Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.
Some diligent rummaging turned up a trio of heatsinks that I think will work nicely for the Hot Box Disinsector. As nearly as I can tell from the date codes, they started life at IBM in the early 70s and are built to take a direct hit; the bent fins show they’ve taken a few shots along the way. Those clips applied direct clamping pressure to the transistor cases: much better than screws that can will eventually deform the aluminum and stop forcing the case against the heatsink.
Heatsink with TO-3 transistors
To judge from the crud built up on the fins and the fact that some of the transistors are now completely open, these puppies were run hard and put up wet.
I wiped off the heatsink grease, cleaned off the bigger chunks of crud, and popped them in the dishwasher for a good scrubbing.
Heatsinks ready for reuse
Lovely! The web across the middle is 1/4-inch thick; they don’t make ’em like that any more. The rear heatsink lost its mounting flanges along the way; I have no idea if I sawed those off or somebody else got there first.
I’ll probably plug the holes, just because it’s a nice idea; the sinks are certainly overqualified for their next job as they stand. TO-3 transistors have power ratings over 100 W: these are at least 300 W heatsinks, albeit with an unknown operating temperature. Egad & similar remarks.
In round numbers, I need maybe 150 W of heat at 140 °F ambient. Each heatsink will dissipate 50 W, which is certainly well under its original rating. Figuring 0.5 °C/W sink-to-ambient (call it 1 °F/W) with decent air flow, dissipating 50 W will raise the heatsinks to 190 °F.
That’s higher than I want, so more doodling & measurements are in order.
I want to measure the air flow from some fans, which means I need an air flow straightener to smooth out the wind enough to make the numbers less error-prone. You can, of course, buy cute little straighteners that bolt onto the outlet side of the fan, but what’s the fun in that?
Air flow straightener – overview
The general idea is to pass the air through a set of thinwall tubes to damp out the turbulence. A downstream gap between the fan outlet and the passages eliminates / reduces the dead spot caused by the fan rotor. About 1 diameter downstream of the tubes, the air flow becomes reasonably uniform and a few more diameters produces the familiar parabolic velocity profile found in HVAC ducts.
A few minutes with a bandsaw extracted a 2-diameter-long tube from a 4-inch diameter heavy cardboard mailing tube. A pull saw and a miter box converted some surplus cigar tubes (which I got a long time ago for just such an occasion; I’m not a cigar smoker!) into 3-diameter lengths. Lay as many cigar tubes into the mailing tube as will fit, jam in one more, and they’ll remain in place with sufficient tenacity for my purposes. I suppose, if you were fussy, you could dribble in some adhesive.
I pushed the cigar tubes to the middle of the mailing tube, mostly because that seemed sensible. As nearly as I can tell, this is one of those things where it’s easy to get a reasonable result (as witness the variety of straighteners used by overclockers) and nearly impossible to get a truly trustworthy quantitative setup (as witness the bizarre vanes used in real wind tunnels by actual engineers). An overclocker discussion lives there.
Air straightener – cigar tubes
A quartet of board spacers screwed into 90-mm (92-mm, whatever) fan fit neatly around the mailing tube’s OD, where I simply hot-melt-glued them into place.
Air flow straightener – fan mount
A cardboard gasket seals off the gaps between the fan and the tube.
Fan gasket in place
The gasket looks like this; the next time I will print this picture and cut it out, rather than repeating some fussy layout and getting it wrong twice. Scissors around the outside, a hollow punch for the four screw holes, and a razor knife for the interior. I considered a CNC project, but …
Air flow straightener gasket
And then it Just Worked.
The “before” flow, measured about 1 diameter downstream of the bare fan standing in mid-air, ranged from 0.8 to 1.4 m/s, with the expected completely dead zone in the center. The “after” flow, 1 diameter downstream of the tube, was 0.9 to 1.1 m/s across the entire width, with no decrease in the middle.
The cross-section area is 12.5 in2 and the flow is maybe 40 in/sec, so the fan is pushing 17.5 ft3/min. More or less, kinda-sorta; it’s a quiet CPU case fan from an ancient Dell PC. I have a box of 60 cfm fans arriving shortly, so we’ll see how they stack up.
The anemometer is a La Crosse EA-3010U, which may be the wrong hammer for the job, but it doesn’t require me to dope out a hot-wire anemometer just to get a few numbers…
CDs being such a useful source of raw material, I cooked this up on the lathe while puttering around in the shop thinking about something else. The general idea is to align a short stack of CDs on the Sherline rotary table, close enough to the center, so that you can gnaw away on the top platter and get nearly concentric results. If you really care about concentricity, this isn’t the way to go, but …
CD adapter bushing in place
The washer clamps the CDs in place with the bushing sticking up a bit from the top layer, so it need not be more than eyeball-aligned; the air gap eliminates the need to get the bushing height Exactly Right. If you’re perpetrating fancy machining on the CD, you probably want a form-fitting metal plate atop the stack to hold it down near the perimeter to prevent getting swarf jammed underneath. Note the stack of washers require to reduce that gaping hole to meet a 3/8-16 bolt threaded into the table.
All by itself, the bushing looks like this:
CD Adapter Bushing
The dimensions, roughly as-built:
Rotary Table CD Adapter Bushing
I used a random plastic cylinder from the scrap pile and cleaned up the edges with a razor knife. Next time, I’d put the fat end near the lathe tailstock, so as to make the chamfer easier.
After doing a repeatability test immediately after screwing the new switch to the tooling plate, I let everything sit overnight and ran the test again. In between, I’d done a few small moves, but didn’t change any of the mechanical positions.
The initial position is 0.07 mm, about 3 mils, higher than before, which may well be due to the limited amount of fiddling I’d done in between.
The corresponding picture shows that the values are well and truly quantized to far fewer positions than the number of digits would lead you to believe:
What’s of interest is that the regression line is perfectly flat, which means the switch has pretty much stabilized. I have absolutely no reason to believe it’s repeatable to anywhere near that accuracy, particularly from day to day, but the switch is normally used to set tool lengths relative to a specific tool that’s touched off against the work surface at the start of what passes for a machining job around here.
This relay-like object appeared while shoveling off the Electronics Workbench. Most likely, it started life in the white-goods world, where recurring cost is everything:
Original relay
Now, doesn’t that look just like a tool length probe? It’s certainly less hideous than the one that’s been working fine on my Sherline mill, ever since I figured out how to make tool length probing work.
Here’s what caught my eye:
Plenty of switch overtravel
Nice metal bracket with screws
All the vital pieces in one convenient assembly!
Some brute force removed the spring and actuator, a few shots with a chisel broke the adhesive holding the coil in place, and this collection of parts emerged relatively unscathed:
Disassembled relay parts
Another shot with a pin punch removed the post from the frame. I intended to un-bend the L-shaped feature that held the post, enlarge the hole, and screw it to the mill. Alas, they formed the angle by notching the steel and it cracked when I un-bent it. No great loss.
The two bumps on the frame held the (now defunct) restoring spring. I simply filed those off while cleaning up the broken edges.
Drill a 10-32 clearance hole, solder a cable with a 3.5 mm stereo plug to the switch, add a plastic cable clamp, screw it to the end of the tooling plate, and it’s all good. That’s the butt end of a broken 2 mm end mill poking down from the spindle…
A G-Code routine that displays the Z-axis coordinate where the switch trips looks like this:
(Tool length probing test)
(--------------------)
( Initialize first tool length at probe switch)
( Assumes G59.3 is still in machine units, returns in G54)
( ** Must set these constants to match G20 / G21 condition!)
#<_Probe_Speed> = 400 (set for something sensible in mm or inch)
#<_Probe_Retract> = 1 (ditto)
O<Probe_Tool> SUB
G49 (clear tool length compensation)
G30 (move above probe switch)
G59.3 (coord system 9)
G38.2 Z0 F#<_Probe_Speed> (trip switch on the way down)
G0 Z[#5063 + #<_Probe_Retract>] (back off the switch)
G38.2 Z0 F[#<_Probe_Speed> / 10] (trip switch slowly)
#<_ToolZ> = #5063 (save new tool length)
G43.1 Z[#<_ToolZ> - #<_ToolRefZ>] (set new length)
G54 (coord system 0)
G30 (return to safe level)
O<Probe_Tool> ENDSUB
(-------------------)
(-- Initialize first tool length at probe switch)
O<Probe_Init> SUB
#<_ToolRefZ> = 0.0 (set up for first call)
O<Probe_Tool> CALL
#<_ToolRefZ> = #5063 (save trip point)
G43.1 Z0 (tool entered at Z=0, so set it there)
O<Probe_Init> ENDSUB
(--------------------)
( Set up length)
G21 ( metric units)
(msg,Verify G30.1 above tool change switch, hit Resume)
M0
(msg,Verify blunt tool installed, hit Resume)
M0
O<Probe_Init> CALL
(debug,Initial Z trip = #<_ToolRefZ>)
O100 REPEAT [10]
O<Probe_Tool> CALL
#<DeltaZ> = [#<_ToolZ> - #<_ToolRefZ>]
(debug,Z trip=#<_ToolZ> DeltaZ=#<_DeltaZ>)
O100 ENDREPEAT
M2
Notice that the results have six figures after the decimal point, but they’re really less precise: you’ll find four pairs of duplicates, which seems highly unlikely. I think the values are quantized to about 25 µ-inch and displayed as whatever the metric equivalent might be.
The corresponding plot looks like this:
Probe Repeatability
The trend line is highly suspect, but the slope shows that the trip point gets lower by one wavelength of violet light (393 microns) per trip. The total difference is a whopping 0.004 mm during the test, call it 160 millionth of an inch.
Both of those are better, by roughly a factor of two, than the previous probe switch.
Bottom line: That’s OK for the sort of machining I do… ship it!
This Peltier cooler just emerged from a pile o’ stuff on the Electronics Workbench, so I combined it with a scrap CPU heatsink (using plain old water as “thermal grease”) and fired it up to get some quick numbers for future reference.
Peltier cooler test lashup
It draws 3 A (the bench supply’s current limit) at 5 V. The cold side got down to 19 °F with the hot side at 75 °F: ΔT = 56 °F.
That’s with zero thermal load, other than whatever arrives from plain old air and those two plastic clamps. It looks like a nice one, so it’s maybe 10% efficient and could pump a watt, barely enough to cool a simple circuit.
Freezes a drop of water just fine, though.
The I-V curve is nearly bar-straight over the first five volts: call it 620 mΩ. The thing would draw 7.5 A at 12 V, call it 90 W, and could pump maybe a whopping 9 W from the cold side.
Actually getting good numbers would require some serious work that I’m not up for. In particular, everything has a serious temperature coefficient, so nothing would be the way it looks. I have doubts about the efficiency guesstimate; I’d like to actually measure that sometime.
But it confirms my opinion of Peltier coolers between hundred-watt CPUs and water-cooled heatsinks: pure delusion.
So I swapped in the snow tires and did the fall oil change a few days ago. Everything went smoothly, although the oil filter, as usual, blooshed oil over the front of the engine and, despite my padding the area with rags, onto the exhaust plumbing.
Digression: I don’t understand why the Toyota engineers felt they had to tuck the oil filter below the exhaust header, behind the front downpipe, and over the flexible coupling to the forward cat converter, with the mounting tube pointed upward. It might have something to do with their rotating the entire engine rearward to get a lower hoodline. It seems to me that angling the filter so it can’t drain and must dump its contents atop the exhaust system isn’t Good Design; I’ve been muttering about it for the last decade.
Anyway, the new filter screwed on easily, its seal ring (seemed to) seat against the block, and one final turn snugged it up just fine. The last fraction of that turn felt gritty, as though part the shell kissed the block, but I attributed that to the fact I was using a different filter style.
I added the usual 5 quarts of oil, wiped up the spills, cleaned off the exhaust pipes, declared victory, called it a day, and put away the tools. Later that evening, I checked for leaks, found nothing, and we drove to a meeting about 12 miles away. As you might expect, the van smelled strongly of hot oil: you cannot wipe all the oil off those pipes.
Oil trails on driveway
The next morning, Mary drove to an all-day class about 15 miles away and, about noon, I rolled out my bike to go grocery shopping… only to discover what you see in the picture (minus the sawdust patch) on the driveway.
This is what we call in the trade A Very Bad Sign.
There are three oil tracks:
Right-front track = outbound to evening trip
Rightmost heavy track = return
Leftmost track = outbound to morning trip
Now, the fact that there’s no huge oil slick means the drain plug is in place and properly sealed. The oil evidently leaks out only under pressure, so the filter isn’t sealed against the block. This can be due to a number of causes, the most common of which is leaving the rubber ring from the old filter stuck to the block. I checked the old filter, which was still in the trash: the seal was still in place, so that wasn’t contributing to the problem.
Regardless, the car was bleeding to death. I called Mary and she reported a dry dipstick.
So I loaded a 5-quart jug of oil into the right pannier, dumped all the tools that might possibly come in handy into the left pannier, topped both off with many rags, stopped at an auto parts store along the way for a new filter, and rode those 15 miles at a pretty good clip. When I got to the parking lot, it was easy to find the van: simply follow its trail. The van sat atop a disturbingly large slick, evidently caused by oil draining off every local minimum inside the engine compartment and under the forward half of the chassis.
The filter was still firmly screwed in place, but when I got it off and compared it with the new filter, they were different: the offending filter was slightly larger in diameter and the threaded hole was noticeable larger. Although it threaded on, the threads weren’t properly engaged, the larger diameter shell did hit the engine block, and it most certainly wasn’t sealed properly.
I installed the new filter, poured in 3 quarts to the get the oil level midway into the dipstick’s OK range, wiped off some of the oil that coated essentially every part of the engine compartment, and we drove home trailing a cloud of hot oil fumes.
As it turned out, the old filter was the same brand as the one that didn’t seal, but with different numbers and a different prefix: the correct filter is a 3614, the wrong one was 3593. Of course, the boxes and illustrations are identical, with slightly different contents. I’m sure they’re adjacent on the shelf and migrate into each other’s slot. It’s worth noting that the filter I bought while on the way to fix the problem was a different brand sporting a part number totally unrelated to 3614.
The butt end of the van was covered with oil, as though the droplets blew out under the chassis and got sucked up against the rear surface; the window was a mess. I sprayed on stout detergent and wiped it clean, but I think we must treat the poor thing to an all-over car wash with the special undercarriage scrub option.
No harm done, as nearly as I can tell, although it’s an exceedingly good thing we weren’t driving off to the grandparents!
The Smell of Molten Projects in the Morning 