The Quarky Engineer

I normally use a regular telescope setup with an equatorial tracking mount and dedicated astro camera for astrophotography. It is complicated to setup, heavy, and a bit time consuming to get up and running. It has the advantage of being larger aperture and far more capable in terms of exposure time, with the tradeoff in complexity.

At the other end of the spectrum are the newest fully automated electronic telescopes. These are devices you turn on, connect to with an iPad and image away.. virtually no setup at all. A few years ago a company called Vaonis made a device in this category called Stellena, and in 2021 they releases a 2ng gen device called Vespera. I recently got a Vespera unit and fired it up for the first time last night.

I turned it on, hit an init button in the software, and it spun around and did a few captures, plate solves, and generated the internal calibration. I then selected the dumbbell nebula in the catalog, and 10 seconds later it appears on my screen. It could not have been easier. It continually shoots and integrates, so the picture gets better and better as time goes by. I picked a few more good targets including the Pleiades, the Veil, and Horsehead, and Orion.

Keep in mind this is from my house in the city of Portland, so Bortle 7 in terms of darkness[terrible light pollution], and it was a 1/3 moon. These are jpgs directly from the device, no processing at all. Just as it appeared on the screen.

Downright amazing results for such a small 70mm aperture with a fully automated stacking system. You can download the original subs and do your own stacking with much improved results, and I’ll give that a try tonight.

I have the Vespera specific HA/O3 filter on order, and once that comes in I should be able to really improve the nebula shots.

This really is an easy way to get into astrophotography with very fast return on your time investment.

I’m trying out a new mount for astrophotography – A ZWO AM5. It is somewhat unique in being a harmonic drive instead of a worm gear. It doesn’t need a counterweight, and there is no backlash. It is also really quiet. I’ll be using an ASIAir Plus to control it as well as the guide camera and primary camera. With autoguiding and good polar alignment I should be able to get some good photon collections.

Looking forward to some clear skies and experimenting with this setup.

This was an interesting picture – there is a shadow on the upper clouds created by the lower cloud and the sun. Not a configuration I see often.

While working on a cool retro super computer project with Christopher Neil Bradley I wanted to add a VGA interface to our board. I designed one on a proto board that has both a 640×480 graphics mode as well as a character mode with multiple font support and a neat way to do a cursor. A fun little side project. For the actual PCB one I’ll add a 256 color pallet.

While in California for Audrey’s horse show we stopped by an interesting artistic light exhibit. It was apparently built a few years back as a temporary thing but was popular enough that it was made more permanent.

The lights in the fields were using fiber optics that came from small transmitters, and all of the power for the lights comes from batteries charged by solar panels. An interesting use of fiber.

Picked up my new laptop today. Screaming fast! And the original 10MB hard drive still works, and has no bad sectors.

In this weekends vintage activities – I played around with what is my favorite of all of the x86 processors, the Intel Pentium Pro. The Pentium Pro delivered a few key milestones – It was the first x86 processor that employed out-of-order execution, register renaming, and speculative execution. Those microarchitecture features would be carried forward far into the future.

It was the CPU in IBM’s ASCI Red supercomputer, the first supercomputer to surpass the 1 teraFLOPS mark. (Note the graphics card in the computer you are probably using right now is faster than that).

It was also a very unique physical packaging with two dies (CPU and cache) interconnected. The in-package L2 cache was a first for Intel.

I was working at Intel when this was released, and it was an exciting time in microprocessor architecture. I remember reading the data sheet and being amazed at the dynamic execution flow.

For many years following its release it was the processor of choice for workstations and servers, with direct support for 4-way SMP. There were a number of 2-way workstation motherboards available as well. The Pentium Pro was probably the platforms I used for the longest amount of time.

This particular variety was a 440FX(‘Natoma’) Intel motherboard with a 200Mhz, 256KL2 Cache CPU and a ATI Mach 64 video card.

Of course I played Quake on it, and yes I used FastVID. The 90s live on.

It is interesting to see how AI focus generation neural networks have evolved.

This is a scan and zoom of an old picture. In the zoom you can see the limited resolution of the original scan. (This was a picture my dad took of me, certainly on 35mm film, in the mid 70s).

The second is the AI generated focuses image. All of the ‘extra’ data is synthetic and didn’t really exist. It is surprisingly good. Perhaps we are entering the era of ‘Enhance!’

I thought it would be interesting to share some thoughts after my first year with the Tesla Plaid. I drove about 6000 miles in the Plaid, and about 6000 more miles across 8 other ICE cars, not including my wife’s BMW X5. Clearly I drove the Plaid the most, in part because of convenience (always full of power), and in part a deliberate choice.

The ‘always being full of power’ is certainly something you grow to appreciate. I charge the car every night, and every morning it is full of happy-go-lucky leptons. While it isn’t particular hard to get gas, it is still a 10 min drive from my house in the West Hills.

I didn’t drive the car on any long trips (more than an hour), and probably never will. I have other cars for that. This car has the purpose of getting me around town as well as producing smiles. Range is not the most important factor, but the 390 miles provided is certainly adequate. I make extensive use of the foot potentiometer, and make it a practice to launch the car with the factory launch control every day.. at least once, often twice. I have discovered with the launch-control I can do 0-60 in my driveway, which is awesome. I am getting close to needing new tires…. I suspect I’ll get 8000 miles out of the Michelin Pilot 4Ss I have on there… about what I would expect given the abuse.

Any conversation about driving the Plaid starts with the power. Indeed the 1000hp feels like 1000hp, but more like it is coming from a V16 engine with nearly perfect torque. 0-60 in 2 seconds just doesn’t get old… and the 0-150 in 9.4ish seconds is equally amazing. Even more valuable is the availability of most of this acceleration in sub-optimal conditions. Even with a terrible road surface you can manage a mid to high 9 second 1/4 mile time. In a straight line the traction control algorithm is downright fantastic. You rarely get any wheel spin noise at all, and if you do it is for a few hundred ms at most. Even in pouring rain I can get mid 3 second 0-60s, which is mind boggling.

I have found that most passengers are just uncomfortable with the acceleration, especially above 60mph. On the flip side, most 10 year olds thinks it is spectacular. The change in power from 90% SOC to 50% SOC is noticeably, but not disastrous. If you were blindfolded and didn’t drive the car a lot I suspect most people couldn’t tell the difference. Below 50% SOC you can notice it more.

The traction control while turning is downright terrible. It has been calibrated such that if the wheel is turned more than about 5 degrees the power output is significantly cut. It still has power, but compared to the mind altering straight line power it is very noticeable. It is no doubt tuned to keep people from loosing control due to excessive acceleration while cornering. Track mode changes those dynamics a bit and allows more shenanigans, but still much less than many other fast cars.

The brakes work, and they can stop the car, but their biggest limitation is in total energy dissipation capacity. One or two hard stops from 150mph works just fine, but the system heat capacity is low enough that any more added heat will overwhelm the system, and you can quickly find your self running out. Tesla is offering a Ceramic Brake option, but it isn’t available yet. I would opt for that in the next version. Around town you probably won’t notice, but if you do some serious spirited driving and track days you will.

The yoke steering wheel was a conversation starter from the beginning. I would summarize one year of using it like this: It is a steering wheel. It steers. Beyond that, it is not anything magical, nor is it something horrid. The better view of the dash and the road is nice, and noticeable. Parallel parking in tight areas is worse. It mostly fades away as not that important compared to other dynamics in the car.

The turn signal buttons are at the very top of the suck list, and they are perhaps the single most negative thing after 1 year of driving. They just are not intuitive enough, or common enough, to replace the familiar stalk. They work, and you can use them, but it takes more concentration and effort, and slows down transitions. It is just not a very well executed idea.

Interior quality is exactly on par with what I expected it to be. It is nice Lexus level, not Mercedes. It has actually worn well, but then again I don’t have much to wear it with. My 10 yr old daughter has spent lots of time riding in the back seat without complaint, and the white interior still looks perfect.

Maintenance has been a total of .. well .. nothing. I plug it in, it drives. Every day. Perfectly. I have not been back down to the Tesla dealer since I picked the car up new.

There are a few things that have surprised me:

(1) I am amazed I have not broken an axel. I launch the car all the time, and the acceleration is just brutal on the drivetrain. 1.3gs of acceleration needs a lot of twisting force, and I’m just surprised it has not snapped.

(2) The roll on acceleration is also surpassingly harsh and snappy. It seems like if I did this in most of my other cars something would break.

(3) The lack of noise – It is fantastic in terms of not yelling to everyone within a 2 mile radius that I’m being a jackass, but it is also unceremonious. A positive more often than a negative, but a disappointment in the soul of the beast. When I drive my 1100whp GTR that soul is ever present and window shaking. 😉

Perhaps my biggest takeaway from this last year is enthusiasm for the future. Lepton powered cars need not miss the excitement. This is a Tesla, in Tesla form and finish, but the potential for others to do DIFFERENT and awesome things with electric powertrains seems limitless. Just like many other petrol-heads I can see the decline of the internal combustion engine as the end of an era, but one can also see the beginning of another, and it doesn’t have to suck.

Today had a little playing around with the Pentiums. I started with the 90MHz SX909, which is the version that has the famous FDIV bug, and ran up through 100,133,150,200, and finally 233 Mhz. It is quite a production range that can all be used on the same motherboard and chipset.

Playing Doom and Doom2 really highlights the significant speed difference over the 486DX2-66. The combination of the faster local L1 and L2 cache, and the superscalar execution is noticeable.

It is nice that all of these still work fine, and a few of them were overclockable.

Good memories across these as well. I have a great memory buying a Pentium 133 when it was first released.