5 ms·

The mechanical Bendix Air Data Computer, part 3: pressure transducers

60 points by picture 3y ago

kens 3y ago
Author here, if you have analog computer questions...
- rck 3y ago
  This is so cool! Do you know if anyone ever wrote a general-purpose "how to program with mechanical parts" kind of resource? I've seen technical manuals for specific machines (like the navy's Basic Fire Control Mechanisms, OP 1140), but haven't found anything more general.
  - kens 3y ago
    I haven't seen a guide like that, but it would be interesting. It amazes me that they were able to get the CADC to work, since it is a complicated 3-D task to get all the gears to mesh without colliding. Since the gears need to be in particular ratios, there are a lot of constraints. The clearances in the CADC are very tight, by the way. It was hard to reverse-engineer because many gears looked like they meshed, but they didn't quite mesh.
  - buildsjets 3y ago
    One excellent encyclopedia of mechanisms is "Mechanisms in Modern Engineering Design", Ivan I. Artobelevsky, USSR Academy of Science. Translated by Nicholas Weinstein. 7 volumes in the Internet Archive at the attached link. https://mirtitles.org/2018/10/13/mechanisms-in-modern-engineering-design-artobolevsky/ https://mirtitles.org/2018/10/13/mechanisms-in-modern-engine... Also, Cornell University hosts the Reuleaux Collection of Mechanisms and Machines. They used to have an excellent online resource with descriptions of each mechanism, animations of how they operate, and even downloadable CAD models. But then Web 3.0 happened and static websites were no longer in fashion, so it was taken offline for a refresh which was never completed. There are still some remnants available on archive.org. https://engineering.library.cornell.edu/kmoddl/ https://engineering.library.cornell.edu/kmoddl/
  - tubetime 3y ago
    you will enjoy https://507movements.com/ https://507movements.com/
  - amatic 3y ago
    There is Soroka 1954 - Analog methods in computation and simulation. Includes mechanical and electrical analogs
- agumonkey 3y ago
  What started your journey into analog computers ? I remember seeing some navy videos from the 40s explaining self tracking guns (through projected geometry parts) and was so enamoured by it. Since then I found out about hydraulic and pneumatic electromechanic devices. Also, don't you find that programmers are fighting the imperative / von neumann model to recreate what those machines were ? systems of coupled parts "reacting" to each others (looking at you web world).
  - kens 3y ago
    What started my journey? Mainly that a collector gave us this CADC to examine, and it turned out to be more interesting (and complicated) than I expected. As far as programming models, yes, the imperative model isn't a good fit here. You can think of the CADC as "purely functional" since the outputs are direct functions of the inputs.
- caf 3y ago
  Presumably the sealed housing of the pressure transducers had to be pressurised to some reference standard?
  - kens 3y ago
    No, it's the other way around. The housing is pressurized to the input pressure. The sealed bellows are presumably pressurized to some standard pressure.
bloopernova 3y ago
The kind of article that I read several times to savour it. The work and skill the original designers and manufacturers must have put into this is awe inspiring. And the work the author put in, what a fantastic job well done!
- kens 3y ago
  Thanks! Yes, it's amazing what the designers could do with the technology that was available. It's the sort of device that you would think couldn't possibly work, especially when it is experiencing the vibration and forces of a supersonic plane
chiph 3y ago
I don't know if the Bendix was considered field-repairable. The Sacramento Air Logistics Center (McClellan AFB) sticker on it as seen on CuriousMarc's video indicates no. But there should be either a manual from Bendix and/or an Air Force Technical Order for it that covers repair & servicing. Not a Bendix expert (but I was stationed at McClellan in the mid-1980s) - I would think that cleanliness and proper lubrication are going to be frequently mentioned. Also I wonder if they had a process to seal the unit up while in a atmosphere chamber. When the diaphragms expand they will be pressing against the air in the unit, so it needs to always be at a known altitude (pressure). The altitude of McClellan Field (KMCC) is 76.8 feet, so outside the tolerance range if they just sealed it up while on the bench.
- kens 3y ago
  I'm sure a repair and servicing manual existed, but I haven't found one, or even a mention of one. Even obtaining the mil spec for the CADC was a big effort. As far as servicing, the official specification is "The computer shall have a reliable life of 1,000 hours with reasonable servicing and shall be designed to have a minimum operating life of 5,000 hours with reasonable servicing and replacement of parts. The computer shall have a mean life of at least 300 hours."
  - jacquesm 3y ago
    Can you explain why the mean life is so much smaller than the reliable or operating life? Or was that meant to be 3000 hours?
    - kens 3y ago
      "Mean life" is essentially MTBF, the average time during testing until the system malfunctions, at which point the system is serviced and testing continues. The "reliable life" is the life with servicing, so it is longer.
      - jacquesm 3y ago
        Ah I see, thank you. Given the complexity of the device and the amount of mechanical bits and pieces in there these numbers are very impressive even if they might seem low compared to solid state.

60 points by picture 3y ago

kens 3y ago

Author here, if you have analog computer questions...

rck 3y ago
This is so cool! Do you know if anyone ever wrote a general-purpose "how to program with mechanical parts" kind of resource? I've seen technical manuals for specific machines (like the navy's Basic Fire Control Mechanisms, OP 1140), but haven't found anything more general.
- kens 3y ago
  I haven't seen a guide like that, but it would be interesting. It amazes me that they were able to get the CADC to work, since it is a complicated 3-D task to get all the gears to mesh without colliding. Since the gears need to be in particular ratios, there are a lot of constraints. The clearances in the CADC are very tight, by the way. It was hard to reverse-engineer because many gears looked like they meshed, but they didn't quite mesh.
- buildsjets 3y ago
  One excellent encyclopedia of mechanisms is "Mechanisms in Modern Engineering Design", Ivan I. Artobelevsky, USSR Academy of Science. Translated by Nicholas Weinstein. 7 volumes in the Internet Archive at the attached link. https://mirtitles.org/2018/10/13/mechanisms-in-modern-engineering-design-artobolevsky/ https://mirtitles.org/2018/10/13/mechanisms-in-modern-engine... Also, Cornell University hosts the Reuleaux Collection of Mechanisms and Machines. They used to have an excellent online resource with descriptions of each mechanism, animations of how they operate, and even downloadable CAD models. But then Web 3.0 happened and static websites were no longer in fashion, so it was taken offline for a refresh which was never completed. There are still some remnants available on archive.org. https://engineering.library.cornell.edu/kmoddl/ https://engineering.library.cornell.edu/kmoddl/
- tubetime 3y ago
  you will enjoy https://507movements.com/ https://507movements.com/
- amatic 3y ago
  There is Soroka 1954 - Analog methods in computation and simulation. Includes mechanical and electrical analogs
agumonkey 3y ago
What started your journey into analog computers ? I remember seeing some navy videos from the 40s explaining self tracking guns (through projected geometry parts) and was so enamoured by it. Since then I found out about hydraulic and pneumatic electromechanic devices. Also, don't you find that programmers are fighting the imperative / von neumann model to recreate what those machines were ? systems of coupled parts "reacting" to each others (looking at you web world).
- kens 3y ago
  What started my journey? Mainly that a collector gave us this CADC to examine, and it turned out to be more interesting (and complicated) than I expected. As far as programming models, yes, the imperative model isn't a good fit here. You can think of the CADC as "purely functional" since the outputs are direct functions of the inputs.
caf 3y ago
Presumably the sealed housing of the pressure transducers had to be pressurised to some reference standard?
- kens 3y ago
  No, it's the other way around. The housing is pressurized to the input pressure. The sealed bellows are presumably pressurized to some standard pressure.

bloopernova 3y ago

The kind of article that I read several times to savour it. The work and skill the original designers and manufacturers must have put into this is awe inspiring. And the work the author put in, what a fantastic job well done!

kens 3y ago
Thanks! Yes, it's amazing what the designers could do with the technology that was available. It's the sort of device that you would think couldn't possibly work, especially when it is experiencing the vibration and forces of a supersonic plane

chiph 3y ago

I don't know if the Bendix was considered field-repairable. The Sacramento Air Logistics Center (McClellan AFB) sticker on it as seen on CuriousMarc's video indicates no. But there should be either a manual from Bendix and/or an Air Force Technical Order for it that covers repair & servicing. Not a Bendix expert (but I was stationed at McClellan in the mid-1980s) - I would think that cleanliness and proper lubrication are going to be frequently mentioned. Also I wonder if they had a process to seal the unit up while in a atmosphere chamber. When the diaphragms expand they will be pressing against the air in the unit, so it needs to always be at a known altitude (pressure). The altitude of McClellan Field (KMCC) is 76.8 feet, so outside the tolerance range if they just sealed it up while on the bench.

kens 3y ago
I'm sure a repair and servicing manual existed, but I haven't found one, or even a mention of one. Even obtaining the mil spec for the CADC was a big effort. As far as servicing, the official specification is "The computer shall have a reliable life of 1,000 hours with reasonable servicing and shall be designed to have a minimum operating life of 5,000 hours with reasonable servicing and replacement of parts. The computer shall have a mean life of at least 300 hours."
- jacquesm 3y ago
  Can you explain why the mean life is so much smaller than the reliable or operating life? Or was that meant to be 3000 hours?
  - kens 3y ago
    "Mean life" is essentially MTBF, the average time during testing until the system malfunctions, at which point the system is serviced and testing continues. The "reliable life" is the life with servicing, so it is longer.
    - jacquesm 3y ago
      Ah I see, thank you. Given the complexity of the device and the amount of mechanical bits and pieces in there these numbers are very impressive even if they might seem low compared to solid state.