Exploration into Somender Groove Theory

To start:
THIS IS NOT ME ASKING YOU GUYS IF SOMENDER GROOVE THEORY WORKS OR NOT, THIS IS ME WANTING TO PERFORM A SERIES OF TESTS OF WHETHER OR NOT IT DOES

To elaborate:
With my recent dyno tests of my engine that established that it runs just as well as it did when it came out of the factory, and my upcoming head replacement in the next few months, I have realised that my engine is the perfect testbed for a specific engine modification that the TB community has been arguing about for the past decade and a half.

A be-all-end-all test for Somender Groove Theory.

I would like to start a thread documenting end user testing of people driving my car as it is right now, the lowest comfortable shift points, the max advance it can have before knocking, etc. All sorts of tests that, after a sample size of 10 or 50 or 100 or whatever test drivers, the cylinder head will be removed, modified with 2 or 3 Somender Grooves (nothing else), and rerun through the testing. The end question being, of course, if the Somender Groove Theory results in any changes from these previous results, and if it does, how it now differently acts.

Somender Groove Theory has, in the past, been a contentious topic. I am just a man looking to create a thread that will finally answer the question of whether it has an tangible effects that are noticeable for people who don't know the modification has taken.

Instead of many people testing many modified cars, giving anecdotal evidence, I want to do a test with many people testing one car, across a significant sample size. If you have tried grooves, tell what you experienced, and that will be added to the list of tests to see if it can be substantiated and isn't just an anecdote. If you have doubts, say your doubts and they'll be tested for.

So, before the experiments begin, anybody got any recommendations for experiments that can be done on a running car, before and after the modifications?

EDIT:
As additional test recommendations are given, I will be adding them to a list in the bottom of this post.

0. Tune MAF CO mixture to make sure it's operating properly, change oil, run injector cleaner, replace fuel pump.
1. MPG over time (at least 10 fill ups) (yes my odometer works)
2. Dyno pull from idle speed
3. Normal dyno pull with Coleman Gas/White Gas
4. [A TBD amount of] people drive the car, saying the slowest speed that they're comfortable shifting at, and the general impression.
5. EGT at cruise. (I have cruise control)

Good luck. I would just make sure the car was at a solid stage zero with a new head gasket.

I could tell you my experience doing that already but you don't want to hear it apparently.

Do I still get your engine when the experiment is over?

2turbotoys said:

I could tell you my experience doing that already but you don't want to hear it apparently.

Click to expand...

It's not that I don't want to hear it. It's that I want to create a series of tests that can verify whether certain affirmations about the modification is true, while avoiding turning this thread into a feces-slinging competition, like every other Somender Groove Thread has become. Give your experience, and maybe we can come up with tests based on them.

Regarding stage zero: Go to my thread on the dyno numbers, the engine is the definition of stage 0.

fatcatbestcat said:

Do I still get your engine when the experiment is over?

Click to expand...

If my engine dies, yes. If not, no.

Will you be milling the head to compensate for the volume the grooves take up? Easy to improve detonation resistance if you're dropping compression.

redblockpowered said:

Will you be milling the head to compensate for the volume the grooves take up? Easy to improve detonation resistance if you're dropping compression.

Click to expand...

I was not, for 2 reasons:
1. Milling the head would require first CC'ing the current head and then shaving appropriately. I lack the funds and/or means to do so, as die-grinding the head to a certain cc would be significantly more expensive than shaving a specific height from the head. If I can receive support in this venture however, this would be a great addition.

2. Milling the head retards the timing, which would move the torque curve up, making the testing unequal.

Number 2 might actually be okay, though? Since Somender Grooves allegedly improve bottom end, if there were low end gains despite the cam timing change, it would be a real credit on the side of the Grooves actually working.

I don't think you'll see as much gain in the Somender groove as other similar methods might give... Baffled intake, swirled valves, etc

But that's just my guess. Here's some textbook material that might interest you.

The main question is whether Somender Grooves work as well as proponents of it claim it does, not whether something else might help more. That's why I want to do testing, and why I'm looking for recommendations for potential tests that can test these claims to confirm either or both sides of the argument. Somender is the easiest modification to do out of what you've listed, but if you can think of a way to implement these in a 530 head for a reasonable price, I'd love to try them.

IansPlatinum said:

I don't think you'll see as much gain in the Somender groove as other similar methods might give... Baffled intake, swirled valves, etc

Click to expand...

I think to make it worthwhile we need some hard data. There could be 50 people that drive the car before and after and say it is better with the grooves, but basic stuff like iat, humidity changes, azimuth of the moon, etc between trials really affect the way LH operates. Easy data to collect like ambient temp, humidity, atmospheric pressure and also not so easy stuff like iat, spark angle, calculated load, and knock data I think would be almost necessary to say for certain it makes a difference. Or else it's just people repeating the same stuff they have for the past 20 years where they think it does or doesn't make a difference.

I know nothing about the chrysler ignition system, but from posts here it seems like it is very aggressive in its ignition advance algorthm. If it is actively trying to push maximum advance before detonation, we could use that advance number as a simple data point. The grooves are rumored to allow for less advance, due to the increased turbulence and flame front speed, right?

My opinion based on zero evidence either way is that it makes no difference. We see stuff like swan necks and cut down valve guides in engines new from the factory, but never has there been a groove. (as far as i know) I would like to be proven wrong though

I would definately be down to help collect some of this data

I would want a full standalone ECU with datalogging and multiple knock sensors, and cylinder pressure monitoring to determine if this is something that is working or not.

Id also think you would want it to be done on an engine dyno with more easily controlled and consistent environment.

blkaplan said:

I would want a full standalone ECU with datalogging and multiple knock sensors, and cylinder pressure monitoring to determine if this is something that is working or not.

Id also think you would want it to be done on an engine dyno with more easily controlled and consistent environment.

Click to expand...

That sounds expensive. Any recommendations for ECUs capable of this?
I'm guessing the cylinder pressure monitoring is to see whether or not the cylinder pressure is actually the same whether using the grooves or not? To make sure that the grooves are actually inhibiting knock or just lowering cylinder pressures.

Know of any places that would do engine dyno tuning cheap?

Smorgesborg said:

That sounds expensive. Any recommendations for ECUs capable of this?
I'm guessing the cylinder pressure monitoring is to see whether or not the cylinder pressure is actually the same whether using the grooves or not? To make sure that the grooves are actually inhibiting knock or just lowering cylinder pressures.

Know of any places that would do engine dyno tuning cheap?

Click to expand...

Megasquirt should be capable.

blkaplan said:

Megasquirt should be capable.

Click to expand...

I see. Can it do the cylinder pressure monitoring using ion sensing?

I guess I'm going to need to start tearing out my 93 parts car engine and harness. I'll be back in Massachusetts at my parts car over Thanksgiving, if anyone reading this has enough experience with megasquirt to do this with me, the effort will be greatly appreciated and will be rewarded with beer.

The point of using ion sensing seems excessive, though. If the intention is to see how high the timing can be pushed, that ends up being dependent on the specific groove application, if they do work as people say. Might there be a way to implement these tests on a practical level, in a way that can be observed for its efficacy (or lack thereof) with a more basic application? To show that just having the grooves either will or will not provide these benefits, rather than the degree to which the benefits are provided?

But if it's going to be an end-all-be-all test, it might be necessary?

I would like to do it, but there are some methods that are very difficult for me to gain access to.

cheap and engine dyno normally don't go together, neither does cheap and cylinder pressure monitoring (although it's been some time since I've looked at anything that can do that)

Smorgesborg said:

I see. Can it do the cylinder pressure monitoring using ion sensing?

I guess I'm going to need to start tearing out my 93 parts car engine and harness. I'll be back in Massachusetts at my parts car over Thanksgiving, if anyone reading this has enough experience with megasquirt to do this with me, the effort will be greatly appreciated and will be rewarded with beer.

The point of using ion sensing seems excessive, though. If the intention is to see how high the timing can be pushed, that ends up being dependent on the specific groove application, if they do work as people say. Might there be a way to implement these tests on a practical level, in a way that can be observed for its efficacy (or lack thereof) with a more basic application? To show that just having the grooves either will or will not provide these benefits, rather than the degree to which the benefits are provided?

But if it's going to be an end-all-be-all test, it might be necessary?

I would like to do it, but there are some methods that are very difficult for me to gain access to.

Click to expand...

Not sure, you would have to do some digging. Maybe some of the Saab guys were able to get the stock ion sensing stuff working with a standalone?

linuxman51 said:

cheap and engine dyno normally don't go together, neither does cheap and cylinder pressure monitoring (although it's been some time since I've looked at anything that can do that)

Click to expand...

Amen to that! This is a place where you have to pay to play.

Would it be feasible, practical, or acceptable to come up with a series of tests or benchmarks that do not involve expensive aftermarket ECUs and engine tuning, and can be done while keeping the engine in the car?

If over a long period of time, and the car remains the same before and after besides the cylinder head, could that be enough to be considered a controlled environment?

I didn't realize that turbobricks was still one of the top web sites for Somender Singh Grooves. There's a really long thread here (and amazingly, some of the pictures are still intact): http://forums.turbobricks.com/showthread.php?t=66779

My off-the-cuff opinion is that the grooves can help a little, but for Redblocks, tight squish is much more effective and proven.

As a hobbyist, you're pretty limited in the testing you can do. If you want to go all out, you need 2 new head gaskets and lots of time. You'd need to pull the head, clean it all up, do whatever else you're planning on, then put it back together un-grooved.

Adjust your timing (LH2.2 NA, right?) to factory angle or desired angle before knocking. Drive it and see how it performs, mileage, etc. This will give you a baseline on the cleaned up engine.

Next, pull the head, clean up any deposits (shouldn't be much, if any), add the grooves, and put it back together. You can try re-adjusting timing based on knocking with the grooves added. Hopefully, you'll have better mixing and reduced knock, letting you get a little more power out of the engine (or better mileage due to less gas needed for same power level).

You might also buy a KnockSense board to help measure knock. I don't know where you'd mount the 2nd knock sensor, or if LH2.2 (right?) can run without its knock sensor. You'd want to adjust the knocksense so that it just starts to flicker on some repeatable section of roadway/speed.

blkaplan said:

I would want a full standalone ECU with datalogging and multiple knock sensors, and cylinder pressure monitoring to determine if this is something that is working or not.

Id also think you would want it to be done on an engine dyno with more easily controlled and consistent environment.

Click to expand...

This, but you could also do it with EZ11xK, if you could Ostrich it. Run it on a dyno, crank up the advance did it knocks, then groove it and do the same. Like redblockpowered said though, if you don't compensate for compression loss, it's a flawed comparison. 1-2 cc of groove would be enough to make a difference.

I would also like this topic put to bed.

Smorgesborg said:

I see. Can it do the cylinder pressure monitoring using ion sensing?

I guess I'm going to need to start tearing out my 93 parts car engine and harness. I'll be back in Massachusetts at my parts car over Thanksgiving, if anyone reading this has enough experience with megasquirt to do this with me, the effort will be greatly appreciated and will be rewarded with beer.

The point of using ion sensing seems excessive, though. If the intention is to see how high the timing can be pushed, that ends up being dependent on the specific groove application, if they do work as people say. Might there be a way to implement these tests on a practical level, in a way that can be observed for its efficacy (or lack thereof) with a more basic application? To show that just having the grooves either will or will not provide these benefits, rather than the degree to which the benefits are provided?

But if it's going to be an end-all-be-all test, it might be necessary?

I would like to do it, but there are some methods that are very difficult for me to gain access to.

Click to expand...

MS 2 and 3 can accept a knock detection signal and adjust the advance to make the knock 'go away'. MS does not have the algorithm to interpret raw data from a piezo style knock sensor or the Saab trionic ignition cassette to determine that knock is occurring. That is a complete separate time consuming exercise. Plus, adapting the Saab cassette to provide ignition and knock monitoring on a non Saab engine would require significant effort. the later Saabs relied on the cylinder pressure signal as a replacement for the cam sensor so you are in to significant amount of coding to try to get that to work with MS. There is no plug and play here.

Knock sensing and using the ignition timing at which knock occurs as an indication of performance is not a particularly useful tool. There are modern (pre DI) engines where the turbulence is such that at high engine speeds knock is a non issue. You can advance the ignition timing and it will not induce knock (at high engine speed). All it does is reduce horsepower.

The Somender grooves might alter turbulence and burn rate which might make the engine more knock resistant in its range of operation where knock is a potential problem. That may be beneficial in terms of octane sensitivity in those regions where the engine displays detonation. That may or may not have any effect on peak hp or torque numbers which are the typical indicators of performance.

The 'wisdom' suggests that you want to target peak cylinder pressures to occur somewhere around 20 ATDC. If adding more ignition advance moves the peak pressure point back closer to TDC you will reduce power. Absent a detonation problem, there is an optimal ignition advance value and going beyond that is not going to yield more power.

If you want to try the Somender grooves on your engine, go ahead. But your testing described testing method is mostly subjective, not quantitative and will most certainly not be a 'be all, end all' test. You would need a dyno test of a fresh unmodified engine with stock advance curve to establish a base to be compared to a fresh engine with a modified head where you have run enough dyno tests to evaluate an optimal advance curve with the modified head. To further complicate things if the Somender grooves truly affect turbulence then you will likely need to alter your fuel map. Most port injected engines operate with sub stoichiometric AFRs at peak output to try and ensure all the O2 is burned up. There is a fine line between using up all the O2 versus adding so much fuel that it cools the combustion mix and reduces peak cylinder pressure. If the Somender grooves work then raising the AFR to reduce the unused fuel cooling problem may be where the performance gain comes from. You won't know that unless you are prepared for the time and money to dyno test to determine the engines new optimal ignition and fuel maps.