000lb car then complain when GM can't make that.
No it isn't. Having a few in spec doesn't mean design is robust. Like I said, you have tolerances. Just making numbers up, but if you have a range from 8-12, and you find several at 10-12 that are fine and work, but some come in at 8-9 which is still in spec, according to the design, taking a sample and finding a few in the 10-12 range doesn't mean the design is valid and robust. Not even close.
Like I said, you don't know it is a design flaw or not without seeing the design. If heads are outside of the allowed specs, then it is a QC issue. If heads are within the allowed specs and failing, it is a design issue.
No part has zero tolerance. Everything has a range. You can say head has to weigh 35lbs, but without a tolerance that is impossible. So you put 33-37 or whatever the design requires, and test at the limits to ensure the design is valid.
Having a few in spec means those are probably on the highend or low end of the process, whichever leads to less wear. Plus it probably isn't one spec, but a stack up that leads to failure.
Think of if you have a basket that can support 10lbs of load. So you have 5 ***** that go in to the basket. The weight is spec'd to be 1.9lbs a ball with a +/-.2lbs. So most of the time if you have a normal process and the spec of the ***** is centered around the center of the spec, you will have 5 ***** total weighing less than 10lbs. But if you get 5 on the high side, you can exceed the weight limit. Sometimes you have ***** weighing over 2lbs, some times less. Now if the vendor that makes the ***** finds it is cheaper to machine off less material and the mean weight of the ***** is now 2.05lbs, meets spec, all good, they just measure each ball, QC passes them all, no issues. But you start getting 4-5 ***** over the 2lb weight in the assembly, and you now have failures. Not all of them pass, not all fail. That is a design issue because you stack up allows the failure.
All passing, all failing, some passing, some failing doesn't give you any idea if it is a design flaw or not. You need a root cause analysis and see if the failure is within the specified design. Once again, since we don't know what the specified design is (besides certain published specs like <.037 for valve to guide clearance) we don't know if this is a design issue. I'm sure the actual spec on the drawing isn't <.037" for guide to spec clearance. It could be .005-.010", and you get parts in that range. The parts with tighter clearances may work fine and last a long time while the ones sent out on the loose end fail.
You are trying to have it both way's. You seem to have some ability to postulate bad head design for vehicles that are operating well and in spec. What more would one need to say the design is fine, even in the absence of drawings and spec's and clearances.
Nice to have more data but what we do have are vehicles , that when they meet the GM in service spec , operate just fine..
I would think that when GM sent out prints to get quoted for manufacturing and everyone came back and said we cant hold those tolerances that GM would have to go back to the drawing board. I don't think GM would say we NEED .001 tolerance and the machine shop came back and said we can only do .010 and GM would say that would be fine. I think we have come to assume that the valve guide is what is wrong, if that is in fact the case and the holes in the guides are out of tolerance which I would think it would either be the diameter or the angle I don't see either on of those being a stackable
tolerance. The locations would be stackable but I don't see were that would cause failure. People have been making heads a long time I don't see why they cant now.
We will never know if it is a design flaw without all the drawings and the design. Can't tell if something meets the design without knowing the design. Simple as that. Some failing, some not failing doesn't tell us anything.
A design flaw doesn't mean every car has it. Did every single toyota with the stuck accelerator flaw fail? Nope? Was it a design flaw, sure as hell yes it was.
Look at a process capability chart like this
Say the design says says 2-8 like in the graph. The ones below 2 or above 8 are a QC issue. Now if parts are failing when they are in the 2-4 range, that is still within the design, and you are going to have a lot of parts 4-8 range that perform fine. That is a design issue.
GM service spec is not the specification for the heads design.
I would think that when GM sent out prints to get quoted for manufacturing and everyone came back and said we cant hold those tolerances that GM would have to go back to the drawing board. I don't think GM would say we NEED .001 tolerance and the machine shop came back and said we can only do .010 and GM would say that would be fine. I think we have come to assume that the valve guide is what is wrong, if that is in fact the case and the holes in the guides are out of tolerance which I would think it would either be the diameter or the angle I don't see either on of those being a stackable
tolerance. The locations would be stackable but I don't see were that would cause failure. People have been making heads a long time I don't see why they cant now.
And that is exactly how it works, you go back and forth, and work with the vendors. The vendors don't know your end results or design outputs in many cases. So you can tell them you need .010" you agree on that, put it on the drawing, but maybe to avoid a failure you really need the .001". The design verification testing may not have had enough samples, or tested the correct outputs. Maybe they took 5 heads and life tested them, but all those heads were nominal so the flaw didn't show up in design verification testing. Or maybe the output out the DV testing was no dropped valves, and none of them dropped valves, but after life testing the guides were all worn, but they never looked at that, or since the design output was met, it didn't matter.
No it isn't. Having a few in spec doesn't mean design is robust. Like I said, you have tolerances. Just making numbers up, but if you have a range from 8-12 [.....]
Yes, you're making too many numbers up in order to justify your theory. We know what the numbers are, we know that they are attainable, and we know that Linamar is failing (in some cases) to meet those numbers.
Bad machine work. Period. Sole cause? Who knows, but one to be addressed/assessed before engaging in totally unfounded speculation.
Like I said, unless we know what happened we don't know if it is a QC issue or design flaw. Could be either. If it drawing says concentricity between 8-12 and Joe programs it, has a bunch between 8-10 and the ones at 8 are failing, that is design. If he gets a bunch between 6-10 and the ones in the 6-7 range are failing, that is a QC issue. If someone has the actual released drawing with all the specs for the head, and takes a brand new one out of the box and does a full inspection and it has dimensions out of the drawings specs, that is QC, but it could be on the low or high end of a range that is in spec. We will never know unless someone from GM releases the drawings.
And it could get more complicated with process capability.
If the spec is 8-12, and the machine is only able to hold a +/- 5, then that is a design issue too because your process isn't capable of hitting the spec. It would be an engineers wet dream to call out perfect specs with no tolerance but we can't do that. You can't say head flatness across the deck of .0000000001 inch and then complain when manufacturing can't do that and say it is a QC issue because they can only do +/-.0001
This is exactly what I've tried to relay in past posts on this issue - it is refreshing to see my point made so succinctly, by a person with the experience and background to lend credibility to it.
IMO, LS7 heads do have a design flaw in that the design is so knife-edged that it could not be RELIABLY executed by GM's contractors, or with the indicated part specifications (I am thinking about posting a short write-up with pics, driving home just how much "slop" exists within the stock rocker arms - they are anything but a "precision" part..).
I chose to attempt to correct the design shortcomings with a roller tiped rocker, one manufactured with greater precision, along with a core design that in theory will improve the longevity of other components in the system.
If the heads that are in spec don't fail and the heads that are out of spec are failing that
is not a design flaw that is a machining and qc flaw. I have heard that some people bought brand new heads from GM and they were out of spec on the valve guide clearance. It would be nice to know if the ones that are in spec stay in spec or no,t if not then it is for sure a design flaw.
Yes, you're making too many numbers up in order to justify your theory. We know what the numbers are, we know that they are attainable, and we know that Linamar is failing (in some cases) to meet those numbers.
Bad machine work. Period. Sole cause? Who knows, but one to be addressed/assessed before engaging in totally unfounded speculation.
No we don't mark. What are the specs then on the head? If you know them list them all out? Not the service limits, but the specs for a new head?
Guide to Valve clearance from X to Y
Concentricity to seat from X to Y
And all the other specs for the head.
We know the machining isn't adequate to have a guide last, what we don't know is if the machining is within GMs specification. Knowing how well supplier quality is monitored, and how out of spec parts are charged back to the vendor, I really doubt that Laminar is just sending a ton of out of spec heads because GM would just dump them as a vendor, they would pay for it all, and they would move on.
More than likely they went back and forth on how dialed in they could get the process, got it to a price they liked, and at the price GM is paying they are just expecting some drop out. Money in the end drives the decisions.
Say they needed .005-.006 for no crazy guide wear, but Laminar quotes $2000 a head to meet that spec because they run the numbers and know they can only hit that 20% of the time and would have high scrap, or would need new machinery, etc. So they come back and say they will do .005 to .010 at $800 a head. GM may know that a certain number will fail that are on the high end, but they run the warranty analysis and figure it is cheaper to let a few slide and pay out under warranty than pay an extra $1200 per head. Sadly that is how many specifications are set. Engineers rarely get their way, the accountants do and you find a happy middle, or in this case a ****** middle.
As for heads out of spec stay in spec, there are plenty of heads that were in spec, and later measured out of spec as far as the service limit. Now the spec on a new head, I've never seen that posted. I know what AH and WCCH shoot for because they provided that, and they typically set them tighter than factory.
The spec for guide to valve clearance to laminar is not <.037". I 100% guarantee that. They have a range they are suppose to meet.
And I'm not presenting ANY theory. This is basic engineering, not theories. I don't know if it is a QC issue, Design issue, sampling issue, use life issue. Maybe it meets spec, and QC is good, but they didn't account for all the owners revving them to 7k all the time and did their design verification in a manner similar to a normal car, so they didn't catch the guide wear. Maybe the design verification testing didn't catch all the variables and a 1.8 rocker non roller just isn't good enough, but they didn't inspect guide wear after life testing, or did and didn't care since they have no failures of their design outputs.
My whole point of these silly rants is we don't know. Is it a design flaw? Could be. Is it a QC issue, maybe. Is it something else, maybe. You can't rule out a design flaw without knowing the design and why they are failing, so people saying a few made it 100k+ without failure so therefore it isn't a design issue are just wrong. I just did a design verification on 30 parts, 29 of them made it to end life zero issues. 1 of them exhibited stress cracks from the decontamination wipe chemistry breaking down the material. We didn't notice that chemistry of the wipe changed, didn't pick a compatible material, and now have a flaw that shows up. That isn't a QC issue, it 100% is a design issue that will show up on parts that are on the low end of thickness. We could respect the parts to make them all ~10% thicker, add reinforcements to the cracked areas, or change materials, that is all part of the design.
[...] I am thinking about posting a short write-up with pics, driving home just how much "slop" exists within the stock rocker arms - they are anything but a "precision" part.. [....]
Please spare us; there is nothing wrong with the stock rocker arm for most applications. A desire for "precision" does not mean it is required for proper performance in a particular area.
No we don't mark. What are the specs then on the head? If you know them list them all out? Not the service limits, but the specs for a new head?
Guide to Valve clearance from X to Y
Concentricity to seat from X to Y
And all the other specs for the head.
We know the machining isn't adequate to have a guide last, what we don't know is if the machining is within GMs specification. Knowing how well supplier quality is monitored, and how out of spec parts are charged back to the vendor, I really doubt that Laminar is just sending a ton of out of spec heads because GM would just dump them as a vendor, they would pay for it all, and they would move on.
More than likely they went back and forth on how dialed in they could get the process, got it to a price they liked, and at the price GM is paying they are just expecting some drop out. Money in the end drives the decisions.
Say they needed .005-.006 for no crazy guide wear, but Laminar quotes $2000 a head to meet that spec because they run the numbers and know they can only hit that 20% of the time and would have high scrap, or would need new machinery, etc. So they come back and say they will do .005 to .010 at $800 a head. GM may know that a certain number will fail that are on the high end, but they run the warranty analysis and figure it is cheaper to let a few slide and pay out under warranty than pay an extra $1200 per head. Sadly that is how many specifications are set. Engineers rarely get their way, the accountants do and you find a happy middle, or in this case a ****** middle.
As for heads out of spec stay in spec, there are plenty of heads that were in spec, and later measured out of spec as far as the service limit. Now the spec on a new head, I've never seen that posted. I know what AH and WCCH shoot for because they provided that, and they typically set them tighter than factory.
The spec for guide to valve clearance to laminar is not <.037". I 100% guarantee that. They have a range they are suppose to meet.
And I'm not presenting ANY theory. This is basic engineering, not theories. I don't know if it is a QC issue, Design issue, sampling issue, use life issue. Maybe it meets spec, and QC is good, but they didn't account for all the owners revving them to 7k all the time and did their design verification in a manner similar to a normal car, so they didn't catch the guide wear. Maybe the design verification testing didn't catch all the variables and a 1.8 rocker non roller just isn't good enough, but they didn't inspect guide wear after life testing, or did and didn't care since they have no failures of their design outputs.
My whole point of these silly rants is we don't know. Is it a design flaw? Could be. Is it a QC issue, maybe. Is it something else, maybe. You can't rule out a design flaw without knowing the design and why they are failing, so people saying a few made it 100k+ without failure so therefore it isn't a design issue are just wrong. I just did a design verification on 30 parts, 29 of them made it to end life zero issues. 1 of them exhibited stress cracks from the decontamination wipe chemistry breaking down the material. We didn't notice that chemistry of the wipe changed, didn't pick a compatible material, and now have a flaw that shows up. That isn't a QC issue, it 100% is a design issue that will show up on parts that are on the low end of thickness. We could respect the parts to make them all ~10% thicker, add reinforcements to the cracked areas, or change materials, that is all part of the design.
Again, thanks for explaining the engineering aspects of design flaw vs QC. Very enlightening !
Its refreshing to hear from an engineer that engineers can screw up like everyone else. A while back I got lots of crap for pointing out that GM engineers are perfectly capable of screwing up.....
Again, thanks for explaining the engineering aspects of design flaw vs QC. Very enlightening !
Its refreshing to hear from an engineer that engineers can screw up like everyone else. A while back I got lots of crap for pointing out that GM engineers are perfectly capable of screwing up.....
DH
Mr. D.H. who said engineers cannot screw up ?? Post's please..
And BTW this series of posts does not prove engineering screwed up the design. All speculation and without a shred of data. Anecdotal inferences/refrences totally unrelated to what we are dealing with..
A mechanical design engineer is the professional you should seek for advice and guidance.
Please spare us; there is nothing wrong with the stock rocker arm for most applications. A desire for "precision" does not mean it is required for proper performance in a particular area.
..Awwh shucks, I might do it just for your entertainment now..
When is more precision a bad thing, and don't forget aftermarket rockers offer more than just improved bearing tolerances.
Valvetrain Engineering 101 - rollers are always easier on guides than non-rollers.
Mr. D.H. who said engineers cannot screw up ?? Post's please..
And BTW this series of posts does not prove engineering screwed up the design. All speculation and without a shred of data. Anecdotal inferences/refrences totally unrelated to what we are dealing with..
A mechanical design engineer is the professional you should seek for advice and guidance.
Yep, no blame, theories, or anything presented. No data, no speculation. Just basics showing why it can still be a design issue, or any other issue, but we will never know unless someone releases all the internal GM documents. In the end, as the consumer, it doesn't matter if it is a QC or design, or whatever issue, we got screwed one way or another.
It was simply to refute that it can't be a design issue because a few cars made it to 60k+ miles without wear. Plenty of people didn't die, or crash, or have any issues with all the recalled ignition switches, doesn't change the fact that it was a design issue.
No we don't mark. What are the specs then on the head? If you know them list them all out? Not the service limits, but the specs for a new head?
[1] Guide to Valve clearance from X to Y
[2] Concentricity to seat from X to Y [...]
1. It's in the service manual, which I don't have handy. It's been posted here before (something like .0014 - .0020).
2. Concentricity spec, by simple engineering principles, will be less than new spec guide clearance. Again, desperation to prove a fantasy has led your argument into the Twilight Zone.
[...] Valvetrain Engineering 101 - rollers are always easier on guides than non-rollers.
Maybe, but they will always lower the RPM at which severe bounce / float occurs, which can potentially destroy your cam (and your engine, for that matter).
Yep, no blame, theories, or anything presented. No data, no speculation. Just basics showing why it can still be a design issue, or any other issue, but we will never know unless someone releases all the internal GM documents. In the end, as the consumer, it doesn't matter if it is a QC or design, or whatever issue, we got screwed one way or another.
It was simply to refute that it can't be a design issue because a few cars made it to 60k+ miles without wear. Plenty of people didn't die, or crash, or have any issues with all the recalled ignition switches, doesn't change the fact that it was a design issue.
02-10-2015, 12:26 PM
