Register Free

Announcement

Collapse
No announcement yet.

De Havilland decoder part one

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

    #41
    One observation, though - and it's something I am going to look into. It does seem that somebody quite high up did not want Rotol to be making metal bladed propellers. It's a common factor in the Spitfire and WW story.
    www.whirlwindfighterproject.org
    It's all good. Probably.

    Comment


      #42
      Well Fedden had upset the Bristol heirarchy in getting Rotol off the ground and would be ousted by the board in the early years of the war. Yet Fedden had powerful connections in the Government and was soon found a senior position. So I would start the search with Bristol's top people and their leverage at high level.

      Comment


        #43
        Very interesting reading of your detective work, it does though raise a couple of questions in my uneducated mind:
        What was(is) the rate of change of twist & was it different between DH & Rotol? I would assume also that it was different for a truncated/telescoped blade?
        Who defined the base angle settings for the aircraft prop assy? The report showing a better speed at 2800 rpm rather than at 3000 rpm would simplistically indicate to me that the basic setting was wrong.
        ( I've read a report on the limited flight test of the Miles Hobby, where it did not achieve its expected max speed, this was put down to a leaking seal so the prop did not operated full range correctly.)
        None of this though would affect the sudden drag rise/ loss of thrust that started this .......or would it?
        Keith

        Comment


          #44
          Ha! My uneducated mind is running on the same lines. The twist is key, and it's an unknown - I am not at all sure what the variations were yet. I have the twist of the Spitfire's 55409, and a Typhoon four-blade (less), and can see the twist of the US Hamiltons varying with design speed.

          The 'book' says that telescoping a blade will change the twist, as well - but if one is basing a new basic design on an imported telescoped blade, as apparently happened with the Spitfire, I imagine one would iron that out? I just don't know yet. Truncating it shouldn't alter anything - the important thing is twist-per-inch, ie. rate of twist, of course.

          Base angles were defined by the prop manufacturer responding to the specifications given by the aircraft designer - rpm, optimum speed (ie forward speed of maximum efficiency), speed range, priority for take-off and low-speed acceleration against maximum speed. The propeller supplier would offer up the model they thought most appropriate (and design one if it didn't exist, if they thought there was a chance of mass production). That would include the base angle.

          It is hard to get the 'wrong' angle on a constant speed blade. The idea was that the blade would find the right (optimum efficiency) angle for 2,800 rpm as it would for 3,000 rpm. The 2,800 rpm would be a higher-drag configuration, ie coarser pitch. This would normally be a less efficient setting, not because of the higher drag necessarily (the lift(thrust)/drag ratio generally stays proportionate) but because the engine was rated at 3,000 for maximum output (higher revs, more horsepower up to this point).

          As the pilot you set the rpm - by moving the pitch control (which is not gradated beyond <- coarse -- fine ->) and watching the rpm counter. If you find that 2,800 gives more thrust than 3,000 - in other words, setting the 'wrong' rpm and thus pitch for maximum speed, then something else has gone wrong. You have induced a form of drag with the blade at 3,000 rpm that also removes lift (thrust), an effect which is reduced at a lower prop speed. This is compressibility wave drag.

          In all this I am not giving propeller rpm, of course - there is a reduction gear involved - but it's a constant.

          Hopefully this shows how an 'inaccurate' blade-setting mechanism isn't necessarily going to be a problem - the mechanism does what it has to do to reach a particular rpm. The only problem might occur when it can't reach one extreme or another - it either 'hits the stops' or there's a leak preventing full actuation. Then you'd get rpm fluctuations, of course. Its possible to hit the stops within the 'envelope' of the WW in the climb, caused by there being more drag - and thus necessary blade-fining - than was designed for. There were indeed rpm problems above 27,000ft. This could also be caused by the blade still not giving low enough drag to maintain climbing revs even in the minimum-drag angle of attack (about -1 degree). The 'dumb' mechanism would carry on 'fining' the blade, passing through the minimum drag angle and beyond..

          I digress. Basically, you encounter prop compressibility and its a 'world of pain', as they say.
          Last edited by Beermat; 11th October 2017, 11:49.
          www.whirlwindfighterproject.org
          It's all good. Probably.

          Comment


            #45
            I must admit that I'm having a problem fully understanding how the prop angle is balanced if its not on the stops. when I flew with my father in his Proctor (many, many years ago), he of course took off in fully fine pitch in order to get the best acceleration, but once airborne would bring both the throttle & pitch control back to about 70% to achieve cruise speed, so presumably the prop was 'floating' between fine & coarse in order to keep a constant rpm via the csu, not necessarily a constant airspeed? I only ever flew fixed pitch jobs, so could not experiment myself, but I did get my father to go to full coarse in flight when at full throttle - rpm went from 2400+ down to about 1900, but he didn't leave it there long enough to see if speed increased or decreased.
            Now when you are saying that drag increased depending on blade angle, this presumably can be affected by blade planform & centre of pressure position relative to the blade rotation (radial) axis - if thats wrong then increased or reduced drag can result......ohh it gets very complex.
            All I know about props is messing about with curves in Flight Sim 2004 to get J near a reasonable value for HP & airframe drag - presumably the fixed software caters for the rest in theory!
            Think I should stop now before I dig a big hole for myself, but thank you for your explanation, & I will go & try Wiki-ing!
            Keith

            Comment


              #46
              Wiki-ing will give mixed results at best. The best thing is to find either old prop manuals or 'Flight' or 'Aeroplane' articles. A lot of over-enthusiastic 'commentators' have produced a lot of flash-looking material on the subject which at best is half-arsed and at worst is wrong. That's the internet for you.

              Yes, the prop angle is always 'floating' if you have a constant speed unit - if it's not, and you are airborne - there's something probably amiss. Constant speed means constant engine speed.

              If you have a two-pitch arrangement, that's something else. It could even have the same prop - but no constant speed unit governing it. That's a more brutal arrangement in which the pilot decides the more appropriate 'gear' - fine or coarse. Then you are against the stops, either way.

              Blade drag - yes, but bear in mind that drag and lift are two sides of the same coin, and on most prop aerofoils in most conditions more of one just means more of the other, in a fairly constant ratio. So a coarser angle to your blade means more thrust ('forward lift' from the blades) as well as more drag. It will slow your engine down, though. A shallower angle - less thrust, less drag, and your engine speeds up. You want to get an optimum engine speed, and this happens at a particular angle, which of course varies with aircraft speed. This is what the CSU will do for you. It doesn't calculate the angle, it just tweaks it until the engine is running at the right speed.

              It sound's like the Proctor had a manual system, in which your father was 'being' the CSU. I believe this was common in Russia, but pilots found the workload too much and Western-style CSU's were introduced in 1944-ish.

              All things being equal the aircraft will settle at a speed too. At full throttle, if you set the engine speed to be that at which it delivers the most horsepower, the physics and a CSU-governed blade mean that will be the maximum speed of the aircraft at that height.

              A little bit of additional drag (without proportional additional lift) due to compressibility will have the effect of making the blade reduce angle slightly to compensate for the additional drag. There is no problem while the blades are still producing lift, but you can see the effect on the maximum level speeds around full throttle height on aircraft like the Spitfire I of changing the propellers, with different compressibility efficiency losses. Sometimes the maths works out that you can reduce the RPM, remove this effect at the height in question (full throttle height), get that lift/drag ratio back and thanks to a constant speed mechanism which is now free to coarsen the blades again increase maximum speed - as with the Rotol Spitfire.

              It all breaks down when you get greatly increased drag but DECREASED lift from the blade aerofoil - which happens when it goes properly transonic, complete with transverse shockwaves etc. This is 'divergence', where the lift and drag curves diverge. Then your CSU can get it wrong, though very quickly it could turn the blades inside out and it wouldn't make a difference. Unless you have a prop designed for this condition you will be bouncing along the edge of compressibility with an intermittently windmilling prop and your RPM and boost needles dancing away merrily. This is what was described by Whirlwind pilots above 27,000ft - and understandably they called it engine trouble.

              It does hurt my head, but it gets easier to visualise the more you do it.
              Last edited by Beermat; 11th October 2017, 15:38.
              www.whirlwindfighterproject.org
              It's all good. Probably.

              Comment


                #47
                I guess this is fairly obvious but I'll post it anyway. The twist on a blade is optimised for a specific design speed, so the helical advance for each part of the blade will be the same, as shown by the convergence in the attached image (sorry about the quality, I had to photograph the page as it can't be scanned without breaking the spine of the book). If you rotate the blade on its axis, as in any VP prop, the lines will no longer converge and the prop is going to operate at a lower efficiency. Presumably the fiddling around with the various propellers tested on the Spitfire included playing with this optimal speed/twist relationship.
                Click image for larger version

Name:	Img_1507.jpg
Views:	1
Size:	451.1 KB
ID:	3676979

                Comment


                  #48
                  That's a great diagram!

                  I guess any blade that you intend to rotate has to be a compromise. I was wondering how they 'got away' with twisting the outer portion of a Hamlton / DH blade by telescoping it, de-focussing it in the process. I guess if you no longer have the geometric convergence assured anyway it matters less.

                  Having said that, I would presume an aircraft intended for speed would have the twist optimised for coarse pitch? Or, thinking about it again, for that part of the coarser end of the range at which the engine is at max RPM around max V?
                  Last edited by Beermat; 11th October 2017, 16:02.
                  www.whirlwindfighterproject.org
                  It's all good. Probably.

                  Comment


                    #49
                    You would hope so but with so much unexplained here you just never know. Its all geometric optimism anyway as the different aerofoil sections will have different l/d curves, Mcrit and so on. Ultimately evolving the 'ideal', or at least the 'minimally doggish' blade was a real case of learning on the job, all while the engine guys were pushing out more power, raising revs and driving for increased airspeed.

                    Comment


                      #50
                      Which brings me to the latest on Hamilton / DH crossover.

                      I have a theory that much of the nomenclature was based on taking the US blades at various stages of telescoping, and calling them different designs without a 'T'.

                      To explain - a basic DP54400 at 10' 9" diameter is thus a 6127

                      54300 is a 6127-6T, 54200 is a 6127-12T and a 54100 is a 6127-18T

                      This has the magical effect of increasing the geometric twist as you go down the series, and speed range.

                      You can still cut to length without affecting anything else. This explains the profile of the 10ft WW's 54409 being quite different from that of the 10' 3" 54350 (Blenheim). The 54409 isn't telescoped to begin with, and has 4.5 inches lopped off. Thus it is broader-tipped than the untruncated but telescoped Blenheim blade.

                      It looks like they did something slightly different with the 5,000 series. Here they twisted the same blade arbitrarily, though they did give a different nomenclature to a telescoped version, the Spitfire's. I think because they had to redesign the Spitfire's twist anyway (it would have been much too great) that trick wasn't possible with this series.

                      I am sure that after calculating the twists of the various 4,000 series blades (I have some partial data for the 54409, so it is possible to work backwards from there) they can be applied in the same series for the 5,000's, for which we have profiles via the 6353. The profile of a 4,000 series - 6127 - exists as data for the Bu Aer 5868, on which it was modelled.

                      In other words, it is truly possible to calculate the shape of any de Havilland blade in the 4,000, 5,000 and 55,000 range* from its number and it's US equivalent now. Damn, I should be charging for this.

                      *and also (if you're good at maths like I'm not) back to the 3,000 and 2,000 series, with knowledge of activity factors - see below - it's worth enlarging, so to speak.
                      Last edited by Beermat; 11th October 2017, 17:45.
                      www.whirlwindfighterproject.org
                      It's all good. Probably.

                      Comment


                        #51
                        Click image for larger version

Name:	US - UK.png
Views:	1
Size:	1.58 MB
ID:	3676983

                        ..of course, the littlest one and the biggest one were home-grown.

                        That the twist outboard of the widest point was what mattered at the time - ie. the bit that could be changed simply by telescoping - is shown by these experiments - https://digital.library.unt.edu/ark:...9930091733.pdf - into precisely the effects of twist and pitch that Ralph highlighted.

                        This is how DH cleverly adjusted the basic 6127 design via telescoping outboard of the widest point - standard procedure for length changes - to adjust pitch distribution - at least for the 4,000 series.

                        They just took a 6105 and twisted it for the 5,000 series. The also telescoped the profile to get the 55409, but didn't call it 'T' because they didn't increase the twist in the way telescoping normally would.

                        Other changes to thickness, aerofoil or twist meant that the range went beyond 400 by the time of the Hydromatics. As such the details of the Lancaster's 455800 matched those of the 6353, a later addition to the basic 6105 family.

                        I will look into that twist, provided in a handy table in HS 130B, and see if I can't work out the full extended sequence of twists across the Hydromatics along with the Brackets.. but not tomorrow. I need to do some work, and say hello to my family!
                        Last edited by Beermat; 11th October 2017, 22:00.
                        www.whirlwindfighterproject.org
                        It's all good. Probably.

                        Comment


                          #52
                          Mods, can you change the thread name, as someone suggested, to something like "De Havilland blades matched with Hamilton Standard equivalents and numbers explained"? It might prove something of a resource in the future.
                          www.whirlwindfighterproject.org
                          It's all good. Probably.

                          Comment


                            #53
                            Going through the Hamilton Standard data in the copy of HS130 that a kind forumite sent, it appears that Hamilton didn't telescope the twist with the rest of the blade profile when they shortened their blades. De Havilland definitely did (Ref the manual 'clocks provided: "Telescoping a blade will affect its shape characteristics in a different manner them will cutting off the tips. [Since the angles of the original stations are not changed in telescoping, the pitch distribution will be affected in a different way.]"

                            The Americans not doing this, at least for their 6105, would mean that it was completely OK to take a 6105-18T 'off the shelf' and make it a Spitfire pattern blade. They would definitely not add a 'T' at the end as it wasn't 'T'd in the British way! The others in the series were also 6105's, but full-length, presumably twisted to match speed regime.

                            Apologies, I am just writing stuff down as I work it out.

                            Click image for larger version

Name:	Spit comp.jpg
Views:	1
Size:	133.2 KB
ID:	3677001

                            Left - 6105-18 Right - 44509 (each Spitfire blade is round-cropped by 4.5 inches)
                            Last edited by Beermat; 13th October 2017, 12:04.
                            www.whirlwindfighterproject.org
                            It's all good. Probably.

                            Comment


                              #54
                              Variations on a theme:

                              Left, WW 54409 (un-telescoped but cropped 4.5 inches per blade), Middle 54350 (I read the number on the shanks of this aeroplane's blades before they went on), Right - the only semi-visible image I have found in four years of looking of a Hamilton Standard 6127, the pattern for the other two (Ju52/3m SE-AER, Swedish Airlines, Malmo 1937).

                              Click image for larger version

Name:	4,000.png
Views:	1
Size:	1.29 MB
ID:	3677002
                              Last edited by Beermat; 13th October 2017, 13:04.
                              www.whirlwindfighterproject.org
                              It's all good. Probably.

                              Comment


                                #55
                                Admitting defeat on the twist. The Hamilton data isn't making sense to me - the 'Lancaster' 6353 doesn't have enough geometric twist to make it viable as any kind of efficient propeller - less than 10 degrees over the crucial three feet (the norm is nearer 20). What am I reading wrong? Any propellerheads out there want to help me - this is the only stumbling block in getting a complete 'programmable' model. The Spitfire has 18% twist here, near optimum for its speed. 10% isn't near optimum for anything (apart from either 25 or 1,100 mph), and makes no sense.Click image for larger version

Name:	65.png
Views:	1
Size:	267.8 KB
ID:	3677008
                                Last edited by Beermat; 13th October 2017, 17:20.
                                www.whirlwindfighterproject.org
                                It's all good. Probably.

                                Comment


                                  #56
                                  [Since the angles of the original stations are not changed in telescoping, the pitch distribution will be affected in a different way.]

                                  In the report this text is crossed out. Seems it was crossed out as an after thought having decided the statement did not apply?

                                  Comment


                                    #57
                                    Matt, a question that I'll ask here although I think the figure I refer to is actually in another thread.
                                    On you graph of the design curves for prop 54409 I'm a confused by the r/R axis. Looking at the pitch curve it should read t=0 at r/R=0.7, but it doesn't. Is the curve out of position or is the annotation on the axis wrong? As the other curves extend beyond r/R= 1.0 maybe the latter?

                                    Comment


                                      #58
                                      Ah, that twist was done in the way some US charts did it, measuring twist from an arbitrary 'zero' at r/R=0.75, rather than one at 0.7. You can just shift the axis to put zero at 0.7, the curve is right. Sorry, I should have standardised on 0.7.

                                      Word of warning, that twist is not to be taken as gospel along its whole length in that it comes from the 5868, the apparent pattern blade for the 6105, so two 'generations' removed from the 54409 (though sharing a t/c curve with that measured on one), backed up by a tiny fraction of the curve showing it's pitch around 0.7 visible in a DH document. You will understand that we didn't try measuring twist - extremely difficult to do on a blade in the first place - on a blade that had suffered an impact.

                                      That's interesting about the crossing out. Something going on there, definitely. Its like DH decided 'on the fly' not to twist with telescoping after all.

                                      Working a few examples I have realised that I was thinking in high speed terms when I assumed faster meant less twist. I may have been over-complicating things. There is no doubt that each 4,000 series is telescoped in profile to a different degree, while the 5,000 series generally isn't but the Spitfire's 55409 is. But the twist? Maybe the hundred series wasn't twist related at all - but then why the significant correlation to design speed in both the 4,000 and 5,000 series? The only other thing I can think of is 'Activity Factor'. Still in the dark, really.
                                      Last edited by Beermat; 14th October 2017, 13:15.
                                      www.whirlwindfighterproject.org
                                      It's all good. Probably.

                                      Comment


                                        #59
                                        Thanks, normalised on 0.75. Just trying to establish how 'efficient' the twist was for the WW's top speed and how the (assumed) 20deg variable pitch may have been set up.

                                        Comment


                                          #60
                                          Here you go - 28 to 48 degrees (from the pilot's notes full version, not the one on the web):

                                          Click image for larger version

Name:	pitch.png
Views:	1
Size:	348.5 KB
ID:	3677024

                                          5 degrees AOA was considered optimal at the time..
                                          Last edited by Beermat; 15th October 2017, 00:15.
                                          www.whirlwindfighterproject.org
                                          It's all good. Probably.

                                          Comment


                                           

                                          Working...
                                          X