machining accuracy

[gbritnell]

What is an acceptable machining tolerance for model building? The short answer is it depends on what you're making. If it's something that requires a close tolerance, shaft fit, cylinder bore square or cylindricity, wrist pin fit or such then you make it fit like it should. If you're making a connecting rod and you machine the profile (outside shape and it's withing .005) what difference will it make. Now the main journal bore and the spacing between the bore and wrist pin hole that should be held to print although if the center to center distance varied by .002 who's going to know and it will never affect performance. The next variable that comes into play is the size or scale of what you're building. The smaller the model the closer you have to hit the numbers, .002 on something that's 3.00 diameter and .002 on something .300 diameter make a big difference in fit. Some builders will scrap a part if they don't hit a number right on the head. That's the level that they work to and are happy with. Most home shop tools, lathes and mills, depending on quality are capable of .0005 that is if the user knows the machine and how it cuts. Here again side cutting with an end mill will only get you so close. Try taking .0005 from the side of a block of steel in the mill.
The next variable is what kind of measuring tools are we using to establish size, digital calipers versus digital micrometers, telescoping gauges compared to bore gauges.
What is done in a high tech shop with millions of dollars of tooling isn't quite the same as what is being done is someones basement or garage.
The bottom line is strive for the best you can do and learn what dimensions are needed and what aren't.
gbritnell

 

[DanP]

Hi all
I'm not a machinist just a hobbyist. I was always doing wood work until I got involved with Live steam Gauge 1 trains. I inherited a Myford lathe with most attachments and tooling from my Uncle. That started the game. I picked up a very nice knee mill which belonged to a friend's father who passed away, it came with lots of tooling. OK not ever having any training on machining I jumped right into the deep end. First year all I accomplish was being able to make big pieces of metal small enough to fit in the garbage can. But as time when on I was able to make progress.

OK I know the subject of this thread is accuracy. Right from the start I knew I was not going to be able to meet people like Gail Graham's (a good friend) level but I would try to get as close as possible and as time passed I got better and better at my accuracy. I guess what I am saying is keep doing your best and soon you will surprise yourself, the day that one piece fit's so perfectly you will know you are on your way. The one thing I have never lost is the amount of fun I have with all of this and that's what is important to me.

I do not post on here often but I read as many of the posts as I can for the knowledge.

 

[Jhawk]

This thread was started in 2008. Has a consensus been reached yet?

 

[Ken I]

All good advice - do your best as a matter of good practice.

How good are your measuring instruments - especially bore gauging.

Do you have any slip gauges for reference.

Plan your manufacturing sequence so that if you make a bore first and it goes oversize you can correct the mating part accordingly and vice versa.

Its easier to ream a holes and adjust the shaft to suit etc. etc.

All my plans (posted on this site) are part numbered in the sequence I made them (or later wished I had) to this end.

Regards,
Ken

 

[mnicholson]

What do you consider to be an acceptable level of accuracy for model making? I have always strived for .001" as a standard where it counts, some times I make and ......

In the golden olden days of Model Engineering when the main measuring tool was a ruler marked out to 128 Th's and parts where made to fit each other, 5 or 6 thou was considered good but we now have much better tools today.

The pros. keep talking about finishing parts to a couple of tenths of a thou. which is beyond my capability so what is a good and reasonable level of accuracy to try and achieve.

 

[mnicholson]

What do you consider to be an acceptable level of accuracy for model making? I have always strived for .001" as a standard where it counts, some times I make and ......

In the golden olden days of Model Engineering when the main measuring tool was a ruler marked out to 128 Th's and parts where made to fit each other, 5 or 6 thou was considered good but we now have much better tools today.

The pros. keep talking about finishing parts to a couple of tenths of a thou. which is beyond my capability so what is a good and reasonable level of accuracy to try and achieve.

Good morning, Just had to chime in for the first time. I been a journeyman Toolmaker for +40 years, own a general machining and fab shop, and have been turning handles most of my 60 years.I have built many laminated injection molds for the military. These molds are required to be within .ooo1 or less for accumulative error purposes. If your shop is not held at a temperature of 68.8 degrees, your are not using a A+ granite surface plate, a .oooo5 indicator and Johannsen blocks as a reference. Tenths are only wishful words. .oo1 is fine for anything of most machining jobs. Thanks for listening. Mark

 

[BryanK]

I think I have an analogy from my career as a surveyor that summarizes what the group is saying ( that tolerances depend on the job being done)

Surveyors tend to be very anal when it comes to measuring or setting out reference points for other people to use.

During my career I’ve worked on jobs that have required tolerances of +/- 0.001 meter to laying out cut/fill stakes for earth moving equipment building roads. Obviously if I taken the time and care to layout the fill stakes with as when I was laying out the mounting bolts that needed to be accurate to .001m it would be a total waste of time.....now being a surveyor it could be very hard not to go overboard in making sure all was accurate especially so early in my career. But time is money...and not only my time but the time of the contractors and machinery if they could not work because they were waiting for me to do the layout, and typically to stay ahead of them meant laying out 500 to 1000 work points. On the other hand when laying out bolt patterns, in a 8hr day I might only be able to layout for construction or verify the position of the finished location of a bolt pattern that would be for a highly specialized piece of equipment............
We had a saying that summarizes my longwinded intro.
“We’re building a road, not a piano”
In practical terms, when starting a project we would be given an accuracy ratio to work with, a common ratio used for most work we did was 1:5000, for dirt work typically we would use a ratio of 1:1000, and high accuracy plant work it would be 1:20000 but could be as high as 1:100000. ( in normal people terms) 1:5000 is your tolerance would be (as an example) if your plan says 1 inch your finished piece should not measure more or less then 0.005”.
When starting a job when we put in our control points that everything was measured from at a 1:10000 accuracy to the main control points used for the whole project. Then from there we laid out our work, we made sure our work was within that 1:5000 guideline. Then we could be confident that then the end product
would all work together.
How I have incorporated this in machining is the main piece that incorporates the other pieces I will take great care to insure it is really close to plan measurement. Then depending on how close. The tolerances need to be for the finished product I decide realistically what an acceptable tolerance is to make the end product work well and then measure those to +/- of that.