Hey everybody! We're back with another OSA member owned watch rebuild (I'll leave it to the OSA member who owns this watch to identify himself, if he chooses to do so). This one is a really special Bulova Type A17A Navigation watch from 1961. It was issued by the Air Force to it's OSA owner in 1966 in Vietnam, and has been unopened since that time.
This is going to be another long one, a total of 6 posts.
I suppose in following the format of the previous rebuilds that have been posted, a little history is in order first.
There were three types of Bulova watches supplied to the US military in accordance with either specification MIL-W-3818A or MIL-W-6433A.
The first under MIL-W-3818A was the general purpose watch. It had a dial with 12 hour markers, outer 60 seconds markers and lumed spade shape hour and minute hands. The second, also under MIL-W-3818A (bu with a different stock number), had a dial with both 12 hour and 24 hour markers (but no 60 second markers) and lumed spade shape hour and minute hands. The third variant was under MIL-W-6433A. This watch was designed for navigation use, and labeled the Type A17A. The Type A17A has a black dial with 12 hour markers; 24 hour inner markers; and 60 second outer markers. The Type A17A was the only one of the three variants that also had a hacking seconds feature.
As you can see from the original photo, the dial on this watch is that of the 2nd MIL-W-3818A variant (the one with no 60 second markers) but this watch was originally born with the MIL-W-6433A dial. I found the date code machined into the barrel bridge, which dates this watch to 1961. At some point between 1961, and when the watch was issued to it's OSA owner in 1966, it had been sent back for repair and had the dial replaced with one from a MIL-W-3818A variant 2. I think that's a cool bit of history!
OK, down to the watch,
The watch would run, so the first thing I did was take an initial timegrapher reading with the watch on a full wind.
It was running fast at 77 seconds per day. An amplitude of 223 degrees was surprisingly good for a watch of this age having not been serviced in 60 years. The beat error on this one was at 3.1 m/s. If you've read my previous rebuild posts, you would've seen how this would've been way out of bounds for a beat error reading. On this particular watch, it has an older style free sprung balance, without a quick adjust for the beat, as well as a Breguet Overcoil hairspring. Beat error is adjusted on these by removing the balance wheel assembly, and manually manipulating very small tools in between the hairspring coils to rotate the collet of the hairspring around the balance staff. While a 3.1 m/s makes my OCD go crazy, it isn't outside the lines of acceptable with this type of escapement in a watch. The actual amount of rotation needed to correct a 3.1 m/s beat error is so small, it really could only be seen if viewing with high magnification. It's only a degree or two of rotation. It is the tiniest of adjustments.
I manually moved the balance wheel and watched the pallet fork entry and exit stones engage the escape wheel, and there is not any lag between the entry stone releasing the escape wheel and the exit stone stopping the escape wheel that can be visually seen. When the beat error is way out of whack, you'll see the timing between these to things begin to widen as the balance wheel moves away from center. At 3.1 m/s beat error, this isn't causing any real issue in the watch, and would have been 'acceptable' when this watch was built.
A good animation of how this Swiss Lever type escapement works can be seen here (begins at 4:50 in the video). FYI – it is oriented upside down, so what you're seeing is actually the underside of the parts that face the main plate:
When the rear of the watch was taken off, I found an inner plate that fits very snugly between the movement ring and the rubber seal. This was supposed to act as a magnetic shield (not sure how effective it was), but it also functions as an additional layer of protection from dirt getting in. It was a very tight fit, but once removed I had access to the watch.
Continued on next page.
This is going to be another long one, a total of 6 posts.
I suppose in following the format of the previous rebuilds that have been posted, a little history is in order first.
There were three types of Bulova watches supplied to the US military in accordance with either specification MIL-W-3818A or MIL-W-6433A.
The first under MIL-W-3818A was the general purpose watch. It had a dial with 12 hour markers, outer 60 seconds markers and lumed spade shape hour and minute hands. The second, also under MIL-W-3818A (bu with a different stock number), had a dial with both 12 hour and 24 hour markers (but no 60 second markers) and lumed spade shape hour and minute hands. The third variant was under MIL-W-6433A. This watch was designed for navigation use, and labeled the Type A17A. The Type A17A has a black dial with 12 hour markers; 24 hour inner markers; and 60 second outer markers. The Type A17A was the only one of the three variants that also had a hacking seconds feature.
As you can see from the original photo, the dial on this watch is that of the 2nd MIL-W-3818A variant (the one with no 60 second markers) but this watch was originally born with the MIL-W-6433A dial. I found the date code machined into the barrel bridge, which dates this watch to 1961. At some point between 1961, and when the watch was issued to it's OSA owner in 1966, it had been sent back for repair and had the dial replaced with one from a MIL-W-3818A variant 2. I think that's a cool bit of history!
OK, down to the watch,
The watch would run, so the first thing I did was take an initial timegrapher reading with the watch on a full wind.
It was running fast at 77 seconds per day. An amplitude of 223 degrees was surprisingly good for a watch of this age having not been serviced in 60 years. The beat error on this one was at 3.1 m/s. If you've read my previous rebuild posts, you would've seen how this would've been way out of bounds for a beat error reading. On this particular watch, it has an older style free sprung balance, without a quick adjust for the beat, as well as a Breguet Overcoil hairspring. Beat error is adjusted on these by removing the balance wheel assembly, and manually manipulating very small tools in between the hairspring coils to rotate the collet of the hairspring around the balance staff. While a 3.1 m/s makes my OCD go crazy, it isn't outside the lines of acceptable with this type of escapement in a watch. The actual amount of rotation needed to correct a 3.1 m/s beat error is so small, it really could only be seen if viewing with high magnification. It's only a degree or two of rotation. It is the tiniest of adjustments.
I manually moved the balance wheel and watched the pallet fork entry and exit stones engage the escape wheel, and there is not any lag between the entry stone releasing the escape wheel and the exit stone stopping the escape wheel that can be visually seen. When the beat error is way out of whack, you'll see the timing between these to things begin to widen as the balance wheel moves away from center. At 3.1 m/s beat error, this isn't causing any real issue in the watch, and would have been 'acceptable' when this watch was built.
A good animation of how this Swiss Lever type escapement works can be seen here (begins at 4:50 in the video). FYI – it is oriented upside down, so what you're seeing is actually the underside of the parts that face the main plate:
When the rear of the watch was taken off, I found an inner plate that fits very snugly between the movement ring and the rubber seal. This was supposed to act as a magnetic shield (not sure how effective it was), but it also functions as an additional layer of protection from dirt getting in. It was a very tight fit, but once removed I had access to the watch.
Continued on next page.
Last edited: