The Smell of Molten Projects in the Morning

Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.

Tag: Rants

And kvetching, too

  • Sears Kenmore HE3 Washer Teardown: Tub Damage Assessment

    After removing the concrete weights from the tub, I saw where the piddle of water was coming from: the stainless steel drum (formally: Basket) had been chewing on the plastic Tub for quite some time. That’s most likely the strange new sound Mary heard, but it’s impossible to see the affected area without gutting the entire washer.

    The top left section, below the detergent dispenser.

    Outer Tub Damage - Top Left
    Outer Tub Damage – Top Left

    The top right section.

    Outer Tub Damage - Top Right
    Outer Tub Damage – Top Right

    I assumed the damage was limited to the top section, because the drum would pivot downward under load. That turned out to be incorrect, as I realized when I looked behind the lower concrete weight: the Tub was scored through all the way around.

    Here’s a view of the interior, taken after I removed the Tub from the washer and pried off all the clamps that secure the Front Tub to the Rear Tub.

    Outer Tub Damage - Inside View
    Outer Tub Damage – Inside View

    And a closeup…

    Outer Tub Damage - Inside Detail
    Outer Tub Damage – Inside Detail

    The drum has a ridge around the front circumference where the round perforated shell joins the convoluted front piece that necks down into the opening behind the door. That ridge contacts the plastic Tub and, even though it’s smooth, generates enough friction to burn through the Tub.

    Now, this is the point where the repairman turns to you and says that, although he (it’s always a he) can replace the drum and Tub, the total cost will be more than a whole new washer.

    Sticker price for the parts, direct from Sears, looks like this:

    • Stainless Steel Basket: $364
    • Front Tub: $150

    Onto that, reports from various forums indicate that you will have incurred some labor charges:

    • Labor: $120+ to this point
    • Parts Service Fee: $50 (I don’t know what that is, either)

    A new HE3 washer is $850, more or less.

    Ah, you ask, isn’t there a Lifetime Limited Warranty on the Stainless Steel Drum? Why, yes, there is, and that’s a story all in itself …

  • Sears Kenmore HE3 Washer Drum: The Rot

    Corroded Aluminum Spider - Overview
    Corroded Aluminum Spider – Overview

    Our Kenmore HE3 washer emitted a dramatic KLONK that had all hands racing for the Cancel button. After a bit of to-ing and fro-ing, some Web searching, and a few hours of teardown, I determined that the washer had failed in the usual HE3 way: the cast aluminum spider connecting the back of the lah-dee-dah stainless steel drum to the shaft had corroded and fractured.

    Now, class, let’s review our chemistry. What do we call a pair of dissimilar metals in an ionic solution?

    Very good. Can you spell “battery”?

    Bonus points: what happens to the battery electrodes as the current flows?

    Excellent! I’m sure you can spell “corrosion”, too.

    The stuff that looks (and feels!) like cheese is aluminum corrosion filling every nook & cranny in the back of the spider. The fact that the drum spins at 900 rpm tells you it’s rather tenacious gunk, but evidently we’ve been washing our clothes in corrosion products for several years.

    If you have a Sears or Whirlpool HE washer, so are you.

    Mary noticed the washer made a strange noise during the spin parts of the cycle, starting a few weeks ago, but it wasn’t anything you’d tear down the washer to diagnose. I’ll have more to say about that in a bit.

    The KLONK happened when a third fracture finally disconnected the drum from the shaft and it started whacking against the outer tub. All that’s holding the shaft in place is the remaining thickness of the spider casting and the interlocking fracture pattern; I can move the shaft, but not easily.

    Here are closeups of the three sections near the hub between the arms. Anything that looks like a crack really is one…

    Corroded Spider - Fracture 1
    Corroded Spider – Fracture 1

    The next section has a nice crack running along the circumference, too…

    Corroded Spider - Fracture 2
    Corroded Spider – Fracture 2

    And the third section…

    Corroded Spider - Fracture 3
    Corroded Spider – Fracture 3

    I hauled it to the driveway and hosed off the corrosion. There isn’t supposed to be that little hole where the sun shines through…

    Corroded Spider - After Rinsing
    Corroded Spider – After Rinsing

    The washer is six years old and cost $1100 new.

    Needless to say, We Are Not Amused.

    More on this as I sort things out. Search for HE3 and you’ll find more than you want to know (at least after I’ve gotten it posted).

  • How to Solve a Parking Problem

    The Walkway Over the Hudson has been a resounding success, at least measured by the number of people using it. The Parker Avenue parking lot has about 80 spaces and, during most days, is jammed full.

    The NYS park system now owns the Walkway and, in their infinite wisdom, decided that the parking facilities should have a fee just like the rest of their lots: $5 / 4 hours.

    Here’s what the Walkway lot looked like on the day the fee went into effect…

    Walkway Parking Lot With Parking Fee In Full Effect
    Walkway Parking Lot With Parking Fee In Full Effect

    To quote from the Poughkeepsie Journal:

    State officials hope there will be no decline in visitors with the new parking fee, said State Parks spokesman Kristen Davidson.

    Basically, there’s enough free on-street parking in the area that most folks park nearby and hike in, which makes sense for a park consisting of about two miles of walking path. The parking fee amounts to a tax on handicapped and elderly visitors who find it difficult to navigate streets and ramps.

    On the bright side, it’ll be a lot easier to bike across the bridge…

  • Water Bottle Spring Cap Repair

    One of our nice aluminum water bottles hit the floor and, of course, the tiny little hinge shattered. It’s some wonderful engineering plastic, but just look at the leverage you can apply to those few millimeters of material. This is the sort of repair that can’t possibly be economically justified, but it pisses me right off when something that should be rugged turns out to be this fragile.

    The 2 mm steel hinge pin snapped the molded plastic center post of the hinge off the cap; we found the larger fragment, but the smaller one may lurk under the refrigerator for quite some time. Nothing bonds to this plastic and, if the post broke in the first place, adhesive isn’t going to help.

    Broken hinge
    Broken hinge

    Some doodling showed that a replacement hinge post should be machineable. The general idea was to square up the remaining chunk of the post, then attach a replacement hinge pivot with a screw. The post is almost exactly 1/4-inch thick, call it 6.2 mm, which means the right-angle feature under the pivot ought to keep the whole affair from twisting.

    Water Bottle Hinge
    Water Bottle Hinge

    I planned to leave the left side unmachined and cut it to fit by hand, but then figured, eh, just make it happen. I also expected to leave the area around the screw a lot thicker, with a neat counterbore around the head.

    This being a bash-to-fit, file-to-hide kind of project, I wrote a snippet of G-Code (at the bottom of the post) to chew out the part from a sheet of Lexan, then did the perpendicular hole & countersinking with manual CNC.

    No pix of that; I was working in a white-hot fury. Basically, I double-sticky-taped a slab of Lexan to a sacrificial sheet, clamped it to the tooling plate, and had at it with a 2 mm end mill. Cutting a 6.4 mm sheet with a 2 mm end mill is a bit iffy, as the flutes are just barely that long; the mill was armpit-deep in swarf and I was dribbling water into the cut to keep it cool.

    By the time I stopped for a picture, the situation looked like this.

    Replacement hinge part
    Replacement hinge part

    For what it’s worth, that’s the second part. I had to lower the screw head below the top of the half-round feature on the left end in order to clear the cap. That’s what CNC is really good for in my shop: make another one, just like the other one, only different exactly like that.

    I drilled a #50 (2-56 tap) hole in the cap pretty much by eye, using laser targeting to touch off.

    Laser aligning to hinge stub
    Laser aligning to hinge stub

    The hole wound up minutely too far inboard, but some filing cleaned up the stub edge and it was all good. I started the tap in the mill, held loosely in the chuck and turning it with my fingers, then finished up on the bench.

    The screw hole goes all the way through the cap. I filed the screw down so the end sits flush at the bottom of the cap, where the silicone rubber gasket should seal firmly against it.

    Here’s what the hinge looks like with all the bits assembled. The spring bears on the screw head, which makes the cap open with more snap than before. I put a little counterbore under the screw head, even after lowering it, to reduce the spring tension.

    Rebuilt hinge
    Rebuilt hinge

    The cap has a spring-loaded latch that never worked very well in the first place and this repair didn’t improve it. As nearly as I can tell, the molded ledge on the cap has a rounded edge that the latch simply cannot engage. This is beyond even my level of interest; Mary was accustomed to using the wire snap to hold the cap closed and that practice will continue.

    Works well enough for us and I got some Quality Shop Time on a rainy afternoon.

    The G-Code uses a slightly modified & simplified version of the tool length probe routines. I’m not convinced that using the G59.3 coordinate system is the right way to go, but everything else seems worse.

    
    (Water bottle hinge repair)
    (Ed Nisley - KE4ZNU - June 2010)
    (Rough-cut 1/4-inch plate with clamp at +Y)
    (Sacrificial plate below, double-stick tape to secure)
    (Tool change @ G30 position above length probe)
    
    (-- Global dimensions & locations)
    
    #<_Stock_Thick> =        6.5                    (overall thickness)
    
    #<_Traverse_Z> =        1.0
    #<_Safe_Z> =            30.0                (clamp clearance)
    
    (-- Section controls)
    
    #<_Do_Outline> =    1
    #<_Do_Drill> =        1
    
    (-------------------)
    (-- Initialize new tool length at probe switch)
    (    Assumes G59.3 is still in machine units, returns in G54)
    
    #<_Probe_Speed> =        250            (set for something sensible in mm or inch)
    #<_Probe_Retract> =        1            (ditto)
    
    O<Probe_Tool> SUB
    
    G49                    (clear tool length compensation)
    G30                    (to probe switch)
    G59.3                (coord system 9)
    
    G38.2 Z0 F#<_Probe_Speed>        (trip switch on the way down)
    
    G91
    G0 Z#<_Probe_Retract>            (back off the switch)
    G90
    
    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
    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
    
    (-------------------)
    
    (-- Get started ...)
    
    G40 G49 G54 G80 G90 G92.1 G94 G97 G98        (reset many things)
    
    M5
    
    (msg,Verify clamp to +Y, stock taped down)
    M0
    
    (msg,Verify X=0 at left edge, Y=0 on finished centerline)
    M0
    
    (msg,Verify tool touched off at Z=0 on surface)
    M0
    
    O<Probe_Init> CALL
    
    T0 M6                            (ensure first tool change pauses)
    
    (-- Drill the hinge pin hole)
    
    #<Pin_X> =                7.0
    #<Pin_Y> =                0.0
    
    #<Drill_Dia> =            2.06    (Drill diameter)
    #<Drill_Num> =            46        (Drill number)
    #<Tool_Num> =            146        (Tool number)
    #<Drill_Radius> =        [#<Drill_Dia> / 2]
    #<Drill_RPM> =            3000
    #<Drill_Feed> =         [#<Drill_Dia> * 100]
    
    #<Drill_Depth> =        [#<_Stock_Thick> + 2 * #<Drill_Dia>]
    
    O<Doing_Drill> IF [#<_Do_Drill>]
    
    (debug,Insert Num #<Drill_Num> drill)
    T#<Tool_Num> M6
    
    O<Probe_Tool> CALL
    
    (debug,Set spindle to #<Drill_RPM>)
    M0
    
    F#<Drill_Feed>
    
    G0 Z#<_Traverse_Z>
    
    G83 X#<Pin_X> Y#<Pin_Y> Z[0 - #<Drill_Depth>] R#<_Traverse_Z> Q[2 * #<Drill_Dia>]
    
    O<Doing_Drill> ENDIF
    
    (-- Mill outline)
    
    #<Hinge_Radius> =        3.75                        (half-width of hinge body)
    #<Cutout_Base> =        2.75
    
    #<Cutout_Screw> =        1.50
    #<Cutout_Screw_Y> =        [#<Hinge_Radius> - #<Cutout_Screw>]
    #<Cutout_Screw_A> =        ASIN [#<Cutout_Screw_Y> / #<Hinge_Radius>]
    #<Cutout_Screw_X> =        [#<Hinge_Radius> * COS [#<Cutout_Screw_A>]]
    
    #<Passes> =                3
    
    #<Mill_Dia> =            1.98            (end mill diameter)
    #<Tool_Num> =            20
    #<Mill_Radius> =        [#<Mill_Dia> / 2]
    #<Mill_RPM> =            3000
    #<Mill_Feed> =            100
    
    #<Entry_XL> =            [0 - #<Mill_Dia>]
    #<Entry_YL> =            [0 - 2 * #<Hinge_Radius>]
    
    O<Doing_Outline> IF [#<_Do_Outline>]
    
    (debug,Insert #<Mill_Dia> mm end mill)
    T#<Tool_Num> M6
    
    O<Probe_Tool> CALL
    
    (debug,Set spindle to #<Mill_RPM>)
    M0
    
    F#<Mill_Feed>
    
    G0 X0 Y[0 - 2 * #<Hinge_Radius>]    (get to comp entry point)
    G0 Z#<_Traverse_Z>
    
    G42.1 D#<Mill_Dia>                                    (cutter comp right)
    G1 X#<Pin_X> Y[0 - #<Hinge_Radius>]
    
    #<Step_Z> = [#<_Stock_Thick> / #<Passes>]
    #<Current_Z> = [0 - #<Step_Z>]
    
    O<Outline_Passes> REPEAT [#<Passes>]
    
    G2 J[0 - #<Hinge_Radius>] Z#<Current_Z>                (ramp down to cutting level)
    
    G3 Y#<Hinge_Radius> J#<Hinge_Radius>
    G3 X[#<Pin_X> - #<Cutout_Screw_X>] Y#<Cutout_Screw_Y> J[0 - #<Hinge_Radius>]
    G1 X0
    G1 Y[0 - [#<Hinge_Radius> - #<Cutout_Base>]]
    G1 X#<Pin_X>
    G1 Y[0 - #<Hinge_Radius>]
    
    #<Current_Z> = [#<Current_Z> - #<Step_Z>]
    
    O<Outline_Passes> ENDREPEAT
    
    G0 Z#<_Safe_Z>
    
    G40
    
    O<Doing_Outline> ENDIF
    
    G30                    (back to tool change position)
    
    (msg,Done!)
    
    M2
    
  • Epson R380 Printer: Resetting the Waste Ink Counters

    So a few days after topping off the continuous ink tanks on my Epson R380 printer, we had a series of thunderstorms that prompted me to turn everything off. Upon turning the printer back on, its fancy LCD panel showed a message along the lines of

    Service is required. Contact Epson Customer Service.

    Oddly, it continued to print perfectly with no further complaints. The error message appeared only at power-on, then politely went away when I pressed the OK button.

    Well, that puppy is long out of warranty, even if I wasn’t using a continuous ink system, soooo… what to do? The printer produces absolutely no diagnostic codes other than that error message.

    A bit of searching gave me the Maintenance Manual for that family of printers. That message isn’t among the ones listed.

    Further searching suggests that at least one of the two waste ink pads / tanks is nearly full and that ignoring the problem will cause the printer to shut itself down, lest it dribble ink. The listed messages warn that the printer is approaching the “end of its service life”, which isn’t the message I saw, but it’s close enough.

    The Maintenance Manual suggests that it’ll be cheaper and better to simply buy a new printer, as replacing the waste ink tanks may cost more than the printer is worth. The website points out that providing a customer-replaceable tank would drive up the cost of the printer, because most customers would buy a new printer before filling the tank.

    In order to get to the waste ink tank, you must remove:

    • Paper Support
    • Printer Cover
    • Front Cover
    • Right Housing
    • Left Housing
    • USB Housing
    • Upper Housing
    • Panel Unit
    • EMI Frame

    I can see why it might take a trained tech a few hours to get all that done… and then reassemble in reverse order.

    The Epson website has a link to a program that will reset the waste ink counters for one of the tanks. Downloaded & ran it on the Token Windows Laptop; it tells me there’s no problem.

    Hmmm

    So I ordered an external waste ink tank from the usual eBay supplier. The hardware is grossly overpriced ($20 delivered) for what it is (large tube with sealed endcaps, some tubing & barb fittings, a syringe), but the deal includes links to programs that will reset the counters. I found several of those programs by myself, so it’s not as if you must actually spend money to reset the printer’s counters. I figured this was in the nature of a learning experience.

    Turns out that the programs are provided by parties having, shall we say, long-term interests that may not coincide with mine. To wit, I’d be batshit crazy to run those programs on a PC I cared about.

    [Update: Something like that.]

    The various program files all passed a ClamAV virus scan, but that doesn’t mean anything these days.

    So, during the next hour:

    • Boot System Rescue CD on my oldest Token Windows Laptop
    • Run partimage to back up the Windows partition to another partition
    • Disconnect from the house LAN
    • Reboot in Windows, which evidently hasn’t seen the light of day in about a year
    • Stifle bleating requests for updates
    • Copy the programs from a USB stick, install as needed
    • Reset one of the ink counters (more on this below)
    • Reboot in SRC
    • Restore the partition from the backup

    All that is straightforward and I’ve written about it earlier. Search the blog for more info using the obvious keywords.

    I attempted to restore the drive’s Master Boot Record from the partition backup file, but partimage complained that the drive size in the backed-up MBR did not match the existing drive size, which suggests something tinkered with the drive’s MBR between the backup and the restore.

    Hmmm….

    You might want to do a bit of reading on Boot Sector Viruses at this point. I have no other evidence to suggest that’s what’s going on, other than to remind you that programs need not do only what they say they’ll do.

    Given all that, I figured this was a great time to update the Token Windows Laptop to Xubuntu 10.04, which installed Grub2 in the MBR and wiped away anything placed therein. The box is heavily multi-booted: Dell Diags, XP, Puppy, and now Xubuntu 10.04.

    Without naming names or providing links:

    • The Russian program seems to not include the R380, but it does include others in that family. I elected to not reset the counters using that program.
    • The Chinese program seems to be a bootleg copy of the Official Epson Adjustment Program, although it’s rife with misspellings and grammatical errors. I told it to reset the “Main Pad” counter and give me a dump of the EEPROM.

    The Main Pad had 16008 counts of the maximum 16200, while the Platen Pad had only 3019 of 54513. Those names do not correspond to anything in the Maintenance Manual, but I suspect the Main Pad is the Waste Ink Tray at the head-cleaning station and the Platen Pad is the Waste Ink Pad running across the printer to catch the overspray from borderless prints.

    Resetting the Main Pad counter to zero cleared the error message; the printer is perfectly happy now. I’ll install the external waste ink tank when I clear the workbench after building the next GPS interface for our HTs.

    The program reported 9922 pages printed. Figuring 7 bottles of ink at 250 ml each, that’s 0.18 ml per page. That’s a slight overestimate because the ≈50 ml tanks were just topped off, but it’s close enough. I’m guessing head cleaning consumed much of that ink, as the printer does plenty of that, and the number of pages seems close to half the number of counts.

    Perhaps it performs a cleaning when more than X minutes has elapsed since the previous print job? That would account for the high number of cleanings; most print jobs are a few pages, at most.

    En passant, I found some totally unofficial ink cartridge capacity numbers:

    • Standard T078x: 7 ml @ $13 = $1857 / liter
    • Large T077x: 11 ml @ $20 = $1818 / liter

    [Update: corrected typo from ml to liter]

    Ain’t that impressive? I love the savings they give you with higher-capacity cartridges …

  • Unit Pricing Obfuscation: Nothing Exceeds Like Excess!

    Walmart Tissue Unit Prices
    Walmart Tissue Unit Prices

    What with this being allergy season, my ladies blew through our tissue stockpile in short order and it’s time to reload. Give that we’re just blowing our noses it in, deluxe-edition tissue paper isn’t a priority, but these Wal-Mart unit-price stickers are not nearly as helpful as they could be…

    Hints:

    • The second label is for a shrink-wrapped block of three boxes.
    • The bottom label is for a name-brand tissue that’s unit-priced per sheet.

    Exercises:

    • Which container has the least expensive sheets?
    • The most expensive?
    • Is the 3-pack more or less expensive than a 1-pack?

    Essay: why do you think Wal-Mart does this?

    More unit pricing grumbling.

  • New Tires for the Van: Overtightened Lug Nuts

    The shop spec says the lug nut torque shall be 104 newton·meter or an equally odd 77 lb·ft. Let’s not get into quibbles about the differences between lb·ft and ft·lb here, OK?

    Anyhow, based on the wildly differing and grossly excessive tire pressures left by the guys who installed the new tires, I figured the lug nuts would be over-torqued… as, indeed, they were. My bending-beam torque wrench goes up to 140 n·m and didn’t even come close to breaking those puppies loose.

    So I deployed a manly breaker bar and applied most of my weight to the far end. A back of the envelope guesstimate says they were well over 200 n·m, with a few grunt outliers.

    Yes, the breakaway torque can be higher than the tightening torque, but they were far beyond even that level.

    Lubed the threads, tightened to spec, and it’s all good. I’ll check them next week just to be sure, but sheesh if we had to fix a flat on the road, it would have gotten ugly.