Control of end breakage rate is the prime requirement for getting better ring frame performance and for achieving higher spindle speed. Repeated occurrence of end breaks in a few spindles often is a source of higher end breakage rate. End breakage distribution is a useful tool to detect and correct defective and disturbed spindles. Deviation of actual distribution from theoretical becomes more marked when spinning conditions are critical (high spindle speed or over spinning). Yarn from spindles which give high repeated breaks have a finer count. An improved method of testing significance of end breakage by any action without being influenced by day to day fluctuations is suggested. Examples are given to demonstrate the superiority of this method
______________________________________________________________________________
Control of end breakage rate at ring frame is the first step for improving ring frame productivity. It not only leads to ends down loss but also restricts spindle speed. Ends down denotes those spindles where end has broken and is waiting for piecer to mend it. Ends down loss is given by
Where,
d = ends down %
e = end breakage rate (breaks/100 spdl hrs)
t = patrol time of piecer in hrs
Further patrol time of piecer also increases with end breakage rate. As a result ends down loss increases exponentially with increase in end breakage rate as shown in Fig 1.
In addition if end breakage rate goes beyond manageable levels, idle spindles will increase. Tenter (piecer) will make spindles, where breaks occur repeatedly, as idle.
Repeated occurrence of end breaks in a few spindles is often cited as reason for poor ring frame performance. Ridgy build of bobbin, as shown in fig 1, will be found on spindles where repeated end breaks occur.
Ridgy Bobbin
Fig2 : Ridgy build caused by repeated occurrence of end breaks
Roller lapping also increases with increase in end breakage rate. Since roller lapping involves higher time for mending, patrol time of the tenter increases and results in a vicious cycle. Some studies were therefore undertaken to find out the distribution followed by end breakages under different conditions and compare it with theoretical. An improved method for assessing the significance of improvements in end breakage rate by any action is also proposed.
End Breakage Distribution
End breakage occurrence being a rare occurrence, distribution of end breakage rate can be expected to follow Poisson distribution. In actual practice distribution differs from Poisson due to variety of reasons like 1.Variability in probability of breaks from spindle to spindle, 2. Disturbances and defects in spindles 3. day to day variability in mixing 4 disturbances in working of preparatory 5 Poor maintenance 6. Variability inR.H. temperature. A study of distribution of end breakage rate will provide useful clues in regard to the quality of maintenance and process control.
Spindle Speed
Closeness of fit of end breakage distribution to Poisson depends on conditions of spinning. End breakage distribution was determined on 80s combed yarn on a ring frame spun on normal and 20% higher spindle speed. Actual distribution was compared with Poisson in Fig 3 and 4 under these conditions. Actual distribution is close to Poisson at normal spindle speed (Fig 3). When spindle speed is increased not only end breaks increase but also departure from Poisson is very marked. (Fig4). Another study was conducted by ‘over spinning’ the mixing to 100s and determining end breakage distribution (Fig 5). End breakage distribution again deviates markedly from Poisson in 100s. This leads to the inference that the differences between spindles (in regard to probability of breaks becomes more pronounced when spinning conditions become critical. This arises from disturbances in settings, defects in parts and back material variations. The yarn spun on spindles with repeated occurrence of breaks is found to be consistently finer in count than that on spindles without breaks. High occurrence of spindles is because of lower yarn strength at the front nip in these spindles because of finer count. So high count variation is one of the reasons for repeated end breaks.
End breakage distributions in well maintained and poorly maintained ring frame sections on the same count and spindle speed were determined and are given in Fig 6 and 7. Distribution as per Poisson is also plotted. End breakage distribution deviates markedly from Poisson in poorly maintained section (Section 1 Fig 7) . While no spindles give more than 4 breaks in section 2, as many as 1.85% spindles give breaks more than 4 in section1. These breaks obviously come from disturbances and defects in spindles, rings and drafting system on these spindles. This shows that clues to quality of maintenance can be obtained by comparing actual end breakage distribution with ideal. Ring Data system by Uster is a useful attachment to ring frames as it gives spindle wise distribution of end breakages.
Common causes for repeated end breaks a on a few spindles are
1. Ring frame defects and disturbances
2. Preparatory deficiencies
Ring Frame defects
1. Disturbed spindle centering is one of the major causes of repeated occurrences of end breaks. This arises because 1.spindle centering schedule is not followed strictly 2. Proper gauges and lighting is not available. Painting top of the gauge white and use of a portable light help to improve accuracy. Electronic spindle gauge can help to reduce subjectivity but requires training.
3. Vibrating spindles and bobbins
4. Worn out rings, spindles and lappets
5. Defective cradle retention spring. Cradle stays in a lifted condition resulting in poor control over fibres.
6. Low top roller pressure, This can arise from worn out hose pipe or plunger or disturbed height setting.
7. Missing bottom apron. Sufficient number of spare aprons should be kept in each staff to facilitate prompt replacement of broken apron.
Defects and disturbances in preparatory
Long thin places in roving due to sliver splittiing in the creel or partial lapping on roller at speed frame and draw frame. Disturbed working in preparatory like roller lapping or frequent breaks.
Proposed Method to estimate improvement
High day to day and time to time within a day variability in end breakages comes as a major impediment in drawing definite conclusions about any actions taken. Proper methodology to be followed in designing experiments to assess improvements in end breakages from any action is first discussed. An improved statistical test whish will help to detect differences to a greater accuracy without being affected by day to day variations is proposed. Examples are given from actual studies to explain this method and bring home its usefulness in interpretation of results. Though the discussion has been restricted to ring frame end breakages, the same principle will hold for breakages in other processes as well.
Experimental Design
The two sources of variability in end breakages that should be taken into account while designing experiment are 1. Day to day and shift to shift variations 2. Machine to machine variations. It is therefore imperative that the two parameters or materials to compared are allowed to run on a pair of ring frames “side by side” and simultaneous study of end breakages is taken to cover all doff positions more than once. Machine difference can be taken care of by interchanging the variables between the machines. An even better method for overcoming the machine effect is to carry out the experiments on more than one pair of machines.
Analysis of Results
To facilitate statistical analysis, the results are divided into units, each of one day or shorter duration. In the usual method, standard deviation and standard error are estimated from the unit test results, from which standard error of difference is calculated. This method has the drawback that this overestimates the variability in difference of end breakages because of day to day to variations. Day to day fluctuations not only increase variability of end breakages but also causes the breaks in the two experimental set ups ( normal and modified) to move up and down in unison. The difference in end breakages on the other hand is not affected to the same extent by day to day variations.
A better method under such conditions would be to calculate the difference in end breakage rate for each unit test and estimate the standard deviation and standard error of the same and check the average difference against this to find the significance. Even if statistical test is not done, such a method will show from visual examination the likelihood of the difference being real. If the difference in unit tests is of the same sign in most of the tests, the difference is more likely to be a real one and not a ‘chance’ one.
The following examples will help to substantiate the merit of this method
Ring Cleaning
Ring frame performance gets affected by deposition of wax, fly, dirt and metallic substances over a period of working and ring cleaning at periodic intervals is suggested to overcome this. To assess the improvements from ring cleaning 2 ring frames were chosen. Rings on one side of each frame was cleaned while no action was taken on rings of other side. Simultaneous study of end breakages was done on cleaned and normal side of ring frames for a period of 9 days, with 3 hours study each day. The average end breakage rate for the two sides for the 9 days are given in Table 1 and Fig 8
Table 1: Reduction in end breakage rate by ring cleaning (Breaks/100spl hrs)
Day No
Ring cleaned, x
Normal(without cleaning), y
Difference
y - x
1
11.0
13.3
2.3
2
9.3
10.4
1.1
3
7.7
11.4
3.7
4
8.8
12.8
4.0
5
7.5
8.5
1.0
6
7.9
11.6
3.7
7
7.9
8.0
0.1
8
7.7
10.2
2.5
9
8.5
10.0
1.5
Average
8.48
10.69
2.21
Standard deviation(SD), Standard error(SE), ‘t’ value and confidence level of significance are given in Table 2
Table 2
Test of significance
Normal Method Improved method
Variance σx2 = 1.248
σ2y = 3.179 σ2y-x = 1.924
Standard error of difference
t = Difference/Standard error of difference 3.15 4.78
Correlation Coefficent r between x and y 0.628
In the normal method, SD of breakage rate for ‘cleaned’ and ‘normal’ sides are calculated separately from which SE of difference was estimated. In the improved method, difference in breakage rate between sides for each day is computed from which SD and SE of difference is calculated. ‘t’ value for each method is calculated and is given in last row. SE of difference of a lower order and ‘t’ value of higher order is found with improved method compared to normal method. As a result, reduction in end breakages by ring cleaning comes out to be significant at a higher level of confidence limit with the improved method. This is because day to day variations in ring frame performance affect ‘cleaned’ and ‘normal’ sides equally. As a result, end breakage rate on both sides move up and down in unison. On days when ring frame performance is poor, both sides tend to give a higher breakage rate. This will be amply clear from Fig 8.
Fig 8 Comparison of end breakage rates with ring cleaned and normal sides
It is well known that
σ(y-x)2 = σx2 + σy2 – 2 σx σy r
where σy-x = Stabdard deviation of difference y-x
σx = Standard deviation of x
σy = Standard deviation of y
r = Correlation coefficient between x and y
Since a positive correlation exists between x and y, σ(y-x)2 is lower than (σx2 + σy2). The positive correlation is because day to day fluctuations in breakage have equal influence on x and y. When values of σx, σy and r are substituted in the above equation, a value for σy-x in agreement with that by improved method is obtained.
Better Carding
In the 2nd example, carding quality was improved on selected card by increasing cylinder and lickerin speeds. The material was channelized separately and creeled on one side of 2 ring frames with other side working with normal material. End breakages were compared on the 2 sides for 10 days with 3 hrs study each day. The results are shown plotted in Fig 9.
Fig 9 Improvement in ring frame end breakages by higher cylinder lickerin speed in card
Fig 9 shows that not only end breakages are reduced by higher cylinder and lickerin speed but also that end breakages by the two set ups go up and down in unison
Standard deviation(SD), Standard error(SE) and ‘t’ value are given for the end breaks with normal and improved carding are given in Table 3.
Table 3
Test of significance
Normal Method Improved method
Variance σx2 = 3.06
σ2y = 4.18 σ2y-x = 2.16
Standard error of difference
t = Difference/Standard error of difference 3.98 7.29
Correlation coefficient r between x and y 0.710
Standard error of difference is lower and ‘t’ value higher by the improved method compared to normal method. Once again this is because of the high positive correlation between end breaks for the two carding conditions.
Make of Ring frame
Third example compares the end breakage rates in two makes of high speed ring frames fed by the same back material. Comparison of breakages rates over a period of 10 days is shown in fig 10.
Fig 10 : Comparison of end breaks in two makes of Ring frames.
Standard deviation(SD), Standard error(SE) and ‘t’ value are given for the end breaks between two groups of ring frames in Table 4.
Table 4
Test of significance
Normal Method, x Improved method, y
Variance σx2 =0.548
σ2y = 2.473 σ2y-x =1.398
Standard error of difference
t = Difference/Standard error of difference 1.71 2.51
Correlation coefficient r between x and y .697
Fig 10 shows that end breakage rate is lower on ring frame make II on all but one day. Table 4 once again demonstrates the superiority of improved method in bringing out the significance of between the 2 makes, which is of finer order The difference between the two makes is statistically significant only when improved method is used.
Measures for controlling end breakages in ring frame have been covered in earlier articles1,2
Conclusions
1. Control of end breakages is the prime requirement for keeping down efficiency losses and for achieving higher spindle speed. Repeated occurrence of end breaks in a few spindles is one of the major causes of high end breakage rate.End breakage distribution is a useful tool for detecting defective/disturbed spindles which give high breakages. By taking corrective action on such spindles, overall ring frame performance can be improved. Extent of departure of distribution from theoretical (Poisson) indicates the scope for process improvement.
2. Deviation of end breakage distribution from theoretical (Poisson) becomes more marked when spinning conditions are critical (high spindle speed or overspinning)
3. An improved method for checking significance of difference in end breakage rate brought about by any action is suggested. This method is less affected by day to day and shift to shift to shift fluctuations in breakages rates. And brings out finer order of differences. Examples are given to demonstrate the superiority of this method
References
1. N.Balasubramanian,Measures for improving yarn quality and productivity, Indian Textile J, 1995, zaMarch, p20
2. N.Balasubramanian and V.Y. Chaudhary, Improvements in Ring frame perfoformance and productivity by frequency invertor drive, Indian Textile J, 1994, Nov, p14