Re: reset circuit

On Mon, 9 Jun 2025 22:28:00 -0700, John Robertson <jrr@flippers.com>
wrote:

On 2025-06-09 7:36 a.m., john larkin wrote:

On Mon, 9 Jun 2025 07:07:23 -0700, John Robertson <jrr@flippers.com>
wrote:
 

On 2025-06-07 1:46 p.m., john larkin wrote:

On Sat, 7 Jun 2025 12:29:06 -0700, John Robertson <jrr@flippers.com>
wrote:
>

On 2025-06-07 9:41 a.m., john larkin wrote:

We have a box that has to never make any false outputs. Bad things
could happen.
>
Part of the fix is to have a solid powerup reset signal, to handle
power brownouts or such. This looks OK:
>
https://www.dropbox.com/scl/fi/cejyyhcrdph1bewn8a6nx/P800_Reset_1.jpg?rlkey=leky75poeerbojmd3xjat4z54&raw=1
>
The 12 and 5v rails will have a bunch of downstream bypass caps. The
dump resistors will discharge them.
>

>
Did you allow for the voltage drop across the transistor in your Reset
Supervisor selection?
>
Just asking...
>
John :-#)#

>
Sure. In normal operation it's inverted-state saturated, so the MAX
part sees all the +5.
>
If the +12 dips to about +10, the transistor base is +5, and the
emitter is 4.4, and the MAX says reset.
>
That transistor has an insane beta and a pretty hunky inverted beta.
>

>
A saturation voltage drop of maximum 0.25V CE for the BCX70, BE
saturation is max. 0.85V...
>
https://www.infineon.com/dgdl/bcw60_bcx70.pdf
>
John :-#)#

 
Assume a lightly loaded emitter follower with +5 on the collector.
Drive the base from some variable supply Vb. Ramp Vb up from zero.
 
Initially, the emitter will track Vb with about 0.6 volts drop.
 
It gets interesting when Vb exceeds +5. The emitter keeps following
the base until the B-C junction forward biases, around +5.6. Then the
transistor saturates in inverse mode. As you tease Vb, the C-E
saturation voltage can be made exactly zero, or you can make the
emitter go above +5 with a little more base drive. Backwards
saturation.

>
Thanks for the explanation, I've never been that great with transistors
- too much magic smoke! Too easily released...
>
I'm better with tubes, EM, and 8-bit vintage TTL for some reason.

I started with tubes. My first job interview, I told the guy that I
preferred tubes because transistors were too easy to blow up. He
didn't hire me. The second guy laughed and hired me.

I got tubes free from old TVs, or dirt cheap from Fair Radio Sales,
surplus. Silicon Valley use to be full of cool surplus stores, but the
real estate got too valuable, and ebay took over.

The beauty of Spice is that you don't really have to understand
things, you only have to make them work. I plan to Spice that powerup
reset thing later this morning for the guy that needs it on his board.
I hope LT Spice models this upside-down saturation thing right.

https://www.dropbox.com/scl/fi/wl51ennrq60edw9ina3za/P800-1Hz-Selft-Trigger.mp4?rlkey=ravaj03c330jjfmplylxr2k7j&dl=0

That can glitch on just-the-right brownout, about 0.7 seconds. The
music wasn't my idea.

>

 
I made a bunch of 16-bit DACS that way once, all discrete parts.
 
The Fairchild BCX70 has beta about 500 and inverse beta about 16, both
measured at 2 mA. I like that part, but AoE rates it dead worst for
Rbb at 760 ohms. Hurt my feelings.

>
So complain to Winfield!
>
John :-#)#