Announcement

Collapse

Get more Tips, Specs and Technical Data!

Did you know... this Forum is a service of the Studebaker Drivers Club? For more technical tips, specifications, history and tech data, visit the Tech Tips page at the SDC Homepage: www.studebakerdriversclub.com/tips.asp
See more
See less

R3 heads flow rate

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #76
    Ask for a little work...
    Completely ignored.

    Mike

    Comment


    • #77
      Originally posted by Mike Van Veghten View Post
      Ask for a little work...
      Completely ignored.

      Mike
      I've got a lot of discussions ongoing and a terrible memory - I don't even remember what I've said in this thread. Let me state my objectives:

      First, I was very amazed at the work that has been going on regarding porting. When I saw CAD models of ports, I was excited because the next step, the very next step, is CFD. However it looked like no one was doing that. I don't know if anyone was thinking of doing.
      My background is a senior structures/dynamics engineer for Pratt & Whitney. I've don't NASTRAN flutter analysis. And have worked closely with 3D aero CFD designed highly swept fan blades since there are both structural and CFD requirements of their shape. So I know a how the process works. I once had a compressor split case thermal transient analysis to see how tip clearance changed during an extreme engine transient. It ran for 17 days on P&W's (might actually have been UTRC's) fastest/biggests RS6000 without a power outrage - the real miracle of that run. So I know big non-linear models.

      With the recent "collapse" of MatLab licensing, a la Home Version for $150 instead of $20k, I thought CFD might be cheap enough by now for individual/hobbyist use. So I started surveying what is out there on the web and have also telephoned CFD code outfits - GT Power in particular to see just how costly the expensive stuff is - yes they are still arrogant with their monopolies but that will change. But more importantly, I want to convince people here that the time has come to move to CFD.

      The first CFD test case I want to run will be to demonstrate that a 28" optimized port will NOT be optimized for 10 psi of boost - or 40 or 100 psi for that matter. Or for closed (stock OE exhaust) or open headers. The "port" will be a pipe with a 90* bend probably capped with an annular "curtain area" office - a crude valve.

      (NOTE: Regarding NASCAR claims of high dP, I wonder where this is coming from since ram effects at 200 mph are only 0.7 psi and restrictor plates are like flying at high altitude or negative supercharging. Unless they are just looking at 45" dP. And why is SuperFlo offering benches capable of 65 inches? And if SF is offering it, that means it's been done probably for 10 years already. I know Erb wanted to flow test in 1970 with 10 psi of Paxton boost but where to get the 40 hp motor to drive the Paxton... Things back then were stymied by those that developed methodology to port heads accurately based on actual quarter mile runs - you make a pass and off come the heads to inspect the piston tops, CC, seat, and port areas. A flow bench will tell you nothing regarding that useful information. So much for early FB testing by novices.)


      This will occur after I accomplish what I need to get done: cam selection for my ported R3 engine. That puts me in the world of selecting an engine simulation package.

      If I don't find one or have inconclusive results, based on LS7 and LS9 factory cams and what understanding I think I may have, I'll try a stock R-Nothing cam (I have a brand new one) w/high lift adapted Mopar rockers. I have a new Var-i-cam that I keep threatening to sell, but it is a very good way to vary fixed cam timing quickly once the timing cover is off - you just set the Hi and Lo limit screws to lock the cam in position. Even making a 2-piece timing cover to expose the adjustment w/o pulling crank pulley.

      When will this happen? I have a monumental amount of big home related tasks that must be accomplished before I get too old (I'm 66) to do them. I figure earliest time will be fall/winter 2020. By that time, maybe others will have picked up the CFD ball and advance the Studebaker state-of-the art. Keep in mind that to run CFD, the flow bench now becomes extremely important to check the quality of the CFD model. Once CFD is trusted, THEN THE FUN BEGINS. Even 2-5 years from now, it will be significantly cheaper and easier to develop this CFD. Sort of like rush hour, the longer you wait - sometimes you get there at the same time.


      EDIT: As I find out more about CFD, I'll post it here. Maybe there should be a CFD thread? and a Rapid Prototyping one for scanning geometries?
      Last edited by tomhoo; 11-14-2018, 11:39 AM.

      Comment


      • #78
        When I bought B-110 from Joe Granatelli in 1969 I learned he no longer had R3 heads. He told me the heads on the first R3's (299 CI) were ported stock heads. He said I could send him my R2 heads and he would make them better than a stock R3 head. I never did.
        In 1991 Corbin had to add material to the intake port floor of Ron's R3 heads to make them flow decent. The heads used on Ron Hall's Bonneville car are 57 GH castings. That engine made 405 HP naturally aspirated. We make 630HP @ 7500 RPM with 6#'s boost. If we would have modified the carburetor to move more fuel, we could have added boost. Those heads must exceed 200 CFM in order to make 405 HP from 304 CI. Evidently Joe Granatelli knew this back in 63.
        james r pepper

        Comment


        • #79
          As it turns out, I have flow-rates ALL from the same flow-bench by the same shop. Some were published in Dick Datson's rag decades ago.

          I've data for Ron Hall's R3, Corbin Walter's(Dave Bloomberg's), Jim Lange's, Lionel Stone's aluminum R3, stock R2, and an assortment of other portings. I'll not bother with all the numbers here, but suffice it to say from best flowing down they were in this order: Lange, Walters, Hall, Stone, stock.

          Comment


          • #80
            Originally posted by Xcalibur View Post
            As it turns out, I have flow-rates ALL from the same flow-bench by the same shop. Some were published in Dick Datson's rag decades ago.

            I've data for Ron Hall's R3, Corbin Walter's(Dave Bloomberg's), Jim Lange's, Lionel Stone's aluminum R3, stock R2, and an assortment of other portings. I'll not bother with all the numbers here, but suffice it to say from best flowing down they were in this order: Lange, Walters, Hall, Stone, stock.
            Do you mean stock R3 or stock "259" heads? Any chance a resort based on estimated HP?
            Last edited by tomhoo; 11-15-2018, 05:30 AM.

            Comment


            • #81
              Originally posted by nels View Post
              Bez, The number stamped is the cc of the combustion chamber and I’ve never seen an R3/4 head that didn’t have the cc’s stamped in it. As for porting, very, very few were ported. Not that none were touched but most were as cast with flashing removed. The print calls for porting, which is nice, especially at Bonneville and for legality in competition but that may have been referring to the early pre R3 heads that were essentially reworked standard heads?
              On another note, are you going to MCACN next week?
              The prints calling for porting are the R3 heads not reworked standard heads.
              David L

              Comment


              • #82
                We have used a large number modeling tools at work for many years. Some are accurate some are more difficult to setup and get good results. Some of the modeling and simulation guys in the company don't like it when I say "all models are wrong and some are useful."

                Good FEA programs when setup with proper boundry conditions, proper mesh etc. are good and produce good results. CFD programs on the other hand are difficult to use require a lot of experience and are very computationaly intensive. We use a HPC cluster with 256 processors and some simulations run for days. Two different CFD programs will produce 2 different results. The use of CFD will point you in the right direction but again it is difficult to get accurate results. When we run tests there are significant differences from the CFD runs.

                In F1 they spend a lot of time and money on CFD. They get good results by calibrating with wind tunnel results as well as taking pressure readings on cars during practice. That is the only way they can get good results with CFD.
                David L

                Comment


                • #83
                  I just found a pretty technical thread talking about testing above 28" dP. (I'm going to copy that thread in case they take it down.) One comment was about the typical dip in flow @ 34 inches. (This is typical when you optimize a system at a single point. Any step you take in any other direction results in sub-optimal.). Lots of talk about Mach Number too.

                  ALSO: I'm very disappointed that no one corrected me that it's "inches of water" not mercury (Hg) - Duh!
                  (Homer Simpson)

                  What's really frustrating is that I should have known better since I worked with water manometer pressures daily at Robertshaw Controls Co.

                  With this in mind, it's clear why people are talking about massively higher dP's for flow benching.

                  Comment


                  • #84
                    A few thoughts and comments:

                    Why CFD and Analysis: "Time" and "Money" cost too much.

                    Hitting water when porting: If you let the burr "rattle," you will hear a difference in the sound when you start to get thin. And who wants smooth surfaces anyway.

                    Simulators:
                    Engine Analyzer Pro Enterprise ($750) allows you to input compressor maps. I'm waiting on 3 questions (they're at PRI - everyone should go there once): 1) predict back pressure? 2) predict cylinder pressure vs crank angle (and what else can you predict vs crank angle), and 3) can you input turbine maps?

                    Comment


                    • #85
                      Tom wrote -
                      Hitting water when porting: If you let the burr "rattle," you will hear a difference in the sound when you start to get thin. And who wants smooth surfaces anyway.

                      Ahh...that's waiting too long. Just break down and get a sonic tester. Not that much money and can save a LOT of heartache from too thin a wall(s).

                      Predict "backpressure"... What's that ? That's one of the things that we try to ELIMINATE..! I hope you aren't one that thinks an engine NEEDS back pressure..!?

                      Mike

                      Comment


                      • #86
                        RE: Ultrasonics

                        When I did my thickness surveys, I used Krautkramer Branson (spelling) Ultrasonic Flaw Detector. If you are trying to "see" the distance to the oil gallery, that is a secondary reflection that can only be seen by looking at the actual waveform. But that is cylinder wall thickness which is easy coupling. But still, to punch the signal through "lossy" cast iron, I was using a 1/2" diameter transducer with a curved (same as bore radius) face plastic extension piece (can't remember the proper name for it.)

                        To measure port thicknesses is worse yet, valve bowl when beside the valve guide, it would require a very small transducer and again radius matching. I never inquired about what would be required (since I wasn't porting) but from all the catalog information I was going through, it wasn't clear that such a transducer even existed. For something to work in the bowl, it's almost like you would need a spherical transducer head - but that would not couple with the surface.

                        RE: Backpressure:

                        Yes, to eliminate it, you first need to measure it. So if doing engine build simulations and you have turbos or just a street exhaust system, the software must be capable of modeling the exhaust system since performance is highly dependent on it.

                        I had a 425/427 BBC with 1.7/8 long tubes w/3" collector running into 3" exhaust pipe 4' long, each splitting into a pair of 2.5" pipes feeding turbo mufflers (4) and then into 4 2.5" tail pipes and that still produced 3-5 psi back pressure measured a couple inches aft of the collector flange.
                        ______________

                        QUESTION 4) for Engine Analysis Pro/Ent: will it predict Reversion? I believe it predicts intake port pressure vs crank angle but I'm guessing at only 1 location which means you cannot see the progression of the reversion pressure wave.

                        Comment


                        • #87
                          In a running engine, the reversion wave is speed dependent. It will be different at various engine RPM's.
                          james r pepper

                          Comment


                          • #88
                            Yes, and when the speed is high enough, there isn't any (hopefully.)

                            However, IF the reversion reaches the carb, the already fueled air, now going backwards through the venturis, entrains more fuel and goes pig rich. One loser cam I had in my R3 would produce a reversion cloud 4" above the air horn of the carb w/bonnet off.

                            Comment


                            • #89
                              Originally posted by tomhoo View Post
                              Yes, and when the speed is high enough, there isn't any (hopefully.)

                              However, IF the reversion reaches the carb, the already fueled air, now going backwards through the venturis, entrains more fuel and goes pig rich. One loser cam I had in my R3 would produce a reversion cloud 4" above the air horn of the carb w/bonnet off.
                              That's were the Studebaker V8s siamesed center exhaust ports kill ultimate normally aspirated power production. With separate intake and exhaust ports, a well designed individual intake runner and exhaust header with the lengths carefully matched to the RPM range and cam design, can use that energy to pull more mixture into the cylinder instead of blowing it out the carb.

                              Of course, that's why Studebaker used superchargers. Pressurizing the intake takes all that resonance problem away. That reversion cloud couldn't be forming when it's being blown back into the cylinder by boost pressure.

                              jack vines
                              PackardV8

                              Comment

                              Working...
                              X