Building better molds: can nobake & green sand work together?
The firm's uneven capacity issue first arose when Durametal, a 135 employee firm that produces refiner plates for the pulp and paper industries, saw an increase in demand lot stainless steel and hi-chrome iron refiner plates. The production of these components required higher levels from its phenolic urethane nobake molding line. Normally manned for two shifts of operation, the increase led to overtime work every week for both nobake shifts in order to keep up with demand.
Meanwhile, a green sand automatic matchplate molding line was significantly under utilized. The matchplate molder was traditionally a medium volume operation, used specifically for ni-hard refiner plate production. Over the past 15 years, due to better performance in the field, there has been an increase in demand for stainless steel and hi-chrome iron alloy refiner plates poured in phenolic urethane nobake molds.
As parts were steadily moved to the nobake line, the reduction in demand left the matchplate molding area with open capacity. In an effort to minimize the overtime requirements on the firm's employees, several different methods to produce these traditionally nobake parts using matchplate molding techniques were evaluated and tested. For all tests, various sizes of the iron refiner plate castings (which fir in the 20 x 24-in. flask) were used.
Straight Green Sand
Several different scenarios were discussed to off-load some of the production on the nobake molding line to the under-utilized green sand molding line. It seemed that the easiest solution would be to simply transfer patterns from nobake to green sand production.
Unfortunately, several issues arose with the pattern equipment. The largest problem was that the pattern boards could not be used in the matchplate molding machine, due to the the frame size of the nobake molds being bigger then molds used on the matchplate molding line. Casting size was also a limiting factor, as some designs were too large for the 20 x 24-in. flask size.
Another issue was the presence of loose pieces in some patterns. These pieces could be easily lost in the green sand molding operation. Pattern design changes would be needed or these patterns would not be candidates for transfer.
Any changes to the molding process still needed to result in castings of equal or greater quality and scrap rate and surface finish as with the original nobake method. The current nobake practice has delivered surface finishes between N8 and N9 levels (using the Fowler Microsurf 334 casting scale). On the same scale, a typical finish for green sand castings is N11. The difference between a ranking of N10 and N11 is that the surface finish is twice as rough for the N11 ranking.
In anticipation of a surface finish problem, Durametal decided to run current green sand patterns in alloys other than hi-hard to see results without altering patterns. A 25% chrome iron and a 17-4 stainless steel were poured into "normal" green sand molds.
The resulting finish was typical of the finish seen on ni-hard castings. This was barely acceptable by engineering standards and unacceptable by expectation standards. Since the surface finish requirements were unacceptable, this option was put on hold for the trial of the other possibilities.
Core Type Insert
The second method tested used a core made of nobake sand. In this concept, a slab matchplate molding pattern was made, acting as the core print upon which the nobake core of the casting was placed. In order to minimize the future pattern work required, three standard size core prints that could each hold several different casting sizes were determined. One of these was selected to be the test size and was made in the pattern shop. At the same time, a frame was made to place on the current pattern equipment so that the drop in cores could be made on the core bench.
The difficulty in this system was the casting's gating system. The alloys being transferred required filtration, but the traditional method of filter placement did not fit this system. After several design changes, the filter placement was solved and a shell sand pouring cup was used as the sprue. The same shell sand-pouring cup is used on the nobake molding line so that no increase in cost was incurred.
An issue that arose with this method was the mixing of nobake sand and green sand. Ideally, all nobake sand would be burned and considered the new sand addition to the green sand molding system. If the cores were heavier than the typical new sand addition, sand system control issues would occur.
In an effort to help minimize the nobake sand weight issues, it was determined that only the cope or working face of the casting needed to be nobake. while surface finish is critical on the working lace of the refiner plate casting, the back is a non-working surface, and therefore could be produced via green sand. This option now provided two different opportunities--one where only the facing portion of the mold was a nobake core and one where the entire casting was a nobake core.
When the entire casting was produced within the nobake core, it resembled the firm's current nobake cast product the best. There weren't alignment problems and the surface finish was identical to the current practice. However, there was additional cost due to the core bench labor and operations and additional time in the molding operation to set the core.
When only the working face of the casting was tested with a nobake core, there were alignment problems between the cope and drag of the casting. The surface finish of the face was identical to current practice and the surface finish of the non-working surface was acceptable.
With the moderate success of the nobake cores in green sand, Durametal attempted to face the green sand with nobake sand at the molding machine with some simple pattern designs already set up for the matchplate molding machine. The nobake sand was spread out over the face of the pattern, compacted by hand, and the green sand dropped on top of the nobake sand. The molding machine then went through the rest of its process normally.
The first problem was evident immediately. The nobake facing sand and the green sand did not knit together. The nobake sand remained on the face of the cope when the machine lifted the top flask for pattern removal because there was no bond between sands. This problem was overcome by using sand cores to hold the two different sands together. The same sprue configuration that was used in the core insert test was used in order to be sure filter placement was possible.
The initial run showed excellent promise with surface finish being equivalent to the nobake castings. It was determined that the high-squeeze pressure of the matchplate molding machine made this surface finish possible. After the success of the initial trials, pattern work was done to mount current nobake mold line patterns to the green sand pattern boards.
Tests on these more intricate designs also proved successful for casting quality. While the knit between the green sand and the nobake sand still was inconsistent, a better sand jagger was designed to hold the sands together, eliminating almost all of the knit problems.
Several hundred castings were run in this manner. Since it was inefficient manually carrying the nobake facing sand from the core mixer to the molding machine, the firm installed a nobake sand mixer at the location of the molding machine.
Included in the installation was a sand heater, pump control package and small silo. No attempt was made to tie this machine into the sand transport system due to the expense. The exit end of the mixer was placed as close to the green sand molder as possible, and a slide was fabricated from the exit to the pattern area.
At the same time, a pattern changing device was devised, as the patterns and pattern boards that were transferred from the nobake molding line were much heavier (more than 100 lb/pattern) than the traditional green sand molding patterns (70-80 lb/pattern). In order to avoid injury, a lifting device and slide were implemented to eliminate manual lifting of pattern equipment.
The first problem that arose was sand sticking to the slide from the mixer. Because of this, cured sand would fall onto the face of the pattern in the next mold.
The time required for each mold was the next issue. Since only one pattern is on the matchplate molding machine at a time, the strip time of the nobake sand became critical. With all the equipment in place, it took nearly three times longer to produce a mold with this method. It took 45 sector the sand to reach the pattern on the conveyor, the operator to spread the sand over the lace of the pattern, place the jaggers to be placed in the nobake sand and the molding machine to cycle into the squeeze position. The strip time for this was 65 sec. This equated to the mold being held in high squeeze for the 20 sec.
Even with only facing the cope of the mold with nobake sand, the cycle time from mold to mold was too long at more than two minutes. This was unacceptable in comparison to the time it takes to make a mold on the nobake molding line. Even with the success in the quality of the castings, the extra time required per mold did not justify using this method.
Success Out of Failure
While Durametal was ultimately unable to achieve a solution to move casting production from its nobake line to its green sand molding line without sacrificing surface quality or increased costs, the firm did receive several benefits from its experiments.
In order to run the tests, Durametal's green sand operation went through a complete audit during the trials, covering everything from machine maintenance to casting shakeout, including sand properties. This has helped increase productivity approximately 15% in this area and made the operators in this area more productive.
The pattern-changing device that was designed and installed to aid operators in avoiding lifting has been an ergonomic success with employees, while not causing any increases in casting production time. The firm also achieved surprising levels of surface finish with the high-pressure squeeze of the automatic matchplate molding machine without wash and hopes to investigate possibilities regarding its use in the future.
This article was adapted from a paper presented at the 2003 Steel Founders Society of America Technical and Operating Conference.
For More Information
"Maximizing Productivity of Nobake Molding Operations," T.J. Gilbreath, MODERN CASTING August 2000, p.39-42.
About the Author
John Cory is the manufacturing manager at Durametal Corp., Muncy, Pennsylvania.