Michael A. Covington    Michael A. Covington, Ph.D.
Books by Michael Covington
 
Previous months
About this notebook
Search site or Web
Ichthys

Daily Notebook

Popular topics on this page:
When to discard old but unused electrolytic capacitors
High-voltage electrolytic capacitors in series
Has the earth shifted?
Using a UPS with medical equipment, lift chair, or CPAP
XGecu TL866 II Plus universal programmer
Astrophotos:
Moon and Jupiter
Many more...

This web site is protected by copyright law. Reusing pictures or text requires permission from the author.
For more topics, scroll down, press Ctrl-F to search the page, or check previous months.
For the latest edition of this page at any time, create a link to "www.covingtoninnovations.com/michael/blog"

2019
October
17

(Extra)

XGecu TL866 II Plus universal programmer
Too cheap to be believed?

Back in August, I ordered a handful of Chinese-made electronic items, trying to get some good deals before the trade war made things more expensive or less available.

Picture

One of them was an XGecu TL866 II Plus universal device programmer, which programs many types of microcontrollers and ROMs and also tests digital ICs. By "programs" I mean it will write a HEX or BIN file onto a chip or read the contents out to a HEX or BIN file on your computer; creating and compiling the program is done with other software.

I bought mine from Amazon for under $60 delivered. They still have them at that price.

And it stayed in the box until today, when I finally decided I needed to get it out and test it, before all hope of returning a defective unit is gone.

Some background:

  • This is not a $700 instrument, but it comes (reasonably) well recommended on forums. It does a lot for the price. I'm going to use it for microcontrollers that I use only occasionally and for which I don't have a better programmer. For example, it would have spared me having to build a special jig for my PIC16F84A project a few months ago.
  • The maker, Haikou Xingong Electronics Co., Ltd., also known as XGecu, also known as Autoelectric.cn, apparently caters to the automotive market. Many people want to duplicate the ROMs or microcontrollers that control car engines, or install modified code in them; relatively few people actually create the code.
  • Of course, the TL866 II has a lot more uses besides that. It even supports some in-circuit serial programming protocols as well as programming in its 40-pin ZIF socket. The only thing it won't do is program anything that requires a programming voltage higher than 18 volts. That means some 30-year-old PROMs are out of reach, although newer substitutes for them are just fine.
  • The earlier TL866A and TL866CS were so widely counterfeited that the manufacturer took them off the market. Since legitimate units had been sold under many names, there was a lot of confusion, and, frankly, I'd expect a counterfeit one probably to be as reliable as the original. (One thing a universal programmer can do is duplicate the chips in another universal programmer!) But, of course, the legitimate developers should get the income.

Per Amazon's recommendation, I downloaded the latest software from the manufacturer, and software is this gadget's weak spot, although it does the essential things impressively.

What don't I like? Mainly strange unawareness that Windows is a multi-user operating system and has definite places to install files. That used to be an endemic problem in the microcontroller industry, but most of the United States has grown out of it by now. Not China.

The software doesn't install in C:\Program Files. It wants to install in D:\XGpro, but my drive D is a DVD drive. I couldn't put it in Program Files because it has to be installed in a folder with write permission (and says so); nor could I put it in my own Documents folder, for some reason I don't recall. (In both cases it installed and then didn't work.) Further, it can't be uninstalled by doing anything other than manually deleting all the files and the shortcuts.

I put it in C:\XGpro and it works. I have no idea if another user account on the same machine would also be able to use it.

The other quirk of the software has to do with some of the lettering on the screen. There is a font size problem, like this:

Picture

Only small areas of the display are affected. I wonder if the author relied on a font that is not distributed with English-language versions of Windows. Follow-up: To fix the problem at least partially, find XGpro.exe, right-click, and choose Properties, Compatibility, Change High DPI Setting, Override, System (Enhanced). However, I still think there is a missing font, because when I choose Chinese or Russian as the language, I do not get the appropriate characters on the screen.

Does it work? Well, I read the code in an old PIC12F509A without problems, and I tested a 74HC373. It worked fine. That's as far as I got. It would take a lifetime to put this device through all its paces. But I'm fairly confident that they did the essential things correctly and were sloppy about less essential things, such as installation and user interface.

2019
October
17

Which of these is more multicultural?
Exotic characters in astronomical names

The latest Journal of the British Astronomical Association contains a letter from me, proposing that the IAU's official names for astronomical objects should always be written in the 26-letter alphabet with no diacritical marks or other special characters, though a more decorated version of the name can be offered as an alternative.

I wrote it in response to a letter from Jean Meeus that cited asteroid names containing African click marks and Polynesian long vowel marks.

My concern is that astronomers are not linguists, will not know for certain whether special characters have been transcribed correctly or rendered into different fonts, and do not have a ready source of information about exotic languages.

What's more, they have to make alphabetical lists, and how do you alphabetize a mark improvised by a field linguist 50 years ago? Even if Unicode knows, people don't.

Accordingly, for the sake of international communication, we need to confine ourselves to the alphabet that is on almost every keyboard all over the world.

My purpose is to encourage multicultural naming. There's no reason astronomical names all have to come from Greek and Roman mythology. But to use names from other sources, we have to package them for usability.

In an online forum I got a counterargument (I don't know how serious) that by insisting on the standard alphabet I was somehow supporting the evils of colonialism.

It is hard to know what such a person is wishing for (nor am I sure it was a serious argument; he may just have wanted to see me knock it down, which I did by pointing out that his whole argument was written in a colonialist language, namely English).

My goal is not to make English supreme, but rather to use the most widely available set of alphabetic characters in the entire world, the ones that are on any keyboard that uses any form of the Latin alphabet, and are supported by even the oldest software.

And if that's not multicultural, I don't know what is.



Using a UPS with medical equipment, lift chair, or CPAP

Picture

What if you want power-failure protection for something that is not a computer, such as a small medical device; a lift chair or power recliner that would be hard to get out of without using its motor; or a CPAP machine?

You can use a computer UPS, but there are a couple of things to pay attention to.

First, some older and larger UPSes shut themselves off if no power is being drawn from them at the time of the power failure. (They think the computer has shut down.) This is a problem if you're in a lift chair, which, of course draws no power when you're not moving it. Then, power fails, and you push the button on the chair, expecting the UPS to power it, and it doesn't. Bad news.

I observed that behavior in APC Back-UPS 500 units made 20 years ago, but not with the modern Amazon Basics 400 VA UPS that we're now using. If you have the problem, solutions include keeping a night light connected to the UPS to provide a continuous load, or possibly hot-wiring a switch or button that will wake up the UPS.

Second, you probably don't want the UPS to start beeping when power fails, especially if it's on a machine that you use while sleeping, such as a CPAP. Usually, the beeper can be turned off. For the Amazon unit I just described, the procedure is:

  • Turn on the UPS by holding down the power button until you hear a long and a short beep.
  • Then turn off the beeper by giving the button two brief presses (some experimentation will be needed to find out how rapidly to do this). When successful, you will hear two short beeps.

I wrote these brief instructions on a label and attached it to the UPS.

A more heavy-handed tactic is to simply unsolder and remove the beeper. I've done that several times in the past.

Last, know what to expect. The purpose of a UPS is not to get you through hours without electricity. It's to get you through short power failures uninterrupted and keep longer ones from turning into excessively difficult situations.

When there is a really long power failure, any UPS batteries that are getting old are likely to fail. When power comes back, give all the UPSes 48 hours to charge up, then test them by unplugging them from the wall. You'll probably find your batteries that are more than 3 years old aren't doing the job.

In fact, it's a good idea to replace UPS batteries every 3 years. That implies writing the date on the UPS (on a label or piece of tape) when it's new and whenever batteries are replaced.

One last thought. Some of the outlets on the UPS are only surge protected, not power failure protected. Pay attention to the labeling, and if you're like me, put tape over the outlets that aren't UPSed, to prevent unpleasant surprises.



Impenetrable alarm clock — bad user interface award

Today's Bad User Interface Award goes to my Sony ICF-C707 alarm clock.

Only with microprocessors is it possible to create an alarm clock that a highly educated person cannot set or use without referring to instructions. A hint: The first thing I discovered was that the TIME SET buttons (two pairs) do not set the time. It went downhill from there.

Bad software designers around 1980 relished every project as an opportunity to create a new and challenging puzzle for the user. But I thought that kind had all lost their jobs by now. Apparently not quite all.

While installing a UPS, I had to set up this clock from scratch. For years previously, I had done little but turn the alarm on and off and change the alarm time. The clock itself is very accurate.

Well, the procedure was impenetrable until I downloaded the instructions. Here's part of it; could you have guessed any of this?

Picture

A "trap for young players" (as Dave Jones would say) is that many of the buttons have no effect unless held down for 2 seconds. That made it extremely hard to learn anything by just trying buttons.

This is not a self-setting clock. It doesn't pick up radio signals. The only reason it wants to know the time zone is because it had been set at the factory but did not know what part of the world it had arrived in. From there, you get to a procedure whereby, with a long press of a hidden button on the back, TIME SET will indeed set the time. (Starting with the year, so it can calculate Daylight Saving Time dates.)

Now... I like the functionality of this clock... its great accuracy and automatic Daylight Saving Time save me a lot of work. But why can't its controls be understandable to people who haven't read the manual? Sony needs to think about how they are labeled and what is shown on the screen. Why not display brief guidance on the screen itself? Or even a few sentences saying what these buttons actually do, embossed in the bottom of the plastic case?

I know there are those who will say, "Well, you should always read the instructions." No. This isn't an exotic new machine, it's an alarm clock. Just because they can make it hard to use, doesn't mean they should.

2019
October
16

(Extra)

Is this a circuit-design mistake?
High-voltage electrolytic capacitors in series

Picture

Since I still use my 50-year-old Heathkit IT-11 capacitor tester, mainly for re-forming and for long leakage tests, I wanted to replace its most failure-prone parts, the electrolytic filter capacitors, before they fail. They could burst and make a real mess.

Above is the circuit design I found. It does something we're told not to do — it puts two 350-volt electrolytic capacitors in series to handle 600 volts. This is supposed to be risky because, if one capacitor becomes electrically leaky, the other one will handle the whole voltage and be overloaded.

There are no 800-volt electrolytic capacitors. I was all set to replace the pair with a single 20-µF 1000-volt film capacitor (something you can get in 2019 that they didn't have in 1969). The correct value is 20 µF because two capacitors in series have half as much capacitance as each one individually (strange but true).

But there wasn't room for that capacitor in the space available! Although electrically ideal, and inexpensive, it's too bulky.

So I used two 47-µF, 450-volt electrolytics in series instead. Was this a mistake?

Well, I started a conversation about the whole thing on sci.electronics.design, one of the old Usenet unmoderated forums. I got a very courteous reply from Winfield Hill (co-author of The Art of Electronics) and helpful replies from a few other people, but, of course, as is traditional on Usenet, some of them (not Mr. Hill!) fell to quarreling and insulting each other.

The bottom line? Load-balancing resistors are not needed. Capacitors can be put in series with no special arrangements.

There are a couple of reasons. Modern electrolytic capacitors are extensively self-healing; that means that leakage in well-made capacitors tends to stop on its own, as the leaky spot heats up and (microscopically) becomes an insulator. And leakage goes up more than linearly with voltage. That means that if two capacitors in series leak unequally, the leaky one will carry less voltage, causing it to leak less (by a larger factor than just the voltage decrease), and the system will self-stabilize. The nonlinearity is documented in this Tadiran application note although they still advocate use of balancing resistors in the traditional way.

Also, my capacitors are rated for 450 volts each, and handling 600 total, so quite a bit of imbalance could occur without pushing them past their limits.



The October 15 crunch

Have you noticed that important activities are scheduled as close to October 15 as possible, every year, everywhere?

2019
October
16

Has the earth shifted?

At last, some fake news that isn't political. A web page reports that according to Inuit elders (Alaskan natives), "the Earth has 'wobbled' or shifted, and that their sky is changed. The elders claim that the sun does not rise from the same position as before. Moreover, the day is longer and the sun is positioned much higher than before."

I haven't heard this from any reliable source and assume it is a complete hoax. The article's assertion that NASA scientists are "very concerned" more or less proves nobody talked to a NASA scientist.

Apparently the writer of the story doesn't realize astronomers observe the sky all the time and have to be able to find things in it. Even a tiny shift would be noticed by the people who aim telescopes (using computers or, in earlier times, setting circles with angles marked on them). And sailors navigating with sextants would notice a shift of even a small fraction of a degree.

Actually, quite a bit of astronomy is devoted to precise position measurement. One of the most sensitive tests is a solar eclipse — if the earth had shifted, the moon's shadow wouldn't hit the earth in the predicted place. Similar measurements, of the moon passing in front of stars rather than the sun, are made frequently.

Finally, the sun rises and sets on time, as calculated with formulae that were last refined in Victorian times and haven't changed substantially for two thousand years.

If it isn't a hoax after all, I can think of two things that might have happened.

Maybe an Inuit elder is comparing the sky today with a detailed description that is several centuries old. Such things are handed down even in societies that lack writing.

In that case, maybe he has noticed precession. That is a well-understood slow change in the earth's axis that was known in ancient Greek times and has been well understood for centuries. It is why most of the ancient Greek constellation Centaurus, for example, is no longer visible from Greece. But precession does not affect sunrise and sunset.

Or maybe he and his people have simply moved to a different latitude. That is the easiest way to make the sun rise and set at the wrong times and change the length of the day. A traditional description of the sky for one latitude won't fit another, even just a few hundred miles north or south.



Two farewells

I must also bid farewell to James D. Hall, D.D.S., who was my dentist when I first came to Athens (1973). His son, Brian P. Hall, D.M.D., has been my dentist ever since he took over the practice.

And I lament the imminent closing of the Athens Sears store, soon to be followed by the collapse of Sears, Roebuck & Co. as a whole.

Sears tool departments have been important to me ever since my father took me to the one in Valdosta regularly around 1965. They've looked about the same for more than 50 years (with more red in the color scheme in later years). Sears' insight was that even in small towns, people buy specialized tools; that was where I saw micrometers, plumb bobs, and other unusual things. In 1994 I bought a Sears 8-inch table saw, not quite realizing it was the last incarnation of a very early product; it still serves me well and is much quieter than most table saws.

Here's what was left in Athens a few weeks ago. It has dwindled much further. Sadly, many of the prices are only discounted 30%, which doesn't make them compelling; Sears had gotten pricey in recent years. I don't know how much longer the store will stay open, or how much longer there will be Sears stores anywhere. See also my earlier note.

Picture

2019
October
7

Edward Van Peenen II, 1940-2019

Picture

Word has reached me that one of my early mentors has died. When I was in 7th to 9th grades, 1968-71, Edward Van Peenen helped me use laboratories, the observatory, and the planetarium at Valdosta State College, where he taught physics. This was my first extended acquaintance with a professional scientist, and it was instrumental in launching my academic career.

To see his Valdosta obituary, click here. The color picture above was taken in 1976, when, on a visit to Valdosta, I encountered him bicycling through the town in his usual way. I last saw him in 1985, and my understanding is that by minimizing his cost of living, he had been able to retire in his forties and continue living the solitary, orderly life that he preferred.

May his memory be eternal.

2019
October
6

When to discard old but unused electrolytic capacitors

Picture

Semiconductors and resistors last forever when not used, but electrolytic capacitors deteriorate with non-use. How fast? Should I throw away my old ones? It used to be standard practice for repair shops to throw them away after two years on the shelf.

The question that came up while I was cleaning up and modernizing my electronics lab. I asked it on two electronics forums and got very divergent answers. So I did some of my own testing.

What follows is a ramble through a lot of information, some of it uncertain. First, some background facts:

  • Many of us have 50-year-old equipment that still works with its original capacitors. The catch is, it hasn't been sitting unused for 50 years. Even charging a capacitor once every 5 years will greatly extend its life.
  • Electrolytic capacitors aren't used as much as they used to be. Nowadays they're used almost exclusively for power-supply filtering. Those big audio coupling capacitors have been superseded by op-amp circuitry, and large capacitors in RC timing circuits have been superseded by digital counters. Even for power supply filtering, there are now 20- and 30-microfarad film capacitors.
  • In the industry, electrolytics are not sold as fresh if they are more than about 3 years old. Since the new equipment built with them also has to have a shelf life, I figure they're actually good for maybe 6 to 10 more years, at least, and this agrees with the specifications I can find, which are short and vague.
  • Of course, you're not going to find measured 10-year reliability for what's being made now; improvements to the process and materials come along every few years. At best, you'll find accelerated aging tests. 1000 hours at 105 C are comparable to 30 years or more at room temperature, depending your preferred version of Arrhenius' equation and the fact that at 105 C, the capacitor is at its thermal limits and experiencing a stress it would not experience at room temperature no matter how much time went by.

    And if the process and materials are continuing to improve, that's actually another reason to distrust old capacitors — not only have they aged, they weren't as well-built in the first place.
  • Some of the best electrolytics now have no date codes, or at least none in the conventional format. That suggests they may have a longer shelf life than the manufacturers actually want to announce, when they are stored at the factory or distributor under controlled conditions.

    Unfortunately, dirt-cheap electrolytics also often lack date codes so they can be sold forever...
  • The industry became afraid of cheaply made electrolytics when a lot of computers failed around 2003-2006 and the problem was traced to cheap manufacturing based on a mis-copied stolen formula. (Strange but true. I'm not making this up!) Nowadays, premium brands such as Panasonic, Cornell-Dubilier, Nichicon, and Rubycon are highly respected.
  • Good electrolytics no longer cost much, partly because of a better supply chain. Instead of $2 at Radio Shack, think 20 cents at Digi-Key or Mouser. I faced a cost of about $15 to completely restock — which I gladly paid — one bad capacitor would cost me more than that in time and frustration.

With all that in mind, I tested about thirty capacitors of various ages. Here is what I now believe, based on those tests:

  • Deterioration of old capacitors is slow and unpredictable. Basically, the percentage of bad ones goes up gradually. I have a lot of vintage-1997 capacitors (22 years old), and most of them pass all tests (after re-forming) as if they were new. A percentage of them do not, and that percentage increases with age. Individual capacitors deteriorate suddenly, while others, the same age, don't show any aging effects.
  • Leakage (electrical) is the most valuable test for old-but-unused capacitors. The deteriorating ones have much more leakage than the healthy ones. The capacitance and ESR don't change much. ESR and capacitance testing are for capacitors that may have deteriorated in use.

    To test leakage, connect 4 things in series: an adjustable power supply (set to the rated voltage of the capacitor), a microammeter, a 1k resistor, and the capacitor itself. You'll see a burst of high current as the capacitor charges, and then, if the capacitor is fresh, the current will settle down within about a minute. If the capacitor is older, it will drop more slowly and continue to improve for as much as an hour or two. Re-forming the capacitor in this way is important whenever you use an older capacitor. It renews the electrolytic layers. When doing an important repair with a capacitor that has been in stock for a few years, it is appropriate to re-form it for a whole hour.
  • Many capacitors perform much better than specified. For example, Panasonic says that on low-voltage capacitors (100 V or less), maximum leakage is 3 µA or 0.01×C×V, whichever is larger. That means a 100-µF, 20-volt capacitor can leak 20 µA. I seldom saw that much, even in older capacitors; 5 µA was more typical.
  • In my experience, brand is a better predictor of bad capacitors than age. Many capacitors, even major brands, no longer have date codes. But if a capacitor looks old, grungy, or untrustworthy, it is untrustworthy. The most unreliable batch of capacitors that I've encountered was a cheap assortment marketed to hobbyists.

So the bottom line? My current practice is to freely use capacitors up to 10 years old, use them with caution up to 20 or 25 years old, and distrust those that are older. There is certainly no need to toss them when they're just 2 or 3 years old. But note that their appearance has not changed much, and capacitors from the 1980s may look just like new ones if you don't look too closely.

I expect people to write in and tell me I'm wrong. But I expect to be told I'm wrong in both directions (too cautious or not cautious enough), so it will balance out.

2019
October
5

Happy birthday, Melody!

Happy birthday, Melody! And thank you for bringing us some mild weather. 100 F on the 3rd, 99 F on the 4th, and 75 F today. How did you do it?



Impeachment is not a football game

Time out for some political advice. Expanding on what I said on Facebook:

Impeachment of the President is not a football game. You are not supposed to pick a side in advance and cheer for it unconditionally. What is going on right now is fact-finding, and you should be ready to adjust your opinion in response to the facts that come out.

I see people doing several unreasonable things:

(1) Making up a story in which their side wins, and believing the made-up story rather than paying attention to reality.

(2) Simply refusing to believe facts that don't support their side, especially if some guy somewhere — quoted somewhere in the news media — makes the same refusal. To a person in this state of mind, one person with a fringe opinion outvotes a hundred who are well-informed.

(3) Assuming that one side can do no good and the other side can do no wrong, or that the slightest hint of wrongdoing on one side makes everything on that side entirely worthless.

Do not be like them.

If you actually have enough information to know for sure how the impeachment investigation should come out, you shouldn't be posting on Facebook — you should be testifying before Congress.

2019
October
4

Happy July 96th!

We have started giving the date as July 96th rather than October 4th because July's weather never stopped. October 3 was the hottest October day Athens, Georgia, has ever had, with a high of 100 F. Today is supposed to be the last day of it, thank goodness.

The night sky has been clear with scattered clouds, an odd combination, but last night I grabbed a quick shot of Jupiter next to the moon, and the bright satellites of Jupiter in a straight diagonal line. The bright side of the moon is very overexposed; you can see the earthlit side. You can also see a few stars; leaves and branches of a tree; and a lot of streaks and glare from reflections and diffraction in my lens.

Picture

Canon 60Da, ISO 400, Sigma 105/2.8 DG EX lens, 1 second at f/4 on a fixed tripod.


If what you are looking for is not here, please look at previous months .