Stupid polarized capacitor tricks

[Rich Grise]

That's why it's called "stupid capacitor tricks" -- not because it's
practical, but because it's kind of interesting and fun.

I had a professor who, while a grad. student at UT-Austin TA'd a lab
course where they did all sorts of crazy projects -- he mentioned one
where he had a student modify a VCR to record the entire AM broadcast band
at once. Not very practical (since realistically a VCR doesn't have that
great of dynamic range, so you're really only recording the stronger
stations -- the weak ones are lost for good), but certainly fun.

Yeah, it's theoretically doable, but kind of pointless. How about
a J-K flip-flop using nothing but relays? ;-)

Cheers!
Rich

 

[Eeyore]

As a "design engineer", (guffaw!) you should be able to substantiate
your 'megasecond' claims and, instead of imagining, back up your
claims with some numbers.

Since you're so interested why don't YOU see if you can remove the oxide layer on an
electrolytic cap's anode and lets us know how long it takes.

A megasecond is only a bit more than a week you know. Not that long.

Graham

 

[Eeyore]

The paper also prevents damage to the oxide film.

Graham

 

[Eeyore]

Yeah, it's theoretically doable, but kind of pointless. How about
a J-K flip-flop using nothing but relays? ;-)

Some time in the 60's ? some students at my old school build a simple relay
based 'computer' using 'war surplus' parts most likely.

Graham

 

[Mark Zenier]

I was reading a circuits book today that spent more time on the construction
and modeling of components than most do. When I came to the section on
regular old polarized electrolytic capacitors it occurred to me that you just
might be able to turn an electrolytic capacitor into one bit of non-volatile
memory by purposely applying the correct or reverse polarity to form or remove
the dielectric ("write"), and then test for this (by checking to see if the
cap behaves more like a cap or a short ) later ("read").

Does anyone know if this is feasible? In my mind it would make electrolytic
caps sort the "dual" to core memory, which of course leave ferromagnetic
materials near one end or the other of their B-H curves for later recall.

There were "Elapsed time" sensors that were a small electroplating
cell, where after some rated number of milliamp hours, the active
metal got plated off one electrode and the voltage went up.

There are also electrochromic(sp?) displays that work by electroplating.
Maybe you could combine the two and have a non-mechanical version of
a dipswitch. Configuration memory you can read...

Mark Zenier mzenier@eskimo.com
Googleproofaddress(account:mzenier provider:eskimo domain:com)

 

[John Fields]

Since you're so interested why don't YOU see if you can remove the oxide layer on an
electrolytic cap's anode and lets us know how long it takes.

---
Since you're the one who made the claim, it's up to _you_ to prove
it.
---

A megasecond is only a bit more than a week you know. Not that long.

---
Hmm...

There are 3600 seconds in an hour, so the number of hours in 1
million seconds is:


1000000 seconds
T1 = ----------------- ~ 277.8 hours
3600 seconds


Then, there are 24 hours in a day, so:


277.8 hours
T2 = ------------- ~ 11.6 days
24 hours


Then, since there are seven days in a week:


11.6 days
T3 = ------------ ~ 1.7 weeks
7 days


Which is, in my book, hardly a "bit" more than a week...

 

[Terry Given]

Terry Given wrote:




Design engineer not a technician thanks.

odd then that you shy away from any and all detailed technical
discussion. Hell, you admitted you couldnt even follow (ISTRIW) John
Fields simple analysis of a rectifier-capacitor filter.

If you can tolerate read/write times in the megasecond range I suppose you could
say it might work.

That doesn't strike me as being practical.

perhaps thats why the title didnt read "practical capacitor memories"

I can imagine the electrolyte would mess things up too.

Graham

Cheers
Terry

 

[John Fields]

The paper also prevents damage to the oxide film.

 

[Eeyore]

No it's not.

 

[Eeyore]

---
Since you're the one who made the claim, it's up to _you_ to prove
it.
---


---
Hmm...

There are 3600 seconds in an hour, so the number of hours in 1
million seconds is:

1000000 seconds
T1 = ----------------- ~ 277.8 hours
3600 seconds


Then, there are 24 hours in a day, so:

277.8 hours
T2 = ------------- ~ 11.6 days
24 hours

Then, since there are seven days in a week:

11.6 days
T3 = ------------ ~ 1.7 weeks
7 days

Which is, in my book, hardly a "bit" more than a week...

It's certainly less than 2 weeks.

Graham

 

[Eeyore]

odd then that you shy away from any and all detailed technical
discussion. Hell, you admitted you couldnt even follow (ISTRIW) John
Fields simple analysis of a rectifier-capacitor filter.

Well .... considering his 'analysis' gave the wrong answer !

I understand it perfectly actually thanks. Fields was as ever just trolling for an
argument.

Funnilly enough, I was thinking about that example earlier today. I'm sure he's got
his sums wrong somewhere. I may return to this with an mathematical analysis of my
own. That's he's wrong is readily shown with any simulation btw.

Graham

 

[Eeyore]

---
Only in a secondary kind of way since it keeps the foils apart when
the cap's being wound.
Precisely.


It'll hardly survive a hammer blow or keep
the oxide from coming off the anode if its polarity is reversed,both
of which will cause a short.

Polarity reversal won't cause a SHORT. It'll be some lowish value resistance
determined by the properties of the electrolyte.

Graham

 

[Bob Myers]

It's certainly less than 2 weeks.

I spent several megamillkilonanomicrogigaseconds trying to
figure that out, but then realized that my time was far better
spent going for a beer....

Bob M.

 

[John Fields]

No it's not.

 

[John Fields]

It's certainly less than 2 weeks.

 

[John Fields]

Well .... considering his 'analysis' gave the wrong answer !

I understand it perfectly actually thanks. Fields was as ever just trolling for an
argument.

Funnilly enough, I was thinking about that example earlier today. I'm sure he's got
his sums wrong somewhere. I may return to this with an mathematical analysis of my
own. That's he's wrong is readily shown with any simulation btw.

---
If you understood it perfectly and it's at odds with reality, which
you also seem to think you understand perfectly, then it should be a
trivial matter for you to spot and report where I went wrong.

I don't see that happening any time soon, even though I think there
was a small error in my work which I'm still examining.

I'd tell you what it was, but that would be cheating.

 

[Eeyore]

I spent several megamillkilonanomicrogigaseconds trying to
figure that out, but then realized that my time was far better
spent going for a beer....

I'm sure your priorities are right !

Graham

 

[Terry Given]

Terry Given wrote:




Well .... considering his 'analysis' gave the wrong answer !

I understand it perfectly actually thanks. Fields was as ever just trolling for an
argument.

Funnilly enough, I was thinking about that example earlier today. I'm sure he's got
his sums wrong somewhere. I may return to this with an mathematical analysis of my
own. That's he's wrong is readily shown with any simulation btw.

Graham

I didnt study his analysis in any detail - because I really dont care; I
seldom if ever use mains transformers, and never for any power level
above a few watts, making anything but the crudest analyses unnecessary

when operating a cap/bridge from mains, resistance doesnt really enter
into it, but inductance sure does.

Barton has a seriously detailed exposition of the humble
rectifier-capacitor filter (one chapter worth), if you do it properly
its surprisingly complex.

I just find it frustrating that you often answer posts a-la this thread,
and seldom (if ever) provide any detail whatsoever - I dont recall
seeing any detailed analysis from you, on any topic. IMNSHO those who
can, do; those who cant, generalise.

Cheers
Terry

 

[Eeyore]

If it means so much to you , you can do it yourself. I personally don't feel the need to
perform an experiment, the outcome of which is predictable and obvious.

Of course you *know* I'm right. If you thought I wasn't you would perform the experiment to
prove me wrong. Since you're not doing that, it's clear you're just mouthing off for the
sake of it.

Graham

 

[Eeyore]

Let me see.

11.6 days IS a bit more than one week. It's also less than 2 weeks. Problem ?

Graham