The Quickshot II USB Joystick

The QuickShot II

I bought this Quickshot-like USB stick because I'd heard good things about it.

I tried it instantly with Vice emulator and Buck Rogers. This game is very revealing about joysticks, because the movement is so twitchy. Lag and poorly working diagonals make themselves known. It's also a flight game of sorts, so this joystick would feel appropriate.

The short verdict: I have to say I was disappointed. The result was not comparable to the self-made Arduino 9-pin joystick adapter and a proper Tac-2 joystick.

Let's rewind a little.

Retro Games have had a reputation of making so-so joysticks. The stick bundled with The C64 was unremarkable, even if physically quite appealing. I've heard better things said about the "black" edition joystick, but I haven't had the opportunity to test it.

Good old Buck Rogers

Now that Retro Games put out a new dedicated joystick, I expected it to be based on feedback from the earlier products, and as a few reviews seemed to give it praise I took the plunge. I perhaps ought to have read a few more reviews, as they tend to be mixed.

The Quickshot directions feel like the non-micro switch joystick the original black Quickshot was, and I can commend it for replicating that feel. There are in fact micro switches inside, it's just they are less noisy.

The grip is good and makes me wonder why I haven't been looking for the original stick before. The suction cups work, but to me this is a secondary feature, I can hold the joystick on my lap.

The real problem is that there's very little center zone for the stick, so in Buck Rogers the ship keeps moving left or right almost automatically, or even ping-pongs when I'm trying to gain control. Diagonals tend to get confused. This also makes it difficult to see if there's lag, as a micro switch might remain on while my senses tell it ought not to be on.

Trying this on The C64 and The Spectrum, as intended, doesn't make it better. In case it doesn't work out of the box, apparently you will need to hold X,Y and L,R for three seconds to change they joystick mode, after which it will register as a kind of joystick the device accepts. With later products, like The Spectrum, it should not matter.

On sluggish sim type games, like Elite on the ZX Spectrum and Stunt Car Racer on the C64 the problems are not as pronounced and the feel is more appropriate for these games anyway.

jstest-gtk

As a bonus, there are 6 additional buttons and the two fire buttons even register as a different button, so something can be made out of these when reconfigured for emulators. For example, I could use the buttons for accelerate/decelerate, arm/unarm/fire missile in Elite, which is kind of fun.

jstest-gtk is good for checking all the buttons and axes. Other software might have different numbering scheme, for example Fuse emulator starts from 1 whereas jstest-gtk starts from 0.

Looking at the insides

Opening the cover. The plate that houses the springs, is still in place.

One saving grace might be the joystick could be easy to mod. It opens easily enough, just remove five screws from the bottom and the part comes loose. 

There's a plate that houses springs for re-centering the stick. Four screws later, the plate can be removed. This actually needs to be removed for the circuit board to come loose.

After that, six more screws and the circuit board can be detached. The internal cables can be easily disconnected, they are not glued in.

The spring plate removed

Now, there are a few things that might influence how well the stick performs.

The four light springs that help return the stick to the center could be stronger I guess. Just to do something I pulled them out a little. This is not likely to achieve much.

The board doesn't look very straight when taken out of the stick. However, there are enough screws to keep it in place so I don't think it's bent when in position. So I'd rule that out.

The circuit board is easy to remove for inspection

The micro switch boxes might not be as flush against the circuit board as well as one might hope. I tried to re-solder them some, perhaps gaining a tiny improvement, but this alone didn't do anything either.

The underside of the board had something looking like re-soldering, before I even touched anything myself. It could be an indication of fixing the placing of machine-inserted parts, or the switches are put in manually as they are not SMD parts.

The other thing is that if they are not well positioned in their surface coordinates, could also result in an inaccuracy. My opinion is that this isn't very likely.

It's not exactly flush with the circuit board

The plastic parts that actually push the micro switches, could be reduced a little. After a little surface scraping I concluded I shouldn't do this as there isn't going back.

So, I didn't see any one thing that would be very crucial, but taking them all together there could be uncertainties when coming out of the manufacture, contributing to diverse experiences the reviewers and commenters have had.

The circle highlights the part that pushes against the micro switch

My adjustments didn't seem to do much, as they were so minor. It's possible the involuntary "ping-pong" effect was slightly reduced, but as I don't have metric for that it could be just my mind working away the problems.

I believe the stick can be further improved, and I think Retro Games can be at least credited for making devices that are easy to open and tinker with.


Towards a solution?

It occurred to me that I could add washers between the circuit board and the holders. This would create some distance between the micro switches and the shaft elements. After a couple of different attempts I ended up using 0.5mm thick washers.

The washers were a little tricky to add, some temporary glue could have helped, but I only needed to add 4 washers there. Fortunately it all still came back together.

I'd already noticed the stick doesn't stay exactly in position. It kind of rotates a tiny bit along its axis. But now I saw it can also be lifted about a millimeter or two. With the washers installed, and joystick lifted, the controls improve, whereas if it lies in its lowest position, I get far more of that "ping-pong" uncertainty.

To counter this effect I added one washer (0.5mm thick) under the central spring. This did create the desired effect, the stick no longer lifts.

There isn't a better picture, after adding them there's nothing to see

This resulted in a more balanced experience. I did introduce some kind of occasional cracking sound, which isn't very welcome, but I can perhaps look into it more. 

Also, if I want to be precise the L/R directions are still too sensitive compared to the U/D directions. This is where removing some material from the micro switch pushers might achieve something, but I'll address that later.

But I'm finished with the Quickshot now, and will return if I come up with any improvements.

The Buck stops here

While my modded Quickshot doesn't still quite pass the Buck Rogers test with flying colors, it's already a fairly usable joystick. It works pretty nice with the selection of games that comes with The C64. Boulder Dash, Monty on the Run, Paradroid. Why not.

On emulation front, Saboteur and even Atic Atac were passable, but Giana Sisters, maybe not so much. Blue Max felt ok.

Like I said, Elite and Stunt Car Racer go well with this. It adds something to sim and 3D games, especially when the extra buttons can be redirected to keypresses.

Dyson V15 now

After year and a half of using the Dyson Detect V15 with Optic/Fluffy head, how has it been?

Generally, I've been quite happy with it. It does require some more consideration and care than the old vacuum cleaners. It's not built like a tank at all, but so far I've not broken anything.

It was quite clear I'd mostly use the ECO mode, as the battery lasts longer and all normal dirt and hair can be removed with it. AUTO can be useful when using the other nozzles for sofas and things like that. BOOST, like I initially observed, depletes the battery in a matter of minutes.

Emptying the dust "bag" should not be left for the last moment. The emptying should be performed with the Dyson mostly inside a trash bag, so the fine grained dirt doesn't float around.

More often than not, it's not enough to kick out the contents, the entire container usually needs to be removed. But I guess more often it's done, the less hassle it is. Laziness prevails and I find myself emptying an overfilled container.

Removing the filter is just a twist away.

There are a couple of warning messages that come up after prolonged use.

The "before" pictures would have been too horrible to bear, so I only added "after" pictures.

Clean filter: Gently screw out the violet back end with hand, knock out visible dirt and particles, wash the cloth part under tap water, let dry for about 24-30 hours.

Weirdly enough, I did not need to do this until after a year of use, but then I had to do it again after a couple of months. Hmm...

Airflow blocked: More mysterious, but can generally mean the dust and hair has built up to the inner intake hole. One symptom is that even if the Dyson seems to take in particles, when I stop pulling the trigger they fall down the tube again. Just clear everything inside.

The roller dirt and hair build-up does not produce an error message, although it's bound to happen. I guess it would have been wise to do once a month or so, but I didn't do it for nearly one and a half years! I think Dyson could have added some kind of hour count and a reset button.

Locations where the hair easily accumulates.

It's far more easy to open than I thought. There's a sort of lever at the side and the part comes off without any use of force. 

The axle outside the roller had collected so much hair it was rather tricky to remove. I used scissors and a sharp paper knife, digging and cutting into the furry torus. I don't have much confidence in the strength of the part, so I tried to be careful.

The furry striped bar can be pulled out. Any visible hair and dirt is removed, then the furry part can be washed under the tap, taking care not to wash the connector part. I used Q-tips to clean that end. Again, wait until next day before re-assembling.

Only after cleaning the roller I saw how slow it had been before, because now it's quite nippy again. Ideally, the yellow-black stripes shouldn't be visible, but lost under a zoetropic blur. 

A vacuum cleaner is perhaps not the most ecological way to kill banana flies. But I was happy to see Dyson with the short nozzle is good for this task. The BOOST mode positively disintegrates them and they end up dead, not sure exactly why. With the less aggressive modes, not all flies end up inside. Those that do mostly remain alive. They won't get out, though. A house fly is not so easily fooled.

Nedis USB-C 3-in-1 Multi-Port Adapter

This caught my eye in Tokmanni, and thought why not try yet another adapter. I already have the USB-C to HDMI adapter for using Samsung Dex with an external display, but I was kind of stuck with wireless mice and keyboard so I largely forgot the whole thing.

Back then I was also worried that a USB hub would be too large, but this is not much bulkier than that video-only adapter.

So, apart from HDMI it has one USB-A style port and a USB-C.

The happy outcome here is that the USB-C to HDMI works with the Samsung Dex phone desktop for displays, and the Samsung S23FE also joyfully accepts USB keyboards, gamepads and mice. This is not something that was explicitly promised on the box, but I think it can be taken granted these days. Or can it?

I have to make the choices carefully as I have to use both port types. So one C-connector mouse and A-connector keyboard, or other way round, but not both of either type.

Does this make Samsung Dex much better? Well, at least the experience improves. Perhaps I'm more intrigued about the possibility of using a keyboard without actually invoking the Samsung Dex desktop.

It's still not a very coherent experience. For example the Slack app doesn't work well with a physical keyboard, after typing one letter it insists on opening the virtual keyboard after all.

I could have connected a keyboard without the adapter, but somehow my Samsung was picky about sticking some USB-C keyboards into it directly. Besides, today cheap small keyboards tend to be either wireless or with a hardwired USB-A cable. Admittedly, a simple USB C/A adapter might do the trick too.

What about games? Asphalt 8 Airborne had worked with the Shrimp USB-C connected keyboard, so I could expect Logitech F350 to work too. And yes it did, but I had to do some other things before I could properly try the game.

Samsung Dex works in two modes, either as a dedicated desktop that uses the external display resolution and sets the touch screen as a trackpad, or as 100% mirrored phone screen, showing exactly what I'd see if I only used the mobile. 

In the Dex desktop mode, if the game is launched and put into full screen, all kinds of weird screen resolution hiccups would appear. I didn't find a solid way to launch the game in the external display resolution, although it's obvious the game can do this. (Because it runs on different resolution Android phones and tabs).

Worse than that, the gamepad direction keys stopped responding until I refreshed the Dex trackpad screen.

Things were better with the full mirrored mode, as the game screen is set up nicely, if not altogether taking advantage of the external display resolution. There were no control anomalies, probably because the Dex trackpad is not invoked in any way.

It's silly to mirror the screen, and it uses battery power too. No doubt the better option would be to have the phone screen shut, but so far I didn't find a suitable configuration. For games it might be wiser to use the phone or tablet screen directly anyway.

If only I just could find an interesting gamepad-controlled mobile game. Don't get me wrong, the actual Asphalt game play was kind of nice but all that wading through menus and opportunities and "coins", in short the full-on mobile game experience, got to my nerves rather soon.

In any case it looks like car and flight games have lenient controls, negating any lag and inaccuracies, but as soon as I try some retro shooters the limits are more obvious.

Apart from that F350 gamepad, the Arduino 8-bit joystick adapter worked too, making this one of my most cross-linked blog posts ever. Not that it works well with any modern games, but I might explore some emulators later. My expectations are not high.

Trangia Triangle

Attached to the "X-Briv" Trangia-compatible burner

The Trangia Triangle is made of four pieces that can be dismantled for maximal flatness. It's not especially heavy either at ~105g, counting the carry bag. Putting the three walls and the circle together doesn't take much effort and it's not slower to set up than the standard Trangia.

Four pieces

Using a Trangia-compatible gas burner, it does give some "spider" vibes, but perhaps with better wind protection. This could be one rationale for the product. Also, it is not fixed to some particular burner, so for example the spirit burner can be used with it too.

Correct way up

Even if not immune to wind, the wind shield works quite well and the food kept bubbling quite evenly. I have a hunch that this combination worked better than the cheap "spider" I've used previously.

The instructions show the three protrusions at the top, but it shouldn't make much difference if the Triangle is upside down? That might even in some situations work better. At least one Fire Maple pot with the heat sink didn't sit that well on the suggested orientation.

The Knorr test

But, practice sometimes proves these ideas wrong. One of the walls ended up slightly bent from heat, possibly because there's a pocket where the flame could be pushed to. Or, somehow the heat is more easily distributed to the wall this way as the pot is in more contact with it.

It's hardly broken, and can be bent to shape again, but it's another small lesson learned.

The not so correct way up

Carrying a pot, the burner and some gas, it's all bound to take up some space anyway, so it might not be that much more space efficient than the Trangia 27. Why does that configuration need to be so darn perfect, leaving no room for invention?

Bullseye Bore

Box, container, laser

Waiting time in crowdfunding is sometimes problematic. I forgot why I wanted this Kickstarter in the first place. In any case I was probably one of the last people to receive this. And it's been some years when it arrived, I've kept waiting to do more with it before posting, but it doesn't seem to happen.

The Bullseye Bore is a laser guide attached to a power drill, with the purpose of assisting straight drilling.

Bullseye Bore connects to the drill chuck with a strong magnet. The lasers are powered by two batteries. When the drill rotates, so does the device, emitting two beams projected as circles. Drilling straight, the circles are concentric. If not, their positioning is lopsided.

Easy to connect. The chuck has to be magnet-friendly, but aren't they all?

There's another version of the device, with three circles rather than two, but I chose this simpler one, "Core Red".

So, bullseye or BS? Ha, ha, I wonder how many reviewers made that joke. Am I being a bore? 

Before I even tried the Bullseye, I envisioned a few problems with the concept. Firstly, there are many drilling tasks where straight drilling might be desirable, but the circles might not be clearly visible. For example, drilling deep into the side of a narrow board.

Secondly, the tool has to rotate quite fast to see the lasers as circles. So I have to go full speed before I even start drilling the hole. This could be confusing.

Can't even photograph it just like that

A little disclaimer. Though I have some educamation on the topic of woodworking, I don't do it as a profession. I know that drill type, diameter, RPM go together with the target material and the task at hand. But currently I only use an ordinary power drill occasionally for small tasks, and go with whatever feels fine.

After inserting the batteries, I tested the laser functionality and was ready to go. There's a really tiny on/off switch in an easily reachable position.

My DeWalt drill has an automatic LED light, which casts shadows with the rotating part, but this isn't too distracting and the LED can be covered if I really want to.

The operation reminds me of the docking sequence in Elite. I watch these glowing circles while drilling, going closer in and hoping without any great assurance, that all will be well.

Looks sturdy enough.

In practice, I was able to drill even a weirdly-shaped piece of wood. The fast dry-rotating felt suspect at first, but it's something one could get used to.

I was surprised to find that even with the first try the result was quite correct, certainly less than 0.5mm off when measured from the edges of opposite sides. You can argue that's not a great achievement for a 21mm thick material, but believe me I've been able to botch comparable freehand drill holes.

As the drill gets closer to the wood, the circles become smaller and it's no longer as easy to compare them. But at that point the drill hole ought to be growing in the correct direction anyway.

It won't fit everywhere, that's for sure.

Bullseye Bore is well made, the magnet is strong enough and the lasers are a fun addition to the drilling process. It's not an universal solution, and as with any tool and drilling assist, it likely has a specific role. Perhaps a repetitive task where mounting a jig or assist would be bothersome.

Since this arrived, I've had a number of household-type drilling tasks, but I didn't really feel a need to use this laser. It's not that often I need a deep, accurate drill hole that couldn't be done with a dowel jig.

Multipaint 10th anniversary of release

The 10th Anniversary of the first public release of Multipaint (24.2.2016) slid past without me noticing.

I actually started Multipaint in late 2013. It was called "kuppapaint" internally! The impetus at that time had to do with Marq's PETSCII editor, which showed practically how things like file selectors and preferences could be handled in Processing. I also contributed to the PETSCII editor and the rapid development was inspiring.

PETSCII editor fit a very clear gap, but there were already cross-platform pixel paint programs that worked with C64 and ZX Spectrum screenmodes. So even if I had started on a pixeling program, I wasn't initially convinced it would need to be released.

I did notice the existing software mostly worked with Windows only and their development had in some cases ceased.

Grafx2 appeared closest to a multi-platform program that could handle various legacy modes. Yet, to me making pictures with attribute limitations was difficult, and the program itself looked complex with many features I really didn't need. Despite being a Deluxe Paint-inspired software, it actually didn't remind me of Deluxe Paint all that much.

ZXPAINT Processing sketch, 2009

Prehistory: ZXPAINT

I did something drawing-related already on a ZX Spectrum, during 1988 or thereabouts. It was a BASIC program that involved moving a pixel around with a joystick. The line would go on and off with the fire button. A checkerboard dither was available.

On Amiga and Atari ST, I dabbled with BASICs such as ST BASIC, STOS and AMOS, toying around with creating editors and beginnings of tiny paint programs. But under the colossal shadow of Daniel Silva's Deluxe Paint, this never felt really necessary.

Only during the PC era, around the year 2000, I began to see the necessity of making my own editors again. Deluxe Paint was a thing in the past and the new PC programs were cumbersome and seemed to focus on the wrong things. My efforts were mostly sprite- and tile-oriented, written in c and worked directly with key shortcuts and supported my modest game programming ambitions.

With Allegro or SDL? Not sure anymore

The "dream" over the years was really concerned with either a full Deluxe Paint style program, or a program that could work with ZX Spectrum attribute limits on the fly. I seem to have been working on at least one version in 2009, called ZXPAINT.

The source code suggests that Processing changed the way file selectors worked, as the source no longer ran without modifications. It might have disappointed me and discouraged me from continuing with this effort.

It didn't look like any of the parts of the 2009 ZXPAINT source were reused in Multipaint, it's more like the experience helped in making slightly better choices later.

ZXPAINT outcome stupidvampire, November 2009

There was also the minor episode of making a doctoral thesis on the topic of self-constructed creative design tools, but I'll leave it at that. It's more that after finished with it, I could again concentrate on hobbies and fun things without the heaviness of heart that comes with having an unfinished thesis looming somewhere in the future.

Multipaint early days

After the premise of the ZX mode had been built into what was to become Multipaint, quite soon I began adding a Commodore 64 Hires mode. I saw myself making C64 hires graphics rather than ZX Spectrum. This was probably because of the 2013 PETSCII wave, I had seen how vivid the C64 scene was and wanted to contribute graphics. Back then it was not even obvious where I could submit Speccy graphics.

I also had a nefarious plan of using my own tool to gain some advantage over competition, but I soon saw it didn't help me that much, and that it would be better to release the tool into the wild. Two images, Inside Job and Countryside, were released before Multipaint itself was out.

An MSX mode was added too, so all the three first modes were attribute-limited in a nearly similar way. After that arrived the multicolour C64 mode, which was already tricky as the program had been hardcoded to support those Spectrum-style graphics. At that point I was supporting all the three major 8-bit platforms I had access to back in the 1980s.

Some of the first tests of the MSX mode, pre-2016

Around this time the idea of a "fantasy console for multi-platform graphics" came to being, meaning that Multipaint pretended to be a computer whose architecture was good for simulating various 8-bit modes. This was too much of a high concept, and I had already been coding myself into the corner, but this idea stimulated my imagination and I kept adding more modes.

It looks like I can't find any pre-2016 Processing sketches, which is a pity. The development diary exists though, and late 2013 entries suggest I worked quite intensely on it from September to the end of the November of that year. After that the ZX and C64 modes and their exports were in place. When I returned in the beginning of 2014, I begun working on the C64 multicolour mode.

Surprisingly, the beginnings of an "Amiga" mode were already in process during 2015. Atari ST was first in a release, though, and the Amiga mode had to wait.

Milestones

Not to make this overlong, there have been some larger milestones in Multipaint: 

During the first years, there was a lot of wrestling about how "intelligent" the colour limit adaptation should be. As the program advanced, I dropped off the attempt at providing any kind of heuristics for the colour adaptation, and the program simply changes the underlying colour. After that nobody really complained much about the colour model.

Adding drop-down menus, Deluxe Paint style, made the program far more accessible. I wasn't especially excited about this first, and no-one had really requested it. But many user requests and new features would have been difficult to add if there were no dialogues or menus. 

With the menus, it was now possible to "see" what features the program has by glancing at them. I could also now add incomplete and experimental modes to an "Extras" window, and get rid of the separate system dialog for choosing the platform.

Pull-down menu in Multipaint 2025

The first 16-bit modes provided further challenges, but after adding Atari ST and Amiga, and a possibility of switching between modes without restarting the app, Multipaint really became more "multi". Before that, it wasn't even possible to load a file of another platform without restarting Multipaint first.

The recent years have brought app window resize, separate preview window and external scripts for handling the various target platform formats.

One of the newest additions has been the "recentfiles.txt" list, which lets the user load a previously saved file from a short list. This was surprisingly tricky to do, what with the different save formats and scripted save modes. It has likely resulted in some obscure file-related bugs that I may still be hunting for the better part of the next decade.

So, you thought about making a retro paint program?

There's been contenders over the years, but now there's almost a deluge of new paint programs supporting C64 and ZX Spectrum modes.

It's probably not too hard to make a single-platform paint program that's better than one of Multipaint's modes. I still believe there's something valuable in Multipaint's simplicity, and the idea that if you've learned one mode you might want to explore others too. A lot of great graphics have been made with potato-level editors, because the basic drawing functions were good enough.

I'll share some advice, some obvious, some sarcastic, but here goes:

• It's a good idea to show the attribute/colour effects already as the cursor floats over the canvas.
• It's NOT necessary to show a "transparent version" of those effects.
• It's ABSOLUTELY NOT necessary to have a "flashing cursor". Granted, I've not seen this in the pixel programs, but in some PETSCII editors it's a pest.
  
  
• Can we have a grid that's not garish and overblown? If the grid is something else than white/grey/black, it will affect the colour perception of what's under.
• Personally I don't understand the need for a pixel-level grid, but C64 programs used to have things like that in the magnify mode. If you do, make an on/off switch.
  
  
• Supporting every possible C64-palette known to man? I tried through my own actions to limit the number of palettes in the wild, but to no avail.
  
  
• Obviously everyone has to come up with their own keyboard shortcuts, or assume that Adobe products' keyboard shortcuts are a widely accepted standard.
• If I could go back in time, I'd make all possible key shortcuts align better with Deluxe Paint.
• I've tried, when really necessary, look into how GIMP and Krita do things.
  
  
• Be assured that if a tool is not very visible in your interface, people will assume it does not exist.
  
  

Extras in Multipaint 2026

It can be smart to add features people suggest. I doubt Multipaint would have been what it is, if there had been no requests or I would have ignored them all.

But after a point you'll have to try to figure out what's your "vision" and boundaries for the software.

• In Multipaint, I decided not to go forward with multipage, full animation or layers. I came to realize Multipaint "pretends to be an old paint program" which helped keep it limited.
  
  
• Importing could be better, but giving it a thought I don't see it a huge priority. If I can discourage people from making conversion-based images, the better.
  
  
• Feature Creep is a thing. Adding a feature seems innocent enough, just make it so it can be turned on/off, man! But even this would add yet another dialog or option in a program that was meant to be simple.
  
  
• Some have requested adding really obscure platforms, in the hope their platform would get more support and visibility. But from what I've seen so far, the support doesn't really arrive that way.
  
  
• "It doesn't have feature X, therefore crap blah blah" is often talk from people who look at it from the outside, probably without really trying the software.
  
  

And the usual: It's often much more fun to start a project than it is to maintain it... but if the project has major updates and good new features, these can again be fun to tinker with.

Multipaint's future is still assured, but not endlessly bright. It depends much on what happens on the Processing/Java front.

...

http://multipaint.kameli.net/

PDsid

I Love SIDs

The PDsid, or Public Domain SID. I bought two as UNI64 sold these for the very reasonable price of 8.90€.

These are drop-in replacements for the Commodore 64 SID sound chip. The PDsid can do 6581 and 8580 and no other models.

The mode can be switched by holding the reset button for three seconds on the C64, I suppose everyone has a reset button? The status should stay even after power off. The mode can be read and altered through software too.

We Love the 6581/8580

Installing PDsid on my everyday C64C turned out to be a longer process than expected. 

The metal shield cover on the circuit board doesn't exactly fit with the PDsid USB connector part. This board extension can be snapped away but I chose not to do this as my first action. The USB may be needed for possible firmware updates.

To the horror of purists everywhere, I instead mutilated the metal cover, drilling holes and bending it away. I had misplaced my hacksaw so I had to resort to this ugly business. If going with this route, I recommend to be generous with the opening as the chip is somewhat wider than a SID.

Oh and in this C64C the metal cover is necessary because otherwise the keyboard holders don't work very well.

Vicious!

After this, I turned on my computer and found the chip working. Using the "reset for 3 seconds" I could hear the double-beep, indicating 8580. I then tried the usual suspects, Lightforce, Giana Sisters, Rambo, Delta, Master of Magic, Commando, 720, Stunt Car Racer and could not hear any problem with the sound at least on my Commodore 1084S loudspeaker. 

When listening to my Digiloi in-game tune, which I've heard ad nauseam, I imagined there could be some additional rumbling at the low end. However I've not heard the song often on the 1084S display loudspeaker so it could be simply related to that. 

Alternative theory is that the PDsid is somewhat more "precise" than a real SID, revealing certain type of sound in more clarity.

One more thing is that the firmware I received might still have a tiny filter initialization bug, but as far as I understand it shouldn't affect anything after reset.

As I can't compare a real SID side-by-side in my setup, I should perhaps shut up.

I could have put the chip to play more difficult SIDs, but as I'm not super familiar with demoscene songs, it might tell nothing to my ears. More importantly, my SD card didn't have any demos at hand.

Detect status and read chip model, write them at top left corner of the screen

With the API it's possible to detect the presence of PDsid, read the selected chip (6580/8580) or change the chip emulation between 0=6581 and 1=8580. I used the Action Replay VI machine code monitor to test a few code snippets.

One of the PDsid instruction sheets is ambiguous, as it tells to write 'P', 'S' and 'D' to addresses, and as we know these letters are different in C64 PETSCII, Commodore BASIC and ASCII! It turns out ASCII values are intended, so it is $50, $53 and $44 for P,S and D respectively. The included code example shows how to do it properly.

The chip uses the $D41D, $D41E and $D41F as API, these are unused addresses and the 'P'+'D' scheme means it's highly unlikely any existing software would accidentally mess with the chip status.

All in all, my first impressions of PDsid are very good, hopefully it continues to be available in the future.