Carl's Notes

When I'm trying to discuss something, I kind of need to be able to provide pictures demonstrating the point I'm trying to make.

And since this site basically refuses to support that directly, all your images need to be hosted somewhere else. Meaning you have to manage them yourself.

Hence the problem.

I have, in the past, been able to handle that with a whole lot of effort on my end. But my patience has begun to run out on that, and I just can't bring myself to bother to bridge what they refuse to directly support here.

If anyone knows of a free, lightly ad-supported platform that does let me upload images directly (and doesn't try to feed them into an AI model), I'd be very interested in hearing about it.

So, I've been stressing over nothing, it seems. I'd been under the impression that I'd need to build a whole control circuit, with multiple switches (power, heating/cooling, and intensity), on a circuit board, putting a PC power supply on my desk, it turns out none of that is necessary for testing purposes!

This realization has led to me dropping a PC power supply on my chair, plugging a PC power supply adapter board into it, unplugging the Peltier device from the original control board, and somehow connecting it to the adapter board.

And it worked just fine!

I don't know if it got colder, but it definitely got cold faster! A very worthwhile effort!

Unfortunately, I didn't realize that the PC power supply had a fan in it, nor that it would be loud enough to be distracting.

Fortunately, I now know that I don't need a huge array of controls (after all, I only use it to cool drinks, not warm them. Nor do I need to change the voltage the Peltier device runs at), as well as knowing that I can use almost any 12V supply for this project. Time to look at my other options!

Varnishing is hard. Especially since almost every woodworker I've been following on YouTube entirely skips over the whole process, going straight from "this is what it looks like before" to "this is what it looks like afterwards" in one jump cut.

What I really need is a text tutorial (NO, video tutorials do NOT work for me! I should write a post on why eventually...) that explains every step in the process. Also what to do when things look particular ways.

The very first thing I'd varnished was a small (four-inch in diameter) attempt at a wooden coaster, so a full paint brush basically drank more varnish than I was able to spread onto the surface. So I'd used a sponge brush instead, and it worked better in that situation.

In this situation, it didn't lay down enough varnish, and two thin coats didn't look like much.

Given the larger surface to cover, I brought the paintbrush back out, and it actually laid down enough varnish to almost look professional!

If it had been a smooth enough layer, I would have left it at that, but it needs sanding smooth again. Which means I need to know what grit of sandpaper to use to avoid making it worse...

A few years ago, I'd found a dual-function drinks coaster. At the moment, the house I'm living in doesn't have anywhere near enough insulation, so I'd been looking for a drinks coaster that was capable of cooling my drinks. Having a glass of juice with lime in it was fairly effective at keeping me cool, but only when the drink wasn't allowed to get that hot.

Anyway, while most electric drinks coasters just heat your drinks, the one I'd found also included a cooling function. Though, I suspect it's less effective given that it's USB powered instead of wall-powered.

Now, in larger heat-moving systems, a highly compressible gas is used to literally pump heat (hence, "heat pump") around, but in much smaller systems like this, you can buy solid-state devices that move heat around to a less efficient extent. Called Peltier devices, if you hook it up to power the right way around, it uses the flow of electricity to move heat from one side to the other. A higher voltage moves more heat around, in this case, and while it comes with a USB cable by default (thus running at 5V), you are supposed to be able to buy a 9V power adapter separately, which increases the amount of heat it pulls out of your drink.

On the opposite side of the coin, if you apply the power in the opposite direction, it instead generates heat, and the power switch on the front of the coaster toggles which direction the power is applied.

Unfortunately, the cooling function has two separate flaws. The first being a small fan. This fan blows air across the bottom side of the Peltier device, with the intent to blow the heat pulled from your drink away, but it's noisy enough to be a distraction most of the time.

The second being that cold doesn't rise. Which means that, if your drink starts out being warmer, you end up with a temperature gradient in your cup, with the bottom layers being colder than the top layers.

With the modicum of success I'd found using this device, and finding that its price had dropped afterwards, I'd bought a couple more. The idea being to experiment with it. For example, fixing the fan noise problem would be simple enough hardware wise, by replacing the original fan with a larger module, which is quite easy to obtain and I've already done so. Structurally, however, I'm still trying to figure out how to support the whole thing, as there's nowhere near enough empty space for a fan of this size in it, so I basically need to build a new frame to support it. Given my limited set of tools and experience, a lot of thought is required here.

But, that doesn't solve the second problem. How do you make an entire glass of water (for example) colder at the same time? Do you mount the Peltier device so it makes contact with the side of the glass instead of its base? Being spring-loaded such that it pivots back and forth as necessary to make contact with the side of the glass, no matter what angle it's on? Maybe with a flexible air duct from the base to the bottom of the heat sink already attached to the Peltier device, so the fan in the base forces a lot of air through it?

Unfortunately, that concept doesn't quite work here, as all of my glasses are round in two dimensions, such that a flat object pressed against them would make contact at a single point, not even a line. Thus, heat transfer would be highly inefficient.

You can see my dilemma, and why I haven't tried doing anything else with this concept just yet. Granted, with how many of these I have right now, I could probably make one that just has a larger fan and contacts the bottom of the glass, then keep fiddling with another to see if I could make it any more efficient.

Well, now that I have both pieces of wood cut roughly to size, now I get to have the "fun" of sanding them even with each other.

Currently, they're clamped together and lined up, and my sanding block is sitting on the project.

But both cuts have turned out to be uneven. One side needs to be sanded down more than the other, and somehow both slopes matched up when they were clamped side-by-side. And since they're going to be mounted some distance from each other, those slopes aren't going to match in the end, so I need to be careful as I sand them smooth.

Unfortunately, it's an entirely manual process, as I don't have a power sander of any kind.

Actually, this is just about one of the higher priority tools I'll want when I have the space and the time. Some kind of stationary belt sander, with supported and "loose" sections (for flat and rounded sanding, respectively), and variable speed control. And I have absolutely no idea where to start looking for such a thing.

Anyway, I should probably take this outside to do my vigorous sanding where the dust can go anywhere it wants.

On a more positive note, the aluminum Pocket Square I bought from Lee Valley has been invaluable for all the alignment I've needed to do on this project. I've used a cheap metal ruler as well, but the square has been great for at least marking angles, even if it doesn't have the length necessary to reach all the way across the pieces.

When my maternal grandfather passed away, he left no Will of any sort (or, he left everything to my mom before he went mental with the last thing he was sick with), so we cleaned out the place ourselves. This meant we could take things we wanted, as otherwise it was all getting thrown out anyway.

I got a few tools, including a vice, a drill press, a bunch of electronics parts, and an old Dremel-brand skill saw.

The vice needs some more bolts, as it was missing a few of them when I found it. Fortunately, I also found a box that looked like it fit (might even be its box, I'm not sure), so it's nicely contained on top of a shelf right now. But it's a really solid vice, enough that I'm pretty certain it'll be able to hold metal for drilling operations, so I'm pretty happy I have it.

The drill press works well. I've already used it to drill holes in an aluminum plate (the work plate for my Snapmaker Original, so I have more holes to clamp things down through), and had absolutely no problems with it at all. Definitely glad for that, too.
(I think the vice was actually screwed down to the drill press when we found it, so they should, apparently, work together)

The electronics parts have already been added to my collection, though I haven't done anything with the backlit LCD displays yet.

Now, the scroll saw. That has been a weird one, to be honest. First, since I don't know the history of Dremel, the company, and my first introduction was the rotary tool that (intentionally or unintentionally) bears the same name, I didn't know that they'd done anything else before that.

It's technically more than "just" a scroll saw. They've labelled it as a "Moto Shop" tool, as it does both scroll saw work, and has a power take-off for a sanding disk and a flexible-shaft tool that I'm pretty certain would eventually characterize their products.

Anyway, the first time I'd looked for its manual online, I'd found something that initially appeared to be what I was looking for, but certain details didn't match up. Eventually, I noticed that the depicted parts diagram had way more parts than mine had (including a blade release lever on top of the arm, among other things), and the document's title actually mentioned a different model than what I was looking for.
(also, to make things worse, the manual actually directed you to read the inside of the box it came in for certain things, like attaching the work table)

A further look (and ignoring all the links that pointed back to my initial find) found me a manual for a slightly older model, but all the parts I could see matched up very well, and it explained everything far better than the first manual I'd found. And with more actually descriptive words, too.

So today I finally took it to the backyard, took it apart slightly (mostly taking off the table, the lower arm cover, and the motor), and used a can of compressed air to blow all the sawdust out of it. Well, most of it, at least. I was able to blow at the majority of the mechanism, but the upper part of the arm isn't exactly accessible without turning the whole machine upside-down and taking the bottom plate off to reach everything. Then I lubricated the only joint I could find, that also had a felt pad to hold onto extra oil.

Then I brought out some wood I've been trying to cut with insufficiently-adequate hand tools, a pair of safety glasses, and an extension cord and gave it a try.

It's, well, apparently not intended to cut plywood as thick as three-quarters of an inch. The blade doesn't move quite far enough to properly clear the sawdust, so it was like pulling teeth to get it to cut. I'm pretty certain the blade actually isn't straight, either, and I don't have any spares at all if it'll be a problem.

A tool I'd also obtained from my late grandfather's place was an old, adjustable mitre saw system. And when I mean "adjustable", I mean that it's not your modern box, but a back plate, a wooden surface to cut through (should that happen), and an angle-adjustable brace to hold onto the back of the saw. Unfortunately, despite all my attempts, I haven't been able to get it to get close enough to a proper ninety-degree angle, so I got as close as I could, then stepped the cut away from where I wanted the cut to happen, and made the manual cut.

So I only had to "straighten" out the existing cut, but even then, the blade kept getting caught enough to vibrate the material somewhat alarmingly. Oh, and did I mention that we didn't find the spring-loaded brace that is supposed to hold the material you're cutting down? And that my grandfather had the blade in upside-down?

Anyway, to keep the blade from being trapped by the length of the cut I was trying to make, I notched into the area beside the cut I was trying to make, and would cut off each piece to make sure I wasn't trapping the blade for very long. Hopefully it also kept the sawdust moving more easily as well.

But still it wasn't quite enough for me to think that it was a feasible option. I did get nearly halfway across the eight-inch piece, and while the cut quality in the end was pretty good, I didn't have anywhere near enough accuracy to think that this was going to be the right option. My fingers had to get relatively close to the saw blade, and while I know for certain that the blade doesn't have enough power to pull the material in, this is the first time I've ever tried using power tools to cut wood, so it definitely made me a bit nervous.

I'll probably give it another go at some point, as it probably wasn't as concerning as I'd thought it was, but it's a bit more work than just having the initial mitre saw cut being at the right angle in the first place...

Way back in 2016, I'd backed the Kickstarter campaign for Jide Tech's Remix Mini, a hand-sized (including fingers, so not palm-sized) Android PC, claiming to be the first Android desktop PC. Being very small and lightweight, but supporting most Android apps at the time, the idea was to get a pretty full desktop experience from the device.

I enjoyed fiddling around with it once I got it, but at the time I hadn't figured out how I was going to use it on a regular basis, as well as them launching a second campaign for a newer device that was supposed to be way more powerful. So I'd shelved it for a while, as I thought it was going to be replaced with a newer model that should have been better.

Unfortunately for me (but I guess fortunate for them), they drew the attention of a thin-client enterprise company, and they were basically swallowed whole. The campaign was shut down, and everybody got their money back, but nobody else got the product in the end.

A few years ago, I'd tried to boot my Remix Mini again to see if it still worked, but it didn't do anything at the time, so I put it even deeper on the shelf and forgot about it again.

Well, I came across the power cables again, and precariously perched it on my desk to give it another try. Unfortunately, I don't have any monitors with a spare HDMI port, so I plugged it into my HDMI capture device, and viewed the stream with VLC. And it booted all the way up, even to the desktop again, with no problems. Even connected to the WiFi again (we might need to change the WiFi information at some point) and realized that I need to update my login credentials!

Anyway, I've shut it back down for the moment so I can figure out my next step. I'd like to be able to have it running on my desk alongside everything else, as I do a lot more things now that would actually work on a system like this. Only, it's a really old OS now, I'm not sure how many of my apps would actually support it anymore.

Oh well. Back on the shelf again. But more towards the front this time, I think...

I mentioned that one of the things I want to do is install an FM radio module into a set of PC speakers, but didn't actually say what kind of radio module I had. While there are a lot of small circuit boards with the necessary chips (and sometimes the crystal for frequency stability), but no antenna, audio amplifier, display, or control scheme. For boards like that, it's just a small part (literally!) of a larger system.

Instead, the modules I have (and I have a few, they're cheap from China, where they're made) include an LCD screen to show what frequency they're on, signal strength, mute setting, and more. Most of them have buttons for volume, channel, and mute, but one even has no buttons and a bunch of through-holes for external control signals. Mostly buttons, admittedly, but the option is there for controlling it with another device.
(at the time of writing, <url="https://www.aliexpress.com/item/33038696986.html">this set of FM radio modules is a good example of the variety available. Searching AliExpress for "FM radio lcd" also should show most of these options, should the one I've linked to disappear)

Anyway, one of the annoyances of all of these modules is that they all use a blue LED to backlight their LCD displays, and it's only on for a short while after being interacted with. And since I'm sick and tired of blue LEDs nowadays, and would prefer the backlight stays on while the device is powered, and it turns out, the very first FM radio module I'd bought had an LED I could swap out. Nowadays, the LED is glued in place, with very short legs that go directly into the PCB. Fortunately, I was able to swap out the blue LED for a (supposedly) yellow one, then connect it to power and ground through a current-limiting resistor.

That looks like this now:



And I'm quite happy with the results, especially because it's not looking yellow, but orange, a much softer, more welcoming colour that I'd actually wanted in the first place. So it was a pleasant surprise to get that colour of backlight out of all the options I currently have access to.

Now, one of the things I'd noticed on all of the FM radio modules I've bought thus far is a set of three smaller through-hole pads labeled RX, GND, TX. It turns out, the control chips include a serial communications protocol that I'd hoped would allow me to pull the RDS (Radio Data System) information out of the module. After all, looking up the datasheet for the actual FM radio chip used on these boards (the Quintic QN8035, in my case) does support decoding this information.

It took me on a very weird dive, but I eventually found someone's own report on their dive into this, with a load of half-Chinese, half-English screenshots of command sets, and the specifications for the UART settings. The commands I've found thus far mostly includes every variation to the buttons already on the devices, including setting the tuned frequency directly, setting volume levels directly or just up or down, the backlight setting (turns out, it can be changed, just not through buttons), nearly full control over scanning for stations, and squelch settings.

But not the RDS information. Rather disappointing. But at least it means I won't need to devote any more time to thinking about trying to add this function to my current plans.

One advantage of the particular board I'm using, is that it includes two three-Watt speaker amplifier chips, so I don't need to add any more hardware to driver speakers.

On the other hand, if I wanted to make sure I had access to this information, I had bought a demo board showcasing the Silicon Labs Si4702. Either that or the Si4703. And unfortunately, that last digit is not listed on the chip itself, so I don't know which chip it actually is. And the -02 version doesn't have RDS support. But, on the other hand, it also doesn't have a display or buttons, so directly interfaces with a microcontroller. It just has the massive disadvantage of not having a separate antenna input, as it uses a headphone cable instead, so no driving speakers either.

Anyway, as you might be able to tell from the picture I've posted above, the LCD panel has lifted off the backlight diffusing panel (due to me fiddling around with it for years without mounting it into anything), so to install it into the speaker housing, I need to clamp it down to be flat again. Actually, I was going to write about that task here, but I think I'll write a separate journal entry about that...

So, some progress, I guess?

While the headphones I use the most are fully wireless BlueTooth ones, I started off with Sony's MDR-G45 (not the in-ear G45LP that came later, I gather) and used them a lot. I listened to a ton of music on my MP3 players through those!

Enough that I have at least two pairs right now, and both should have broken wires near the plug end by now. One is stored in a box for further experimentation (I'd wanted to convert them to pure BlueTooth at one point, but never quite figured out how at the time), and the other is hanging over my desk as a backup option for my gaming headset.

This latter pair definitely had a problem with its cable, as I'd tried to listen to something on my laptop, but its cable definitely caused problems.

It took me a while to figure out how to solder to the new plug (as there's no tabs like on older style plugs), but it's working again.

Slightly annoyingly, I have a whole bag full of four-contact 3.5mm headphone plugs, but appear to be entirely out of the three-contact ones. Much as I'd have loved to use a standard three-contact plug (due to it only being a pair of headphones with no microphone), as all I have are four-contact, I'd wanted to include both a microphone jack and a play/pause button. Because the option is there, of course.

I'd even found the information necessary to add up to three additional buttons for other functions, but didn't have the confidence to give that a try at the time.

But it works now, and I have ideas for how to add functionality for next time, so we'll see how long this repair lasts.

The general concept for this project is having a collection of wireless temperature (and more, hopefully) sensors I can place in various locations (outside front and back, garage, etc) that we can monitor from multiple locations, via a web server somewhere in the house.

Currently, I have two cheap wireless temperature sensor systems from the same company, and it turns out both use the same remote sensors, with no way to prevent their base stations from listening to the other sensor. And technically, the same brand does sell a "double" base station with two sensors, but the price jumps a lot from any of their one-sensor models, and I couldn't afford it. Conversely, I already have a bunch of pieces that could, theoretically, be turned into a small network of wirelessly-connected temperature sensors.

I even have a pretty solid idea of how to handle the hardware of the remote sensors. The inexpensive microcontrollers by Espressif are both inexpensive, and all include WiFi capabilities. Then there's a massive list of easily connected tempearture, humidity, and pressure sensors for very little money.

On the positive side, the actual sensor modules are both cheap, and quite easy to connect in hardware, and talk with in software. I've chosen which module I want to use, but if I need to change to another one, it won't be that difficult to swap out software libraries to make sure it still works.

The more interesting part is the microcontroller. While I have a variety of them (I've been buying samples in a wide variety for quite while now, mostly out of curiosity), I'm leaning towards the ESP32 series. They also have BlueTooth capabilities, not that I think I'll be using that, but also are all dual-core, with at least one real-time clock, and that's about perfect for what I want to accomplish with it.

The other half of the system is actually a lot trickier than I'd anticipated. A single-board computer (shortened to SBC, apparently) that can run "normal" software (as opposed to embedded software) is my aim, for many reasons. These include compact size, quiet operation, built-in WiFi and wired network capabilities, and hard storage of some sort. The Raspberry Pi is a very attractive option to me for all of these reasons, but I didn't realize I'd wanted to experiment with something like this until very recently, and not only do I not have the funds for it, but the supply has actually been drying up due to many reasons. There may be "compatible" alternatives, but I'm not comfortable with them right now due to not knowing how well-documented they are.

Unfortunately, because it's essentially a system in two parts, I can't really build one half at a time, as the other half would need to be in place for testing. And I work so much better sort of building the whole thing up at the same time.

One thing that's holding me back right now is due to not understanding Linux's Network Manager system. Technically, if I was writing the central server code on my laptop (for eample), I could write the remote sensor module code to talk to the laptop's server code. But I'm not comfortable putting the password to our home WiFi network directly into my source code, even for testing.

Instead, I want to have a separate WiFi card that hosts its own WiFi network, without linking it to any other networks connected to the computer. Hopefully that improves additional security measures. I have a secondary USB WiFi adapter, and it works, but I don't remember how to make it start working again. Unfortunately, Linux WiFi drivers are in a terrible state of affairs, and nobody seems to have a central repository for USB WiFi device drivers.
(from recent experience, I'm not even convinced that Windows can easily do what I'm trying to do, either)