Bridgeport-Powered Component Pinch Roll Project

Discussion in 'Foundry tools and flasks' started by Melterskelter, Oct 3, 2022.

  1. Melterskelter

    Melterskelter Gold Banner Member

    Very often in my foundry work I need to roll sheet metal into arcs or round cylinders. A couple years ago I made nice small hand-powered pinch roll (as opposed to pyramid roll) that will roll light gage steel 16ga 12 inches wide and lighter as well as aluminum and copper.

    But I have felt quite limited at times needing to roll heavier gage steel. Right now I am in need of a 4 inch wide 2 foot in diameter .125 ga hoop for my muller. So, I have the crazy? idea of making a modular unit that I can power with my mill head and assemble in 3 pieces on the mill table. I have been busy the last couple days drawing in my spare time. And it looks like it may be feasible. I plan to either cast or weld up the main frames for roll support and cast the rollers. Here is where the plans are so far. For scale the rolls are 3.5" in diameter and 6 inches long. Shafts are drawn as .75" diameter and will likely be carrried by bronze bushings. .500 Pitch Chain drive with an idler (no idler drawn yet) will drive the pinch rolls. The deflector roll is 2.5" diameter and is free-turning.

    upload_2022-10-3_17-13-2.png

    upload_2022-10-3_17-14-23.png

    These are early sketches. Lots of trimming, chamfering, drafting, and general refinement are still needed not to mention rotation reversal of the opposed rolls. I intend to have some provision of controlled adjustment of the gap in the pinch rolls and for the degree of dflection casued by the deflector roll. The blue segment above is a to-scale drawing of a segment my mill table.

    Suggestions are welcome.

    Denis
     

    Attached Files:

    Last edited: Oct 3, 2022
    Tobho Mott likes this.
  2. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    For one-time or low use I've made a male/female wooded die and just pressed/walked it through the circumference (actually 1/2 the circumference because I didnt have stock that long) in my hydraulic press.

    Best,
    Kelly
     
    Melterskelter likes this.
  3. Melterskelter

    Melterskelter Gold Banner Member

    Yes, and I have made a couple bending bars with two pins set about an inch apart and connected to a long arm that makes round stock or squarish stock bending fairly straight forward. But, I really love my smaller pinch rolls because the bends are so fair, fast, and adjustable. The need for thin bar bending comes up often enough that I want to try this. I will likely groove the rolls to allow rolling round/square stock too.

    Denis
     
  4. Melterskelter

    Melterskelter Gold Banner Member

    Tops likes this.
  5. Melterskelter

    Melterskelter Gold Banner Member

    I've been busy printing pattern parts (in the midst of other projects). Here are some pics

    Since last posting, I aquired a SLA printer for printing letttering on patterns. I intend to print the main bosy of the patterns using filament (PLA) printing and then inset or attach name plates printed with SLA. The reason is that resin printers can produce so much finer resolution. The filament printer is for larger format stuff for my ppurposes. The accuracy of both is quite remarkable being within a couple thou or so on a 4-inch-wide filament print and within a thou on a 2" resin print.

    The resin printer I got was the cheapest on the market proactically. I did not buy because it was cheapest but rather since it s reviews were so good and ireviewer's experience was that it worked well right out of the box. I was not disappointed in my ourchase of the Phrozen Mini (not the 4K version). It works well with no tweaking and produced the greenish prints shown. It reproduces the draft I drew into the letters using Onshape. Med2.JPG Med3.JPG Medallion.JPG

    Hereare some of the pieces printed for pattern making. The grey ones are PLA. I will glue the parts with cyano glue. These parts just exceeded my size max on the filament printer. So, I had to split the drawings in the computer and print in segments. They require light sanding and then I will apply high-build primer and sand a bit more. Then I will finish with lacquer. They include all needed draft angles and chamfers. It does not take lon
    Patterns.JPG

    A typical large part shown above takes about 4 hours to print in PLA. And the medallions take about 20 mins.

    Denis
     
    Mantrid, Tobho Mott and Tops like this.
  6. Melterskelter

    Melterskelter Gold Banner Member

    Today I learned the value of adding a brim to a PLA print that has sharp corners and is somewhat thick in cross section. I printed a pattern for my net 6X3.5" bending rolls in PLA but did not use a brim around the periphery to improve adhesion of the print to the plate. The corners lost some adhesion as the print built resulting in this sort of fit up. I am fixing the problem with Bondo by just placing a bead of Bondo around the edge and pressing the part down onto a flat surface to allow the Bondo to cure.

    The parts are two-toned as my silver filament spool ran out mid-print and I used some black. For those non-printers in the group, the smaller coars appearing cylindrical shape is the support material temporarily printed into the hollow pattern. It avoids needing to (likely unsuccessfully) string long strands of molten filament in the air as the print builds. It is loosely attached to the actual print and is intended to be pulled out after the fact---as shown in one of the pics.

    I am presently printing a deflector roll pattern for a net 2.5X6 inch roll. For it I provided a brim and it appears to be maintaining firm adhesion throughout the printing process. It will be done in a few hours and I will post images later.

    These prionts were made using a largish (.6mm) nozzle with a fairly thick (.3mm) layuer height thus saving hours of print time at the cost of mildly rougher surface finish.
    Roll Pattern No Brim2.JPG

    1/2 the support material shown between the two print halves.
    ROll Patttern no Brim.JPG

    Denis
     
  7. Melterskelter

    Melterskelter Gold Banner Member

    So, I used a 5mm brim inside and out on my 6 x 2.5” pattern. The print and brim remained 100 % attached throughout. And the bowing was better, but NOT eliminated. E820B88F-A013-489F-808C-350E3909BF40.jpeg 484D42D0-F484-4444-91DA-B953CB8F5856.jpeg
    4D83F21B-0AD5-4386-8CAE-1FCF5B764B28.jpeg


    Anyone know how to reduce bowing to near zero?

    PLA 60 base. 215 nozzle. .6mm nozzle. .3layers.

    Denis
     
    Last edited: Oct 13, 2022
    Mantrid likes this.
  8. Mach

    Mach Silver

    Looks like layers are cooling too fast. Are you using a cooling fan? Change the orientation to print vertically as a open ended cylinder, possibly use an enclosure or increase bed temp/extruder temp?
     
    Melterskelter likes this.
  9. crazybillybob

    crazybillybob Silver Banner Member

    Denis,
    I put my printer in an enclosure. This keep drafts off the print as they can cause uneven cooling. Lowering my cooling fan duty cycle can help when printing large items. I also print things like patterns with a full flat bottom (where the two parts meet) if possible. This eliminates the wasted support material and helps keep parts from warping. If i have a part that just seems to warp no matter what I do I'll add a straight section on to the drawing as a sanding allowance then sand it off as i sand the mating joint flat. Depending on the part a few tenths of a mm.
     
    Melterskelter likes this.
  10. Melterskelter

    Melterskelter Gold Banner Member

    Thanks for the suggestions above. The easiest thing for me to try would be to adjust my base temperature. My goal is to print the parts so they require minimal post-printing fettling. So, I think I will print these cylinders with parting planes oriented axially as they are. Making them print as “solid” hemi-cylinders rather than shells makes some sense as it is the contraction of the cooling plastic that must induce the stress that warps the part. As printed, there is much more plastic drawing in the upper portion of the printed cylinder than the lower portion.

    I also contacted a friend of mine who prints professionally. He suggested raising the bed temp 5 C after the first couple layers. I’ll try that. That should cause the first layers to be laid down a little shorter (as the build plate will be relatively contracted) and may induce a bit of tension to counters the unbalanced contraction force of the cooling upper layers.

    Chat with the Prusa help desk recommended against fan speed changes. So, I’ll hold that idea, for now. Might try in the future. The room in which the printer resides is very draft free and temp stable.

    It seems that conceptually ANY process is simple. It is the nuances that will get you! ;-)

    Denis
     
  11. Melterskelter

    Melterskelter Gold Banner Member

    I am getting down to the end of my pattern work and hope to be able to cast some metal in the next couple of days. I am sure that when I try the patterns in the sand that they will teach me how I overlooked certain areas of the pattern and will require some rework to make them draw cleanly. 5AB5BA76-604D-43A8-8A2A-241741DCB8EC.jpeg Denis
     
    Tobho Mott likes this.
  12. ddmckee54

    ddmckee54 Silver

    I'd print the parts as a solid with just a couple of bottom layers, use 3-4 perimeters, 10-20% infill and 4-5 top layers - or more. That should stand up to ramming the sand. The bottom layers take forever to print on large parts, and the way you are using them the bottoms don't need to resist the ramming. You might want to experiment with the infill pattern too, some are stronger than others. The infill will help tie the part together, this should help resist the warping.

    Did you let the parts cool on the print bed, or are you like me and impatient enough that you pop them off the bed as soon as they are done printing? You did say that the part and brim stayed attached to the print bed. It could be that the parts are warping after they've been removed from the bed. Maybe you could clamp them to a flat surface as they cool?

    Don
     
  13. Melterskelter

    Melterskelter Gold Banner Member

    Thanks for the suggestions, Don. I made a mistake printng them hollow. My intent was to save filament. But it is likely little filament was conserved compared to printing them solid as support material was printed into the hollows anyway. I have been using Gyroid pattern infill of late. It seems to be very strong. I wish I had just used it throughout these half-patterns. Ramming does not seem to threaten damage to the patterns I have made so far.

    Of course, I yanked them as soon as the printer head raised! I think you may be right that allowing them to cool clamped to a flat surface MAY have helped. But, as soon as I noted they were warped---within a minute of pulling them, I clamped them to my mill table to cool. I actually suspect that stresses similar to those encountered in welding occured in the layers as they built up from the base. There was much more contraction stress in the solid top surface than the hollow bases===warpage. I am presently reprinting these patterns to be cast vertically. More on that in my next post. These "simple" patterns have taught a fair bit both on the pattern printing side and on the casting side.

    Denis
     
  14. Melterskelter

    Melterskelter Gold Banner Member

    I have done some casting and had mixed results.

    With respect to the rolls, I have tried twice and failed twice. I expected casting the erolls to be simple. But, I got slapped down for that hubris.

    I set the rolls up in a pretty conventional horizontal orientation. And I used a riser to counter expected shrinkage in these "chunky" castings. Well, the riser was not in the right place. I usually find that shrinkage is more likely near the gates as that should be the hotter part of the casting. But, with these thick castings, the center of mass plays a more significant role. So, the result was ugly. There was some spillover at the end of the pour from the pouring basin into the open riser, but that did not cause the shrink problems.
    Rolls with shrink.JPG Shrink Detail.JPG

    Right now I am reprinting the patterns with draft for vertical orientation when casting. I will place a decent riser right in the middle of the top surface and hope...

    1.5 degrees of draft each side or 3 degrees included angle.
    upload_2022-10-24_12-15-22.png

    I will have to put together a cope and drag for this. It will be pretty basic made from OSB as I expect (hope for) very limited use of the cope and drag.

    Denis
     
  15. Melterskelter

    Melterskelter Gold Banner Member

    The deflector roll support, on the other hand, came out great. I expected trouble with it as I worried there were half a dozen good candidates for causing trouble. But, none bit me this time.

    The goal was to make a casting consisting using these basic patterns:

    upload_2022-10-24_12-27-53.png

    That rectangular hollow presented some sand work problems. Though I printed it with 2 degrees of draft, it was very hard to draw the pattern wna not pull sand with it. So, i decided to go with a core to form that hollow. The core was made from new casting sand and 6% by weight West System 5:1 epoxy. I did not vent the core or include coal (the epoxy provided carbon to prevent burn-on).

    To accomadate the core I plugged the hollow with filament-printed plugs. The cope plug was just flat and the drag plug had a protruberance to form a core print in the drag.
    Split pattern with core print pattern in place..JPG

    This is the core box which was simply coated with vaseline petroleum jelly and wax. The white cap had a recess printed to form a key on the end of the core. The core print pattern was thickened in the computer by 30 thou all around to allow some clearance for the core and it had 45 deg shoulders to allow the core to nestle into the 45 deg recesses of the print.
    Core box exploded.JPG Core box assembled.JPG

    Here is the key end of the core. Note there are brass pins in the core which will allow the core to be pinned in place. The pins are splayed (not parallel) to help them have better retention properties.
    Key end of core.JPG
    The cope end of the core. Note how much closer the tops of the pins are than at the bottom. I pushed the pins about 2.5" into the drag after placing the core in its print. They provided good stability for the core. I did not wnat it to topple while I transported the mold out to the furnace for casting. And I wanted to prevent float of the core should iron find its way under the core. As it turns out, iron did not get under the core, so there was no bouyancy trying to float the core.
    Top end of core.JPG
    The rectangular print in the sand to receive the core.
    Core print key impression.JPG
    The core (coated with graphite---likely unnecessary) in place and pins pushed into the drag. The pins are 1/8" diameter and telescope inside 5/32" tubes placed in the core. They fit really nice and snug but still slid easily.
    Core in place.JPG

    I'll take a break now and post pics of the results later. You will like the resulting casting.

    Ooof, long post....

    Denis
     
    Last edited: Oct 25, 2022
    Tops likes this.
  16. Melterskelter

    Melterskelter Gold Banner Member

    And here is how the casting for the deflector roll turned out.

    Right out of the sand. You can see 5 gaggers stuck to the casting. That sand is hanging unsupported on 3 sides and the gaggers were needed to keep it from having a "ceiling" collapse. The gaggers were simply 4" pieces of concrete tie wire that was bent with a right angle 1/2" from an end. THe wires hold better if they are not quite straight and have a slight curve over their length. In the photos they are mangled due to me shoveling out sand. But they did protrude out of the sand a bit and, so, stuck to the iron. I used a needle-nose plier to insert them, being careful to bring the bent end down and in good contact with the sand but not so deeply embedded as to cause a fracture of the sand. It takes a bit of trial and error to get it right. Despite their small size, they make a big difference. If you look closely you can see they are just barely attached to the skin of the casting by little bumps of iron. I was able to pull them free them with a plier. Next time I might try to push them JUST under the surface and push a bit of sand over them.

    Overview of defl support casting.JPG Gaggers on hanging surface of mold.JPG

    Gagger detail.
    Gagger detail.JPG

    The core was intentionally made so that it did not quite touch the cope lest there be fracturing of sand. So there was a 1/16" thick kiss of iron covering the front of that rectangular opening. Here the kiss is broken out. Incidentallty, you can see that the surface finish left by the core is better than that of the casting that was in contact with the green sand.
    Front of deflector support   Kiss broken out of window..JPG

    I guess I do not have a photo of the back of the casting, but take my word for it being nice and clean.

    So, next up is casting a couple large (pinch) rolls and a small (deflector) roll. And casting a left and right hand pinch roll support. No cores needed for them.

    I was very pleased with the performance of the epoxy core. It was very strong and solid and was easy to remove post casting.

    I had not used gaggers like this in the past. There will be more of them in my future.

    I just built a pretty flimsy cope and drag for casting pinch rolls out of 1/2' OSB remnants retrieved from the scrap pile at a nearby construction site. I think it will be enough to allow a couple of castings to be made. I don't anticipate using it again after that.

    Denis
     
    Last edited: Oct 24, 2022
    Tobho Mott and Tops like this.
  17. Melterskelter

    Melterskelter Gold Banner Member

    I had better results yesterday. I cast the cylinder vertically with a riser centered on top. That worked well. Here it is in the el-cheapo flask molded and ready to pour alongside the mold for the support bracket. The cylinder mold required 25 pounds of iron including pour basin and sprue and the support bracket required 15 being 10 pounds itself. Casting a cylinder and supportcasting.JPG

    Being relatively chunky with not a lot of surface area the cylinder remained hot for a long time. This is it 1 Hour after casting with the covering sand knocked off the top. It is about 1250 according to my heat gun. 3.5inch dia cylinder 1 hour after casting.JPG

    A view of the coooled cylinder, runner, filter, sprue and basin. The sand on the top and bottom is oxidized as I knocked off the covering sand.
    3.5 Cylinder.JPG

    That chunky cylinder caused plenty of piping in the blind riser as it cooled and the riser fed it. Riseer piping on cylinder.JPG

    But the casting is sound with no contraction/vacuum defects. Riser sliced off. Riser cut off.  Sound.JPG
    I got lucky on the support bracket too. No voids, cold shuts, sand boogers etc. Just a nice clean casting.
    Support casitng fresh out fo the sand..JPG Back of support.JPG

    I hope tomorrow's pour goes as well as this one.

    Side note: I am also very pleased with my "new" furnace performance. It took 1 hour and 26 mins to get to 2600F yesterday with an A20 holding 50 pounds of iron. 17L of diesel. I also noted the iron is very soft and cuts fast on the horizontal bandsaw.

    Denis
     

    Attached Files:

    Last edited: Oct 28, 2022

Share This Page