What does an embedded beginner need to know?

Just don't build medical or aircraft stuff that way or that'll haunt you some day :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

As others have pointed out, the common answer is: "The specifications were incomplete at the beginning."

Mark Borgerson

I just went back to some old notes. Running on a 3V supply, the low side output driver goes to 0.5V at 10mA and 1.5V at 25mA measured, so

50 ohms going towards 60 ohms over that 10mA to 25mA compliance. On the high side it was 0.5V at 8mA, so about 62.5 ohms. I don't have notes for trying to pull more than that from the high side.

But this is "relatively equal," I'd guess. Perhaps because this was by design and the LxW values are intentionally different?

Jon

... snip ...

Now consider what is going on inside that chip. At 10 ma, the drive component is passing 10 ma through 0.5 V, for 5 milliwatts internal dissipation. At 25 ma, it is passing 25 ma through 1.5 V, for 37.5 milliwatts internal dissipation. Which is most likely to destroy the output transistor?

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

Careful, as Chuck hinted, that's pushing your luck. Tzzzt ... phut ...

*POOF*

A newer device I checked ended up near 35ohms both H and L. It is best if digital stuff is designed that way because it eases fast multi-phase PWM controls, dead time control and so on. The H to L difference is mostly a remnant from the TTL days. That stuff really could not source worth a dime.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Well, I wasn't suggesting this as practice. At the time, I was just playing with a new part to monitor about where it hit mid-point and where it hit 0.5V. Reporting my notes isn't a suggestion and I didn't intend it as one.

I was mainly supporting Joerg's comment, by looking back at early notes I took. In particular, note that at 10mA/0.5V for the low side drive mosfet and 8mA/0.5V for the high side drive mosfet.

It might have helped if I hadn't reported the 1.5V figure for the low side. But it was there in the notes, so I wrote it out. Oh, well. No good dead... and all that.

Jon

Because of what I already know about P and N channel mosfets, when I see balanced outputs like this I tend to imagine they are designed to be that way. And PWM DAC is what comes early to mind.

Don't I know it. I will never forget Ioh=400uA.

Jon

[...]

I'll never forget the time when I pushed one real hard driving a diode sampler, touched it and ... tssssst ... had seven blisters on my index finger, 1/100" spacing.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

That's how to avoid ever losing your pin spacing ruler, I'll grant you!

You reminded me how I didn't like the LM1488. That thing was hot just sitting there. Hot enough in normal operation that I couldn't keep my finger on it for long. The LM1489 was cooler, but still hot. I was very glad when the MAX232 became available.

Jon

Not necessarily. "Once burnt, twice shy" is ancient wisdom that fully applies here. Countless generations of parents have consistently found that there's only one way to teach caution: first-hand experience of actual pain caused by lack of it.

Egos, like toes, have to be stubbed a couple times to convey the value of caution. Once the necessary number of painful experiences have taken place, the lesson can be completed and renewed by scary movies or war stories ---- but the initial stage will invariably be accompanied by an "ouch!".

So yes, beginners will be a little full of themselves. Sooner or later that will lead to a bloody nose. Since you know it will happen regardless what you try, and it's actually in everybody's interest that it does, the best you can do is keep some control over when and how it does, so nobody gets hurt too bad. Book the costs under the heading "educational expenses".

There's also CMOS versions. I have never embraced the MAX232 although it now has 2nd sources. Single sourced Max parts won't be used on my designs because clients had too many issues with delivery problems.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

At the time, it was just the MAX232 (not the A version) and not much else as an option. I suspect it was one of the few early parts giving MAXIM a boost. Today, I agree. But back then I didn't have a lot of options, other than hot, power-hungry LM1488's and LM1489's. So anything was better, single source or not.

Jon

Why didn't you use the 14C88 and 14C89? They are around since the 80's.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

My first significant hobby project that I designed myself was when the MAX232 didn't exist and the only options for me was the 1488/1489. This would have been 1981. My next significant one requiring RS-232 was when the MAX232 had recently come out from Maxim and it made more sense to use one part than two and fewer rails, so I don't recall even knowing about the C versions at the time (or looking to find out, having had already seen such benefits to getting away from the

1488/1489 pair.) Do you know when they first came out as a commonly available part?

Jon

I don't remember but I just looked and they are in the 1986 National databook.

Also, I am not a great fan of charge pumps since they can produce quite an EMI racket. I love smooth inductive power transfers versus banging capacitors around.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Yes. I distinctly remember one incident, back when we were still losing the war and the US hadn't gotten involved yet. I had a 100 watt iron on something on top of my dresser, and I knocked it off. I caught it, protecting the floor etc. from being burnt, but I rapidly learned the smell of scorched flesh. After restoring it to the holder I could yell and get parents involved. I haven't caught another iron since. My palm eventually healed.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

It depends, most price-sensitive devices use what I'd all area- matching on their P-N drivers, so a typical uC has 10mA spec, for 220mV Vol and 500mV Voh drops. To get better matching , you need to skew the area, which costs money

Even LVC logic, which has a nominal 24mA symmetric spec, gives 550mV NF and 800mV PF at 24mA,

Then, you can find the Die impedance, and bonding impedances, favour the N Side even more. (Which digital designer things about the Vcc impedance ?)

I recall checking an early AVR, and the 'rubbish' (!) that came out of the ports driven HI, was indicative of quite high Power impedances inside the chip.

-jg

Well, we are now getting deeper into the application, I suppose....

They can. In the applications I was designing for my hobbyist needs, there was no difficult analog requirements and I would have eagerly used the MAX232 instead of those two HOT LM1488/LM1489 pairs without any impact on the rest. Horses for courses. (I think you wrote that, no so long ago.)

Yes, me too. But as a hobbyist without something serious going on with the analog side, the extra power supplies would have been a pain and inductive switchers for two new rails (+12-15V and -12-15V) not something I wanted to grab after, either. These were one-offs, you know.

Jon