De Havilland decoder part one

Read the forum code of contact

Member for

14 years 5 months

Posts: 3,447

*****EDIT**** Please read to posts #50 and 51 for the conclusion!

This is the beginning of a table that should become, with the application of some geometrical rules, a method for plotting any de Havilland metal blade by its five digit number.

It only covers the brief flourishing of metal bracket-type props, but covers an array of late thirties and wartime designs.

It is fairly concrete around 4,000 and 5,000 series blades, but gets a bit fluid at the edges. What I need is people out there who know of definite examples of 2,000, 3,000 and 6,000 blades, their diameters and applications to let me know - blocks can be shifted around until everything fits.

The key was discovering that blade twist defines the third digit - the less twist, the faster the aircraft is intended to go forwards relative to rotational speed (which fortunately was kept fairly constant by design in 1936-1940).

I plan to extend this out 'forwards' to Hydromatic types, and 'backward' to wooden types (with 'Aircraft clocks' help), in due course.

Note that this assumes RH rotation. Direction is indicated by the 4th digit - '0' becomes a '5' with LH rotation, '1' becomes a '6' and so on.

Attachments
Original post

Member for

14 years 5 months

Posts: 3,447

As promised, here's part two, Hydromatics. Clocks, what do you have on the wooden props?

[ATTACH=CONFIG]256126[/ATTACH]

Attachments

Member for

20 years 3 months

Posts: 1,494

Hi Matt,

This is an excellent piece of work on your part. Well done for collating them all together, and spending the time to decode it all.

Member for

14 years 5 months

Posts: 3,447

Thanks David!

Your Hornet data helped. Also thanks to the others who chipped in their props - especially Ed with that drawing.

I can now plot the full shape of any wartime DH metal blade based entirely on its number. Ok, it doesn't impress girls (Ed, no one seems to have joined 'youspinmyprop.com', can't understand why) but it might prove useful to restorers in the future?

Member for

12 years 2 months

Posts: 146

Hi,
it looks really interesting.
A lot of work and knowledge behind it, for sure.
I'm impressed.
And I'd really like to know how to 'plot the full shape of a blade'.
I'm no restorer, but I like to do some 3D CAD 'exercises' when it comes to planes...
Where should I start?

Juraj

Member for

14 years 5 months

Posts: 3,447

It's all done with blade design curves, TT. You just need to know the variation of five things along the blade axis - chord, thickness/chord, angle, face alignment and edge alignment. You also need to know the aerofoil type used and the shank type.

As a starter for 10, here's a Lancaster 'needle' blade, expressed as a table in its US equivalent, and minus the last 6 inches! You can safely assume this has a Clark-Y section.

[ATTACH=CONFIG]256142[/ATTACH]

I am working on the others, but as this is where real value kicks in I might just hang fire on publishing any new tables (this one already existed) just yet!!

Attachments

Member for

16 years 2 months

Posts: 823

Matt,

When do you think they swapped over from RAF6-based to Clark Y/Joukouski type?

Member for

16 years 2 months

Posts: 823

Matt,

Looking at the pattern of numbering I noticed that in the 900-1249 group, top line, for 11'6" and 11'9" the third number is 2 while for the rest of the 'set' it is 1. Is that correct or a typo?

Member for

14 years 5 months

Posts: 3,447

Ah, I was wondering who'd spot the deliberate mistake.. no, sorry, just a typo!

As for the other thing.. I am working on that..

Member for

11 years 8 months

Posts: 1,354

Beermat, a seminal work that I trust will one day result in the book 'Props and Pints : Home Brewing and Prop Making by Matt Beermat'. It will be a hipster hit. You might need to start wearing a monocle !
Can you change the thread title to include the word 'Props' - thread might be missed by someone doing a search in 2023...
Can you also dumb it down with a table showing dH number and explanatory term. You need to take me along on the days after I have finished a bottle of red and my brain starts to bake off from too much data. If the topic was a bicycle, I will always need trainer wheels.

In return I promise to continue chasing lost of dH prop drawings to allow you to run stress tests on your opus.
Also, I understood that dH only got into props after securing the Hamilton Standard licence agreements in the late 30's, previously only sniffing the poodle behind of Ratier.
What, Sir, are you doing chasing wooden props ? Are you saying that dH evolved an internal engineering classification system for wooden props PRIOR to investing into HS?
(I have found drawings for Wapiti -Jupiter, Bulldog - Jupiter, Demon-Kestrel if that helps)
Surely you are not talking Jablo type props ? I understand that dH were Duraluminium folk from the outset. Dammit, Sir, what fiendish wood props are you talking about? Did they make timber props for Moths ? I have some drawings of these too.

For me, the abiding avenue I wish to stride confidently down one day is a translator between an identical HS hydromatic blade and a dH hydromatic blade. I have found a RAAF manual which does this for C47 and Mustang, so hopefully this dataset can help develop this aspect of the topic. I have to believe that dH, as HS licensee, must of had a systematic coding system that somehow matches HS blade numbering. Invariably, you can find HS 'station' tables, so maybe this is the 'HS' way, while dH went for numbers. They are both the same thing? I think HS used more ink but catered for dummies, which is a bigger and more reliable market. In the long term, building an understanding is, I offer, to support the making of new blades or adaption of other postwar blades to WW2 and other historic applications. Ideally your publication develops into a primer which thoughtful folk can use to guide and manage the selection of blades that allow historic aircraft to stay in the air safely and affordably.

I do wonder if it is possible to forge a master LH and master RH tractor blank, E shank, D shank, F shank, then machine this to ANY profile, within an 'envelope of profiles' that develops from your work. Ideal.
All strength to your arm, and I raise a frosty glass in salute to you, paddling your canoe into the malarial, jungle depths of this topic.

Member for

17 years 5 months

Posts: 334

Beermat
There is nothing like a challenge to produce results.

Currently working on something for P & P which will take a couple more weeks to complete (I hope its only takes that long!).

Having a quick look at the easily located info I have and comparing it to your tables, I can offer the following part numbers etc. for your analysis.

12' 9" DP4551172TA
12' 9" DP4551778TA
16' 0" DP456250A
14' 0" DP4541062A
13' 0" DH6353A
13’ 0" DP455800A
P454500
P455950
P4551150
P4551200
P4551350
P4551400

A couple of the numbers are already on your tables, I just left them in the list to show the full part number sequence. DH6353A is any interesting part number given its prefix. This data came from RAF documents. There are a couple more for me to look through first before I can give you any more data from them.

As to wooden props that P & P raised, I have yet to OCR a document which I believe covers some of these.

It's not until you put the info into a spreadsheet as you have, that you see the patterns. Keep up the good work. I will provide information as I locate and identify it.

Member for

14 years 5 months

Posts: 3,447

P&P - Had a nice pinot noir last night - In a funny way it helps with this insane pastime.

Seriously, it would be nice to set up a simple blade blank production and a programmable CNC , to produce blades to order both for authenticity and optimisation of performance. The two might even be the same thing, it seems folk knew (or at least cared) more then than they do now.

For example, there is a Spitfire I currently flying with a cut down Battle prop. Half-right, they were from the same blank (more on this later) - but the twist is far from optimal. My guess is it accelerates well, but struggles to meet the speeds 'in the book'.

I will come up with my thoughts on production, and what shared common origins and what was a fresh design, in a bit. It seems it's a lot simpler than it first appears.

HS blade numbers - they do seem to be sequential, not relating to anything about form. Still, one can work out the sequence of adaptions of basic designs from the numbers - so for example a 6105 is a cut-down 6103 (these are immediately consecutive, odd number meant right hand).

There is one match - the 'H' in DH6353A stands for 'Hamilton' - and it's also the DP455800 blade, as per Lancaster. This is a basic blade shape for the 55,000 series, a major clue! I think I have found the US equivalent for the basic shape for the 5,000 bracket series as well.. working on it.

Regarding the numbers - thanks for those! Just what's needed.

DP4551172TA and CP4551778TA both fit in as a continuations of the table, which is satisfying! The 'T' would stand for 'telescoped' - ie. blade shape changed to accommodate severe truncation.

456xxx is a whole new range to me, at 16ft. What was that for? Something 4-bladed, 6-spline, E-shank?? (oddly the 3-blade 6-spline E-shanks were numbered '455' like the 5-spline ones, relying on the difference in lengths to distinguish them).

Member for

17 years 5 months

Posts: 334

Beermat
DP456250A is listed as being fitted to the Manchester I, fitted with a Vulture II engine. I checked if it was a typo, but the assembly is listed as being 705lb in weight, where as the rest are around 500lb.
Another that will interest you is that fitted to the Sunderland II, with Pegaus XVIII, P551953A.
What is interesting is that the spline is not listed as being either a SBAC or SAE type. It is listed as having a 100mm tapered spline.

Member for

14 years 5 months

Posts: 3,447

Those are interesting. Especially the enormous prop on the Manchester. Now, it's only Wikipedia but I have no reason to doubt it - the Vulture was reduced 0.35 to 1. According to Flight in 1942 its max rated power was delivered at 3,000 rpm. This would make the prop tips move at Mach 0.79 at sea level - even while the aircraft was stationary. They would reach Mach 0.88 at 200 mph at 15,000 feet. As the 'correct' transonic profiles didn't exist then, they would suffer enormous compressibility-induced drag. Interesting nobody said the Vulture gave less power than expected on the smaller-bladed (12' 9" diameter) Hawker Tornado, or that the revs needed to be reduced for any reason. Maybe time to re-address the 'Vulture was a failure' common knowledge, as per the Peregrine?

Member for

16 years 2 months

Posts: 823

Still looking for info on Vulture gearing but I did notice that the installation drawings of the Vulture II and IV are labelled "De Havilland Constant Speed Unit Ratio 0.8182 to 1"

Member for

14 years 5 months

Posts: 3,447

Ok, so I am now confident that all de Havilland bracket-type, 5,000 series are of the basic shape equivalent to the Hamilton Standard 6105. This is contrary to an earlier post by me, for which apologies. I could show my workings but it would be tedious..

What threw me was the differing profile of the Spitfire's blades, appearing to belong to a different sub-series from the longer blade variants of the 5,000 series. This was because they were telescoped, not just truncated - as was the practice when shortening blades beyond 6 inches. De Havilland gave them a different 'full pattern' length, upsetting the numbering sequence and confusing everybody for 80 years. What they had done was extrapolate backwards to create a virtual pattern blade which is what the telescoped blade would look like if the Spitfire blade was indeed a truncated version of a design and measured back from there! If you have just followed that the first time then you deserve a Nobel prize. My head hurts.

The clue was here.. from NACA report 3G26, 1943. Note the 6353 as per Lancaster blade, for which we have a table, has the same profile (profile is indicated by Activity Factor, a mathematical descriptor of geometry - if it varies in the same way with length, it's the same profile)

[ATTACH=CONFIG]256211[/ATTACH]

So there we have it - an American pattern blade for very many types of British aircraft carrying de Havilland blades. Of course, P&P has the inner half of a Battle blade, that's good to go. The outer part.. well, at the moment there are some thick-line-drawn graphs in an RAE R&M of the Spitfire blade which one can mathematically de-telescope to 13' diameter from 10'9" (cropped from 11'6"). I've done that, it seems to fit.

And then there is the 'Lancaster' blade table, which matches the extrapolated 5,000 series despite being Hydromatic EXCEPT for a peculiar and very large 'bump' in the thickness to chord curve, which I am trying to understand.

Attachments

Member for

17 years 5 months

Posts: 334

From the DH airscrew report R83, the numbering system is described as follows:

D.H.BLADE PART NUMBERING SYSTEM
(a) Production Part Numbers
The production part numbers of de Havilland blades for counterweight bracket type hubs are given as a five or six figure number preceded by the letter P signifying production.
The first of the numbers is always 5, which indicates that the number following is a blade number. The second figure gives the blade shank size (i.e. 4 for 4000 size, 5 for 5000 size, etc); the
last two figures indicate the direction of rotation and the diameter. For basic blades these last two figures in the production number are always 00 for R.H. rotation and 60 for L.H. rotation.
The figure or figures between the first two and last two denotes the basic blade design.
Cut-off blades.
The most usual method of trimming an airscrew to reduce its diameter is to cut down the blade ends and round the tips to give a reasonable shape.
When this is done, the resulting difference in diameter in inches is added to these last two figures.
e.g. P54409 is a 4000 size blade of R.H. rotation cropped to give 9.0" less diameter than the basic P54400 blade.
Telescoped blades.
Another method of reducing airscrew diameter is to trim the blades in such a manner that the outer stations are all moved inward towards the hub without any change in the actual dimensions.
Telescoping is generally done on stations outboard of the maximum width of the blade.
When the tip is moved in, say twelve inches, then the station half way between the station of maximum width and the tip is moved in half as much, or six inches.
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.]
When the diameter of the airscrew is reduced by telescoping the ends of the blades instead of cropping them, a letter T is added at the end of the blade number.
e.g. P55264T is a 5000 size blade of L.H. rotation telescoped to give 14.0" less diameter than the basic P55250 blade.
Blades suitable for use in the D.H. hydromatic hub are denoted by a figure 4 placed between the letter P and the figure 5. Apart from this, the numbers are similar to those for the counterweight bracket type blades.
Thus P4551150 is a 5000 size hydromatic blade of L.H. rotation and basic
diameter.
(b) Experimental Part Numbers
The part numbers of experimental blades are given as either four or five figure numbers preceded by the letters SKP.
The original experimental numbers were all four figure numbers with the first number indicating the shank size, e.g. SKP5804 was a 5000 size blade, SKP4007 a 4000 size blade, etc.
When any series reached the limit, e.g. SKP4999 a figure six was added in front of the experimental number, and the series started again, e.g. SKP64000.
When this series runs out the six will be changed to a seven, and so on (these numbers being quite independent of shank size). T
hus for the five figure experimental numbers the second figure will give the shank site, e.g. SKP65612 is a 5000 size blade, SKP64243 a 4000 size, etc.
The diameter of such experimental blades is indicated by a dash number after the blade number. Thus SKP66024-12 is a 6000 size blade cut back to give 12.0" less disaster than the basic SKP66024.
Telescoping is indicated by the letter T in the usual manner. Thus SKP65612-16T is a telescoped 5000 size blade SKF65612 giving 16.0" less diameter.
(c) Wooden blades
De Havilland wooden blades bear a three symbol number; the first of the symbols is always P. The second denotes the basic blade design,
and is a number. The third denotes the shank size, and is a letter.
e.g. P3Q is the third basic wooden blade design, and uses a Q
shank.
Note that the basic wooden blade designs are numbered consecutively irrespective of shank size. Thus the second wooden blade design
happened to be on the W and bore number P2W. The shank sizes are as follows:
100 mm. diameter is P shank 110 mm. diameter is Q shank and hence up in 10 mm. increments until 190 mm. diameter is Y shank.

Member for

14 years 5 months

Posts: 3,447

Thanks

A case of 'RTFM'. But still matches.. and it doesn't mention that the 'blade design number (digit 3) is only really twist. There was only one basic blade profile per thousand series.. or two if you count the shortened and telescoped Spit blade separately (as DH half-did).

They didn't add a 'T' to the Spitfire blade is just a quirk - I expect they imported that design wholesale from the US, because of when it was, and gave it a number that described its shape as if it was part of an un-telescoped series. In other words, because it came off the 'shelf' that way (probably as a pre-telescoped 6105A-18, as per the DC-2), and hadn't been re-drawn in the DH office, no T was added. If it had been, it would have been a 55427-T, not a 55409.

Hate to contradict the manual but the fourth digit was 0 for RH and 5 for LH (ie basic blades were 00 or 50) unless a digit carried over from the 'inches removed' column. Thus 12 inches removed would make a RH '1' and a LH '6'. 24 inches would result in a '2' or a '7'.

Goes to show, even the in-house publications need to be checked against real life. Proof for the Mrs, when putting up flat-pack. They're not instructions, they're only the manufacturer's opinion :-)

Member for

14 years 5 months

Posts: 3,447

By the way, Ralph, did you notice this line in the Martlesham trials of a Spit with the thick wooden Rotols?:

"The results show that the maximum level speed is reached with the airscrew controlling at 2800 engine r.p.m. On increasing the r.p.m. to 3000 the speed was reduced, on the average by 4 m.p.h". Similar effect to the Manchester..

Member for

14 years 5 months

Posts: 3,447

I can make the Hamilton figures for the 6353 fit everything else if we start the telescoping from a point considerably inboard of the maximum chord - the 24 inch station.

Member for

16 years 2 months

Posts: 823

Yes I did read that, however it still comfortably outperfromed the Whirlwind