My soft-start boards are being fabbed, but I was checking to make sure everything looks good. I did test the TLP191B PV coupler and found it was a lot hotter than the datasheet typicals. At 10 mA on the LED, the output is 10.5 volts open-circuit and 62 uA shorted. Probably LED chips keep getting better.
I was also interested in how hot my DPAK fet might get on startup.
It weighs about 0.3 grams, and I'll assume that's all copper. That gives a temp rise around 8 K per joule, absent any heat sinking. I guess I could test that.
formatting link
The area under the red curve is about 0.3J, so it won't get very warm.
I'll get a little transient heat sinking from the PCB... very little, I'm thinking.
formatting link
A d2pak would have more mass.
That was a brutal simulation. It took forever to run and I kept getting kilovolt and mega-amp spikes when the ramp started. Made no sense. The obvious silly things are there to tame it.
On a sunny day (Tue, 20 Dec 2022 08:52:32 -0800) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:
Honestky, you do not need simulations for stuff a simple as that?
The circuit is not trivial, and we want the first article, rev A, to work and be sellable. That requires some amount of compulsiveness. I work with one big organization that defines *six* pcb layout iterations up to production release, and they use at least six. It takes them years to do a board.
Simulation encourages thinking and reduces risk. It's fast, so allows lots of experiments. We archive the Spice models in our design notes folders, to explain to future people what we did and why.
I think it's important to leave good design notes, fully backed-up. We have standards for that.
On a sunny day (Wed, 21 Dec 2022 07:27:39 -0800) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:
Yes, maybe it gets to the point where we need to present a working simulation before customer buys?
formatting link
Was this my last simulation or did I use it just to draw the circuit??? Note current source ramp generation circuit.
As you may have heard the Andreas fault or there about was triggered as I predicted in my reply to you in the thread 'thinking at Stanford' on Dec 17:
formatting link
How can I predict, well the answer is here: Study shows how machine learning could predict rare disastrous events, like earthquakes or pandemics:
formatting link
its simple I did see a mechanism. I am but a neural net...
Main point is always that things really work.... The 'proof of the pudding is in the eating' is that not the English saying?
. . . and it's a hard boot to the ego to realize that you've missed the one complication that killed the circuit, if only once in a blue moon, either out of the box, or six months down the road when the environment changes seasonally, or a new techy gets hired with different ideas about 'who's on first'.
Suppliers expect defects and failures in ppm scaling.
On a sunny day (Wed, 21 Dec 2022 15:12:05 -0500) it happened legg snipped-for-privacy@nospam.magma.ca> wrote in snipped-for-privacy@4ax.com:
It all depends,,,, the way I design is mostly using circuits that I accumulated over the years that have proven themselves. Of course you make mistakes. 'Simulation' I use only for filters, but then there are many free filter design programs online, its faster and more accurate than me doing the math. But not even always... Nothing like a proto board and a scope.. As a kid (tube ages) I build tube stuff, did not even had a meter. Transistors made things even easier, things and circuits from magazines.. Everything worked, then later I build an oscilloscope, then I took an old tube TV set (round CRT at that time, school days) and replaced everything with transistors piece by piece including the HV. Gives you RF and high power and magnetic deflection etc experience. There were not even computers and simulations in the fifties and sixties. Everything worked. Now you buy a chip that does everything you need duh What a world. until the chips are no longer available of course:-)
Software bloat these days, I started coding in binary, still love to code in asm (PICs these days) C, what happens now in the Microsoft bloated crap is a disgrace.
As a very young kid I got hold of Van Aisberg's book: "Zo werkt de radio" (Dutch for "That is how the radio works)" from the local library
formatting link
mother needed special permission as I was actually too young to get it myself and the way that was written was clear as glass and I just understood it Van Aisberg also wrote 'Zo werkt de televisie" (That is how television works)
formatting link
I sort of absorbed it in a flash.. So.. later I ended up working at the TV studios here.... Math ??? WHAT MATH?
We are just a neural net What also helped me was the RSGB radio handbook (build transmitters too)....
All part of me it seems Others have made other fields their own. But it was fun! All that with almost no pocket money under sometimes challenging circumstances My interest these days goes way beyond electronics, and internet opened up a wide source of things for learning. Worked in many fields that used electronics, medical, chem, art, space, what not. Maybe one day I will write the book: The Fart Of Electronics All is relative ain't it?
Most of the faults That I've found over the years COULD be understood, or simulated, but only after the fault mechanism was retrofited into the circuit's concept, or later, the model.
Cicuits that protect have their own issues, as they may often be required to perform when everything else around them is doing a geeky 'which way do I go and what was I thinking' dance that is an attractor to chaotic behavior - unpredictable.
FPGAs that deprogram on ground bounce or output voltage/supply line misallignment are often misapplied in power control circuits, where such abnormalities can be considered as only single-fault abnormal. Such power control circuits are, traditionally, expected to recover without requirement for part replacement.
I simulate voltage dividers, RC timing circuits, diffamps, filters, quantizers, control loops, LC circuits, vregs, all sorts of stuff. Spice is usually easier than using a calculator, and leaves a commented record. A Spice file is easier to save than photographing a whiteboard.
Of course every sim is managed as an engineering document. It has a visible title, date, version, and author, and comments as are useful.
We do that. We assign a project number (The Znnn series, currently at Z544) to any serious prototype or test, and document the hardware and the test results on a server. That turns out to be very handy over the years.
I had an Eico VTVM whan I was about 7. And a Heathkit OM-3 scope when I was 12. Lissajous patterns were fun at parties.
I had one job interview when I was a college freshman. It told the guy that I preferred tubes because transistors were expensive and easy to blow up. He said "that won't do" and sent me away. The next guy laughed and hired me; I designed $200e6 worth of stuff for him.
Yeah, I did a lot in assembly. Kids these days love layers of abstraction, so they can go layers deep to set a port bit high.
We had Popular Electronics magazine.
We were poor, but my mom set up a revolving credit account at Allied Electronics for me, $30 a month as I recall. I could order anything. At that time, a CK722 transistor was about $7.
I'd also go to the local Radio Parts store, where the guys behind the counter, and the owner, would usually just give me stuff. I think a
1/4 watt resistor cost 15 cents, more than a Crystal hamburger.
I recently had a $20 hamburger (it was worth it) and resistors now cost a fraction of a cent.
On a sunny day (Thu, 22 Dec 2022 10:37:14 -0500) it happened legg snipped-for-privacy@nospam.magma.ca> wrote in snipped-for-privacy@4ax.com:
...
Yes, have not done a lot with FPGAs, needed some for work, and at home used one to brute force some TV encryption system, do some system conversions etc some stuff:
formatting link
note the card holder ;-) I stopped hacking TV after some female politician in the EU started screaming 'hackers' standing on a chair like the women in that comic scared of a mouse calling Tom the cat (the police). Not worth the risk, others continued and got busted. Good old days.. alt.satellite.tv.europe was full of hacking. Most systems are vulnerable... The experience however may come in handy in a war situation like we are heading for. Still read the crypto group but it is not what it used to be, and >90% spam. Has become more of a crossword puzzle place ;-)
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.