Looking for lathe mandrel test bars.

I would like to buy a 2MT and 3 MT lathe mandrel test bar. Or a combined bar for those 2 sizes.

All I seem able to find are those made by ATOZ of India. Now their specification sheets look good, But I have yet to find much good stuff made in India, apart from the excellent Zither chucks,

Does anyone have either of these test bars made by Atoz and are they as good as the spec sheet claim?

I appreciate that their are other methods of alignment (Rolley's Dad's method etc), but I would prefer a direct measuring system if possible.

Regards,

Andrew.

You could try these **LINK**

Thanks Ian,

I have been told that the RDG test bars are in fact made by ATOZ. Although I don't know if this is true or not.

Andrew.

I bought one from Arc Euro, though I have no idea of who the manufacturer is.

Hello Gordon,

That is interesting as I could not find one when I looked at Arc. It was my first port of call as usual.

Thanks,

Andrew.

Hello Andrew,

Warco sell them too, but not double ended; that would be a special made to order I feel sure

Regards

Brian

Hello Brian,

Atoz make a double ended 2mt / 3mt bar. The specification is very comprehensive and looks good. Trouble is I am very suspicious of Indian made kit, having been stung before now with equally good speecifications that turned out to be wrong!

Andrew.

Hi Andrew,

This is an interesting product. Unfortunately, it falls into the 'mine is bigger than yours' bracket.

We were selling it, and we had great sales for it. However, there is and will remain a question over its use.

The original method of its use was/is to load it into the MT female hole of the headstock, and keep the parallel part free running WITHOUT being supported by any centre in the tailstock. The key idea being to measure the alignment and run-out of the spindle in the headstock (in simple terms).

Then comes the question of: how far away from the head such measurements should be taken. The general answer being between 100mm to 150mm max.average - away from the headstock, depending on the type of lathe/headstock assembly.

However, various issues came to light:

• In free running (i.e NOT supported with a centre in tailstock), there was a question of accuracy of the bar being raised. We started to find some in-consistancy. The bars came from a good maker. No relation to RDG or ATOZ. Also, I do not believe that RDGs comes from ATOZ. We were working with the late John S and another specialist tool maker who specialised in making these, to try and figure out the situation. Neither could provide an answer.
• In the free running state, as the length of the bars became longer - year on year - mine is bigger than yours - new entrants to the hobby started to measure more and more away from the headstock i.e. 150mm+, which in turn will give bigger run-outs, WRONGLY - raising the question of accuracy of the lathe. Trying to explain to new people that this is to be expected over longer distances in free un-supported state, became a mission... one which we wished to avoid.
• If the bar is supported 'between centres' one would get great results over the overall 'long' length of the bar, but you have to question what is the real purpose of the bar in the first place? Checking of the headstock? or Alignment between centres?.

If the real purpose was the first - i.e. checking of the headstock, results were uncertain, and there needed to be some point of reference which was difficult to ascertain. Although we were happy with the bars we sold in most cases, in the hobby environment, the conclusion was that no test bars currently on the market over short nominal lengths were quite correct - keeping in mind the original principal. However, they were fine depending on intended purpose of use - such as between centres.

Although we have removed the bars from sale a while back, based on principal, the maker is happy selling them successfully to customers in U.S., Germany, Israel and Australia, in good volume, without a single question being raised over accuracy.

ATOZ is a trading company. They do not buy from our maker, as they are unable to pay our maker the price for the quality they make.

Andrew, I am sending you a PM shortly.

Ketan at ARC.

Just to clarify 'between centres':

The bars we sold had a drawbar thread at the end of the morse taper. However, many such bars come with a 'centre drill point' (for want of word), at that end instead of a drawbar thread. This allows a centre to locate into the bar at the headstock, with a similar point for a centre to locate at the tailstock end. i.e. between centres. For all of this to be 'true' the 'centre drilling' at both ends needs to be accurate at time of making such bars.

As the bars which ARC used to sell had a drawbar end, one had to locate the male morse into the female hole of the headstock, and support at the tail end with a centre, which still worked for the purpose of 'between centres'.

Ketan at ARC.

Further to Ketan's helpful note: It is worth mentioning that Schlesinger paid particular attention to the design of test bars ... His were taper mounted, but unsupported at the distal end; and they were bored-out in steps [to approximate a taper] so as to reduce the tendency to 'sag under their own weight'.

MichaelG.

.

Sorry, the document is, apparently, still protected by copyright so I don't feel at liberty to share a link.

Have a look at ... seemingly they have just come into stock...

**LINK**

George.

I never thought test bars were useful except for checking runout of the spindle taper.

To check headstock alignment with the ways just chuck up a piece of bar and take a light cut off the OD. The diameter should be the same at both ends. If not, the headstock is misaligned with the ways,

To check tailstock alignment first make up a bar with centers in both ends. Machine away most of the OD to leave an unmachined collar at each end. The collar nearest the headstock needs to be long enough to acommodate a lathe dog. Mount the bar between centers, driving it with a dog. Take a light cut off the collar nearest the tailstock. Advance the carriage to the headstock end of the test bar. Without disturbing the cross slide take a light cut off the collar. If the diameters of the collars are equal the tailstock is in line with the centerline of the spindle (at least it i in the horizontal plane).

Hi All
As an alternative if you have a precision ground bar

25 – 40mm is ideal, of constant diameter say to
within ten thou and as well you have a tenth
accuracy dial indicator you can use these to check
the axis of your machine.

Just set the bar up in your chuck a 3 will do but
better a 4 jaw. Use the dial indicator to centre it. If
you have a 4 jaw chuck.

After cantering one point unless you are very lucky
the bar will not be in line with the machine axis at all
points. Also your precision ground bar may not be
perfectly straight, it does not matter! The dial
indicator is going to sort out any wobble in the bar.

We are going to take several measurements along
the bar, if it is say 300mm long maybe measuring it
ever 2cm would be a good start.

After setting up your dial indicator on the saddle and
moving it to the start point of your measurements
say near the chuck, and checking it is on centre you
are ready.

You will also need a ruler or other means to move
the carriage to the next measuring point.

Rule off rule off a piece of paper it will need 4
columns and as many rows as you have measuring
points. Mark the 4 columns; position, High, Low,
High minus Low

Now we are ready to find if there is any alignment
error between the axis of our lathe spindle and the
path travelled by a cutting tool at the diameter of our
test bar.

Before starting check again that the tip of your
indicator is on centre, and you are at the first
reading position 0.

Then rotate the spindle until the indicator reads
highest and note the reading on your paper in the
high column.
Then rotate the spindle until the indicator reads
lowest and note the reading on your paper in the
low column.

Repeat this for all the other positions.

At this point you have all the other highs and lows,
we can now calculate the error! Subtract the low
from the high and note the result in the last column.
You may get positive and negative numbers.

Repeat this for all the other measurements the
results will either be all the same (very unlikely), or
they will differ hopefully progressively, a straight line
error, you have a straight bed and saddle travel. But
the alignment to the spindle is out.

Alternatively the result will describe a curve this
indicates a straightness error in the alignment of the
saddle and bed as well as any overall error between
the alignment of the saddle and bed. On older
lathes there is often some error detected a dip in
the bed near the spindle is very common.

If you can use Excel you can do all this on a
spreadsheet and graph it.

This method does not check the alignment of the
taper in your spindle to the axis of rotation. However
this is easily done with a finger type dial indicator.

This method is a very good indicator of the
condition of any machine slide to spindle alignment.

To speed this process up fewer measuring points
can be used. Or for extreme accuracy more points.

Regards
John

Sag and manufacturing tolerances are the reason industry moved away from mechanical alignment to optical or laser alignment. I have considered the best way to do this in a hobby environment and think a rifle scope looking down the spindle bore at a target mounted on the tailstock would be a good starting point. I don't know how short a distance they can focus down to.

Martin C

Martin,

I doubt if a rifle scope would have the necessary resolution for lathe alignment.

An autocollimator is more appropriate: Very expensive as a commercial purchase; but there was an interesting DIY design featured in [two issues, I think] of MEW.

MichaelG.

Thanks everyone,

There are some very interesting facts that have come out of this thread. Ketan , as usual, has made some very cogent points, some of which I was unaware of. It is always interesting to hear what an insider in the trade has to say!

John, the description of your alignment method sounds like Rolley's dad's (Sp?) method. Like a test bar, this doesn't really tell you what happens when actually cutting metal, but using a test bar (with appropriate care), is a much simpler operation.

Whatever method one uses, the real test is how parallel can you turn in practice? Some 40 years ago, I used a test bar to align a headstock and it worked a treat. I never gave a thought to the problems that could arise. I now realise that I was very lucky in getting a good result. I simply ran a DTI down the bar and adjusted the head stock for zero run out! I didn't even check the bar run out by rotating it. There must have been someone looking out for simple minded folk.

John, I am not sure that simply turning a bar and checking for the same diameter at both ends is a wise move, unless you know that the bed isn't twisted. I must have checked maybe half a dozen amateur's lathes over the years. Only one was set up with no twist and that belonged to a toolmaker friend of mine who knew what he was about. Maybe someone with more experience than me can chime in here. I think that a combination of a twisted bed and a headstock adjusted to give the same diameter at each end will not necessarily give the same diameter in the centre? I freely admit I could be wrong here!

Thanks everyone for some really good comments,

Andrew.

Hi Andrew Hi All

Lathe Accuracy measurement.

I restored my lathe about 2011, At the time I was thinking on how to align the spindle and came up with the method I mentioned previously. As you can see below I used two dial indicators and a 25mm ground steel bar to check the spindle axis of my lathe. I had had not studied Rollie's method as I thought it related to twisted beds, a case of GMTA.

I attached the dial indicators to a temporary jig fabbed up from scrap and heavily weighted I wanted better control and no errors caused by the saddle that at that time was worn. I posted the fix here posted at MEW: Please ignore the preamble, all was and is OK. **LINK**

I also used a stretched 0.29 Metric piano wire and microscope to measure the 2.4mt between centres bed straightness. optically aligning the wire to a gap (It did not touch the gap) by eye, and reading off a micrometer
It gave repeatable accuracy to .0002", I think accuracy in Imperial! over such a long distance that sort of accuracy is better than many lasers.

If I had seen this paper recently written I would have tried it instead of the microcope
**LINK**

The CERN Accelerator is to this day mainly aligned by wire.

**LINK**

Images:
An Excel spreadsheet for one of the tests I did is below

Also a couple of images of a wire mounting point, there were 2, one at each end of the bed and the microscope stage kludged up from scrap and a 50 power student microscope, nothing hardened! just scrap It worked well. I don't have measurements it was done on the fly the v groves were just milled. It was however rock solid and sturdy. once adjusted the whole jig was clamped tight that is important.

Oh and a tenth accuracy Mitutoyo microscope head and a linear bearing formed the measuring stage that had to position the grove on center before a measurement was taken.

Also the precision counterweight used to hold the wire straight, yes it is hanging, you can barely see the wire in the photo! the gear is about 250 dia by 40mm thick. Amazing!

As it was a "one of" everything was done as a one off. Get the job done and nothing more.

The lathe bed was subsequently ground in situ, It worked well and I use the lathe to to this day, is it perfect? no, but almost. I do have a write up but too many pages to post here.

Regards
J