User-Agent: Mozilla/5.0 (Windows; U; Win95; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0
X-Accept-Language: en-us, en
Subject: Re: zero-power toggle circuit; was, how to master electronics
References: <3DC517EB.firstname.lastname@example.org> <3DC57B4E.email@example.com> <firstname.lastname@example.org> <email@example.com> <3DD39D90.firstname.lastname@example.org> <3DD4E13B.email@example.com> <firstname.lastname@example.org> <3DD63323.email@example.com>
Date: Sun, 17 Nov 2002 10:50:15 GMT
NNTP-Posting-Date: Sun, 17 Nov 2002 05:50:15 EST
Organization: Road Runner
Winfield Hill wrote:
> Wafer wrote...
>>Winfield Hill wrote:
>>>>>>... The diode is only needed if the load includes a capacitor...
>>>>>A more serious omission is a lack of any load-fault protection.
>>>>>The 10k resistor becomes useful when adding a current-limit
>>>>>feature, as I've done below. ...
>>>>> [ snip Win's circuit ]
>>> I'm not very happy with the above scheme. For one thing, combining
>>> short-protection with the flip-flop memory leads to awkward issues
>>> like we saw when charging capacitive loads, etc. Further, momentary
>>> current-limit conditions could reset the flop. And then there's the
>>> issue of the series diode, wasting power. The circuit badly needs
>>> further improvement, and I can picture it all in my mind.
>>> But right now I'm going to mow the lawn and rake leaves.
>>I would like to see what you have in mind to take care of load
>>capacitance. After all, your circuit doesn't address this problem
>>at all. A 0.2V drop seems a fair price to pay...
> More like a 0.5 to 1V drop at the higher currents a general purpose
> design might need. Hmm, your remarks sound faintly like a challenge.
> Here's a circuit that's tolerant of load capacitance without a series
> diode and includes robust load-fault protection.
> : V+ <--+-----(O) V+
> : V+ V+ | 6 to 18v
> : | Q4 | R6 0.1 5W
> : R4 V+ v\| |
> : 10k | 2n |------+ 5A limit
> : | Q2 | 4403 /| |
> : | BSS84 | | |
> : +---||--+ +----||--+ Q3 pmos
> : | ||>-' | ||>-' IRF9540N
> : ,--------------------+ |--, R5 |--, 0.1 ohm
> : | R2 | | 22k |
> : R1 ,- 220k -- | ------+ | |
> : 1M5 S1 | +------ | ----' +-----(O)
> : | _|_ | 2n7000 | | _|_ \
> : +---o o---+ Q1 |--' R7 /_\ |
> : | | ||<-, 1k | LOAD
> : | 0.1uF +------||--+ | green LED | |
> : === C1 | | '----|>|-------+ /
> : | R3 gnd ON indicator +-----(O) -- return
> : gnd 100k |
> : | gnd <-'
> : gnd
> Here I've simply isolated the toggle flip-flop from the power switch.
> No ugly time constants, etc. The MOSFET needs a heatsink to handle
> serious shorts, but if two resistors are added for a foldback current
> limit, the heat sink could be reduced or eliminated. If a fault LED
> is desired, R5 could be reduced (R4 too) and an LED placed in series.
>>I'm not happy with the idea of using a series resistor for current
>>monitoring though, since THAT can waste MORE power.
> OK, picky picky, perhaps I can deal with that issue - in a later post.
>>The idea of using a MOSFET capable of 15A seems silly to me. If
>>you can source that kind of power, why would you need zero power
>>in the off state? Automotive batteries go dead long before a CMOS
>>chip could drain them much at all.
> 15A maybe, but a low-resistance FET has two benefits: more delivery
> to the load and less self heating so a heat sink can be eliminated.
> The benefits occur at practical currents like a few amps, which can
> be easily drawn from a set of 5-year shelf-life Alkaline batteries.
> Also, CMOS chips draw no current; one can use them if they prefer.
> - Win
This is the basic idea I had to handle the capacitance issue as well,
but the extra parts bug me, as I'm sure you had hoped to avoid. I would
like to make one suggestion though, with regard to the ON indicator. I
would use a super bright or high efficiency LED, so you can then have a
much higher series resistor, thereby reducing the current required.
As for the use of a larger FET, I see your point. When the currents will
go higher, sure it's nice, but for low power circuits like a little
pocket radio or something, it would be a waste of space and cost. I
guess where we disagree has to do with whether the circuit is designed
for general purpose use, or will vary depending on the requirements.
CMOS chips may not draw much, but they do draw something, otherwise they
would be 100% efficient. We both know that can't be.