A Scraping Newbie Has The First Of Many Questions

[M Kiser]

I am hoping for some help from some folks who have more scraping experience than I do.

I recently decided to teach myself how to scrape in order to restore a Hardinge HLV-H that I bought. After watching a few videos online, it seemed like such a simple thing to do. Instead, it has turned into a project with rabbit holes and sub rabbit holes! And due to a personality trait (flaw?) that I have I am actually enjoying myself. What follows is a summary of what I have done so far.

I bought a hand scraper ($140.00 without blade!!!!!), the Connelly book and the Richard King video from Dapra.

I needed a straightedge but am cheap by nature so I decided to buy a casting from Martin Model that I would machine and scrape myself. Had no clue how to scrape yet, so I also bought an angle plate to practice on. Both castings were nice and machined very easily with no hard spots.

After trying the hand scraper, I quickly realized that there was no way I was going to scrape the entire project that way. I don't know how people do that. Or why. Checked the prices on Biax scrapers and found that they are around $5,000!!! So I went down another rabbit hole and designed and built my own. It is based on a right angle drill from Menard's, a short length of linear guideway and an eccentric within an eccentric for an adjustable stroke length. Pic attached.

I read about the need to stress relieve after machining the castings, so I wrapped them in stainless foil and put them in my precision heat treat furnace. Which is a wood stove that we use to heat our workshop. Using an infrared gun thermometer I watched the temp as best I could and it got to a max of 1,300 F and was allowed to cool overnight as the wood was consumed. Both parts were a nice uniform grey color.

I started scraping on the angle plate and found that the uniform grey was a coating that was pretty tough to get through! Got it all off, but due to being a newbie I dug some ugly scratches in the process. Oh well. I then focused on one of the large surfaces to establish flatness and after about twenty cycles I ended up with what is shown in the picture below.

My questions so far are:
What was the terrible grey layer post stress relieving?
Is the blued pattern in the pic ready for flaking or should I scrape more?

 

[Bakafish]

I think you are describing the same road many of us have taken. Wow, scraping sounds cool! A a hand scraper is how much?! A Biax is what, WHO THE HELL COULD PAY THAT? I'll just build my own... Damn, building a power scraper is hard, that's why they charge that much. Great, now I need a diamond wheel carbide sharpening lap.

Quite a tidy little power scraper you built, I'm sure people who want a Biax but haven't capitulated would be interested in how you put that together. That angle plate surface looks good to me (also newbie) although the depth of color you are getting makes me worry that that might be too heavy of a 'print' that there is too much ink there and it is giving a false impression of the actual contact area. Maybe you are just a better photographer than me though :-)

Stress relieving in a wood fire would seem to expose the parts to a lot of chemistry that an oven wouldn't, excessive carbon and oxides. I know you eveloped them in SS foil, but unless you are sure that the seal was hermetic that surface layer may be caused by such contamination. Otherwise it's from the heat, and the possibility that the temps weren't what you thought they were. The hard layer may have always been there and just changed color due to your treatment as well.

Here in Japan they still sell Tungaloy handles and blades in many sizes. The blades are particularly nice, although you seem to be using factory Biax one piece blades on you unit. I have a post where I showed converting a couple Tungaloy hand blades to (stiff) Biax compatible units.

 

[M Kiser]

I think you are describing the same road many of us have taken. Wow, scraping sounds cool! A a hand scraper is how much?! A Biax is what, WHO THE HELL COULD PAY THAT? I'll just build my own... Damn, building a power scraper is hard, that's why they charge that much. Great, now I need a diamond wheel carbide sharpening lap.

Quite a tidy little power scraper you built, I'm sure people who want a Biax but haven't capitulated would be interested in how you put that together. That angle plate surface looks good to me (also newbie) although the depth of color you are getting makes me worry that that might be too heavy of a 'print' that there is too much ink there and it is giving a false impression of the actual contact area. Maybe you are just a better photographer than me though :-)

Stress relieving in a wood fire would seem to expose the parts to a lot of chemistry that an oven wouldn't, excessive carbon and oxides. I know you eveloped them in SS foil, but unless you are sure that the seal was hermetic that surface layer may be caused by such contamination. Otherwise it's from the heat, and the possibility that the temps weren't what you thought they were. The hard layer may have always been there and just changed color due to your treatment as well.

Here in Japan they still sell Tungaloy handles and blades in many sizes. The blades are particularly nice, although you seem to be using factory Biax one piece blades on you unit. I have a post where I showed converting a couple Tungaloy hand blades to (stiff) Biax compatible units.

Thanks for your input! I did another print with thinner ink and posted a pic below. Generally similar, but with smaller blue spots. I think it is time to start some finish cycles.

 

[Bakafish]

Thanks for your input! I did another print with thinner ink and posted a pic below. Generally similar, but with smaller blue spots. I think it is time to start some finish cycles.

You didn't mention any details about your reference surface plate, obviously that's a major factor in how true your surface can be. Most of the online instruction emphasizes the importance of Hinging, that Richard King instructs. The print getting a bit lighter in that one corner gives me a little concern that you may have a bit of a convex surface going on. Obviously a convex surface can misleadingly print as if it is flat, but hinging should help confirm the truth. Ideally you have alternate means of checking this, a blade straight edge or a Spherometer (home built should work fine) are traditional means of confirming it, but a sensitive indicator at a corner should be able to show any rocking. You may even be able to hear it by tapping the corners as a surface in contact sounds different than one with a gap.

Alternatively, as you are making an angle plate and you will likely need to do step scraping to ensure you get a true perpendicular plane on the other surface, you may start working on that setup and if that is done properly (even with an unfinished surface in contact with the plate) it can indicate the flatness across the plane (as long as the axis of the curvature isn't parallel with the reference surface.) This complexity is why a simple straight edge is usually a better first project.

I mention all this not to second guess your work, it looks good to me, just my thought process of doing something similar.

 

[M Kiser]

You didn't mention any details about your reference surface plate, obviously that's a major factor in how true your surface can be. Most of the online instruction emphasizes the importance of Hinging, that Richard King instructs. The print getting a bit lighter in that one corner gives me a little concern that you may have a bit of a convex surface going on. Obviously a convex surface can misleadingly print as if it is flat, but hinging should help confirm the truth. Ideally you have alternate means of checking this, a blade straight edge or a Spherometer (home built should work fine) are traditional means of confirming it, but a sensitive indicator at a corner should be able to show any rocking. You may even be able to hear it by tapping the corners as a surface in contact sounds different than one with a gap.

Alternatively, as you are making an angle plate and you will likely need to do step scraping to ensure you get a true perpendicular plane on the other surface, you may start working on that setup and if that is done properly (even with an unfinished surface in contact with the plate) it can indicate the flatness across the plane (as long as the axis of the curvature isn't parallel with the reference surface.) This complexity is why a simple straight edge is usually a better first project.

I mention all this not to second guess your work, it looks good to me, just my thought process of doing something similar.

The plate is a new 12x18 grade A. Should be good there. I have noticed though that when printing it is very hard to move the angle plate over the blued surface plate. Is this normal or is the ink too viscous? It makes it very hard to hinge. And I get your point about the complexity of the angle plate. Probably should have started with something else.

 

[JST]

If those are the same surface and view, the first one seems heavy in some areas where the second seems light, while the middle is about the same. That could indicate some rocking. The areas are bottom left and top right edges.

And the spots are in some places more like a smear, with streaks. I don't understand that, is it something about the surface? If so, you might need to scrape it straight down all over to get the surface to show nice spots.

What ink? If it is Canode, maybe the streaks are just the ink, it does not seem to thin down like HiSpot, and it gets weird as it dries (which it does fast), and if sprayed lightly with Windex can do just that same sort of spot.

I hate blue hands, but I may go back to HiSpot blue just because of the way Canode acts for me.

 

[Bakafish]

I have noticed though that when printing it is very hard to move the angle plate over the blued surface plate. Is this normal or is the ink too viscous?

The hinging step should be done with no ink and both surfaces as clean as possible. The uneven weight distribution of an angle plate makes predicting the points more complex, but you are essentially trying to ensure that it isn't rotating around the center (which would indicate a spherical convex surface) and that it has equal and opposing rotation points ~1/3 of the width of the object. On a roughly square object like yours, this should be rotated in respect to all four sides to detect a cylindrical curvature, but again the weight distribution has me worried that this may be hard to interpret without a known good example, hence the alternative methods provided.

 

[M Kiser]

If those are the same surface and view, the first one seems heavy in some areas where the second seems light, while the middle is about the same. That could indicate some rocking. The areas are bottom left and top right edges.

And the spots are in some places more like a smear, with streaks. I don't understand that, is it something about the surface? If so, you might need to scrape it straight down all over to get the surface to show nice spots.

What ink? If it is Canode, maybe the streaks are just the ink, it does not seem to thin down like HiSpot, and it gets weird as it dries (which it does fast), and if sprayed lightly with Windex can do just that same sort of spot.

I hate blue hands, but I may go back to HiSpot blue just because of the way Canode acts for me.

I made another pass and things are somewhat better. See pic below. I think in general that it is heavy along the right front and right rear edges, which might explain why the left corner is printing light. The ink is Charbonnel water washable printing ink. Is that like Canode? Would it help to thin it with a very light oil to slow the drying? I really hate the idea of using Dykem and putting up with the mess.

 

[M Kiser]

The hinging step should be done with no ink and both surfaces as clean as possible. The uneven weight distribution of an angle plate makes predicting the points more complex, but you are essentially trying to ensure that it isn't rotating around the center (which would indicate a spherical convex surface) and that it has equal and opposing rotation points ~1/3 of the width of the object. On a roughly square object like yours, this should be rotated in respect to all four sides to detect a cylindrical curvature, but again the weight distribution has me worried that this may be hard to interpret without a known good example, hence the alternative methods provided.

Hinging dry seems pretty obvious in retrospect. What was I thinking? The spherometer idea is pretty interesting. I might just cobble something together if dry hinging is inconclusive.

 

[M Kiser]

Here is a pic of the latest cycle. A bit heavy along the right edges?

 

[TGTool]

You could take a marker and make a U shaped line around the lighter area just right of center, as well as a little along the top edge. This is the exclusion zone

Then go after all the blue spots on the remainder and it may even up the marking for you.

 

[JST]

I don't know anything about Charbonnel. Being water based, I would not think that oil of most any sort would mix well or do anything good.

The HiSpot, as you know, is non-drying, since it has a grease type base, no water. What like about it is that I can put it on the reference once, and keep using it for many cycles. As the cycles refine the surface, the HiSpot gets thinner and seems to generally match up to give good resolution.

By contrast, the Canode gets sort of "chalky" as it dries. I use it still, because it is easy clean-up, but I always wonder if I am getting good information from it, because it always seems much thicker than HiSpot. Is the Charbonnel like that? It looks like it behaves a bit better.

Looks like those stringy "splashes" or whatever they were have cleaned up on the latest prints. I am wondering if that was really just telling you that you had low spots that a somewhat thicker layer of marking was filling in on. I see the middle light area still looks much more like that than the heavy areas.

Can you put the Charbonnel on thinner? Not necessarily thinning the ink, but putting it on the reference in a physically thinner layer, or rubbing it out thinner.

Do you use a roller or a dauber to put it on the reference? I always have used a dauber, so maybe that is why I seem fussier about thickness. Don't know.

That's the thing with HiSpot. It will thin down to a haze as you continue taking prints over time, but the prints are always clearly defined. And you can put on as much or as little as you want, and rub it out to any sort of thickness. Sure is messy, though.

EDIT: With Hispot, and maybe with the Charbonnel (never used it) you can put on a thin layer easily, and then you get a very clear distinction between areas that touch, and areas that "almost but not quite" touch. You can avoid having "droplets" that make the stringy marks.

I would agree to the idea of hitting all the areas except the light areas and see if that evens up the prints in a reasonable time. It should.

 

[M Kiser]

Per TGTool's suggestion, I did establish an exclusion zone and scraped again. The results were inconclusive which got me to thinking about Bakafish's comment about a cylindrical surface. Rather than cobbling a spherometer type device to check it, I set it on the surface plate and shimmed it until I had the scraped surface as parallel to the surface plate as I could. Readings were taken with a tenths test indicator through a gage block at 25 locations in a 5x5 grid. See attached pic. Units are ten thousandths of an inch. I was pretty surprised at how flat it is! Should I proceed to finish scraping?

JST- regarding the Charbonnel ink, I have no experience with anything else so I can't say how it compares to other types. Thinning with light oil provided a nice viscosity but by the time it was rolled thin enough for clear marks there was almost no color left.

When I was getting ready to buy my scraping supplies I spoke with Richard King about what I needed. He was very helpful in steering me in the right direction and recommended the Charbonnel inks. Apparently Canode is not available anymore. He would be a great source of info regarding the characteristics of the various inks. Richard...........are you reading this? Care to give us your two cents worth?

 

[LexD]

Could you post some pictures of the internal mechanism of your power scraper please?

 

[M Kiser]

Could you post some pictures of the internal mechanism of your power scraper please?

I'd be happy to. If you have questions I will gladly try to answer them. My designs tend to make use of whatever I have laying around. The only parts I had to buy were the right angle drill ($45 at Menard's) and the thin section bearing at the large end of the connecting rod. And of course the Biax blade which cost more than the drill and bearing combined. It took a while to design and build.

 

[jwmelvin]

Are those cast parts from 3d-printed patterns, or 3d-printed parts?

The eccentric/crank allows adjustable stroke? What are the notches on the outside of the pivot?

 

[M Kiser]

Are those cast parts from 3d-printed patterns, or 3d-printed parts?

The eccentric/crank allows adjustable stroke? What are the notches on the outside of the pivot?

The covers are 3D printed from PLA resin that were sprayed with silver hammertone Rustoleum. Not a strong resin, but I made the housings pretty thick. If it gets dropped, something is likely to break. But I can just print another if that happens.

Yes, the eccentric allows adjustable stroke. The knurled knob holds the eccentric to the crank. The eccentric has 24 notches that engage 4 pins in the crank. Loosen the knurled knob and the notches disengage from the pins allowing adjustment. The notches were filed to provide grip for my thumb while adjusting. The small end connecting rod bearing has to be a self aligning type to allow the rod to move during adjustment.

 

[JST]

It doesn't matter a lot, I have about a half pint left of the Canode, in two colors. And plenty of HiSpot as well.

There was a discussion a while back, so the two cents are in. But you are actually using the Charbonnel, and we see the results, so it was interesting to see how you had done.

About the pics.... How many passes of scraping do you think have been done since the pics above? There is the one with a number of what looked like streaks in post 1, the one in post 10 where the streaks are much reduced, and the last one one where you marked the tenths? There is a good deal of change between these three pics, and I'm trying to get a handle on what made the change.

 

[jwearing]

There's something about the shape of the blue spots that looks wrong to me. They are long, and the gaps between them are long. JST -- is that what you meant by "streaks"? This is how my prints look when I'm done roughing. I'm guessing that your strokes are too long. Or maybe you're making short strokes but connecting too many of them together. When you're finishing you really want distinct separated strokes. You don't have to scrape off all the blue every time, you can split the spots into smaller ones instead. What's the range on your stroke adjustment?

The other thing is that these three prints are significantly different:
- The first is heavy on the edges, especially the lower left and upper right.
- The second is light on the lower left, pretty even otherwise.
- The third is heavy on the edges again, a little light in the middle.
How many cycles were between these passes? In general are you getting consistent results spotting? If not you need to work that out before moving forward, you could have something thermal going on, or maybe a convexity.

 

[M Kiser]

It doesn't matter a lot, I have about a half pint left of the Canode, in two colors. And plenty of HiSpot as well.

There was a discussion a while back, so the two cents are in. But you are actually using the Charbonnel, and we see the results, so it was interesting to see how you had done.

About the pics.... How many passes of scraping do you think have been done since the pics above? There is the one with a number of what looked like streaks in post 1, the one in post 10 where the streaks are much reduced, and the last one one where you marked the tenths? There is a good deal of change between these three pics, and I'm trying to get a handle on what made the change.

I lost track of how many passes were made. I've set aside the angle plate and started working with some old Brown & Sharpe cast iron indicator bases that I found in my stash of treasures. The uneven weight of the angle plate is a complication I don't want to deal with now. I took a closer look at the two tubes (red and blue) of Charbonnel and realized that the labels are not the same. The red is clearly marked "Etching Ink" and "Aqua Wash" and the blue says "Block Printing & Etching" and "Water Washable". Maybe I have the wrong stuff? I'm going to try the red on the surface plate and see if it is easier to get clear marks without smearing. I've had a really hard time getting the ink thickness right with the blue.