rj newbould
Diamond
- Joined
- Aug 29, 2005
- Location
- Hernando, FL
We’ve all heard the expression “Necessity is the Mother of Invention.” and in my case it’s true.
My very first “real” invention was due to my employer’s need for a variable grinding wheel dresser to create an involute form, which would be variable.
I was working for Alan Tool in South Amboy N.J. at the time, about 1961. They had a mold to make for an involute gear. Not knowing precisely what the shrinkage would do to the form, they wanted to be able to vary it slightly according to test results.
This was long before computer generated forms. The available dressers used gearing to generate the form, and were not easily varied. We didn’t have a Diaform available, or the means to make a master for one if we did have it.
The owners of the company and the design engineer were stumped for a solution. I asked if I could have a go at it. That night I designed an involute dresser, which used no gears, and would be infinitely variable within it’s size range.
I gave my sketch to them the next day, and the engineer’s response was “That’s a world beater.” They had me make it, and it worked perfectly.
That gave me the confidence to tackle the problem which led to the Newbould Indexer.
Fast forward to 1971. I had started my shop in 1965 in ½ of a one-car garage with $200 as my only cash. By 1971 I had two Model 242 Excello EDMs and a host of other equipment in a 3,200 sq ft loft in Kenilworth N.J..
Both EDMs had built-in rotating spindles with friction locks. If you used anything other than a round electrode, aligning it necessitated using the table travel and an indicator on the electrode in order to lock it in the desired position. Once unlocked, the whole procedure needed to be repeated, often meaning a time consuming re-positioning of the table.
I didn’t like the inconvenience and waste of time doing all that, and decided to install some sort of locking system where I could just lock the rotational position where I wanted it.
My first though was a shot-pin setup, but even with a two-pin vernier setup that was too limiting. That would be OK for going back to an original position, but I often wanted to make just a very small-predetermined angular adjustment. Yes, a sine adjustment could work, but that seemed awkward to me besides not being all that accurate.
A worm gear in such a small size would not be accurate enough for the precision I sought.
An optical ring with microscope could do it, but was very expensive and would have been tricky to install. This was back in the days before precise magnetic encoders.
A serrated-tooth indexer with multiple numbers of teeth combinations (such as made by AA Gage in Michigan) had all the accuracy and mechanical characteristics I wanted, but were not direct-reading.
By this time, a week had gone by and I had decided a differential serrated-tooth mechanism was the only way to go.
Now… How to make it easy to use? That was the question. The following week was spent hour after hour, sometimes till the wee hours in the morning, punching different combinations of teeth into my Frieden rotary calculator, looking for a combination that would be easy to use. Finally, after all that time I was sitting at my desk about 10AM, with a nice warm sun on my back, looking all those notes scattered on the desk, and decided there was no combination to be had. All combinations required a mathematical calculation to determine the final angle, and none would provide a movement of exactly one second of arc.
Damn… I said to myself. I don’t want to have wasted all that time. I’ll find another way. What if I make the second set of teeth 61 minutes to give me a one-minute differential from a 60-minute set, what happens?
That gives me 354.09836 teeth leaving a small tooth at the end of the circle, now what? OK, I’ll just clear a spot on the other disk for that odd tooth. But I can’t move it, it won’t fully engage… Well, if I take out another tooth it will. Now I have two positions with exactly one minute differential.. That’s handy, but what good are two positions? Not much, so I’ll take out another tooth. Now I have 3 positions. Damn… If I keep this up, I won’t have any teeth left on the opposing set. But wait.. If I clear 60 degrees worth of teeth, I can go back to my original start and will have moved exactly one degree. Eureka… That’s it. I can make the differential whatever is easy to read as long as I subdivide the previous set’s division in equal parts.
That thought process took about 10 minutes. I was able to find a mechanical solution to a mathematically impossible problem, by not limiting the number of teeth to an integer.
A serrated form of engaging members is about 2,000 years old. The first patent for a differential set was made in Netherlands before I was born. All those which were designed to be engaged in a variety of positions, used a whole number of teeth (integers) in the circle.
There are only 3 elements, or characteristics if you will, to the Newbould Indexer. Serrated teeth, the differential principle, and the use of an non-integer number of teeth with a corresponding open space on the engaging member. Remove any one of these three elements, and it no longer works as a direct-reading indexer. That makes it basic, and that’s why it was put in the Smithsonian.
How it came to be in the Smithsonian is another story if y’all are interested.
My very first “real” invention was due to my employer’s need for a variable grinding wheel dresser to create an involute form, which would be variable.
I was working for Alan Tool in South Amboy N.J. at the time, about 1961. They had a mold to make for an involute gear. Not knowing precisely what the shrinkage would do to the form, they wanted to be able to vary it slightly according to test results.
This was long before computer generated forms. The available dressers used gearing to generate the form, and were not easily varied. We didn’t have a Diaform available, or the means to make a master for one if we did have it.
The owners of the company and the design engineer were stumped for a solution. I asked if I could have a go at it. That night I designed an involute dresser, which used no gears, and would be infinitely variable within it’s size range.
I gave my sketch to them the next day, and the engineer’s response was “That’s a world beater.” They had me make it, and it worked perfectly.
That gave me the confidence to tackle the problem which led to the Newbould Indexer.
Fast forward to 1971. I had started my shop in 1965 in ½ of a one-car garage with $200 as my only cash. By 1971 I had two Model 242 Excello EDMs and a host of other equipment in a 3,200 sq ft loft in Kenilworth N.J..
Both EDMs had built-in rotating spindles with friction locks. If you used anything other than a round electrode, aligning it necessitated using the table travel and an indicator on the electrode in order to lock it in the desired position. Once unlocked, the whole procedure needed to be repeated, often meaning a time consuming re-positioning of the table.
I didn’t like the inconvenience and waste of time doing all that, and decided to install some sort of locking system where I could just lock the rotational position where I wanted it.
My first though was a shot-pin setup, but even with a two-pin vernier setup that was too limiting. That would be OK for going back to an original position, but I often wanted to make just a very small-predetermined angular adjustment. Yes, a sine adjustment could work, but that seemed awkward to me besides not being all that accurate.
A worm gear in such a small size would not be accurate enough for the precision I sought.
An optical ring with microscope could do it, but was very expensive and would have been tricky to install. This was back in the days before precise magnetic encoders.
A serrated-tooth indexer with multiple numbers of teeth combinations (such as made by AA Gage in Michigan) had all the accuracy and mechanical characteristics I wanted, but were not direct-reading.
By this time, a week had gone by and I had decided a differential serrated-tooth mechanism was the only way to go.
Now… How to make it easy to use? That was the question. The following week was spent hour after hour, sometimes till the wee hours in the morning, punching different combinations of teeth into my Frieden rotary calculator, looking for a combination that would be easy to use. Finally, after all that time I was sitting at my desk about 10AM, with a nice warm sun on my back, looking all those notes scattered on the desk, and decided there was no combination to be had. All combinations required a mathematical calculation to determine the final angle, and none would provide a movement of exactly one second of arc.
Damn… I said to myself. I don’t want to have wasted all that time. I’ll find another way. What if I make the second set of teeth 61 minutes to give me a one-minute differential from a 60-minute set, what happens?
That gives me 354.09836 teeth leaving a small tooth at the end of the circle, now what? OK, I’ll just clear a spot on the other disk for that odd tooth. But I can’t move it, it won’t fully engage… Well, if I take out another tooth it will. Now I have two positions with exactly one minute differential.. That’s handy, but what good are two positions? Not much, so I’ll take out another tooth. Now I have 3 positions. Damn… If I keep this up, I won’t have any teeth left on the opposing set. But wait.. If I clear 60 degrees worth of teeth, I can go back to my original start and will have moved exactly one degree. Eureka… That’s it. I can make the differential whatever is easy to read as long as I subdivide the previous set’s division in equal parts.
That thought process took about 10 minutes. I was able to find a mechanical solution to a mathematically impossible problem, by not limiting the number of teeth to an integer.
A serrated form of engaging members is about 2,000 years old. The first patent for a differential set was made in Netherlands before I was born. All those which were designed to be engaged in a variety of positions, used a whole number of teeth (integers) in the circle.
There are only 3 elements, or characteristics if you will, to the Newbould Indexer. Serrated teeth, the differential principle, and the use of an non-integer number of teeth with a corresponding open space on the engaging member. Remove any one of these three elements, and it no longer works as a direct-reading indexer. That makes it basic, and that’s why it was put in the Smithsonian.
How it came to be in the Smithsonian is another story if y’all are interested.