2019-12-22

Low-voltage Christmas tree lights

Battery-powered Christmas tree lights bought at Intermarche for ~$4.
50 white LEDs are connected in parallel, battery holder for 3 AA cells contain also mechanical switch and resistor in series. LEDs seem to be pretty well matched, with no observable differences in emitted light.


When powered from 3 NiMH cells this draws ~60mA. When powered from 2 NiMH cells (one pair of battery contacts shorted) - about 2.5mA, with very low light intensity.
These lamps have nominal parameters specified as 3V/20mA, but it does not make much sense with battery holder for 3 cells.
I have to warn that this one was dead on arrival with cable not soldered to contact. Quality control might be not perfect, but it is still great value if you have soldering iron prepared.
As improvised power source I've used old Nokia charger (ACP-7E) with additional diode and 27Ohm resistor in series.

2019-12-10

Fixing "The Application was Unable to Start Correctly (0xc000007b)" error


  1. Check if executable is 32-bit or 64-bit.
  2. Download http://www.dependencywalker.com of same type
  3. Open executable in Dependency Walker. Most likely there would be error(s), but it is not important. The important thing is list of DLLs:
Look at the "CPU" column for red text - this is missing DLL. In this case - vcruntime140.dll, part of Visual C++ 2015-2019 runtime for x86. Missing VC++ runtime libraries are quite often, in general, depending on dll number:
"80" -> VC++ redist 2005
"90" -> VC++ redist 2008
"100" -> VC++ redist 2010
"110" -> VC++ redist 2012
"120" -> VC++ redist 2013
"140" -> VC++ redist 2015 (or 2015-2019)

2019-09-28

Dell E6440

Keyboard: from my perspective worse than in T500. Finding F5 of F9 without looking would not be easy. No separated Del/Home/End/PageUp/Down block. Similar, yet somehow less intuitive cursor block indentation. The problem is no one if making "proper" keyboards since T420/T520. On a plus side: separated volume up/down buttons - something that is missing in T440/T550 (equivalent generation from Lenovo).
One pretty annoying thing is squeaky sound from right Shift button - I'm guessing this button might be slightly too wide for used mechanism.
Bottom cover is made from metal, but when detached it is rather thin and flimsy - this is not like T500 back cover that could basically be skeleton. It is held by 4 screws (3 on the HDD/left side and in the center) and - at least on my unit - it does not fit perfectly on right side. I wouldn't mind few extra screws.



Rather weird looking speakers, with two membranes each. Magnets are quite strong and they are catching metal dust. One annoyance related to built-in speakers is pretty loud pop sound each time integrated amplifier goes to active state. Fortunately I'm not observing this when using active speakers connected to headphone output.
The good: battery life at low backlight level seems excellent - for my 48Wh (effectively) battery work time is estimated as 9 hours.
I hate that battery is not identical to battery from E5440 - I have spare battery of this type, it has the same electrical connector, almost identical size, but different notches and latches in plastic case are making it electrically incompatible.
It is very easy to recover Windows - Dell supplies tool that creates bootable USB.
Buttons for trackpoint are disappointing (buttons below trackpad are OK) - they need to be pushed all the way down but they are flat and have no tactile feedback.
My previous laptops were Thinkpad R500 and T500 and comparing to them E6440 seems little loud and hot - fan is spinning in situations when it was idle on thinkpads. Currently I'm using ThrottleStop I'm experimenting with undervolting using ThrottleStop with voltage offset -95mV for CPU Core, Intel GPU and also for CPU Cache. This might be helping a little, but R500/T500 still were better.
Display viewing angles are poor, but on the plus side there is no visible flickering starting from third (out of 16) level of brightness. Brightness is also noticeably higher than on '500 series thinkpads.

Note: to get service number run wmic bios get serialnumber

Aneng AN8008 multimeter

Aneng AN8008 multimeter - 9999 counts.
 Differences from AN8009:
  • no NCV (Non-Contact Voltage detection)
  • no temperature measurement
  • square wave generator output (not present in AN8009)


Bottom part looks like rubber holster, but it is actually single piece of plastic.


ICL8069 as reference voltage source.

24C02 - most likely for calibration storage.
PCB markings: ZT109 V0.4 170213 => http://zotektools.com/products-2/zt109/
Range switch is not the best I've seen.
Buzzer is latching and reasonably quick, but personally I would prefer "analog" one.
My other complaint would be rather poor display visibility when looking from angles above it (from below it is fine).
Note: auto power off can be disabled by holding SEL/HOLD button while powering on. Multimeter would still beep after each 15 minutes or so as when auto power off normally would be about to trigger.

  • 1 VAC is 5% down at 2.2kHz (RMS will not work at the frequency)
  • At 1Vrms input frequency range is from 1Hz to 3MHz
  • Input impedance is 10Mohm on DC and AC
  • mV range is high input impedance for DC up to about 2V then it drops to 2kOhm, mvAC is similar, except it is capped with a 10Mohm resistor
  • Frequency input needs about 600mVrms to work at 1kHz.
  • Frequency counter and duty cycle works without zero crossing in Hz range.
  • Duty cycle works from 2% to above 99% at 100kHz with 2Vpp (1Vpp do not work), precision is within 0.7 (About 0.6 too low).
  • Hz input impedance is 10Mohm up to about 12V where it is clamped after the PTC
  • 10A range is badly calibrated at high current
  • 10A range will change some percent at 5-10A current due to heating
  • 10A range will give an audible alarm when current is above 10A (uA will not)
  • Ohm needs about 2.5s to measure 100ohm
  • Ohm voltage is 1.0V open and 0.42mA shorted (same as continuity)
  • Continuity is fast (About 20ms); note: it actually seems slow to me and I would prefer continuity test from cheaper meters
  • Continuity beeps when resistance is below 50ohm.
  • Diode voltage is 3.2V with display of up to 3.000V with 0.2mA, maximum current is 1.8mA shorted
  • 70000uF takes about 8 seconds to measure.

2019-09-18

Disabling HDD APM

APM (Advanced Power Management) feature is quite problematic for my 2.5" laptop disk - 500GB HGST HTS545050A7E680. Depending on disk activity it might unload heads way too often - every minute or so. This is annoying (clicking noise), increases mechanism wear and probably even increases power usage when it semi-randomly parks head just to activate again few seconds later. I have no idea why so short idle time was selected as default in firmware - just an error plus lack of testing or planned obsolescence.
CrystalDiskInfo can be used to disable APM (menu Function/Advanced Feature, AAM/APM Control), but it gets little tedious as this has to be repeated each time disk is powered on and requires few clicks each time.


Here is another tool - APMdisable.exe from https://www.codeproject.com/Tips/808412/Disable-APM-for-Hard-Drive-in-Windows. Changing disk parameters require administrator rights - set this in file properties.
To run APMdisable.exe each time disk is powered on I've used Task Scheduler, creating new task triggered by "On workstation unlock of any user" event - as described on https://superuser.com/questions/15596/automatically-run-a-script-when-i-log-on-to-windows:
 Important: "Run with highest privileges" - set this also in APMdisable.exe properties.
Add also second trigger: "At log on"/"At log on of any user" - apparently logging first time after restart does not count as "On workstation unlock".
 Command line arguments: "-dis"
Starting both when using battery and AC power.

Also: set "Require a password on wakeup" in Control panel -> "Power Options".

2019-09-14

Halogen oven

Cheap (~$27) halogen oven. 220-240V, 1200-1400W, single circle-shaped heating element.
Nominal capacity 12l, with extender ring - 17l.













Bottom of the bowl diameter: ~28cm.


 It seems like it also suitable for baking (about 45 minutes for plum cake, using higher grate, 170C, lowered to 150C after about an hour).


2019-08-17

Electric kettle repair

Cheap electric "cordless" kettle with round base, sold probably under hundreds of brands.
Advertised and looking like made from brushed steel, although weirdly this is non-magnetic metal.
Symtoms: randomly not working at all (no neon light near the switch, no heating). Mischievously working once or twice after disassembling and reassembling.

 Faulty element: electric switch with integrated thermal switch (aka thermostat switch) - original marking DY-03G T125.
 It might be not worth the effort, but replacement switch is easily available on ebay (TM-XD-3 T125 and various other names), costing $0.99 with free shipping. Disassembling, replacing switch and reassembling kettle is pretty straightforward.
This metal plate at the bottom of the switch is actual element responsible for switching off when boiling starts - it pushes plastic rod inside when temperature arises.

Kettle stopped working again after two years, this time thermal switch at the bottom failed: LIANG JI LJ-06. There was no continuity between central pin and "L" connector. Actually I'm not 100% sure if it was switch that failed or if metal bottom of the kettle got deformed as slightly loosing nuts that are holding are switch was restoring continuity. I don't think this is permanent solution though and it is very likely that it would fail again at least expected moment.




Replacement switch does not seem to be available at reasonable price (price + shipping would be basically same as price of second similar cheap kettle).

Given my previous good experience with Strix-based product this was one of my main criteria when choosing new kettle (here: KeAi-series).
Plug/switch (Strix R-Series) from older kettle, used for nearly 20 years. I've stopped using this kettle because plastic parts started to deteriorate, switch is still fine.