Hey Max. Your danged PM box is full again!!
Wow, I missed a lot. Max, I wouldn't take away anything in terms of a comparison with cam changes as indicative of a manifold causing a drop in torque. Isolate the variables, as you already know. That said, it's hard to say what I'd go with on a throttle size, because A) is there a compressor map for the K04-02x floating around? and B) even if there was, you're so far off of it with your chemical cooling that I don't know how sound it would be to base flow rates from power output from it. The reality is you'd be staying small port head, and therefore small port manifold (those transition spacers are such a joke). Again, just like the dual tip 250 hp comments in the other thread, (and I know you already know this) removing restrictions from the intake path lets the turbo not work as hard to make the same pressures. Larger plenum and runners designed with the injector out of the charge path are going to give gains across the rpm range and not affect low end torque in a negative way, assuming the plenum isn't retarded large. In fact, it may allow boost threshold to be hit SOONER in the rpm range. The biggest variable, as you know, would be throttle size, but a few passes on the dyno with different throttles via adapter plates would tell you what you needed to know. That said, you seem to think fueling would need adjusting to fully realize what each throttle did, whereas I think WOT = WOT. My inner knowledge of ME7.5 has probably grown outdated as I stopped paying attention a few years ago.
The only basis to go off that I know of is the forumla from Corky Bell's Maximum Boost regarding throttle size:
Velocity = CFM/(throttle cross sectional area in^2) X (1 min/60sec)/(1 ft/144 in^2)
This doesn't take into account airflow at different rpms obviously, so your midrange monster may not fit here, as this looks solely at peak power airflow rate. Now a graph could easily be created at different flow rates to see where it may or may not impact your engine. The only guideline given is to not exceed 300 ft/sec peak velocity. I think I have a 65mm Audi throttle that I'd be willing to donate though.
Last edited by 20v master; 10-02-2012 at 11:47 AM.
After crunching a few numbers, a stock displacement motor running at 35 psi (I think that's the number I've seen you throw around lately, not sure if that is spike, midrange, at redline?), a volumetric efficiency of 100% (yes, I know this is a mystery number, but we are F/I so in reality, using boost a pressure ratio vs performance of the NA 20v 1.8 is a vague guess on CFM), and a 65mm throttle, this setup would just be bumping over 300 ft/sec velocity at 7K rpms. I'm guessing you are getting 35 psi in the midrange, but less at redline? A quick boost plot and a MAF log to convert to CFM would tell us with decent accuracy what size throttle to run. Let me know if you want the spreadsheet.
Last edited by 20v master; 10-02-2012 at 04:30 PM.
Audi TT mk1 FAQ - Demokratikally Elekted Undersekretary of Peter Panhood of the Independent People's Republik of Offtopikstan
We have 1.8l of displacement or 108.7 cubic inches. With the 4 strokes, we have the intake valve on a single cylinder opening once every two revolutions. I know you already know all that but I want to keep everyone in the loop, so every two revolutions our engine moves 108.7 CI
So, we are moving:
(using this formula: RPM X CID / 1728 X 2)
94.61 CFM @ 3K
126.15 CFM @ 4K
157.69 CFM @ 5k
189.23 CFM @ 6K
220.77 CFM @ 7k
This is at unrealistic ideal situatuations and not even taking into account volumetric efficiency. As you pointed out, there is no real way, but an educated guess to pinpoint the exact volumetric efficiency that we are operating at. Obviously we are not operating at 100% VE, and optimistically I'd like to think we are somewhere around or above 75%
With this crude data in hand, we can compare it to known tests performed on a flow bench for various manifolds.
The OEM units flowed roughly at 28 psi
150-156 CFM across the board
SEM at 28" Hg
Dahlback at 28" Hg
ABD at 28" Hg
APR at 28" Hg
(there are two newcomers in IE and 034, I am not including them since they haven't been independently tested yet).
So my conclusion on paper is that, in my goldilock region of 3K to 5k, anything above a stock unit is going to loose TQ. It would take 6-7k rpm for most aftermarket units to start to shine, and at that point, not really worth it. What's your take on my crude calculations?
Last edited by Marcus_Aurelius; 10-05-2012 at 08:24 PM.
They are nothing but a map showing projected flow numbers at various known efficiency levels. For everyone else to understand, the goal of the compressor wheel is to create pressure, nothing else. The efficiency map shows how much energy (in percentage) is used to create pressure/boost as opposed to creating heat. For example, a compressor operating at the "ideal" 75% efficiency island is using 75% of the energy (which comes btw from the exhaust driven or turbine wheel) to make pressure and the rest is used to create heat. This is where you get the temperature Delta between compressor inlet and outlet, the percentage of the compressor wheel used to create heat determines the temp variance between inlet and outlet.
In my case, as long as I can make pressure (aka boost) I'm fine. The air temperature will be cooled down at great levels before it is used. It is not unusual for rally teams to create compressor efficiency maps that fits their applications. Just like me, they run way off the manufacturer's compressor map by maxing out the compressor's ability to make pressure, but use water injection and race fuel to get good usable cylinder air charge. Old rally trick or formula that is also implemented by the turbo autocrossers like myself.
So wind the piss out of the turbo...cool the exiting Lava with hardware and chemicals.
"...I recommend books. People who don't know what they're talking about are less likely to write a book about the subject...."
Edit: I just popped over and asked our R&D guy (who has 20 years experienc at GM before working here) and he's in agreement that it would be better to look at your actual flow rates from the MAF (assuming there is no scaling going on there, ask Gonzo) and not make assumptions based off NA VE.
Edit #2: You don't even apply VE to your calculated flow rates. If you used .75 or .8, you'd almost be below what the OEM manifold flows, even at 7K. And those flow rates for manifolds were done with no throttle attached, aka you're going to reduce inlet cross sectional area by a small percentage, and therefore affect flow through the manifold. I go back and forth between pitching you to try the SEM and holding it for my 180Q (there's already one on my stroker setup), but it may come down to sending it to you for testing with the throttle body. Dyno's (load bearing) don't lie. The only problem there is you'd have to tap it for your w/m nozzles, which could be done and would force me to run w/m injection if you decided the manifold didn't suit your needs.
With the flow you are putting through it, especially at the pressure you see, I just can't believe that the OEM manifold isn't a choke point, especially with the injectors hanging out in port flow path.
Last edited by 20v master; 10-03-2012 at 01:56 PM.
From what I got from formal training (more focussed on tuning aspects), you can never operate at 100% or higher for various reasons, and that's on a turbocharged car. Theoretically, say I am making 30 psig of boost in the intake manifold and open the throttle valve, at 100% VE I should be able to cram in 31-32 psi of boost. The reality however, is that remaining exhaust gases (we are nowhere near perfect exhaust evacuation on a stock head/manifold/turbo/downpipe in the 1.8t), valve design restrictions, and heat of the charge, all comes into play to ruin the fun.
In general, the VE will not change much unless alterations are made to the valves, combustion chamber and such, what will change is the absolute temperate and mass air. A bigger turbo, or a more efficiently intercooled car, will have much higher mass air than stock but the motor will still flow the same 94.61 CFM at 3000 rpm. The main factor being the much more efficient absolute temperature (actual temp+ 460F) and higher mass air. Keeping in mind that the volume V is constant for the motor, the variables in the equation becomes the absolute pressure PSIa (actual+14.7) and the absolute temperature and mass. The ideal and actual CFM flowed will change as a factor of rpm, absolute temp, absolute pressure, mass air etc, but the VE will remain under ideal regardless.
Real life example is:
I force the K04 to pressurize 30 psig which is 44.7 psi absolute, it is the same any turbo, bigger or not, making 44.7 psia (BT advantage is producing lower absolute temp at fixed mass, and being able to pressurize that amount at higher revolutions). At 3k I will be limited by the rigid 94.61 volume of the motor and functioning under max VE, regardless of turbo and components used.
If you didn't operate at higher than 100% VE, then an FI engine wouldn't make any more power than an NA engine. The heating of intake air over ambient reduces the mass entering the cylinder, but the pressure overcomes this. Whether that fact is reflected in the density of the air entering the cylinder or the higher VE, it's got to show up in the basic engine power formula. You left VE out of your 1.8T CFM calculations, but didn't reference pressure ratio (PV=nRT). Again, I know you know all this, but spelling it out for anyone that wants to contribuite, but you have to account for it somewhere.
Based on your calc....
94.61 CFM @ 3K = 167 g/sec @ .75 VE = 125 (you're flowing roughly 95 g/s here, ~57% VE)
126.15 CFM @ 4K = 222 g/sec = 167 (you're flowing roughly 210 g/s here, ~95% VE)
157.69 CFM @ 5k = 278 g/sec = 209 (you're flowing roughtly 240 g/s here, ~75% VE)
189.23 CFM @ 6K = 334 g/sec = 251 (you're flowing roughly 245 g/s here, ~73% VE)
220.77 CFM @ 7k = 389 g/sec = 292 (you're not even close, 63% IF you held the 245g/s )
You can easily see, you're close to 100% at 4K.
Yes, there will always be residual gas in the cylinder once the exhaust valves close, but taking into account that VE should rise with rpms due to swirl, tumble, squish, but primarily due to higher piston speeds creating quicker changes in pressure on valve opening, yours is dropping rapidly past midrange. A lot of this is due to cam specification, but this does lead me to believe you've obviously still got flow restrictions somewhere. You can't change the turbo by class rules, and you've obviously got a much more optimized intake tract that most if not all, but there will be pressure drop with the 2 bends, across the IC core, through the throttle, intake mani, and valves. Out of those which are the ones that haven't been modified, are the easiest to change, and weren't designed for what you're using them for? If there's something I'm leaving out, as I haven't brusehd these cobwebs off in a while, I'm all ears. I miss Dr. Jones's engine class.
Some very good points Adam! The tested flow rates at the OEM manifold didn't account for TB and most likely the restriction created by injectors protruding in the flow path of the OE design. I'm not running a Gonzo tune (at least not yet), he only implemented some features over the existing file for me. That file has a MAF offset for some reason, so I wouldn't fully trust the logged readings. On top of that, I run a housing with a larger cross section which scales down the actual values (although by a known amount).
Reading back into it, it is becoming more obvious to me that modern turbocharged engines operate closer to 100% VE than I was saying (still won't go over 100% VE ). Certain standalone Engine management systems arbitrary even use 90-95 % in their VE table as a base for forced induction. Now, the question for you is how can I accurately use mass air rate values (lbs/min or G/s) to assess various known manifold air flow rate (CFM)?
Since we are switching to air mass rate and incorporating pressure to the equation, this is what I'm getting for the 1.8l motor at the Psia I'm running (for everyone else this is no longer applicable because we all run different pressure. To use the formula you'd have to input your own Psia and corresponding absolute temp)
Using the formula n(lbs/min)= P (psia) X V(CFM) X 29 / (10.73 X abs temp)
Ideal lbs of air/min:
3k = 22.94 lbs/min
4k = 27.69 lbs/min
5K = 34.30 lbs/min
6K = 38.75 lbs/min
7K = 42.97 lbs/min
I am thankful for the help and the time you're taking to bring some much needed technical oversight. I see that there are restrictions left and room for improvement (both intake and exhaust side). Like you said, my intake track is flowing pretty well now, and I agree that the manifold/TB could be the remaining restrictions on that side of the motor, I just don't want to loose any TQ in the red highlighted range that is used the most while racing.
I would feel bad drilling your SEM and not being able to use it (although your future setup would be sweeter with direct port injection anyway ). As soon as I schedule a dyno day, I am going to contact you to see what we can do, but the 65 mm TB is a definite.
Last edited by Marcus_Aurelius; 10-03-2012 at 11:25 PM.
Audi TT mk1 FAQ - Demokratikally Elekted Undersekretary of Peter Panhood of the Independent People's Republik of Offtopikstan
Eric, the graph you posted is somewhat rigid as it doesn't account for engine speed aka rpm. The first 1.8l CFM numbers I posted took rpm into account at a NA level for the motor, but we are taking it a step further and bringing pressure and air mass into the the picture.
Linky didn't work for me, so I searched it
There are two graphs posted that are somewhat usable but at very low psi (18 psi is half what I run and doesn't do me any good).
Let's see if I can consolidate everything we got so far in one post:
Adam got this below, based on our 1.8l motor volume at various RPM. He calculated ideal where VE = 1, then VE = .75, and finally what I was seeing in terms of VE% basing his results on one of my G/s logs.
formula is n(lbs/min)= P (psia) X V(CFM) X 29 / (10.73 X abs temp)
3k = 22.94 lbs/min ==> 173.19 G/s
4k = 27.69 lbs/min. ==> 209.05 G/s
5K = 34.30 lbs/min ==> 258.96 G/s
6K = 38.75 lbs/min ==> 292.56 G/s
7K = 42.97 lbs/min ==> 324.42 G/s
At VE = .95 ------------------------------ At VE = .90
3K = 164.53 G/s ------------------------ 3K = 155.87 G/s
4K = 198.59 G/s ------------------------ 4K = 188.14 G/s
5K = 246.01 G/s ------------------------ 5K = 233.63 G/s
6K = 277.93 G/s ------------------------ 6K = 263.30 G/s
7k = 308.19 G/s ------------------------ 7K = 291.97 G/s
Now I'm just waiting on Adam to tell me the best way to compare the data to the tested manifold flow rates.
And sorry if I sparked a fire in you to start crunching numbers, it wasn't intended. Anyways, I see you're taking flow numbers from the intake mani comparison I did. You are listing those as being at 28 psi, when they were at 28 in of vacuum. They were done with no throttle or injectors in place, only injector bungs and all vac ports blocked off. You'd be hard pressed to correlate those numbers in vac to a certain flow rate in positive pressure on the inlet and a different flow rate on the ports in vacuum at different rpms, and that doesn't take into account the pulses of pressure caused by valve cycling. Also, I never tested an SEM on the same flowbench. I don't see the specs on the SEM that you have numbers for, ie small vs big port, what size throttle bore, etc. So all that said, the intake mani result comparisons are just relative, not absolute enough to be able to predict performance.
You can look at the dyno below to see the effects of changing from a small port passenger side manifold and throttle to a big port SEM manifold with 80mm throttle on an AEB head with a 3071R setup at 24 psi. I don't know specifics of the whole setup. It may have gone from a stock SMIC to a full FMIC at the same time, or it may have just been a piping change with the same core before and after to swap to driver's side throttle. Either way, I think it's safe to say going to a bigger IC and more piping would have contributed negatively, yet it's obvious there was no loss in torque at any rpms you'd be concerned about. First is the actual dyno, second is the same data plotted so it's easier to see (dark blue vs pink in second graph).
You can actually see the boost comes on harder and sooner with the SEM manifold. Remember, peak volumetric efficiency is going to be at peak torque, which is defined by the camshaft profiles. Like I first said, I wouldn't take a loss of torque in low/midrange from a cam change as indicative of it being caused by the intake manifold.
As for VE not going over 100%, that depends on your definitions I suppose. If you're comparing NA to FI peformance, then it most definitely is possible. If you're comparing theoretical, then yes, you'll never hit over 100 due to residual exhaust gas, and fuel occupying space that oxygen could be in for a straight volume at atmospheric pressure standpoint.
28" HG??? That seems fairly high for the flow test, why did you test at such high vacuum that is outside of our motor's operating range? I never saw the result thread (maybe it got blackholed by mods protecting vendors). All I've seen is charts with individual runner, and overall flow comparisons between a host of manifolds used in various threads. I got the SEM manifold flow numbers off of their website and it stated "28 psi test pressure", so that's what probably got me mixed up with your test.
I guess that's where my attempt to make a theoretical comparison ends. I have no data where I can compare the OEM manifold to SEM in terms of flow (not that it would paint an overall picture applicable to my application that depends on transient response more than total flow). If I'm understanding this properly, there is no "real" full comparison under positive pressure (which would make more sense for a turbo car) between the manifolds for the 1.8t? How did the community come to the consensus that brand X (aka SEM) is the top performer for X,Y, and Z applications?
I feel that I'm leaving this with more questions than answers
From the comparison thread...."This does NOT correlate into which will make the most power on a given setup. Intake manifold flow is dynamic and undergoes turbulence, heat change, as well as sudden transitions, like during a WOT lift to shift. All testing is done at 28 in of water to ensure accuracy in comparisons, but flow in vacuum is entirely different than positive intake manifold pressure."
Water, not mercury. This is what the flowbench that was used was setup for, what the flowbench operator always tests at, and was merely a reference to keep tests consistent from manifold to manifold.
This answers one my question, Adam's test was conducted at 28" of water, not 28" HG. Makes a lot more sense to me as 28" of H2o is roughly 2" HG. Maybe if I could find flow tests in positive pressure from the new IE and 034, a comparison can be made between these two and the mighty SEM. What do you say Dizzy (or axlekiller if you prefer )?