De Havilland Mosquito (Wood vs. Metal)

[cvairwerks]

To me the biggest stupidity that both the US and Aus engaged in during ww2 was the total redrawing of imported designs. Packard wasted untold manhours redrawing all the Merlin blueprints to US standard when it would have been far simpler to train the production staff to read and use the RR drawings. These were all new hire staff and had to be trained in reading the US blueprints so sticking with the UK prints would have been easier and faster and have no chance of the inevitable conversion errors that arise when converting mm to US inches etc. Likewise training their draftsman to work with the RR drawing standards would have taken a minute fraction of the time taken to redraw every blueprint.

How would you handle the problem that almost most every machine tool in the Packard and other factories were in inch configuration? Almost every machinist out there worked in inch, and had nothing but inch micrometers, dial gauges, calipers, indicators, last words and so on. It was already hard enough to get small tools. For some things it was so bad, that the government put the word out to amateur radio operators to turn in every electric meter that they could, so that they could be repurposed for the war effort.

 

[Shortround6]

Packard wasted untold manhours redrawing all the Merlin blueprints to US standard when it would have been far simpler to train the production staff to read and use the RR drawings.

Whose production staff?
Did Packard make everything in house or did they subcontract items?
At one point Allison had over 800 subcontractors, granted a number of them were suppling washers, nuts and screws. But places like Maytag (washing machines) were suppling Aluminum casting to Packard, (small stuff, not engine blocks).
Most engine companies subcontracted things like valve springs and even valves. US had several different carburetor companies, that supplied everybody. Same with magnetos/ignition parts.
A large number of drawings would have to be redone in order to give them to the subcontractors.
It makes sense for Packard to do the drawings so all the American suppliers would be on the same page.

 

[MiTasol]

How would you handle the problem that almost most every machine tool in the Packard and other factories were in inch configuration? Almost every machinist out there worked in inch, and had nothing but inch micrometers, dial gauges, calipers, indicators, last words and so on. It was already hard enough to get small tools. For some things it was so bad, that the government put the word out to amateur radio operators to turn in every electric meter that they could, so that they could be repurposed for the war effort.

Given you cannot cut a thread with a 0.0393 pitch on any inch US thread cutting lathe (because the gears are all set for normal US thread pitches) this means they had to modify their lathes to run that pitch, and any other non-inch thread pitch. In other words they had to create metric lathes for the job. Likewise all good tool companies produced metric micrometers and other measuring instruments. Reading a micrometer calibrated in mm is very easy, especially when the measurement is a rounded one rather than an unusual decimal one.

Ask yourself the question - would you prefer to checking 0.0393 or 1.00. 1/100 of a mm is 0.000393 of an inch so is actually a much finer measurement than the 0.001 normal inch measurement level of fineness and much closer to the 1/10,000th of an inch that was considered superfine.

And remember that fairly early in the war, before Pearl Harbor even, nearly all major tooling was custom built because of the large British/French/etc orders that were being produced. Many tools used in both airframe and engine manufacture did not even exist in the product designers mind, let alone in reality, when the part was designed.

To suggest that an American or Australian machinist who has only ever worked in inches is too stupid to learn how to read metric tooling is insulting to all in that profession. Certainly the British and Canadian machinists had no problems working in metric so why would a new machinist fresh out of training school who had never worked in inches find it difficult.

 

[MiTasol]

Whose production staff?
Did Packard make everything in house or did they subcontract items?
At one point Allison had over 800 subcontractors, granted a number of them were suppling washers, nuts and screws. But places like Maytag (washing machines) were suppling Aluminum casting to Packard, (small stuff, not engine blocks).
Most engine companies subcontracted things like valve springs and even valves. US had several different carburetor companies, that supplied everybody. Same with magnetos/ignition parts.
A large number of drawings would have to be redone in order to give them to the subcontractors.
It makes sense for Packard to do the drawings so all the American suppliers would be on the same page.

Yes Packard will have had many subcontractors and in some cases redoing a drawing for a subcontractor would make sense but when a valve spring for example is a round metric measurement, or even to two decimal places, and made from wire that is a round metric measurement, again even to two decimal places, then using purely metric numbers makes more sense that fartarsing around in inch to five decimal places or n/128ths. The biggest problem with Packard using a Brit drawings is not the metric measurements, where used, but the projection angle and in many cases that does need a new drawing unless the subcontractor is purely working for Packard, as many were.

Accessory manufacturers (carb/fuel/ignition/etc) only needed the dimensions of the mounting pad/face/drive and producing that in inches is definitely required

The machinist/fabricator does not overly care what measurement he is using so long as it is easy to measure and "visualize". "Visualising" 55.55 mm with a ruler in front of you is much easier than visualising 2 and 15/128ths of an inch or 2.187 of an inch, even with a ruler in front of you.

As a restorer I can promise you that working with US blueprints that have measurements in fractions - especially when they use 64ths and 128ths of an inch - is a total pain in the ****. The first thing you have to do is convert that into decimal to make it easy to measure. When, as is often the case, there are a string of measurements, all in differing various fractions, converting the lot to decimal is the only way to avoid errors. It is also not unusual to take such a drawing and find that the overall part dimension given is longer or shorter than the sum of the mixture of 1/16ths, 32nds, 64ths etc on the drawing or that when you ask an Autocad expert to convert a drawing to CAD the dimensions given on an inch blueprint are impossible to digitize accurately unless you have a dead part so you can determine where the, often large, error in the original drawing is.

Working with simple round numbers and simple decimal measurements to only two decimal places is far far easier and much less likely to produce the errors that exist in many US drawings.

 

[cvairwerks]

There were two big problems....For the machine tool manufacturers, only a tiny fraction, if any of their products were metric at that time. When you're scrambling to get machine tools out the door, you don't want to add to the mix having to convert part of your production to run metric in parallel to your existing lines. By revising the designs to inch standard, it means being able to use the tooling you have on hand and getting production up much sooner. Lead times for new machines were in area of 6 months or more. My grandfather's company changed over from making pumps to parts for anti-aircraft guns, rifles, tanks and howitzers, all in a matter of days.

The other big problem is that you would have forced machinist to spend a lot of their own money to change over to metric. Until much later in the war, when various production lines started the mass hiring, the companies did not supply the individual machinist with the small tools required for the job. One supplied their own scales, calipers, micrometers, dial indicators and so on. A journeyman machinist at that time, probably had a year's gross pay in tools in his box. Very few could afford to suddenly have to duplicate 50% of his box in one shot.

Been working on restorations too, since 1979, and in aircraft manufacturing since 1986. Rarely have I ever run into fractional drawings for machined parts. See lots of them for folded and flat parts such as tabs and brackets, but mostly with wood components.

AutoCad has made designers stupid at times. Specifying a dimension to 4 decimal places drives production costs way up. Combine that, with then specifying a tolerance table that is too tight for what is actually being built. I've seen stuff designed to 4 places, get a 3 place tolerance and then have the assembly drawing tell you to hand fit to the location.

Error avoidance starts at the design level, but continues right to the production floor. Because so many unskilled people were being brought to the production lines, many companies had to train them up first. To help minimize production losses to defects, there were often 100% QC checks against calibrated check tools, master gauges or fixtures. A fresh trainee could insert the gauge into the requisite feature, or the part in the fixture and it met spec or did not. Look at some of the wartime production films and you see the inspection work.

 

[GreenKnight121]

Yes Packard will have had many subcontractors and in some cases redoing a drawing for a subcontractor would make sense but when a valve spring for example is a round metric measurement, or even to two decimal places, and made from wire that is a round metric measurement, again even to two decimal places, then using purely metric numbers makes more sense that fartarsing around in inch to five decimal places or n/128ths. The biggest problem with Packard using a Brit drawings is not the metric measurements, where used, but the projection angle and in many cases that does need a new drawing unless the subcontractor is purely working for Packard, as many were.

Except that nowhere in the US was wire made in " round metric measurement", or valve springs, or pretty much anything!

All of it was in US Inch standards - which was close... but NOT the same.

 

[Howard Gibson]

dH used a lot of metric measurements in their engines as well, presumably because it is far more precise than inFerial measurements are. Other Brit manufacturers also used meteric so dH were not alone on that. There is only one dimension called a mm but there are multiple inches and other inferial measurements and you need to know exactly which of the many measurements with the same name they are using

For example many inferial users talk about n gauge material. Here are just a few of the gauges in use in ww2. For a good example of how confusing these can be note the lines that say

Birmingham Gauge for Sheets and Hoops. (See Birmingham Gauge).
Birmingham Gauge. Abbreviation: B.G.
Not to be confused with the Birmingham Wire Gauge.​

View attachment 700273

This manual alone has another two and a half pages of different gauges in common use in 1Sheet metal gauges are different for steel and aluminium.

Sheet metal gauges are different for steel or aluminium.

I am a mechanical designer running SolidWorks. I have found I can create library components by copying and pasting the data from the McMaster Carr website. The big problem I have with His Majesty's units of measure is that Microsoft Excel interprets fractions as dates. The very common dimension of 1/16" (1.6mm) turns out to be January_16.

This probably was not a problem for deHavilland during World War II.

 

[MiTasol]

The other big problem is that you would have forced machinist to spend a lot of their own money to change over to metric. Until much later in the war, when various production lines started the mass hiring, the companies did not supply the individual machinist with the small tools required for the job. One supplied their own scales, calipers, micrometers, dial indicators and so on. A journeyman machinist at that time, probably had a year's gross pay in tools in his box. Very few could afford to suddenly have to duplicate 50% of his box in one shot.

I was not aware of that - that totally changes the whole situation.

 

[MiTasol]

There were two big problems....For the machine tool manufacturers, only a tiny fraction, if any of their products were metric at that time. When you're scrambling to get machine tools out the door, you don't want to add to the mix having to convert part of your production to run metric in parallel to your existing lines. By revising the designs to inch standard, it means being able to use the tooling you have on hand and getting production up much sooner. Lead times for new machines were in area of 6 months or more. My grandfather's company changed over from making pumps to parts for anti-aircraft guns, rifles, tanks and howitzers, all in a matter of days.

There was a lot of reverse lend-lease and RR specific tooling would have been supplied that way if demanded as part of the contract

 

[MiTasol]

Been working on restorations too, since 1979, and in aircraft manufacturing since 1986. Rarely have I ever run into fractional drawings for machined parts. See lots of them for folded and flat parts such as tabs and brackets, but mostly with wood components.

I agree on machined parts but NAA and Bell and CW drawings for sheetmetal and cast components are predominantly in fractions. By predominantly I mean almost exclusively on ww2 machines. Usually only 16th and 32nds but 64ths and 128s do crop up occasionally. Same with Boeing and Consolidated from memory - not had to work on them for years.

Consolidated, like NAA, Bell and Curtiss, at least had a logical system to their drawings and part numbers but I am sure that Boeings numbering system was created by an raving lunatic who changed it daily/weekly/monthly on a whim while drunk and/or hungover.

 

[MiTasol]

AutoCad has made designers stupid at times. Specifying a dimension to 4 decimal places drives production costs way up. Combine that, with then specifying a tolerance table that is too tight for what is actually being built. I've seen stuff designed to 4 places, get a 3 place tolerance and then have the assembly drawing tell you to hand fit to the location.

Error avoidance starts at the design level, but continues right to the production floor. Because so many unskilled people were being brought to the production lines, many companies had to train them up first. To help minimize production losses to defects, there were often 100% QC checks against calibrated check tools, master gauges or fixtures. A fresh trainee could insert the gauge into the requisite feature, or the part in the fixture and it met spec or did not. Look at some of the wartime production films and you see the inspection work.

Autocad is not the only thing that makes design engineers stupid at times. Ego and laziness are also major factors. As an example I gave one the job of drawing up a mod that required a slightly longer push/pull rod than originally fitted to a DC-3 (we were upgrading from -92's to -94's) and he came back with this magnificent drawing which included two custom designed end fittings. When I asked why he had not specified the standard AN490 used by Douglas he eventually admitted he did not know it existed and had not known were to look.

Error avoidance continues into the operation of aircraft. I have fairly recently sent two corrections a manufacturer for their manuals for a new model when staff came and asked which of two interpretations of the same paragraph was the right one. Both times I faxed the manufacturer with photos showing both interpretations "worked". The first time we waited for the manufacturer to reply. On the second, a rod end would not have been in safety so we assembled it in safety and waited for the confirmation before flying. Both times we had answers overnight.

Waaayyy back when GAF were manufacturing the Nomad (one small step for Australia and one giant leap backward for aviation as one potential foreign customer wrote in the demonstrator logbook) they had a program where if you made a product improvement recommendation that was accepted by GAF you got a nice cash bonus. I made more one year from GAF than from my job and nothing I recommended was new - much of it was things that had been standard at other manufacturers since the early DC-3. Some never made it to production but I incorporated in some of the three I maintained. For example GAF were building Bell JetRangers under licence when they designed the Gonad and instead of fitting the same tacho generators as the B206 they fitted massive bulky Smiths units that looked like they predated the DC-3 and required making a special tool to remove one nut. They did use the same indicators so it was a simple change and carried out progressively every time a Smiths unit died. Same with the oil and torque pressure transmitters - each transmitter 75mm sq by nearly 200mm long instead of 25mm diameter and 50mm long. That required a gauge change as well so we did one aircraft and kept the other three transmitters as spares. The Bell parts were available "everywhere" whereas the Smiths units had to come from GAF who considered 14 days to ship a spare to be reasonable. The operator naturally had very different ideas.

 

[ThomasP]

I love McMaster Carr.

 

[MiTasol]

Except that nowhere in the US was wire made in " round metric measurement", or valve springs, or pretty much anything!

All of it was in US Inch standards - which was close... but NOT the same.

Things like valve springs were made to the RR requirement so either they made the wire to spec in the US or they imported the wire or the finished product from the UK using reverse lend lease

 

[Howard Gibson]

Things like valve springs were made to the RR requirement so either they made the wire to spec in the US or they imported the wire or the finished product from the UK using reverse lend lease

Did they really?

I understand that International Harvester did a poor job of manufacturing Hispano 20mm cannons. Packard designed and built engines. At some point, if Rolls Royce specifications made to sense to them, they could ask themselves what material they would use if it were their part.

"What would I call up if that were my valve spring?"

The Packard engineer probably would be right.

 

[MiTasol]

But then the spring would not be interchangeable with the RR spring as it would not fit the keeper and base ring perfectly and I pretty sure the requirement was that Packard and RR parts must be totally interchangeable.

 

[BarnOwlLover]

Did they really?

I understand that International Harvester did a poor job of manufacturing Hispano 20mm cannons. Packard designed and built engines. At some point, if Rolls Royce specifications made to sense to them, they could ask themselves what material they would use if it were their part.

"What would I call up if that were my valve spring?"

The Packard engineer probably would be right.

Not directly related to the Mosquito, but the problem with the US built Hispano cannons, at least as I understand it, was that most of them had an excessively long chamber that tended to cause light primer strikes. The French versions had the issue even at times it seemed, but for sure the early British and a lot of US built ones did, possibly due to conversion between metric to customary measurement. The Brits realized the issue and on the Mk2 versions incorporated a shortened chamber, as well as a Martin-Baker belt feed mechanism to replace the drum, and other changes that BSA did. Other than belt feed, the US didn't fix most of the problems until the M3 and M24 versions (based on the British Mk5 version).

 

[Howard Gibson]

But then the spring would not be interchangeable with the RR spring as it would not fit the keeper and base ring perfectly and I pretty sure the requirement was that Packard and RR parts must be totally interchangeable.

Is that issue so obscure that experienced eingine designers at Packard not see it?

 

[MiTasol]

Packard's excellent design staff would know that if they replace 3mm diameter spring wire with 1/8 (3.3mm) it is not going to sit correctly, will chafe on the high spots on the valve keeper and base ring and that will produce minute metal fragments which the oil system will pick up and send to the bearings which will then be destroyed.

They will also consider the possibility that the loads the keeper and base apply where in contact with the spring could well cause spring failures as the spring is operating with loads it was not designed for.

IF the redesign the keeper and base ring then the 100% interchangeability no longer exists so repair and overhaul facilities would have to carry sets of RR and sets of Packard springs and mating parts and that there would still be a strong possibility that Murphy will make someone intermix components which will lead to engine failure.

Remember that Packard staff designed a number of major modifications to the engine that RR incorporated in RR manufactured engines.

 

[ThomasP]

This is just an example.

The L.S. Starrett tool company began making micrometers and other measuring/inspection tools - calibrated in the Metric (SI) system - in the early-1900s (ie pre-WWI). These tools were made for use in the US by machinists/companies involved in manufacturing items dimensioned using the Metric (SI) system. Initially they were available by special order only. WWI changed that. Most (though not all) of the companies producing large volumes of ammunition for the French in WWI purchased large numbers of the SI based measuring tools - in particular for the inspectors/inspection departments. They trained their machine operators, machinists, and inspection people in the use of the Metric (SI) system. They found it reduced errors.

There were US companies during the interwar period that took Metric (SI) dimensioned contracts from foreign countries for material (mostly civil) that continued the practices developed during WWI. This did sometimes include things like specialized wire and sheet metal for Metric (SI) dimensioned parts.

Many of these Metric (SI) calibrated tools were still around in WWII, and were taken out of storage to be used during the pre- and early-war period for supplying material to France. Although on the part of the US there was little need for the use of the Metric system during WWII after the fall of France, there were occasions where it was more convenient to use small groups of people specifically trained in the Metric (SI) system for production of various material.

I used to own a Starrett 0-25mm micrometer manufactured in 1930, a 0-25mm thread micrometer and a 0-30mm dial indicator that were made in 1936. I also had various machinist quality metric gauge blocks and gauge pins for inspection purposes that were made (I think) pre-WWII as they came with the micrometers and had similar markings.

Although it was less common in normal US usage, there were also various production machine tools (lathes, milling machines, boring machines, jig boring/grinding machines, etc) made with metric screws (for movement aka travel) that were either made for foreign contracts or for specialized production in the US. Again, some early examples were made for the WWI ammunition contracts.

 

[MiTasol]

You raise a good point Tom. Wright built the Hispano engines under licence in WW1 and after. I would expect those to have stayed metric as I believe the original contract was to support the Hispano factory who could not keep up with demand. I have seen several Wright Hispano engines but never worked on one so cannot confirm that from personal experience.