PDA

View Full Version : Video of making dials



Tony Ennis
09-20-2014, 09:46 AM
From time to time, members ask about making new dials. Tubalcain (AKA Lyle) has a new 4-part series. Nothing fancy, but sometimes it makes all the difference to just see it. I very much liked his stamping jig though it does beg for a slight improvement. That's in part 4.

New Crossfeed Dials for the Logan Lathe (https://www.youtube.com/watch?v=Y0TwJQ-0FBw)

dp
09-20-2014, 10:55 AM
I like Mr. Ishimura's stamping jig - http://homepage3.nifty.com/amigos/stamp/punch_mark_holder-e.htm as it avoids the problem of off-center stamping that Lyle's has.

Mister ED
09-20-2014, 12:08 PM
I like Mr. Ishimura's stamping jig - http://homepage3.nifty.com/amigos/stamp/punch_mark_holder-e.htm as it avoids the problem of off-center stamping that Lyle's has.
Not really the jigs doing (granted the jig is nice). Mr. Isimura avoids the need to stamp off center by only stamping one digit. Mr Pete only did that with the 0.

RichR
09-20-2014, 12:29 PM
I watched part four of the video and felt that the spacers may have been an unnecessary complication added to the process. A 1/4" slot
should have been adequate. The 1/8" shift of the punch using spacers could be done by lining up the line on the dial with the left or right
edge of the slot.

dp
09-20-2014, 12:55 PM
Not really the jigs doing (granted the jig is nice). Mr. Isimura avoids the need to stamp off center by only stamping one digit. Mr Pete only did that with the 0.

The offset is set by the position of the indexing gear relative to the dial which can be placed and locked at any angle. Once set the offset is consistent around the dial. The impression is made over the dial center in every case.

Tony Ennis
09-20-2014, 01:05 PM
Mr Pete's jig is 10x simpler and gives good results...

Mister ED
09-20-2014, 01:21 PM
The offset is set by the position of the indexing gear relative to the dial which can be placed and locked at any angle. Once set the offset is consistent around the dial. The impression is made over the dial center in every case.
Ah, I see it now. Looking a the completed dial, with only one digit centered over the marks ... I was having trouble understanding the need for the jig complexity. But with your description I see where it would allow you to offset numbers and still be able to strike the punch perpendicular to the circumference, giving more evenly defined numbers.

TGTool
09-20-2014, 01:56 PM
Another fly in the ointment is that the number stamps don't all have their characters well centered - left to right and up and down. Plus, the faces aren't all wonderfully square with the shank so that the impression may be unequal, even when everything is squared. As an example, in Tubal Cain's video he comments on the lower left serif of the number 1 being indistinct because of the curve of the dial. That's an issue, but the flag at the top left of the 1 character was impressed just fine.

I've complained to one of the stamp manufacturers to no avail. They suggested buying the precision characters for multi-character stamp holders. Yeah, maybe, but they're only about 1" long or so and $17 per character. My dodge was to buy a couple extra sets when they were on sale and just select for the best in each to make up one set.

PStechPaul
09-20-2014, 05:39 PM
I just watched the four video series and found it quite interesting and inspiring. I liked the three carriage stops for the different length scribes. I added the following comments to part 4:


Excellent videos as always, even if they are copyrighted a month in the future ;). I like your ingenuity and the saw blade was a great idea. I wonder if it's possible to use the threading gears and dial to get 100 divisions?

For the stamping, you may also have made the slot with the edges along the radii to strike perpendicular and at the desired spacing.

Another idea would be to add a vernier scale to the mark on the fixed plate, if you want to read tenths. Might be overkill, but perhaps interesting to consider how it might be done.

I want to make a larger threading dial for my HF 9x20 lathe, and it needs to be scribed along the radius with 16 divisions. I may be able to use a setup similar to yours, but use the cross-feed to scribe the marks.

Thanks!

MrSleepy
09-20-2014, 05:52 PM
Another idea would be to add a vernier scale to the mark on the fixed plate, if you want to read tenths. Might be overkill, but perhaps interesting to consider how it might be done.

Not too difficult

http://modelengineeringwebsite.com/Vernier_scale.html

Rob

Tony Ennis
09-20-2014, 07:38 PM
For the stamping, you may also have made the slot with the edges along the radii to strike perpendicular and at the desired spacing.


That would be my suggestion also. The spacer would ideally be a wedge, but it would not have to be.

Thomas Staubo
09-22-2014, 07:56 AM
Here's a nice documentation on the same subject:
http://neme-s.org/images/PDF_Files/Making_Graduated_Collars1.pdf

I particularly like the way he makes the lines by pulling instead of pushing.

dneufell
09-22-2014, 01:22 PM
Great video! Thanks again Mr.Pete ........:)

Forrest Addy
09-22-2014, 09:21 PM
Ol' Tubal Cain can't cover all the bases.

I'm pretty sure he was aware his stencilling jig has a few deficiencies but it worked well if not perfectly. For instance only the "0" was struck with the stamp's centerline aligned radially. All the other characters were offset to achieve straddle centers but the stamp centerline when spaced over was offset but not radial. Hence the weak impression on the outboard side of the numeral.

The dial dia was 1.750 making the radius 0.875. If the stamp centerline was offset 1/8" the face struck not square to a centered tangent plane but 8.2 degrees from the radial centerline.

The remedy was to have additional grooves in his stamping jig offset from a radiial line; one RH, the other LH. But that greatly adds to the fiddle factor and probably another episode to the series. Besides, there is the problem of explaining the geometry to the people he's targeted his series to; a few may not be quick studys in seeming abstruse geometry. More than a few will be discouraged by the apparent complexity. OTH three index line one left on centered one right scrubed in a to-sized stamp channel would allow the new dial to be rotated on-position for each stam. Aimpler and give the stamp a shot at a centered tangency.

Tubal Cain is a good teacher, refreshingly concise. He explains his points in a logical order, siimply and fully but doesn't maunder off the topic like many show and tell machine shop gurus.

TGTool
09-22-2014, 09:38 PM
I use a fixture that centers the stamps on the dial, then just align the dial position for where I want the character. It can't be mechanically indexed precisely for a couple reasons. Some characters such as 1 need to be cuddled up closer. And the stamps aren't precisely made. On the set I use I always have to remember that the 7 is offset to the left so I have to position the dial a little differently than other numbers. So it's all positioned by eye in the end to get double digit groups to look good. I usually put layout dye on the dial so I can give each position a light tap, peek at where it's showing, then adjust as necessary until I'm satisfied.

And the other thing I do wrong is to hit stamps more than once. On mine it's easy to take a quick look at the impression depth and whack it again once or twice if it needs it. The fixture makes sure to position the stamp in exactly the same place so the only time I might get an offset double hit or ghost impression is when I've failed to tighten the dial clamp after rotating it.

dp
09-23-2014, 12:19 PM
Using variable kerning for numbers is unusual in typesetting because numbers are normally expected to be columnar-aligned. Tweaking the kerning would create meanders of whitespace between columns. That is not a requirement for dials, of course. Mr. Ishimura's jig is quite simple, made from scrap bin pieces, and allows precision radial spacing (monospaced but you can do all the 1's from the same setting) and provides adjusting the lateral offset from the dial edge.

The problem of imprecision caused by cheap stamp quality cannot be cured with a jig - that can only be corrected with a wallet. My HF stamps are junk in this regard, for example. The earlier suggestion to pull rather than push the cutter for creating the index lines results in a very visually pleasing tapered line.

MrSleepy
09-23-2014, 01:35 PM
I'm glad a thread about dials and graduations came up..I have a metric/imperial dilema to solve with regard to the dials on my lathe...

It was sold in imp/metric versions ,but on the imperial version you could have imp/metric dials as an option. Which is what I have.

http://i897.photobucket.com/albums/ac180/MrSleepy123/okdial1_zps0cf97c03.jpg


The main crosslide leadscrew is imperial @ 200thou per revolution..the metric dial being driven by a compound gear train at 5mm/rev.
At some point the rear metric dial had seized due to corrosion and coolant ingress (but not rust)...and the previous owners had solved the problem by removing and binning the pinion gear..rather than a (very easy) full dismantle and re grease.(lazy futtin bast).

So I'm also making a 1/2ins dia 24T 48dp 14.5pa pinion aswell.

http://i897.photobucket.com/albums/ac180/MrSleepy123/okdial3_zps1b27cfd6.jpg

The compound is the same.. missing pinion.

On this one though I'm going to make a new (3d..3mm) metric LH leadscrew , to fit the original dial as it moves the wrong way for the dials imho.(0-2-1-0 as it advances to the chuck).

Although it may be easier to buy some RH 3d trapezoidal shafting and remake the dial... so that as the tool advances to the chuck..eg the dial increases 0-1-2-3/0 etc.

Forrest Addy
09-23-2014, 06:21 PM
I was thinking. A vernier using the Tubal Cain idea would be very doable if it wasn''t for the fact that its hard to find a saw blade with 90.91 teeth. I worked the math and diddled with the CAD but 90 teeth (they make 'em) won't work. It has to be 90.91

However, if a clever dude could make an accurate pitch determination of the saw blade he's using as an index plate at the pawl pin radius, and determine the pitch radius for the pin diameter... From that he could calculate the chord of adjacent tooth spaces. Then he could make a short length of straight ratchet teeth on the mill where ten teeth are spaced 1.1 times the calculated chord length. Build a hub with an arm to place the short ratchet at the proper radius. Tweak the ratchet in and out to fine tune the lenght of the vernier.

Yes there will be a pitch error between your straight vernier's linear increment and that arc increments you wish to graduate. (They spread out in the middle) My best solution is to make the ratchet thin enough to bend at the right radius. This assumes a meteor struck your dividing head or you don't have one.

That is, if making a vernier cross slide dial for your lathe poses an irresistable challenge

Clear as mud? Maybe some smart guy will figure it out and run with it.

MrSleepy
09-23-2014, 06:39 PM
I was thinking. A vernier using the ...if making a vernier cross slide dial for your lathe poses an irresistable challenge .


Forrest...did you read the Graham Meek link I provided earlier.


http://modelengineeringwebsite.com/Vernier_scale.html

It is a metric implementation of the G.H. Thomas imperial vernier as applied to his drill press.His drill press construction book and the dials and graduation book have been combined into one...

http://www.amazon.co.uk/gp/product/1857611063/


Rob

Forrest Addy
09-23-2014, 07:20 PM
Forrest...did you read the Graham Meek link I provided earlier.


http://modelengineeringwebsite.com/Vernier_scale.html

It is a metric implementation of the G.H. Thomas imperial vernier as applied to his drill press.His drill press construction book and the dials and graduation book have been combined into one...

http://www.amazon.co.uk/gp/product/1857611063/

Rob

Nope. It's more fun to noodle these things out free hand.

PStechPaul
09-23-2014, 07:27 PM
I'll throw out some ideas, and if they don't work, they can be thrown out in the other sense.

1. If you make a second scale with ten tick marks, and then turn it to an angle with a cosine of 0.9, the horizontal coordinates of the tick marks will be on the desired points for a 10:1 vernier. But this would not provide a solid mechanical angle stop for accurate scribing. This is similar to a transversal: http://en.wikipedia.org/wiki/Transversal_%28instrument_making%29

2. If you have a 100 tooth saw blade 10" diameter, and mount it 1/2" off-center, the effective diameter increases by 10%, while the pitch stays the same, and it would be equivalent to a 110 tooth blade. This should work with the same ratchet and pawl indexing tool, just moved out by the same amount. There will be a slight error, but essentially insignificant. It might be a lot easier to find a 110 tooth blade (http://www.amazon.com/Skil-74501-110-Tooth-Steel-Circular/dp/B000WIKS3Y) than one with 90.9 teeth (impossible) or 91 teeth (close enough), and having the vernier scale with finer graduations provides a direct vernier rather than a retrograde vernier which would be made with the wider divisions of a 91 tooth blade. See: http://en.wikipedia.org/wiki/Vernier_scale

3. You could use a piece of elastic material and stretch it by 10%, and then mark the same 10 tick marks as are on the mating dial. When the elastic returns to its original shape, it will have the proper spacing for a vernier. It would probably need to be something like white silicone rubber marked with a fine black pen, and then coated or laminated with a clear material and mounted.

4. A similar method would be to make the markings using a CAD program where the divisions can be made very accurately, and then printed on a suitably durable material and coated or laminated. In fact, the dial itself could be made that way, and it might be even easier to read than a scribed metal surface. It might require a bit of tweaking to get it really accurate, but lathe dials in general should not be relied upon for absolute readings. Mostly you use them for relative readings for DOC by miking, touching, adjusting to zero, and then advancing a few thousandths.

Forrest Addy
09-24-2014, 09:57 PM
Yes, I know this thread is running past its sell-by date but Paul raised an excellent point that I missed entirely.

I can confirm Paul's suggestion #2: It will work and work well according to my calculations. I worked some math and did some CAD. Taking a 7 1/4" dia 100 tooth saw blade and trying it with a 1/8 dia pawl pin in the tooth space gives a pin center to blade center radius of 3.551".

This results in a radial displacement of the saw blade out towards the pawl of 0.356". The blade displaced on a 1.1 radius will index 10 postions in the previously scribed space of 11, making a near perfect vernier just like Paul said.

It sounds lame but it's damn accurate (a very small error in that the vernier scale graduations bunch at the ends and spread in the middle by milli-RPH's) and it's simple to do. YMMV with saw diameter, pawl pin size, and tooth geometry.

Tony Ennis
09-25-2014, 07:44 AM
Are you sure about that, Forest? The saw blade still turns 1 degree at a time regardless of pin size. Sure, the center of the pin may be a closer or farther from the center of the saw blade, depending on pin size, but the scribing is happening on the other end of the arbor. The diameter of the dial is not changing.

Wayne Sippola
09-25-2014, 08:38 AM
Are you sure about that, Forest? The saw blade still turns 1 degree at a time regardless of pin size. Sure, the center of the pin may be a closer or farther from the center of the saw blade, depending on pin size, but the scribing is happening on the other end of the arbor. The diameter of the dial is not changing.

It would work - and a pretty neat idea. The offset saw blade is not turning on it's center - it's turning the arbor the dial is on, but from an offset position. If you had two blades together, one offset from the other by 10%, you can see how turning one tooth space on the outer (or higher) blade won't be a full tooth space on the lower one. It is an approximation, but for short spans, it would be a pretty good one.

The only time I made a vernier, I used a dividing head - which I guess this whole dial engraving discussion assumes you don't have.

Forrest Addy
09-25-2014, 01:16 PM
Tony, you are correct: the pawl pin size has no effect on the angular steps produced by a toothed index wheel. One increment at a time. Index and click. Each click a precise angle stop reflecting the accuracy of the index plate. The pawl pin can be any practical size that provides a positive stop in the reverse direction.

However if you offset the index wheel to alter its index interval over a short arc, you have to determine the amount of offset needed to obtain it. So you have to measure the pawl pin center to index wheel center to determine the effective radius. Since the index interval changes in proportional to the effective radius, you can calculate the precise offset necessary to obtain the desired index interval.

It should be noted that as the offset and/or length of arc increases so does the indexing error.

For extra credit (bragging rights before hostile strangers), anyone care to determine the tooth by tooth indexing error in the above problem using a theoretically perfect 7 1/4" 100 tooth saw blade?

Adding some hours later, I got out my Wolfe & Phelps trade math book I bought for class in 1961 and did the math. I calculated the max indexing error as 0.010 degrees or 36 seconds. YMMV because I'm not much good at this anymore. I confuse 6 with 8 and 7 with 9, swap digits, can't see the bottons on the calculator, and the cleaners shrunk my cape: I'm no more poweful than a switch engine these days.