Ok, I give up. 7 start thread

alloy said:
I just ran a couple of parts. The narrow thread is between thread 2 and 3. Goes all the way around.

The z offset between starts is .1164 with a F.815.

I also need a 9 start gear. I'm going to try and figure the input for that one and see if the same thing happens.

I think it's great everyone here is working together and putting out ideas. I appreciate everyone's help.
Click to expand...


I just had a thought. If the math in the G code is correct then the only thing I can think of is a problem with the lathe controls. Try moving the entire process about 1/2'' to Z+ My thought is that maybe there is a glitch in the motor encoder right at the point where it starts the third thread with your current setup. Off setting everything to the right would put it in a different position.
 
JimDawson said:
I just had a thought. If the math in the G code is correct then the only thing I can think of is a problem with the lathe controls. Try moving the entire process about 1/2'' to Z+ My thought is that maybe there is a glitch in the motor encoder right at the point where it starts the third thread with your current setup. Off setting everything to the right would put it in a different position.
Click to expand...

I will give that a try. I have it out about 2" from the chuck. to work with. The math looks correct. Spacing, feed, everything.

I tried running one of the inserts I had ground up and broke it almost immediately. Was waste of $487. Going to grind up another cemented carbide insert as a spare.
 
Just tried it and same thing. Moved it 3/4" closer to the chuck. Used my new took and the cut looks great.
 
So the tool was the problem?
 
No, cut is nice and clean, but same problem.

One thin thread
 
Have you double checked all of the Z offsets between starts in the gcode for typos or mistakes? I don't know how they'd sneak in there, but it's hard to think of what else would cause the situation you’re seeing.
 
Yes I have. I downloaded a plugin for Fusion that makes XYZ moves different colors so easy to check.

Here are the numbers.

Z.805 Start
Z.9214
Z1.0379
Z1.1543
Z1.2707
Z1.3871
Z1.5036 end

Feed .815.
S200
 
Last edited:
Bizarre. Those *look* right.

I tried to measure the crest-to-crest distances in your picture and extrapolating the short thread showed that it's something like 20% too "short" - close to the previous thread (green arrows are the same length, red arrow is shorter).

But oddly the root of the next thread (orange circle) to the left looks wider than the others.

My (humorous) understanding of the CNC way to fix this is that you could just change the gcode to make that offset bigger to compensate for whatever is messing things up.

Z.805
Z.9214
Z1.0674 <-- this gap now 20% larger
Z1.1839 <-- this gap the same
Z1.3003
Z1.4167

The other thing to do would be to sharpie the thing up, make only the initial cut on each of the threads, leaving them very shallow, and see if all of those lines are equidistant. The short thread in your picture is enough shorter that you should me able to make out the difference in pitch lines.
 

Attachments

  • Untitled 2.png
    Untitled 2.png
    438 KB · Views: 8
I also looked at the numbers you provided and don't find any obvious typos.

I made a spread sheet to better visualize the data and that leads me to a recommendation.
It seems that you have chosen a FEED (.815) which, when divided by the STARTS (7) results in a PITCH (FEED/STARTS) 15 decimal places long (0.1164285714285710).
When you round that number to four decimal places (0.1164), times 7 starts, you get 0.8148, which is .0002 short of your FEED (.815).
That resulted in using Z offsets that are not exactly equal to the FEED/STARTS. You also added 0.0001 to the Z offsets between starts 2 and 3, also between starts 6 and 7. I don't know if any of those small numerical complexities and anomalies matter, however we are searching for the unknown factor, so below is my recommendation.

Instead of choosing a FEED number, calculate the FEED by choosing a PITCH (peak to peak, up to 4 decimal places) and multiply it by the STARTS.
That way, all the increments between STARTS are precisely equal and defined to 4 decimal places (no 15 decimal place numbers).
In the attached spreadsheet, I showed the numbers, in a way that I can absorb (so I hope it's helpful to others). Also, in there is everything you need to calculate the FEED and Z offset STARTs for a 7 start thread plus another for a 9 start thread. All you have to do is enter two desired values. In the yellow cells, enter PITCH and START 1 Z. The FEED and remaining Z offset values will be calculated for you (all precisely defined to 4 decimal places with no remainders).

If you don't have or uses Excel, let us know and I'm sure somebody here can get the spreadsheet into Google Docs(?). The spreadsheet is dead simple to use (insignificant learning curve compared to trouble shooting the thread problem).

Crap! The forum won't let me attach an Excel worksheet. I suspect it's because of no permission for a non-dues-paying slug like me.
Somebody (a paying member) PM me an email address and I will email the spreadsheet to you. Please post it for the benefit (I hope) of all.
 
OK, I took a look at the numbers you're showing above and I think I found the issue. My assumption is that there should be equal spacing between each of those values. But the Second and Third values are off, with the second off by a considerable amount.

Here is what I did. As you can see Starts 2 & 3 are not equal to the rest of the Starts.

0.8050
0.9214 (0.9214-0.8050=0.1164)
1.0674 (1.0674-0.9214=0.146)
1.1839 (1.1839-1.0674=0.1165)
1.3003 (1.3003-1.1839=0.1164)
1.4167 (1.4167-1.3003=0.1164)

If the 0.1164 is the correct spacing then it seems the numbers should read:

0.8050
0.9214 (0.8050+0.1164)
1.0378 (0.9214+0.1164)
1.1542 (1.0378+0.1164)
1.2706 (1.1542+0.1164)
1.3870 (1.2706+0.1164)
 
Last edited:
You must log in or register to reply here.
Back
Top