What's new
What's new

Newbie scraper in Japan

How do you sharpen the blade? It looks as if it is scratching more than scraping.
That's the thing, I didn't have a chance to sharpen the blades I just gave the ones I got a visual inspection. They looked clean and grabbed the back of my nail, but they are likely dull. I have a dedicated carbide grinder (#400 wheel on it) and parts for a low speed lap that I haven't put together yet, but I don't want to grind inside the house and I've got a ton of blades that I want to sharpen, so setting everything up is a project in itself. I just wanted to make sure the Junker repairs looked viable and was hoping I could get some experience, but sharpening practice is the next thing on my list.

looks like the radius could be larger, or the angle adjusted down a little to make a smooth wider scalp.
The radius was very close to 90mm which I know is a bit aggressive for roughing, but I was dealing with a very small surface that was already quite flat. The Biax had a very small angular range where it felt it was cutting effectively and creating chip dust. If I had too much angle of attack it would chatter and bounce (not sure if I should have tried to muscle it, but I assumed that was not the right move) and trying to shallow it out to get a wider cut, it would lose bite and just skate (I'm guessing that's a sign of a dull blade.) I tried using the other edge of that blade with similar results, then switched to the factory Biax spring steel blade (60mm) that came with the other tool. It was also likely dull although I noted it had other nice characteristics. To the eye, even with a loop, the blade edges are shiny and well defined, no obvious chips, but I'm assuming at this point that an untrained visual assessment doesn't cut it when it comes to blade condition.
 
You need to sort out the sharpening to do any serious scraping. I made my rig out of a tiny cheap grinder plus a cast iron disk and a wooden rest. Works well enough for my needs. An inexpensive 400 grit diamond lapping disk from China will also work well, mounted on the same kind of small hobby grinder motor.

A photo is here: https://www.practicalmachinist.com/forum/threads/studer-rhu-450-from-the-1960s.343863/post-3108490

I also made a radius jig: https://www.practicalmachinist.com/...450-from-the-1960s.343863/page-4#post-3111305
 
On Sunday I put on my Big Boy ultrafleece sweatpants and pulled out the restored XB-74 Carbide Grinder (the model number always makes me laugh because they only have really made a single machine for the last 50 years, and you know they were going for some cool spaceship name at the time.) I have a horde of fresh carbide blanks and old tools in need of a resharpen, but having never done it before, I wanted to limit the apprentice damage and selected 4 candidates: Tungaloy 20mm, 25mm and Biax 20mm and 25mm. I started with a Tungaloy CST-20 20mm blank that had come with one of the used hand scrapers I bought. It had been incorrectly mounted sideways so that the flat side was used, I can only assume for scraping paint or labels, and the corners had been badly chipped. So I wasn't worried about trashing it, and it was a good sample to see how fast I could remove material.

I dressed the old #400 grit diamond wheel at 1800RPM with a white chunk of Alumina designed for that task, set the table to -5° and bagan. Happily the removal rate wasn't very fast, there are some very tiny dings in the wheel substrate in a few places, and I could feel them when grinding, but I did my best to pull the work across the whole surface in each pass to make sure the wear was even. I had the VFD set to 30Hz, it was whisper quiet and didn't bat an eye to the cutting loads, so very happy overall with the machine. I would reverse the wheel direction and use the opposite side to keep the blank symmetrical, since the outside edge of the wheel moves faster and the way I felt comfortable drawing the blank across it was favoring the trailing side. It was too slow to easily remove the 2-3mm of chipout that the damaged blade had, so I instead just ground it to have a single and practiced on the opposite side to get a clean double edge.

These are the observations of a self-admitted newb, so grain of salt and all that. When reprofiling a new blank, it seemed to work best to just grind in the arc profile on one face first, then try to use the little centerline peak to match the profile on the opposite face. It basically indicates where there is a mismatch in the two sides curves as the peak will deviate from the center. If that line is wavering all over the place, it is because the two faces arcs don't match. So get the arc right first, then match the faces. Hand making a clean radius seemed like it would be inaccurate, but the results came out reasonable enough that (as I had been told) a jig is likely not needed. The time taken to make the jig is probably better served in just getting the curve right by hand and gaining skills with the grinder. Not that it will stop me from making some overly complicated contraption, but just to keep it real.

The arcs are subtle. I have an accurate arc gauge plate I 3D printed, the factory blades that came with the good Biax seemed to be well used, and reground many times. It's one of the reasons why I didn't immediately recognize that the scraper itself was actually new, the blades that were thrown in there must have been used with some other machine. But they were all ground and sized for finishing work, with roughly 40mm - 60mm radii. The blade holder that came with the Junker, as I mentioned, was shop made, but the carbide blank looked to be from Biax (or Biax OEM) as it has the same milled face finish as the factory spring steel tools, not the super mirror finish of the Tungaloy blanks. The grind on it was pretty garbage technique wise, with one side having the mid-ridge tracing a diagonal across the edge, and both sides the arc was not centered on the blank. It came ground to ~60mm as well, but I reshaped one side into a 90mm and touched up the other to the 60mm it came with. I don't want to change too many variables so that I can confirm the scratching is sharpness related (although I am not assuming this sharpening I just attempted is really sharpening.)

Carbide1.jpegCarbide2.jpeg

Forgive the quality of the photos, I really should have brought out the microscope, but everything is all over the place and I don't want to have to make space for it. You can see the clear grinding marks, they are small, but I suspect that the experienced scrapers would want that more polished. The left blank is a 25mm Tungaloy and the right is a Biax 20/150 spring blade that came with a 60mm grind that I preserved. I just did a few really light passes as it doesn't have a lot of carbide left and I want to be conservative with it. You may also notice that I was a bit more aggressive with tightening the radius on the edges of the Tungaloy blade, as I have come to understand that doing this can help reduce the dreaded edge scratching when one of those corners digs into the work.

Back to the arcs, the difference between 90, 60, and 40 on a 20mm wide segment is quite subtle. Especially when you might be treating the edges differently, and the actual contact area is limited to a subsection of that arc length. My brain keeps telling me that I need mathematical precision, but my eyes are saying, "Dude, that's an almost perfect match to the gauge. Relax." Anyway, my main concern at the moment is dealing with the finish so I want to try these as they are, but will need to work on my lapping solution to get mirrors on those cutting surfaces.
 
You need to sort out the sharpening to do any serious scraping. I made my rig out of a tiny cheap grinder plus a cast iron disk and a wooden rest. Works well enough for my needs. An inexpensive 400 grit diamond lapping disk from China will also work well, mounted on the same kind of small hobby grinder motor.

A photo is here: https://www.practicalmachinist.com/forum/threads/studer-rhu-450-from-the-1960s.343863/post-3108490

I have that covered but I need to adapt some other wheels/lapping disks to it.


Yeah, I have a design I was going to build but was told to try to freehand it by an experienced member before going through the trouble. It didn't seem too hard to get a good arc with just keeping the cutting time smooth and symmetrical. I've not forsworn a jig (because I like side projects and idiot proofing) but for now I will try to do it by hand. I'm fortunate to have several lifetimes worth of new blanks, so I can afford to regrind until it looks right.
 
Your photos look a lot like what I get from a resin-binded wheel before going over to the lap, but possibly with a coarser grit. My plated wheels of nominally similar grit leave a less uniform grit pattern, apparently because I have some particles that stick out.
And I found it surprisingly easy to freehand the radius, if the scraper body is resting at a consistent rake angle. While the diamond is effective at removing carbide, it's very different from rounding off a mild steel flat on a belt sander; the steel sort of melts away, while the carbide endures so you can develop a hand feel for the appropriate pressure, and maintain that as you swing the blade around.
 
This morning I tasked the Ol' Ender to printing up the puck, and while it was whirring away I transfer punched the 60mm hole pattern used by the XB-74 🚀 grinder to a spare wheel I picked up cheaply at auction.

d_wheels.jpeg

The factory wheel spec is for a 125mm dished profile (there's a universal classification system, I looked up the exact descriptor before, but these don't really use it and most of the wheels I see on the used market don't either.) The wheel on the left came with the grinder, it has a nice wide face with a thin squared shoulder (20mm x 2mm) and is #400 grit, the one on the right is a little larger at 135mm diameter, but fortunately still fits, and has a 7mm wide #220 grit face and the edge is a rounded profile. They both use a 31.75mm hub, but the bolt pattern on the bigger one was 55mm, so I had to drill and countersink a new set of mounting holes. I may have to re-balance it, but I did my best to remove the same amount of material for each hole.

d_wheels2.jpeg

The puck has a 12mm hole that matches a similar hole in the motor hub, this allows me to align it concentrically with a small section of shaft and still keep a flat bottom surface which is much easier to 3D print. A 3mm pin is pressed in to engage with a hole drilled in the cheap lapping disk to keep it from slipping. The raised 16mm lip is used to center the lap. 4 magnets are set edgewise and are a very handy 9.6mm wide, which is just shy of the 10mm recess provided in the wheel. I designed it so that I could use 2 longer screws to secure it directly to the hub, it's probably overkill. I may redesign it to be hand removable if it gets in the way, I just didn't want to over engineer it until it proves its utility.

d_wheels3.jpeg

The 100mm #1000 grit lapping disk (which has some kind of thin plating that's constantly flaking off the back) is well supported around the edges by the wheel face itself and is held firmly in place by the magnets, hub and pin, but can be quickly pulled off when using the wheel itself. I'll try everything out tomorrow, weather permitting.
 
I have a simple website with info about getting started openscraping.com, the site needs a bit more work but it will give you an idea about preparing blades etc and the other equipment needed.

As for the router project that is a different thread altogether, for a router using aluminum extrusions there is no need to scrape that plate (and doing so would be profoundly difficult without a large reference surface plate and a crane)

Luke
QT Ops website. Two rotating abrasive wheels are required, a resin-bonded diamond wheel (perhaps a d150 100mm cup wheel), the other a cast iron or ceramic disk loaded with a fine (3 micron seems to work) diamond lapping paste.
Going from d-150 to a 3-micron lap seems a very big jump.
 
The modified wheel fit nicely and the #220 grit was noticeably more aggressive, but still quite controllable. I used it to make a more aggressive 40mm and 60mm radius grind on a small 15mm wide finishing/spotting blade for one of the hand scrapers.

grinder4.jpeg

Then I put in the #400 grit blade and the new puck, using a 12mm rod as a centering mandril. Despite the effort, the cheap honing wheel itself wasn't very concentric, but at the 15Hz speed I set the VFD to for the lapping it wasn't a big deal.

grinder3.jpeg

The #1000 grit lap cut more aggressively than I expected, I suspect going higher may make sense. It didn't give a mirror finish, I'm not going to obsess about it for the moment.

hone1.jpeghone2.jpeg

Will try these on the test block shortly.
 
scrape_1.jpeg
Tested with the rigid blade holder 25mm width, 90mm diameter (5 degree negative grind) 2 cleaning passes, then switched to the spring blade 20mm width, 60mm diameter (Neg 5°) and two passes.

Both blades seemed much easier to to handle, I wasn't chasing the attack angle and didn't encounter any obvious chatter. The scrape pattern no longer had linear scratches, but both blades now exhibited horizontal artifacts that hopefully can be seen above. Not sure if this is chatter (from not holding the tool properly, or incomplete sharpening) or if this is expected. The sharpened blades improved finish quality for sure, and eased operation, just wondering if I should order a #3000 grit honing disk and trying to get a mirror finish on these blades, need to change technique or if this is normal.
 
For one thing, those diamond lapping discs can take a little while to break in. They will get progressively smoother with use, but it takes a while. If the 3,000 grit is not too expensive I would try it out. Do they make a yet finer one? I've got diamond plates up to 8,000 grit. You could also go the plain lapping plate and diamond lapping compound route, that always produced a very smooth cut for me with something like 9μ compound or even a little finer. It does look like the finish left by your scraper isn't quite where it could be yet.
 
I took a closer look using better magnification and I may have mislead myself with exactly how much was really lapped. Tomorrow I will use a marker to color the edges of the blades and make sure it is really doing what I thought it was. I may not have actually removed much material with the lap at all. There are #2000 and #3000 grit plates from the same supplier, they are quite cheap, but I want to make sure the #1000 does something useful before ordering more.

A real lapping plate wouldn't fit the way I'm attaching it as there's only a few millimeters gap between the wheel and the tilting tool rest. The thin diamond plates just squeeze in there. I'd have to machine a whole wheel replacement, and that's a bit out of scope at the moment. I know others have been successful with similar CN laps, I am now thinking I just misread the residue it left on the blade and assumed it was cutting much faster than it was. The bright sun made it look smoother than it was.
 
For general scraping I use a 1500 grit flat lap disc. For shaping I get the coarsest one I can find, a 280 or 400 or so. They don't stay coarse very long but do allow you to crate your initial radius.

A trick I use it I grind the front radius dead square with a coarse lap then put my angle on the edge with a fine lap. I DON'T grind the bevel right to the centre line as it's a waste of diamond and carbide and for a novice it makes it a lot easier to easily keep a keen edge. I only put a fine cutting edge on the first little bit. It sharpens very quickly using less pressure, less diamond and you don't have the problem that many novices get of grinding all of the face EXCEPT the very edge. Of course, the cutting edge gets wider each sharpening but you'll get many sharpenings before the two faces meet in the middle.

Your example above looks like you could do with a little more angle on the grind but it's not bad. Also, I know it's just a practice piece but your scrapes are too close together both side by side and end to end. Spread them out so that the gap betwen scrapes is at least the width of the scrape, and do the same for the ends. You'll have a good checker pattern going in no time.
 
I DON'T grind the bevel right to the centre

Good points on the extent of the grind, that had occurred to me but I didn't want to complicate things by premature optimization, especially if I was asking for input. These were mostly new blanks with the default ~300mm radius, so I had to hit them pretty hard and using the ridge ensured that the two sides matched and the edge was as supported as possible. The grinder was designed to run much faster than I used it, the little VFD makes it silent and smooth at slower speeds. Reprofiling, even with the #400 grit was pretty fast and easy.

I think the wheels I have are both using metal bonded (not electroplate like the hone), it really looks like some kind of sintered bronze like material (check the pictures above), but they may well be the more common resin type, but the resin wheel I have looks much different. Either way, there's a lot of material still on there and it was used for a long time before I got it, I expect them to outlive me as I suspect my usage will be limited. I also have a ton of new carbide blanks, I bought a motherload of them and several hand scrapers for a song. My focus right now is just learning how to make a useful edge, I think I'm getting close.

As far as the scraping itself, I was just focused on getting the feel of the difference from before the blades were sharpened and I was trying to create a fresh base layer to restart the printing process. I didn't take pictures of the first passes done with the Junker Biax and the rigid blade, but it was much easier to work with than last time. The photo shows my second round of passes using the New Biax and the spring steel Biax finishing blade. I didn't change the stroke (again, trying to do a comparison with as few variable changes as possible) and so the photo wasn't trying as much to demonstrate technique as to show the somewhat stuttered cuts the blade was making. The scrape quality was a significant improvement I feel, it scraped more material per pass and wasn't making chicken scratches, but I think we can all agree there is still some room to improve that finish. But your point on the technique is well taken, once I'm happy with the scrape quality I will focus on stroke discipline.

You suggested a bit more grind angle which is confusing to me as it is currently negative 🧐, it is set to -5°, do you mean -6 or 7° or the other direction? If the honing doesn't help I can test regrind a different angle.
 
You've got slight chatter marks but TBH if it's not bothering you then just go with it. Adjusting the negative angle might halp it but it could be you're just not going at it hard enough. Try a heavier scrape and see how it fares. A good thing is that your scrapes are nice and consistant so spread them out a bit and get some hours in.
 
A real lapping plate wouldn't fit the way I'm attaching it as there's only a few millimeters gap between the wheel and the tilting tool rest.
On the Baldor 500, you can adjust the inboard/outboard position of the tilting tables considerably. You might take a peek at the underside of the table on your grinder to see if that's possible on your machine. It's pretty common to have to move the tables outward when you increase the tilt angle.
 
On the Baldor 500, you can adjust the inboard/outboard position of the tilting tables considerably. You might take a peek at the underside of the table on your grinder to see if that's possible on your machine. It's pretty common to have to move the tables outward when you increase the tilt angle.
Unfortunately not. Although the table is easy to remove, it is designed around a specific wheel size. The thick pins it rides on are fastened in v-ways ground in the support collar. The choice of pivot point means it can move through its entire range with little change in the gap, which is nice, but there is just enough room for the thin electroplated laps.

I've got the parts for a low speed lap, it would be the best place for a nice cast iron plate. It's on my endless list of stuff I still need to build.
 
Despite the weather taking a cold turn, I set up the lap again and inked up the face of a blade. Even at 900RPM that #1000 grit hone just chewed through it. So my theory that it wasn't actually removing the rough grinding grooves was wrong, but the finish it leaves isn't a lot better than the #400 wheel. So I ordered the #3000 grit, which is the finest one they have of that size on Amazon. I suspect at this point that the finish quality has more to do with my lack of skill, but $10 is cheap and the whole point of the magnetic puck is to make swapping these laps easy.
 
tldr; Blades must be sharp! As others have said, a course disk ~#240-#400 for initial shaping and #3000+ for honing and sharpening at fairly slow speeds is all that is needed.

After the morning rain cleared up I put a quick 3.5mm (dog?) hole in the #3000 grit lap and set up the grinder. I had inked up all the blanks I've been working on, as it is really easy to lose track of which ones and which sides had been sharpened. The new plate was slightly better made, there was no strange flashing flaking off the back. Boy did it cut though. Again, I've got the motor trickling along at under 1000RPM and these electroplated disks really take off material. The finish was noticeably finer with the #3000 disk, not a mirror, but closer.

hone3.jpeg

Subjectively the blades felt sharper to the touch, they seemed to cut even easier as well and the chatter marks, while still present, were better with a more consistent scrape width per stroke. The previous individual scrapes looked like a bit like a seismograph, which I interpreted as chatter and different depths of cut (blade goes deep, the cut goes wide, blade pops up, cut becomes narrow.) I was bracing a bit harder and trying different amounts of pressure though, so some of the improvement may be attributed to that. No reason not to use a higher grit though as I don't think any interstitial gris are required and so there is no real time impact to just going from roughing to the ~#3000 plate. As I was trying to reduce the variables I just stuck with the Junker and used the rigid blade holder and 90mm radiused chip. The repaired scraper seems to be working great, tightening the wobbler adjustment fixed the stroke length drifting and the Phosphor Bronze plates are showing no signs of wear and are holding up perfectly.

I flipped over the test block to start roughing the rough milled surface, and just worked on trying different pressures and techniques while removing the deep scoring left by the original manufacturing. After about 12 passes I could feel the edge of the blade kind of falling off, It felt more skatey and I could observe some scratching parallel to the stroke. I flipped the blade to use the other edge, and that scratching went away immediately, so I'm pretty sure that's a clear sign of dulling (likely prematurely due to me f'ing around and the rough milled surface it was cleaning up.) The picture below is not me pretending this is any good, just providing documentation of the 'output' of the tests.

scrape_2.jpeg
 








 
Back
Top