Needing more than a spark test?

[graham-xrf]

We used to say, "anyone can build a couple, but it takes skill to make millions".

In this case, if anything like that starts to happen, and nowhere near millions, we would already have trouble. It being all open, we can rely on some (in China) to do a knock-off. For the relatively few folk at HM who might try one, we can help them, so long as we keep it simple and reasonably sure-fire.

(Lots of engineering!) Or our participants will use products from different lots, (or manufacturers,) or inadvertently get counterfeit parts which maybe externally indistinguishable from real ones. Causes a lot of heartache. Any way to eliminate analog tweaking in this X-RF unit is probably a good idea.

I had not thought of counterfeits! Anyway, regarding analog "tweaking", I do agree. Way back, when I first started working in (military) electronics, I would eliminate all multi-turn adjustments. Not even for offsets. It either had to have too little to care about, or the thing did software self-calibration every few minutes. Digitally controlled potentiometers were my thing.

This is not to say that software never has problems! Seen lots of issues. But overall, it is usually faster and cheaper to scale up production of digital designs than analog ones. That's why we see digital overtaking analog.

The noise immunity - either it's on or it's off. It could not last, even at 12V CMOS. There came a point where 5VTTL for computing was sucking the output of power stations, even when we went to LSTTL and FPGA arrays. Drop it to 3.3V, and then to 1.8V, and use MOSFET technology. So much closer to the noise! Getting used to "eye height" displays. Sure, one can clearly see the point where going digital, with the profligate expenditure of bandwidth, gives the advantage.

On that very subject, is what the link in #773 was all about.
Derek Muller (Veritasium) is sponsored, but he is completely up front about who, and why.
The Most Powerful Computers You've Never Heard Of..

 

[WobblyHand]

In this case, if anything like that starts to happen, and nowhere near millions, we would already have trouble. It being all open, we can rely on some (in China) to do a knock-off. For the relatively few folk at HM who might try one, we can help them, so long as we keep it simple and reasonably sure-fire.

I had not thought of counterfeits! Anyway, regarding analog "tweaking", I do agree. Way back, when I first started working in (military) electronics, I would eliminate all multi-turn adjustments. Not even for offsets. It either had to have too little to care about, or the thing did software self-calibration every few minutes. Digitally controlled potentiometers were my thing.

The noise immunity - either it's on or it's off. It could not last, even at 12V CMOS. There came a point where 5VTTL for computing was sucking the output of power stations, even when we went to LSTTL and FPGA arrays. Drop it to 3.3V, and then to 1.8V, and use MOSFET technology. So much closer to the noise! Getting used to "eye height" displays. Sure, one can clearly see the point where going digital, with the profligate expenditure of bandwidth, gives the advantage.

On that very subject, is what the link in #773 was all about.
Derek Muller (Veritasium) is sponsored, but he is completely up front about who, and why.
The Most Powerful Computers You've Never Heard Of..

Sure-fire, is what we need to make.

Got burned on some counterfeit parts. Marked the same, but didn't meet offset and bias specs. Besides that, they seemed ok. Might have been rejects that someone scooped up, or actual counterfeits. Genuine parts were much better behaved. Circuit worked again!

Oh that video was a tease. Lots of stuff, then quits just before showing the punch line.
The part about analog tide computing and Lord Kelvin's use of the new fangled Fourier stuff was interesting.

 

[graham-xrf]

What do you do for batteries? The LED displays tended to run down the batteries pretty quick. I have an old HP45, with the leather case. I'm not sure where the heck the charger is! For general use, when not at a computer, or in the shop, I use a HP15C. Three button cells, lasts for years, if not 10 years. Long live RPN!

I have seen DIY pages, showing how to dump the NiCads, and put in Nickel Metal Hydrides. I don't know if any opted for a single Lithium. On my ancient original HP25, I cut one side of the plastic retaining strip that sits between the AA's, and I found I could use alkalines. External charged Ni-MH could be slipped in, and the battery carrier put back.

There are folk who restore these, and they publish modifications circuits, not only for charging. There are videos on reworking HP45's, and HP67's. If all else fails, you can resort to a look-alike app on your smartphone. I always wanted the non-volatile feature of the successors to my LED HP25, but could not afford them. I was not so keen on the one with the magnetic strips programs store. I let it, and slide rules, etc. drift into past memories. Then I get the sweet Christmas present. HP35 became legend, and now I have one!

You can get in depth stuff, circuit diagrams, etc from sites like this --> HP calculators LINK

 

[homebrewed]

Old lathes seem to be like old houses. Once you start tearing into them you find other problems. In my case, I discovered that my lathe's drive belt was literally hanging on by a thread. Fortunately, it's easy to replace. It's on order from LMS.

I've got the front bearing + spindle pulled. Now to get the rear bearing off....

 

[graham-xrf]

The tiny $15 Pi Zero 2W I wanted.. sold out!
One may as well try for rocking horse poo!

It can do all one wants, although, with 512Mb of RAM, the kiddies who might try and load up Chromium and head for YouTube might be disappointed! Actually, it would do it if run in a less bloated browser.

It can do the scheme I suggested before, where one of these is back-to-back with the photodiode PCB, and of course, on the "user side" of the lead sheet shield. It would have a webserver, just like the one I already have on my Raspberry Pi mailer, where the output plots are running live on a web page. Any PC, or smartphone, that has a browser, could just display the page. The network is standard, and can be via WiFi, BlueTooth, EtherNet (via a little USB adaptor), or plug in a 1080p monitor direct into it's HDMI port.

From Farnell, the Pi Zero 2W are £11.12 each, with a lead time of 120 days, and request for quote for quantities more than one. I guess they might be thinking to ration them. The Pi Hut did that before, one per customer, and they still ran out!
[ Edit: Today from Farnell - all forms of Raspberry Pi 4, whether 2GB, 4GB, 8GB RAM - all sold out! ]

For now, the stuff I have runs on any Pi, so it does not stop development. We shall see..
Hopefully, your real steel hardware is anticipating nice clean bits, fresh oil, superdooper bearings, and a belt that makes it feel like new..

 

[homebrewed]

Hardkernel.com has some SBC Linux boards that have pretty good performance. I have three -- two C1's (no longer offered) and an 8-core job configured as a media server for our TV. One of the C1's is used as a print server and the other is used for a home-brew call blocker. They all have HDMI outputs and a bunch of GPIO's. They all have either wireless or ethernet connectivity to our internal network so I can manage them remotely.

Regarding my lathe, I've got the old bearings pulled and the headstock modified to provide an access port for lubing the transmission gears. The volume occupied by the transmission normally has no access unless the headstock is completely removed from the lathe bed. An oversight, in my opinion. Other owners have cut square holes on the top of the headstock and then added a plate to keep swarf (and tools) out of there, but I like my approach better. In normal use the access hole I drilled is covered by the motor controller box so it's a very simple solution. Just my style

Now I can start putting the lathe back together. My main concern now is pressing the bearings back into place without damaging them (along with not ending up with "extra" spacers, screws etc.). I have my mill and a rotary table so I'm OK if I need to make something the right diameter to accomplish that.

 

[WobblyHand]

The tiny $15 Pi Zero 2W I wanted.. sold out!
One may as well try for rocking horse poo!
View attachment 394309
It can do all one wants, although, with 512Mb of RAM, the kiddies who might try and load up Chromium and head for YouTube might be disappointed! Actually, it would do it if run in a less bloated browser.

It can do the scheme I suggested before, where one of these is back-to-back with the photodiode PCB, and of course, on the "user side" of the lead sheet shield. It would have a webserver, just like the one I already have on my Raspberry Pi mailer, where the output plots are running live on a web page. Any PC, or smartphone, that has a browser, could just display the page. The network is standard, and can be via WiFi, BlueTooth, EtherNet (via a little USB adaptor), or plug in a 1080p monitor direct into it's HDMI port.

From Farnell, the Pi Zero 2W are £11.12 each, with a lead time of 120 days, and request for quote for quantities more than one. I guess they might be thinking to ration them. The Pi Hut did that before, one per customer, and they still ran out!
[ Edit: Today from Farnell - all forms of Raspberry Pi 4, whether 2GB, 4GB, 8GB RAM - all sold out! ]

For now, the stuff I have runs on any Pi, so it does not stop development. We shall see..
Hopefully, your real steel hardware is anticipating nice clean bits, fresh oil, superdooper bearings, and a belt that makes it feel like new..

I was looking for the Launchpad board yesterday to make an electronic lead screw. Nope. TI has none. The official distributors, Mouser and Digikey have 0. Mouser says they may get some by 7/29/2022. This is the chip shortage. It's also all the little cheap peripheral chips that are in short supply. They are often made on old process nodes, and the manufacturers have made 0 investments in those nodes, because there's miniscule payback for making cheap chips cheaper. It's amazing how dependent we are on these cheap chips. Our modern day opium.

So, use the Pi's that you have. I have a few laying around, from an early one with a 26 pin connector, a couple of RPi2's a few RPI3's and some RPI4's. None of the little ones. They just don't quite have enough memory to make it easy to get things done. Have an RPi2 as a print server, and one as a sprinkler system controller. Perfect use for them.

Eventually supply will catch up, I hope.

 

[graham-xrf]

@WobblyHand
Your spare Pi resources are more than mine. I am now unable to mention Pi4(s) in the plural. I have had one Pi4 running for near a couple of years, 24/7. It's my mailser/web server. That one is dominated by the IceTower on it, and so it doodles along at about 36C. The MicroSD only runs a few lines at boot-up, which then gives it all over to a 500GB USB3 SSD drive, which has the OS, etc. on it.

My only "other" Pi4B sadly, looks kaput. Clearly most of it works, but whatever it uses to work the MicroSD connection does not
Beyond that, a couple of Pi 3B+ , which work just fine. I have them right now being used as XRF tryouts testbed.
Hanging around in the "electronics resources "junkbox" is an old Pi model A with the 26 pins, and a "hat" with a wobbulating signal generator on it. No electrons have moved in that one for years!

For guys like Mark, and yourself, the computer part is not a problem. You would hook up what takes your fancy. I was looking to the Pi Zero as something most any other interested user could reasonably get going with this, without any special computer or electronics expertise. That it comes at such a low cost (when available), was the motivation.

The business of chips shortage is interesting. It is having a big hit on new car manufacture, phones, and all sorts of products. It seems that when COVID hit, the last thing folk were going to do was visit socially distanced car showrooms. They feared for their jobs, and hung onto the wheels they had, sending second-had prices way up. Even a small percentage downward shift in car sales throughput would cause layoffs, etc. These places operated with just-in-time deliveries, and they were turning off the chip supply tap. The chip fabricators moved to making other silicon for other clients. Cars have hundreds of processors in them, and a car-maker advance order, placed late 2021, was going to the back of the queue. Something similar happened to chips for smartphones.

An exception was Elon Musk's Tesla. They kept their orders in, and placed new orders for new designs, even if they would have to stockpile the chips. What happened to Raspberry Pis, especially the Pi4B, their Compute Modules, the Pi Pico, and the Pi Zero, is that demand soared way beyond expectations, because they were being widely used for embedded control and robotics in industries outside of their planned target market.

Random Topic Change Tangent
I am browsing through the available soldering iron ensembles on offer (Amazon, eBay, etc) My old 45W 24VAC Weller with the bang-bang "thermostat" switches Curie temperature magnets on the back ends of the tips still works OK, but it is a tad unkind to surface mount tiny stuff. I am also going to have to move on to using heated re-flow like Mark (and you?) do. It's the only realistic way to deal with 0.5mm pitch quadpacks.

Your picture of a LT6237 DFN-8 soldered onto a DIP adaptor (post #702, page 71) was, in essence, us admitting that for practical use, we have to "convert" the chips back into something of a size we can get our fingers around, and apply a Weller to without destroying it! They made the stuff too small, forcing the stupid irony of reflow soldering the things to little PCBs that can have it fit in DIP-size layouts!

 

[WobblyHand]

@graham-xrf, I have a Hakko FX-888D soldering iron. It works very well and you can set the temperature to a degree, no matter what kind of degrees you are working with. It wasn't super expensive, and all sorts of iron tips are available. I use it for all the DIP pin size stuff and some 0.5mm work. Has a holder for a sponge, and also some curly steel wool for removing tip oxidation. It's a nice iron. Have no idea if it is still available.

These days we need to be able to work with the little stuff. For that, an oven or hot air is a must. I couldn't find parts in the bigger packages, (that was two years ago!) and didn't want to spin a board for prototyping a circuit I came up with. Hence the silly adapter PCBs, then wired into perf boards. Not optimal, but it does work.

Next go around, I'll re-teach myself KiCAD and do some boards. I have used it, for schematic capture, but was having troubles adding parts and footprints. Might revisit it again. However, in a strange turn of fate, the ppa-release for FreeCAD uses a library libocct which is a higher revision than that used by KiCAD and my base distribution. My package manager most generously removed KiCAD. Kind of frosted about that. There are workarounds, but honestly, that's rather unfriendly of FreeCAD, breaking other packages. But, have been learning FreeCAD, so not all is lost.

 

[graham-xrf]

@graham-xrf, I have a Hakko FX-888D soldering iron. It works very well and you can set the temperature to a degree, no matter what kind of degrees you are working with. It wasn't super expensive, and all sorts of iron tips are available. I use it for all the DIP pin size stuff and some 0.5mm work. Has a holder for a sponge, and also some curly steel wool for removing tip oxidation. It's a nice iron. Have no idea if it is still available.

These days we need to be able to work with the little stuff. For that, an oven or hot air is a must. I couldn't find parts in the bigger packages, (that was two years ago!) and didn't want to spin a board for prototyping a circuit I came up with. Hence the silly adapter PCBs, then wired into perf boards. Not optimal, but it does work.

Next go around, I'll re-teach myself KiCAD and do some boards. I have used it, for schematic capture, but was having troubles adding parts and footprints. Might revisit it again. However, in a strange turn of fate, the ppa-release for FreeCAD uses a library libocct which is a higher revision than that used by KiCAD and my base distribution. My package manager most generously removed KiCAD. Kind of frosted about that. There are workarounds, but honestly, that's rather unfriendly of FreeCAD, breaking other packages. But, have been learning FreeCAD, so not all is lost.

I had a stumble or two, so I maybe had better let you know what happened with me.
Forgive me if you already got to this.. but anyway..

KiCad 6 is a huge update! It has got so much new stuff, and it seems all real nice. I suppose that there can still be some minor annoyances if one is used to something else, but I can see why it's now being widely used. It has some cool new stuff, like dim-out transparency to layers you are not working on with the current one overlaid up clear.

The new one --> Blurb & Pictures
Even with my latest LinuxMint, the Ubuntu repository it uses is 6 months or more behind in previous versions, so it loads up KiCad 5.
To get KiCad 6.0, I had to use the ppa:kicad/kicad-6.0-releases
You can just follow the downloads page, and select for your distro --> https://www.kicad.org/download/
BUT..
If you ever had a previous KiCad install - although installation of the new one does happen OK, there is a stall!
It won't see any global libraries, even though they do exist, and are all correctly installed.

In my case, I had first used the software manager for getting KiCad 5. Then almost immediately afterwards, I was going after the newly discovered KiCad 6.0 Thus, I no longer looked like a first-time user. For many, they may have a stash of their own library components and symbols. This is still OK. Everything in ~/.local/share/kicad/ can stay. It will create a "6.0" folder in there. You can move your stuff in, and if you did use that path, still no problem.

It is the other hidden folder that needs attention! I purged the old KiCad, including configuration files, but there is a place the installer (sensibly) will not touch. That is the hidden folder in your home directory. It wants to make ~/.config/kicad/6.0/ when you are detected as a "first user", but there is a file in there called sym-lib-table, where the name and type of all the symbols are listed. It will be stuffed full of paths with the wrong environment variables. It needs to see ${KICAD6_SYMBOL_DIR} everywhere.

The solution is easy. Purge the old KiCad. Save any of your own creations safe. Delete the hidden folder in .config. Then install. On first use, you get the pop-ups that guide you, and it all works from there on. Opening any projects from the old version will work, but it saves them as the new version, and they then can't be opened in the old version. It has a facility to re-name files with .lib extension to the new extension it uses.

It just looks all really nice, with 3D views and all. There are a bunch of YT videos on using it.