I have a Korg N1 synth, which I’ve had for many years. Recently, it started having strange symptoms: after warming up for several minutes, the audio would get slightly distorted, but only on certain notes. The effect was rather subtle, but enough to drive my wife absolutely bonkers.
First, I tried to isolate the problem. I tried headphones, connected to the jack on the front left of the keyboard, to rule out my amp. The distortion didn’t go away. Next, I compared the output from the rear audio jack to the headphone, and the distortion sounded exactly the same.
I first tried google, to find out if this was a common problem. It turns out that the N1 wasn’t particularly popular, so there wasn’t much info on problems/repairs, not even teardown photos.
OK, time to take it apart. I first took the screws off the wood side panels, and removed the metal cover on the front, under the keys. Totally the wrong approach. Upon looking at it more closely, I noticed that the bottom is hinged. All you have to do is take out the big philips screws, and then it flips open.
Once I had access to the inside, I looked/smelled for any burned/damaged parts, loose wires, and bad solder joints. Nothing looked amiss. Another thing that’s a common failure point in aged electronics is aged electrolytic capacitors. They often go bad. The most obvious clue is that the tops will bulge up or burst open. I couldn’t see any that looked obviously bad, so I started touching the tops of the SMT electrolytics to feel if any of them were bulging. I found one that had an almost imperceptible bulge on top. It’s in the photo below:
The bulge is so slight, you can’t even see it in the photo above, but I could just barely feel it. Surely, this wasn’t the bad component? Since I couldn’t find anything else wrong, I decided to try replacing it.
At first, I was going to try to remove it w/ a hot air gun, but it’s so big, and close to other components, that I decided that was too risky. I searched the Internet again, and someone suggested just cutting the can in half w/ some cutters, and then yanking out the rest. It turned out to be an easy and safe method. And, it left two short protruding wires attached to the circuit board, which made it easier to solder in a new cap. (Sorry, I should have taken photos, but forgot).
Anyway, here’s what it looks like w/ the replacement soldered in place:
I was extremely skeptical that replacing this capacitor would fix the distortion, but amazingly, when I turned it on, and left it for an hour, the problem was gone! It’s been over a month now, and there is no more distortion, no matter how long it’s turned on. Amazing.
Over the past couple of years,my Weller WESD51 soldering station had been getting progressively flaky. Sometimes, I would have to fiddle with it for a while to get it to heat up. Finally, one day, it just stopped heating altogether. I had trouble finding info on how to fix it, mainly because I couldn’t even find the pinouts for the soldering pencil attachment jack.
Finally, I hit the jackpot last week. I found this thread in AllAboutCircuits discussing a similar model, the WES51. The main difference between the WESD51 and the WES51 is that the WESD51 has a digital temperature display, while the WES51 only has a status LED. While the thread didn’t tell me how to fix it, I found the user/troubleshooting manuals attached! I have linked the manuals at the bottom of this post. Unfortunately, following the troubleshooting guide didn’t help me find the problem, because everything checked out OK.
Then I found this guy’s YouTube video on fixing a WES51 that wouldn’t heat up. As I suspected, the PCB’s in the two different models is very similar. In the guy’s video, he fixes it by replacing a 2.2uF capacitor that’s connected to the heater’s power transistor. I checked the corresponding capacitor in my WESD51, and sure enough it was bad. I found a in my junk parts bin to swap in and bingo, my WESD51 is working again! The step by step procedure is below.
Step 1: Open the case
First, you need to open up the case. Pop out the rubber feet at the bottom of the controller case. Underneath are philips screws.
After you remove all four of them, the blue upper body easily separates. Here’s what’s inside:
Step 2: Remove PCB
Separate the green PCB from the casing. You don’t have to remove the temperature adjustment knob from the faceplate. Locate the 2.2uF/50V capacitor. It’s circled in red in the photo below:
Step 3: Remove existing 2.2uF capacitor
Carefully desolder the capacitor circled in red above. If you don’t have a capacitance meter, just try swapping in a new part, and see if it fixes the problem.
Step 4: Solder in a replacement capacitor
Make sure to pay attention to the polarity of the electrolytic capacitor… the negative terminal faces the line of SMT resistors on the right of the photo above. I didn’t have an exact replacement in my parts bin, so I just used a 4.7uF/25V part… the circuit voltage is 24V, so 25V isn’t much headroom, but it’s easy to replace it again if it fails in the future. Before putting everything back together, I fired it up, and it was heating again!
Commercial mine machines are quite expensive, and often inflexible, containing only a few canned programs. In this series, I will show how to build an infinitely reconfigurable photic-stim mind machine on the cheap. Rather than build an oscillator circuit, we will drive our mind machine with computer-generated waveforms via an audio player. You can use your favorite audio playing program on your computer, or a portable MP3 player.
UNDER CONSTRUCTION: I still have a ways to go w/ filling in details, but I thought I would at least throw up the article first, since it might take me forever. Please pardon the omissions.
Photic-stim goggles cost typically more than $50, but they’re typically built with only <$1 worth of LED’s, and some cheapo ski-type sunglasses.
Below are the lightframes which came with my Sirius AVS machine, which I purchased from MindModulations.com.
Note how they are just as I described above, a cheap pair of ski-type sunglasses, with 8 LED’s wired into a sheet of black plastic. The plastic is attached to the sunglasses with double stick foam tape.
Below are the materials we need for our project:
Bill of Materials (BOM)
Ski goggle style sunglasses. I purchased this pair for under $10
Stereo wiring with 1/8 in phone plug. A cheap pair of earbuds from a local 99 Cents Only Store is perfect for cannibalizing.
Thin, flexible black plastic sheeting. A spare DVD case courtesy of junk mail from AOL will do fine.
LED’s. The type and number of LED’s is up to your own taste. Since my design uses the black plastic sheet insert, I picked 3mm LED’s… they need to be small enough that they won’t bump into your eyes when you’re wearing the goggles. If you want to drill directly into the goggle lenses and eschew the plastic sheet, you can use bigger LED’s. I bought my LED’s from UniqueLEDS.com. They have an extensive selection, and list detailed specs for the LED’s, including brightness. SUPERBRIGHT
Not pictured is a few inches worth of hookup wire. I had some kynar-insulated 40AWG wire wrap wire laying around, which was perfect, due to its extremly thin diameter.
How many LED’s you use is up to you, but the minimum is two, one for each eye. I decided to use 8 … 4 for each eye. Also, I used two different colors, green and blue, on alternating diagonals. Note that my LED’s have 3 leads instead of the usual two. This is because I had a wild idea of using bicolor LED’s so that I could get different color combinations, but I’m not very sensitive to the colors, so I decided to hook them up like one-color LED’s. There are various recommendations on the web for color selection. Personally, I like the bright white LED’s that my Sirius frames used, but I decided to experiment with colors on this set.
First, cut the black plastic which will support your LED’s. Ask I said above, you can skip this step, and drill directly into the lenses, but it will look a lot uglier, since the wiring will be exposed in the front. Also, the drilling method doesn’t let you easily reconfigure the frames if you change your mind about the layout. I used graph paper to make a paper template for my LED layout:
You should lay out the LED’s so that they’re fairly well-centered over your eyes. If you space them out a bit, it will allow for inaccuracy, and eye movement. Note that my layout is rotated 45 degrees with respect to the layout used by the Sirius. My DVD case was soft enough that I could just use scissors to cut out the template, but you may need to use an X-acto knife.You can use a drill for the holes, but I spun an X-acto knife around, instead, and slowly increased the size of the holes until I had a tight fit. (BTW, for the curious, those are not my hands in the photo).
I misaligned one of the holes a bit, as will be apparent in later photos, but it works fine, even though it’s a bit ugly.
Since we want to keep the LED’s far enough away that they don’t hit your eyes, we need to bend the leads as tightly as possible. I used a pair of needle-nosed pliers, but if you’re careful, you can just use your fingers on the edge of your desk:
Here’s what they look like when they’re done. Since my layout uses alternating colors on the diagonals, I first bent two of the green LED’s. Then, I bent another pair of green LED’s in the opposite direction. The opposite polarity of the bending is important, because the LED’s are hooked up in parallel. Here’s what they look like before I cut the leads:
Now, the wiring and layout of the leads will vary considerably, depending on how you want your LED’s to behave. The lightframes I built behave very differently from the Sirius lightframes, which utilize the industry standard wiring. In the Sirius setup, all 4 LED’s on a particular side light up synchronously. The left side responds to the left channel, and the right side responds to the right channel, so the two sides can be controlled independently. This is an example of how they can blink:
My design mimics the IC/D setting of TC-Softworks lightframes, hooks up each pair of diagonals together. The greens respond to the right channel, and the blues respond to the left channel. Therefore, both eyes are always lit, but the intensity and color can vary.
You can use your imagination to come up with your own unique design.
If you are building your goggles to work with a particular mind machine, be aware that there is another variation in the wiring, which affects the compatibility. All of the schematics I have shown up to now have used Common Ground (CG) wiring, where the cathodes (grounds) are all tied together. Some machines, such as my Sirius, use Common Power (CP) wiring, where the anodes (power) are tied together. Before wiring up your goggles, make sure you know if you need CG or CP wiring. As it stands, the goggles I built are incompatible w/ my Sirius.
You must also be aware that which polarity you choose will affect the polarity of the signals that you use to light up your goggles. CG goggles light up when the signal is positive, and CP goggles light up when the signal is negative. Thus, if we use the following signal, CG goggles would light up in the first half of the cycle, and turn off in the second half, where the signal goes negative.
On the other hand, CP goggles will do the opposite, and be dark in the 1st half cycle, then light up in the 2nd half cycle.
Next, I fit them into the plastic, and soldered the wires:
Be careful to solder as quickly as possible, and not to use a high-wattage iron, because LED’s are semiconductors, and can easily be destroyed by overheating.
The inner pairs of LED’s have to be bent a bit differently, so that the leads don’t touch and short out w/ the outer pair of LED’s.
In my case, they were the blue LED’s. Note that the LED’s w/ the extra bend actually go on the inside, so I had to take the first pairs out first. Here’s what they look like up close. Although the inner and outer wires look like they’re touching, there’s a tiny clearance. Here’s what it looks from the other side:
I know, I know, the holes are a bit off, but it’s accurate enough.
Next, I connected up the wiring. Since my goggles are wired up for CG, the center leads (the cathodes) of all of the LED’s are wired together. The anodes of each color are all wired together. NEEDS PIC OF TRS WIRING HERE
I recently acquired a few TP-Link Kasa KP115 smart plugs with energy metering. It was a bit disappointing to find that they barely show any data. There are no graphs, no voltage or current, just instantaneous power and summary energy stats and runtime:
I wanted to find the daily energy consumption of various appliances, so I needed a way to reset the energy counter to zero. Believe it or not, TP-Link doesn’t provide a way to reset the energy. TP-Link support says that you have to delete the plug from Kasa, factory reset, and then add it back to Kasa.
The most ridiculous part is, the smart plugs actually have a command to reset the energy without resetting the whole device! They’re just too lazy to add it to their app!
I found a github repo with a list of the commands available in the TP-Link protocol. It turns out that there’s a command to reset the energy monitor!
Erase All EMeter Statistics
The repo contains a Python3 script that can query the energy stats, but it doesn’t have an option to send the emeter reset command. Fortunately, it has a command line option to send an arbitrary JSON command to the plug. I tried sending the above to my KP115, and the energy meter was instantly reset to 0!
Next you need to know the IP address of the smart plug you want to reset. If you don’t know how to find it from your WiFi router, you can scan for it with an IP scanner, such as Fing. Look for a device named KP115 (or HS110).
At some point in the past several months, probably after a Windows 10 update, my PC started to display a cryptic and annoying dialog box every time I logged in. I started wondering if somehow, a virus had gotten into my computer. I hadn’t spent the time to figure out what it was causing it, and how to get rid of it, until today. A google search of “Smoni Failed to create empty document” didn’t yield any useful results. It’s so annoying to go through Microsoft forums, and find the that the only advice they can give you is reset or restore your system. No thanks. It’s a lot easier to spend a little time troubleshooting on my own, than spending hours backing up and reloading all my programs and files.
I pulled up the Task Manager, scanned the running process list, and quickly found the culprit:
Aha! What’s this BrnlPMon process? Simply right click on BrnlPMon, and select Open File Location from the menu:
Windows then pops up a File Explorer window, with the offending file highlighted:
Hmm.. so what’s BRNIPMON.exe? Right click the file, and select Properties from the context menu. Select the Details tab on the resultant pop-up window:
Bingo! It’s Brother IP Monitor, which I figured out is monitoring software for my Brother WiFi connected laser printer. Something in my recent Windows 10 updates (probably the new Creator’s Edition), caused an incompatibility, which triggers this stupid error dialog.
At this point, I could just uninstall it or delete the files, but since I now know that it isn’t malicious, I’ll just leave it for now, and look for a software update when I have time, and see if that fixes the issue.
Over the years, it has occasionally had issues with low water pressure. I’ve had to take it apart several times, and have narrowed it down to a couple of common problems. In this article, I’ll explain how to quickly diagnose and fix it, so you don’t have to spend hours on it like I did.
First we need to figure out where the problem lies.
Are both the hot and cold water flowing slowly? If so, the problem is likely in the spray head assembly. See Spray Head section below.
Is only one side flowing slowly? If only hot or cold is flowing slowly, and the other side is flowing fine, then the problem is likely that an inlet screen is clogged. See Clogged Inlet Screen section below.
Pull out the spray head, and use a crescent wrench to hold the end still while you unscrew the spray head from the faucet:
Turn on the water full blast. If it flows out of the hose strongly, then your spray head is definitely the culprit. With dry fingers, pull the black plastic buttons off the spray head.
grab the pin located by the red arrow in the photo above with a pair of needle nose pliers. Turn it, and push it in and out until it moves smoothly
use a small screwdriver inserted into the slot in the metal piece denoted by the red circle in the photo above, and slide the metal piece to & fro until it moves smoothly
(optional) if you have silicone spray, spray it into the areas where the plastic and metal interface, and again move the metal pieces to & fro until they move even more smoothly
At this point, you can reassemble and test it out. Most likely, the problem will be fixed already. However, as long as you have it taken apart, you might as well clean off the calcium deposits. There are two places where calcium builds up: 1) the inlet screen
and 2) the aerator and jets
The entire black plastic face containing the aerator should be removable if you grab the flat sides of the inner black core around the aerator (see red arrows in photo above) with a wrench and unscrew it. Unfortunately, mine was glued on by serious calcium deposits (in the photo above, I’ve already cleaned off the calcium but it was still stuck). If you can unscrew the black plastic face assembly, you can disassemble and clean the aerator more thoroughly. Since mine was stuck, I simply soaked both ends of the spray head in hot vinegar, and then used a brush to rub off the remaining calcium deposits. If the gray jet holes are not clean, you can also insert a round toothpick into each hole, and rotate it to clean it out. Reinstall the spray head and test. The water should be flowing strongly now. If so, you are done.
Clogged Inlet Screen
If only the hot or cold side (usually only the hot side) flows slowly, while the other side flows strongly, the inlet screen on the clogged side is probably clogged up with calcium deposits. Go under the sink, and with two wrenches, unscrew and detach the inlet hose from the faucet on the side that’s clogged.
sorry for the horrible photo. The red arrow in the photo above points to the inlet screen. It’s not removable, so you have to clean it in place. Mine was completely gunked up with calcium deposits that had broken loose from my old water heater. I used q-tips soaked in vinegar, followed by scraping with a small screwdriver to clean the screen. Reattach the inlet hose and test. If the water’s flowing strongly again, you’re done!
In my previous article, HowTo: Upgrade Scosche Rhythm+ Firmware, I showed how to update Scosche Rhythm+ firmware via their Fitness Utility iOS app. Some people have had issues with the 3.01 firmware installed by the latest V2 Fitness Utility, notably incompatibility with certain apps, and/or flaky readings.
I contacted Scosche via live chat, and they told me that there was no way to downgrade from 3.01, except for sending the unit back to them. The V2 Fitness Utility no longer has a Firmware Update button, so there’s no way to use it to install any firmware other than v3.01. Instead of sending mine back to them, I decided to try to get a hold of an older version of Fitness Utility, in order to downgrade the firmware. It turned out to be a very laborious and time consuming procedure. I was hoping that I could share the IPA file of Fitness Utility 1.4.1 so everyone else could save a lot of time, but as reader Hap noted in the comments below, IPA files are tied to specific Apple IDs.
If you want to downgrade your firmware yourself, rather than send it back to Scosche, follow the rather lengthy and complicated procedure below.
Current versions of iTunes no longer support app installs, so you need to downgrade to an older version. The linked article states that there’s yet another hurdle, in that as of iTunes 12.5, Apple is using certificate pinning, which nullifies the ability of Fiddler to snoop HTTPS traffic. I tried an older version of iTunes, but it was no longer able to communicate with the App Store (Apple just LOVES to put up hurdle after hurdle for us!). After much searching, I discovered that in December 2017, Apple quietly released iTunes 12.6.3 for enterprise users who still need the ability to do app installs. Because it uses certificate pinning, I had to devise a procedure to get around that.
WARNING: THE PROCEDURE BELOW IS PROVIDED AS A RESULT OF MY OWN FINDINGS. THERE IS ABSOLUTELY NO WARRANTY, AND THERE IS A SMALL POSSIBILITY THAT YOUR DEVICE CAN BECOME BRICKED DURING A FIRMWARE UPDATE. MAKE SURE THAT YOUR DEVICE IS FULLY CHARGED BEFORE STARTING. IN FOLLOWING THE INSTRUCTIONS BELOW, YOU AGREE TO RELEASE ME FROM ALL LIABILITY, AND PROCEED AT YOUR OWN RISK.
How to download Fitness Utility 1.4.1 and use it to downgrade your Rhythm+ to firmware 2.62:
Find your current iTunes folder, and rename it to iTunes.sav, or just move it to a new location.On Windows 10, it’s located at C:\Users\<yourusername>\Music\iTunes. (Don’t worry, after you’re done, you can reinstall the latest iTunes, and restore your old iTunes folder).
Download and install Fiddler. DO NOT START FIDDLER YET
Launch iTunes 12.6.3 and download any random app. iTunes will prompt you to log in with your Apple ID. This is the loophole we use to get around the certificate pinning. It turns out that iTunes 12.6.3 only checks the certificate during the login process, and doesn’t detect when we later swap in Fiddler‘s fake root certificate so that it can snoop HTTPS traffic.
Before proceeding, it’s best to kill any programs on your computer that access the web, because they will pollute your Fiddler capture. If you have your web browser open in order to read this article, kill all of your other tabs that might be accessing the web in the background.
In Fiddler, go to the File menu and uncheck File->Capture Traffic
From the Fiddler menu, go to Tools->Options->HTTPS. Check the Capture HTTPS CONNECTs and Decrypt HTTPS traffic checkboxes. A dialog box will pop up asking if you want to Trust the Fiddler Root certificate. Select Yes to it, and all of the ensuing dialog boxes. Don’t worry, after we’re done, we will remove the fake certificate, and restore your original.
In Fiddler, go to the menu to check Rules->Automatic Breakpoints ->Before Requests
Launch iTunes and search for Fitness Utility in the App Store
In Fiddler, go to the File menu and check File->Capture Traffic
In iTunes, click the button to download Fitness Utility
A few requests with red icons on the left will appear in the Fiddler capture pane. Select HTTP Tunnel to upp.itunes.apple.com:443 and click the green Run to Completion button in the right pane. Next, select HTTP Tunnel to p14-buy.itunes.apple.com:443 in the left pane, and click the green Run to Completion button in the right pane
A new request should appear in the Fiddler capture pane: HTTPS p14-buy.itunes.apple.com /WebObjects/MZBuy.woa/wa/buyProduct Select it in the capture pane, and then in the right pane, click the TextView tab, look for
In Fiddler, go to the menu to check Rules->Automatic Breakpoints ->Disable
Make sure the HTTPS p14-buy.itunes.apple.com /WebObjects/MZBuy.woa/wa/buyProduct request is selected in the Fiddler capture pane, and click the green Run to Completion button.
After iTunes shows that Fitness Utility is downloaded, verify that you have the Fitness Utility 1.4.1.ipa file in C:\Users\<yourusername>\Music\iTunes\iTunes Media\Mobile Applications
Connect your iOS device to your computer, and use iTunes 12.6.3 to install the Fitness Utility 1.4.1 to your iOS device, or use iFunBox instead as described below in Update 20170112
Launch Fitness Utility 1.4.1 on your iOS device and turn on your Rhythm+. WARNING: MAKE SURE YOUR RHYTHM+ IS FULLY CHARGED BEFORE UPGRADING THE FIRMWARE. IF IT DIES DURING A FIRMWARE UPGRADE, IT MAY BE RENDERED UNUSABLE.
Tap the Commands button at the top right of the screen, and then tap the Start button next to Firmware Update.
After the update is completed, power cycle your Rhythm+
You can check that the firmware version is now 2.62 by tapping the Attributes button at the top left of Fitness Utility.
VERY IMPORTANT: Once you verify proper operation of Fitness Utility, on your computer, have Fiddler restore your original root certificate with Tools->Options->HTTPS->Actions->Reset All Certificates.
Copy your Fitness Utility 1.4.1.ipa file somewhere so that you can reuse it in the future if you wish.
Delete the new iTunes folder, restore your old iTunes folder by renaming iTunes.sav to iTunes, uninstall iTunes 12.6.3, and reinstall your original version of iTunes.
Now that you have your own copy of Fitness Utility 1.4.1.ipa, you are free to try any future firmware upgrades from Scosche, because it’s easy to go back to a working version if you don’t like the new one. If you use iFunBox, you don’t even have to mess with swapping out iTunes versions.
If you prefer to downgrade to firmware v2.4, you can use Fitness Utility 1.4.1 and follow the procedure below:
*** WARNING: DOWNGRADING TO FIRMWARE V2.4 DISABLES THE ABILITY TO UPDATE FIRMWARE VIA FITNESS UTILITY. IF YOU LATER CHANGE YOUR MIND, AND WANT TO INSTALL A DIFFERENT VERSION, YOU WILL HAVE TO SEND THE UNIT BACK TO SCOSCHE. ***
send the unzipped HEX file to an e-mail address accessible from your iOS device
open the e-mail you sent on your iOS device, tap the attachment, and then scroll through the on screen icons until you find Copy to Fitness Utility, and tap the icon.
Turn on your Rhythm+ and follow steps 19-22 above.
The above method actually works with any version of firmware HEX file that you are able to obtain.
Update 20180112: I tried installing Fitness Utility 1.4.1.ipa with iFunBox instead of iTunes, and it also works. Launch iFunBox with your phone connected to your computer, and install the app by clicking the Install App(*.ipa) from the main screen. Firmware 2.4: scosche-rhythmplus-2_4.zip
Typically, OpenEVSE firmwares are flashed into the board using a hardware programmer, such as a USBasp. In the past, this was required, because the firmware had grown so large that there was no space left in the ATMega328P‘s flash to fit in a bootloader. However, the latest versions of the AVR tools that come with Arduino have shrunken down the binaries to the point that we now have space for a bootloader. Once the bootloader is installed, OpenEVSE can be programmed in exactly the same fashion as an Arduino Pro Mini, via a USB->TTL UART adapter, such as a FTDI cable, using the stk500 (arduino) protocol.
Before we can program the chip with a bootloader, we need to make a minor hardware mod. After a reset, the bootloader waits to see if a new firmware wants to be flashed before proceeding with booting the installed firmware. It is only during this very small time window that the ATMega328P‘s MCU is ready to accept a firmware. In order to trigger a reset via software, we need to connect the DTR pin of the FTDI cable to the RESET pin of the MCU via a .1uF capacitor.
The DTR pin is on the far left of the 6-pin serial connector. The RESET pin can be accessed at either the left side of R10, as pictured above, or at Pin 5 of the ISP connector (red circle).
Once the hardware mod is in place, we must set the fuses to use a bootloader, and flash in the bootloader, using a hardware programmer. In this example, we will use OptiBoot, because it’s smaller and faster (115200 baud) than the standard Arduino bootloader.
Substitute your FTDI cable’s virtual serial port for COM5 above.
For those who are not comfortable with command lines, it’s also possible to use the Arduino IDE to burn the bootloader, and flash in firmwares.
Set your board to Arduino UNO by using the menu to navigate to Tools->Board->Arduino UNO
Select your hardware programmer via Tools->Programmer
Install the bootloader via Tools->Burn Bootloader
Disconnect the hardware programmer, and use Tools->Port to select your FTDI cable’s virtual serial port.
Thereafter, you may flash in your sketches with the upload button. The above procedure will also work with any DIY or other Arduino clone which is not wired for a bootloader. Note that the bootloader takes up 512 bytes, so your maximum sketch size drops from 32768 to 32256 bytes.
I have a Samsung RF28HMEDBSR french door refrigerator that’s only a few years old.
Several months ago, I started to notice a mild clicking sound coming from it occasionally. The sound would always stop as soon as I opened the door, and then usually restart a little while after closing the door. In the past few days, the noise got considerably louder. It became clear that the sound was coming from a fan that was inside the refrigerator compartment. It started sounding like a fan whose blades were hitting something. Then this morning, it became unbearably loud.. like there was an airplane inside my kitchen!
I thought maybe the fan bearings were just dry and needed oil. After doing a lot of research on the Internet, I figured out that the evaporator fan, which circulates cool air inside the refrigerator, was probably the culprit. It turns out that ice builds up on the evaporator (due to bad design of the defrost circuit), and eventually hits the fan blades, causing the racket. The evaporator fan is a box fan that’s attached to the evaporator cover in the back of the fridge, behind where it says Twin Cooling:
The fan looks like this:
The proper fix is to remove everything from the refrigerator, take out all the drawers and shelves, remove the evaporator cover, and then melt the ice. I didn’t have time to do this today, and just wanted to silence the racket, so I decided to try a quick hack. The ice build up usually occurs on the coolant pipes feeding the evaporator. Notice how there are two large oblong air holes in the evaporator cover (see above photo), above Twin Cooling. The coolant pipes are approximately behind the air slot on the right.
I decided to try blowing hot air into the air slots, to melt some of the ice enough so that it wouldn’t hit the fan anymore. It’s important not to blow air that’s so hot that it melts the plastic cover. I set my dryer to high, and then pointed it at my hand, adjusting the distance so that the air was just a little too hot for me to tolerate. Then I aimed it at the intake slots, at about the same distance, and alternated blowing air into them, 10 seconds at a time, for 2 minutes:
Voila! The noise is completely gone! When I have more time, I will do the proper fix, taking the evaporator cover off, and melt the ice that’s covering the evaporator. Most likely, there’s a lot of ice back there, which blocks air flow to the evaporator, reducing the efficiency of the refrigerator, which wastes electricity, and in the worst case, keeps it from cooling properly. I will make a post in the future, documenting the process as I go.
In the meantime, if you want to tackle the proper fix yourself, here are some YouTube videos which are helpful:
At about 4:35 in the video above, the guy has a good hack for preventing the issue from ever happening again. He moves the temperature sensor for defrosting from the inlet to the outlet pipe of the evaporator, which extends the defrost cycle.
UPDATE 2020-05-20: It’s been almost 3 years since I applied the temperature sensor moving hack described above, and I haven’t had a recurrence of the noisy fan, so it works well as a permanent fix!
The video below gives a lot more details on disassembly procedures:
I’m running some server software in a virtual machine running Debian 8 which is running under VMware on a Windows 8.1 host. While it’s easy to access the server through VMware’s virtual NAT from the host computer, it isn’t at all clear how to access it from outside the Windows host. There are two ways to accomplish this feat:
configure VMware to set the Network type from NAT to Bridged (easy)
keep the NAT and forward the VM’s port to the host, and then open a port in the host’s firewall (hard)
So which method is better? (1) is easier, because when you configure your VM’s network to bridged, the VM will get its own IP address on your LAN, and will be fully visible on your LAN, just like your real computers. (2) is more secure, because you only expose the ports that you need to the LAN, so you don’t have to configure the firewall in the VM’s client OS.
For our example, let’s walk through how to remotely access an apache server running in our VMware VM, listening at port 8080. First, let’s find the IP address of our VMware VM. Since my VM is running Debian, we simply run ifconfig:
The output of ipconfig shows that on my LAN, my Windows host has IP number 192.168.1.115, and on VMware’s NAT, its IP number is 192.168.88.1. To test access to our apache server from within the Windows host, we an simply open a web browser, and point it to http://192.168.88.144:8080. Next, let’s configure things so that we can access the server from any host on our LAN.
1. Network Bridging (Easy Way)
To switch our VM from NAT to Bridge mode, simply go to VMware’s main menu and select VM -> Settings... A Virtual Machine Settings dialog will pop up. In the left side of the dialog, select Network Adapter, and then on the right side of the dialog, under Network connection, change the setting from NAT to Bridged:
After you reboot your VM, it will obtain an IP number from your LAN. My VM came up with IP number 192.168.1.111, so apache is accessed via http://192.168.1.111:8080.
2. Port Forwarding (hard way)
Now, let’s see how to do it while keeping the NAT. In the Virtual Machine Settings above, make sure NAT is selected. If you change the setting, make sure to reboot your VM afterwards. The first thing we need to do is forward the port from our VM to the host. From VMware’s menu, select Edit -> Virtual Network Editor…
(Note, VMware Player, unlike VMware Workstation, doesn’t come with vmnetcfg.exe, the Virtual Network Editor. You can follow instructions here to access it: DOWNLOAD VMNETCFG.EXE & VMNETCFGLIB.DLL FOR VMWARE PLAYER). In the Virtual Network Editor, click the Change Settings button near the right bottom of the main dialog. In the next dialog, select the NAT from the listbox, and then click the NAT Settings… button:
Next, in the Nat Settings dialog, click the Add… button, and fill in the info for your server’s port:
Host port: the port number you want to use to access the server … can be different from the actual port used in the VM if you like Type: select TCP or UDP Virtual machine IP address: the VM’s IP on the NAT Virtual Machine port: the port number used by the server inside the VM Description: arbitrary info
Finally, click the OK button to save your port mapping. At this point, the port forward is functional, but most likely, your have a firewall running on your host computer. You must open up a hole in your firewall for the Host port you selected above.
In a Windows 8.1 host, if you’re using the built-in Windows Firewall, run WF.msc. Select Inbound rules -> New Rule…. Under What type of rule would you like to create?, select Port – rule that controls connections for a TCP or UDP port. From the New Inbound Rule Wizard, select your protocol and port(s):
In the next dialog, select Allow the connection. Finally, you can decide where you want to rule to apply, Domain/Private/Public. Unless you’re planning to use the server while travelling, it is best to leave Public unchecked. Finally, you’ll be presented with a page to enter a name and description for the mapping. After you click the Finish button, you should be able to access your server from any host on your LAN. In our example, a web browser should work when pointed to http://192.168.115:8080.