Dual level holes - Measurement method suggestions wanted

[Ron Rompen]

I have run into a bit of a problem here due to past practices that I am hopefully going to be able to correct. However, the solution to the problem is a little elusive, and I am hoping that someone here has run into a similar situation that they have solved, and can provide some suggestions.

We currently inspect some of our parts using a dual level air gauge (dual level holes in the part) and two columns. The original thought was to check the last hole machined – if it was good, then the supposition would be that the previous holes are ALSO good. This was a great thought, however it is proving to be inaccurate.

I am now looking at having a gauge designed which will check EACH of the holes individually, at two levels. The problem lies in evaluating the ‘measurement’ of the holes. As currently designed, this would require 8 columns for a 4-hole part. This is not really feasible for a variety of reasons.

I am not concerned with the operator being able to evaluate the actual diameter of the hole(s) being checked – just to be able to determine a pass/fail condition on all the holes at the same time.

Does anyone know of a method to evaluate multiple inputs from an air gauge, and signal a pass/fail?

Thanks

 

[Rexton]

What is the size & tolerance of the hole diameters? If you're looking for a quick pass/fail you could use a GO/NOGO pin set if available.

 

[Marc]

My Thanks in advance to any of you who can shed an opinion on this. <Bump>

 

[stevegyro]

Any chance you are using the Edmunds ‘Trendsetter’ columns?

My dad actually designed the “air-to-electric” converters (lower part of the column), along with the column itself!

————————

Like the last person, having hard time conceptually with your situation.

Are the ‘two level holes’ similar to a main bore, with a counter bore near the top surface?

“4 holes at two levels” is that like ‘Z height’ on a Cmm, as far as ‘two levels’?

Thank you for expanding on this.
Steve G.



Sent from my iPad using Tapatalk

 

[Ron Rompen]

I am having a more difficult time than I had anticipated trying to explain this - been one of those weeks

The two holes are vertically above one another - diameter is 11.269 - 11.291mm (yes, that IS pretty tight, thank you very much).

Customer requirements specify that we provide actual measurement date, so a GO/NOGO setup won't work for us. The two holes need to actually be measured with some form of variable gauging.

I am considering air gauging, however this will require (along with the cost of 4 dual level spindles), 8 columns to read-out the measurements. This will have an inherent potential for error, with operators failing to read/understand that many readouts.

 

[stevegyro]

I am having a more difficult time than I had anticipated trying to explain this - been one of those weeks



The two holes are vertically above one another - diameter is 11.269 - 11.291mm (yes, that IS pretty tight, thank you very much).



Customer requirements specify that we provide actual measurement date, so a GO/NOGO setup won't work for us. The two holes need to actually be measured with some form of variable gauging.



I am considering air gauging, however this will require (along with the cost of 4 dual level spindles), 8 columns to read-out the measurements. This will have an inherent potential for error, with operators failing to read/understand that many readouts.

Columns are not bad. Can you get digital readout, along with the ‘analog ish’ display?

Some columns have RS-232 or Bluetooth output.
I see recording of all that data a bigger challenge.

Air gaging has many benefits. If you keep a high and low ‘master’ this is an “up front cost” along with yearly calibration, BUT your day to day operations will run very efficiently.
Personally, I highly recommend any ‘master’ setups where non-contact is possible.
The ‘real world’ is that gage pins wear, and even worse is the fact that a gage pin is a ‘cylinder’ in reality, and does not wear evenly.
Using ‘Deltronic’ type of tenth gage pins yield different outcomes depending upon the person using the pins. Shallow bores are specially difficult to repeat between operators.

Air Gaging:
Just “Cal” check your high and low masters, every morning, noon and end of day (for starters). Then establish a sampling plan (for the GAGE electronics & air system) in case they require more frequent checking of the gage. You can establish this frequency based on the gage stability in your shop, demonstrated over time.
My dad built electronic column gages, and at least 1/2 of the setups were for air gaging.

AIR GAGES:
What I like most is the ‘non-contact’ measurements, ‘self cleaning’ where residue and debris have nil effect on measurement, and the excellent linearity, over a small range ID, such as the tolerance of your specs for this part/ application.

I find using dial bore gages very difficult when it comes to “reliability” between operators. For a tolerance as tight as yours, very likely the only method which will perform well on a GR&R is air gaging, IMHO.

Only challenge (my recollections) is temperature drift, but as long as the units are fully warmed up, they hold nicely.

What kind of fixture do you use: fully manual, or a vertical-slide mechanism? All 4-bores in one shot, then lower to second set of 4-bores?
Or, (8) hand manipulated ID bore with the air spindle (I hope not).
Repeatability and operator to operator consistent readings on the same part are PRIORITY Number one, as you know I’m sure.

Does the specification require checking true positions for the same holes, nearby at another gage or Cmm? Just curious.

Thank you,
Steve G.





Sent from my iPad using Tapatalk

 

[ScottK]

what about a split ball probe on a digital indicator?

I find them much easier to use with better R&R than bore probes and easier to translate for operators.

I'm currently using a split ball for 100% inspection at the machine on a couple critical holes with a tolerance of +.0006"/-.001"

QC samples and checks with CMM and have yet to reject for this.

 

[stevegyro]

what about a split ball probe on a digital indicator?



I find them much easier to use with better R&R than bore probes and easier to translate for operators.



I'm currently using a split ball for 100% inspection at the machine on a couple critical holes with a tolerance of +.0006"/-.001"



QC samples and checks with CMM and have yet to reject for this.

Yes Scott, sounds like you have a practical and cost effective solution there.

Just wondering how often the ‘split ball gage’ needs to be calibrated (ie every 300 parts or 3 days...)v, and if you keep any history on ‘wear or other factors’ which will eventually cause drift or non-linear response.

Thank you!
Steve Giarratana




Sent from my iPad using Tapatalk

 

[ScottK]

Yes Scott, sounds like you have a practical and cost effective solution there.

Just wondering how often the ‘split ball gage’ needs to be calibrated (ie every 300 parts or 3 days...)v, and if you keep any history on ‘wear or other factors’ which will eventually cause drift or non-linear response.

Thank you!
Steve Giarratana




Sent from my iPad using Tapatalk

Hi Steve - here's what I have attached in photo.

Mine is down to 0.00005" so I use the master ring to check/re-set the gauge frequently. What the machine operator does is collect a couple hours of parts, then measure the holes. Each time they measure the holes they confirm the setting with the master-ring and adjust if needed before measuring.

The indicator is calibrated every 1 year.

There will be a bit of a learning curve. I'd recommend putting it on a stand to keep it vertical and remove "shaky hand" variation.

this setup cost me about $1100 without a stand. Mitutoyo Indicator with resolution to 0.00005". Dyer blind hole probe with holder and pin. And master ring.

I also have a through hole probe for the same dimension as a backup. That one looks more like a ball.

 

[stevegyro]

Ok, sounds most economical. Air gaging is order of magnitude higher (ie. Over ten grand with fixtures etc).

1. Have you performed a GR&R? If you are in central CT I could lend a hand. Use ten parts which vary from your low tolerance to high. Using identical part bores will kill the GRR. I’ll explain that later.
Then get three people to measure those ten samples three times.
This is a 3x3x10 GR&R.

If a little patience with this mechanical instrument gives you good consistency (ie. GRR less than 25 pct. of Tol), then you don’t even need a fixture. That would be great!
An added benefit is that you only have one, maybe two instruments to calibrate.

2. From what you said about the operator collecting a bunch of parts, I’m hopeful that he/she keeps them in order.
You may be better off measuring the parts in sequence, just to learn the behavior of your boring process.

3. The frequency that you measure parts can be cut down considerably, when you document the stability of your process with hard evidence.
Someone a lot smarter than me once said “100 percent inspection is only 80 percent effective”. For some reason I agree with that person.

4. The idea is when you learn how many parts you can run before you need to change your drill/end mill/reamer, then divide that quantity by five, and that quantity can be your sample rate.

5. For example: Lets say every 150 pieces you need to adjust or change tooling. Then you can sample a part every 30 pieces, and you will get 5 points to plot between tool adjustment/ changes.
If you or customer is ‘dead set’ to measure 100 percent, then you can record them and set your subgroup size to 30, and your X-bar Range chart will be very representative of your process.

Steps 4 & 5 are on the assumption/ assurance that:

A. Your people and instrument exhibit a ‘passing GR&R’.
B. Gage is stable over a greater span (parts qty. measured), than the adjust/ re-tool rate of your machining process. In example above, gage should remain stable over 300 pieces to be safe (150 x 2) from “Nyquist” theory. That indicator you have is ‘fifty millionths’, but in reality it is a ‘tenth indicator’. The last digit is called a ‘half digit’, and again, the indicator designers are using Nyquist theory. You always want your sample rate minimum of 2x the rate of frequency you are sampling.

[email protected]

Looks like you are on a good path. ;-)
I’m hoping you have MiniTAB software. It will help you ‘grande’ to set up your GR&R, and present results.

Best wishes, SteveG.




Sent from my iPad using Tapatalk