Clearing burrs in small holes--unusual situation

[FromScratch]

I am making some magnetic devices (solenoids) from pieces of steel transformer laminations glued together. They are small, so I'm making rivets from #14 copper wire electrically insulated with a layer of super glue. This requires drilling accurately aligned 1/16" (1.6 mm) holes. Predrilling before gluing looks to be a real headache, but drilling after assembly has a tough problem.
When you drill a hole in solid steel, the burrs just get pushed on through. When you drill a glued stack of steel laminations, however, burrs get shoved out into the glue gaps because the glue is softer than steel burrs. I can't have this because it produces electrical shorts between laminations, which need to be electrically insulated apart to suppress eddy current power loss and heating.
The drill bit doesn't reach these burrs because they have pushed out into the glue to make the hole larger than the drill bit diameter at that point.
I have tried spinning the bit in the hole along with steel wool and sand, hoping the drill bit will spin these materials out into the glue gaps and shear the burrs off, but I'm not having any luck. As I am making devices with over a dozen laminations, it's really hard to see into the hole and hand guide some ultrafine pick type tool along the glue gaps.
Any ideas?? Thanks in advance--this is the toughest problem I've encountered in a very long time.

 

[JimDawson]

I would experiment with some different drill bits. I would start with a 135 degree split point cobalt bit. Then maybe a brad point, not the best for steel but laminations are pretty soft. Higher spindle speed and less pressure might be helpful, but use enough pressure to keep the bit cutting and not rubbing which will form a burr.

The only other thing I can think of is to drill the laminations before gluing, then debur them individually before assembly.

 

[Bi11Hudson]

In a commercial application, the holes are usually punched individually before the plates are varnished. For a home shop, low quantity operation, I would venture to suggest the plates be set up and rigidly clamped, drilled, then glued. With the holes drilled in this method, the only tails would be on the first and last plate. Make the stack full by 2 and discard the outer plates. Or reuse them as "guides" for the next stack.

I would prefer varnish(GE's Glyptol) over glue. But I'm old school about electrical work. When the core is assembled, the plates are again clamped as tight as possible and then pinned. Copper is mighty soft to use as a reliable rivet. It'll work, but I personally would want something a little stronger. In the "old days", screws were common.

The coils I make are much smaller than what you're building. I use aluminium sewing machine bobbins and 1/4 copper tubing to get the strongest field on the steel armature. but then, I only need about 1/8 inch travel multiplied by an offset bell crank.

.

 

[FromScratch]

@JimDawson:
I'll try the drill bits you suggest.

@Bi11Hudson: I may end up doing as you say. It's just going to be a real pain to maintain alignment accuracy after disassembly and reassembly. Core component mating edges need to be finally finished for a very close fit--my solenoids require a travel of less than one mil because they are being used to apply pulsating compression to rocks at the rocks' resonant frequencies, which may be 5 kHz or so. To get high force, the working gap is only a few mils, and gaps elsewhere in the core must be much smaller than that, really pretty much in dark-gap contact.
Actually, my coils may be smaller than yours--the current research prototype has a core cross section of 5/16" square. Hence the 1/16" copper wire rivets instead of readily available screws--#4 is much too large.
I may also end up going out on the net and springing for 1/16" diameter screws and nuts, whatever size that is. If they aren't available in lengths up to 3/4", I may need to buy the corresponding threading die and thread 1/16" piano wire on the ends to take the nuts.
The real stress on the core is not spreading them apart, it appears as a shear force on the rivet, which closely fits the holes. I think it'll be OK. And the rivets are backups for glue joints.
I may try brass wire too, but it may be too brittle to fan out into an acceptable head without cracking. And steel would just require putting too much stress on the core parts, which lose permeability when distorted.

 

[Bi11Hudson]

To put some things in perspective, machine screws are based on a numbering system that starts at 0.060 inch, Size (Nr) 0. Add (or subtract) 0.013 for each size. As in Nr 1 is 0.073, Nr 2 is 0.086, Nr 3 is 0.099, and so on up to Nr 14, at a fuzz under 1/4 inch. Smaller sizes are subtractive. Nr 00 or 2/0 is 0.047, Nr 000 or 3/0 is 0.034 and so forth. There are smaller sizes but 3/0 (120TPI) is the smallest that I use. In that small a size, the metrics are easier to get in quantity.

Then there are the metric sizes, the "almost" or "just about" sizes. There is a 0.7mm that I have occasional use for. The 1.0mm size is fairly common at 0.03937. I teach the smaller sizes as 1mm=~.040. In that size, it's close enough. Much of my usage is 2mm sizes, at ~0.080, just a fuzz smaller than a Nr 2. By the time 3mm is reached, the differential is starting to show, at ~0.018.

There is an archaic system, that is rare at best, of fractional sizes. The only size I have on hand is a 1/16-80 that was acquired by mistake. At 0.0625, it is so close to Nr 0 at 0.060 that the two cannot be distinguished, even by measuring quickly. Only by the fact that the 1/16 will not fit a Nr 0 tapped hole. I know there are(were) 1/32 and 3/32 sizes. But I have never seen any. A 3/16 and a Nr 10 are only a couple thou different. A lot of the hardware coming from generic sources seem to mix the two indiscriminately.

I do use a lot of 1/16 brazing rod. In the lathe, I cut the threads(80 TPI) a little deep. Using a die, by hand, there is a chance of breaking the rod, but it doesn't happen very often. No more so than with sloppy work. Brazing rod looks like brass but is more of a bronze. Tin more so than zinc. Most often I just solder a tab on the end and don't bother threading it. I only thread for fitting parts to a model.

Maintaining registration of the laminations is no big deal if drilled square on a drill press or some other machine. The only factor to consider is time. Take it slow and you shouldn't have any trouble. I have never worked pieces that small, but I have used the lams from transformers for making "stuff".
.

 

[FromScratch]

Thank you very much. You're a treasure trove of valuable information!

I've since considered, as you suggested, soldering a head on one end of a brass rod. Then, squeeze the other end with pliers to flatten it for an instant grab with a little compression, and dab on some JB Weld for a strong permanent head. And come to think of it, I have some brazing rod--maybe even some 1/16". If not, I know where to get it.

I have had to make several custom clamping assemblies for cutting, filing, drilling, and sanding lamination strips only 5/16" wide. Kind of a pain to use, especially getting stacks accurately lined up in the clamp. But glued or not, those lams do need clamping to do anything. I even clamp single lams between sacrificial stock to file them without bending. Silicon steel is a little harder than mild steel, but at the typical .02" thick, it's still fragile.

 

[Flyinfool]

All the transformers and motor field coils we make at work have the lamination stack held in a clamp and then with a tig chase a puddle down 2 opposite corners. Seems to work quite well. Yes the plates are all electrically connected by the weld but that does not make a measurable difference vs when we used to rivet them together. But is sure is faster. I do not know if the eddy currents become more problematic at the tiny scale you are working at. Our lamination stacks were all around 3 inches tall.

 

[FromScratch]

The small scale is actually an advantage--greater surface to volume ratio means easier to cool. However, since I will be running the solenoids at around 5 kHz or more, eddy currents and hysteresis losses will generate a lot of heat. I'm making sure that the laminations don't make contact, in particular at their edges, to minimize eddy currents. I want to clear burrs in the rivet holes but as I study the matter I think they may not make that much difference, since the eddy current loops cannot form over any real extent, with the rest of the surfaces being sealed. Still, I'll probably need to use efficiently designed fan cooling to get the force I want. After all, I can't do anything about hysteresis losses.
I've noticed that transformers have those side welds these days, in particular the MOTs (microwave oven transformers) I'm cannibalizing for lamination stock. I don't have the rig to do that. Maybe epoxy beads? JB Weld sure is strong, and it can take the heat quite easily. A number of narrow beads might do-- need to leave most of the metal exposed for cooling.