Reset uP during Power Up

I have a printer that requires 5VDC and 28VDC power. The original power supply is not available, and I was unable to find an affordable power supply with the required voltage and current ratings. So I'm using two power supplies, one for 5V and one for 28V. The problem is that upon power up, the 5V comes up immediately, but the 28V supply takes between one and two seconds to come up. I need a simple (i.e. low component count) circuit that will hold RESET low on the uP until both 5V and 28V are up. Whatever I use must be through hole. Any suggestions would be appreciated.

You could use an N-FET in the 5V line, with its gate tied to the 28V supply (resistor divider and a cap if you want to set the levels and slow things down some) such that the 5V supply doesn't come up until the 28V supply is stable.

This is a circuit we use for that situation...

ftp://jjlarkin.lmi.net/POR.jpg

The MAX809 (use the OnSemi version!) does a nice clean powerup reset function, and the transistor makes it sensitive to both supplies. The voltage divider can be futzed with to set the +28 threshold.

John

1) RC delay from +5 to drive base of grounded collector PNP; emitter to eithher +5 (recommended) or +28. 2) Diode cathode to +28, anode to resistor to +5V; anode is "logic" tap (sloppy, but gives another option).

You probably have to take the printer apart to access the reset line. Just put a much bigger cap on the reset line. You probably don't care if the printer takes a few extra seconds to power up. Don't get all "logical" when lengthening an existing timer will do.

--- _____ This'll give you RESET low for about 3 seconds after power-on.

View using a fixed-pitch font.

+28V>------------------------>+28V

+5V>----+-----+----+-----+--->+5V | | | | [4M7] [10k] | [10k] | | | | +-----|---|+\ | _____ | | | >---+--->RESET | +---|-/ LMC7225 | | | [1µF] [10k] | | | | GND>----+-----+----+

|| _____ ______________ RESET |||_________|

The pudding ;)

Version 4 SHEET 1 880 680 WIRE 128 -16 16 -16 WIRE 240 -16 128 -16 WIRE 336 -16 240 -16 WIRE 448 -16 336 -16 WIRE 128 32 128 -16 WIRE 240 32 240 -16 WIRE 448 32 448 -16 WIRE 336 144 336 -16 WIRE 240 160 240 112 WIRE 304 160 240 160 WIRE 448 176 448 112 WIRE 448 176 368 176 WIRE 496 176 448 176 WIRE 128 192 128 112 WIRE 304 192 128 192 WIRE 16 256 16 -16 WIRE 240 256 240 160 WIRE 128 272 128 192 WIRE 16 384 16 336 WIRE 128 384 128 336 WIRE 128 384 16 384 WIRE 240 384 240 336 WIRE 240 384 128 384 WIRE 336 384 336 208 WIRE 336 384 240 384 WIRE 16 432 16 384 FLAG 16 432 0 FLAG 496 176 RESET- SYMBOL Comparators\\LT1017 336 176 R0 SYMATTR InstName U1 SYMBOL voltage 16 240 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 5 SYMBOL res 224 16 R0 SYMATTR InstName R1 SYMATTR Value 10k SYMBOL res 224 240 R0 SYMATTR InstName R2 SYMATTR Value 10k SYMBOL res 112 16 R0 SYMATTR InstName R3 SYMATTR Value 4.7e6 SYMBOL cap 112 272 R0 SYMATTR InstName C1 SYMATTR Value 1e-6 SYMBOL res 432 16 R0 SYMATTR InstName R4 SYMATTR Value 10k TEXT 24 408 Left 0 !.tran 10 startup

--
JF

I was waiting for your usual Doh! moment, but it seems to have escaped you this time. Here's a clue: In the real world, slow rise times on comparator inputs make messy outputs.

-- Joe

Time delay relay.? Do they make those in through-hole? Also, might not be the cheapest solution. If you're going to roll your own perfboard or PCB, I do like the other solutions posted better.

-mpm

--
No shit, Sherlock?

I thought about it, but since a reset line is nothing like a clock
input I decided not to run any hysteresis around the comparator since
a little chatter from the comparator output into the µC reset won't
make any difference.

More to the point though, instead of opening a discussion in a
friendly way, why do you jump in all attitude?

[snip]

My POR's usually have a "diode" (*) across 4M7 to ensure dumping the cap during fluctuating or brown-out power conditions.

(*) Takes more than a diode to cover all conditions. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed

Little RC things are OK for unimportant applications, but a serious por circuit should have

hard low during rampup

clean release edge (some gadgets may hang otherwise)

guaranteed minimum reset time

handle browounts properly

check all supplies (depending)

Given how many parts it takes to do this properly yourself, it's usually easier to start with an official por chip.

Lots of regulators these days include a PGOOD pin, which helps in checking multiple supplies. It's not unusual for a board to have six or even ten voltage rails these days. One I'm doing now has

+24 +12 +6 +5A +5D +3.3 +2.5 +1.2 +0.9 (at 4 amps!)

-5

-12

and a few reference rails. The critical PGOOD pins hold off RESET, and the rest are checked by the uP after powerup.

Hey, we need a por chip that has, say, 10 inputs, each with a threshold programmed in nonvolatile memory, and a few outputs, like the hard RESET- and a couple of warning flags. It may as well do high/low limits. Heck, it may as well digitize all the supply rails and report over SPI. Drive some LEDs, too.

There may be something sort of like that already.

John

just get a 24V DC relay and drive the coil with the 28Vs. pass the

5 Volts through a contact.. Problem solved. If you want to ensure the coil does not get over loaded, which I don't think will be a problem, you can put a R in series enough to drop the voltage.. Just do your basic ohms law..

Jamie

Don't think so but there are tons of multichannel PORs

icleid=3D540340 . Digikey carries some 4-channel jobs from EXAR too...

Sorry, but I've often gotten a kick out of watching you correct and then recorrect your own circuit suggestions, sometimes going from bad to worse. Don't take it too hard, we can't all be Jim Thompson. I'm sure I'm just as sloppy as you.

-- Joe

Nice little comparator and since CMVR typically exceeds V+ by a few hundred millivolts, something like this should work: Please view in a fixed-width font such as Courier.

. . . . . +28V>---+----------------------------->+28V . | . | . | . | . +5V>----|-------------------+-----+--->+5V . | LMC7225 | | . [3.9M] .----+ [100K] . | | | | . +--------------|---|+\ | _____ . | | | >---+--->RESET . [1M] '---|-/ | . | | | . '-------------------------' . | . ------------------------+------------- . .

Because you invite that behaviour? You don't think a complex device might not suffer partial reset issues if the dirty was dirty like you make it?

Follow the rules, comparators have a touch of positive feedback for slow inputs. This rule is far more important than whether or not to allow crossing disconnected lines on a schematic :)

One is a convention, the other is required for proper device operation.

Grant.

Is the printer barfing on that arrangement now? Or are you trying to be nice 'just in case?'.

Do you know what's on the reset line now? Just an RC, or does it have a voltage monitor/supervisor chip? It may be as easy as one resistor, or take a couple transistors. I'm a bit late to this one.

Curious. Why didn't you add a 28V -> 5V converter, rather than buy two separate PS?

Grant.

--
Geez, to me that sounds a lot like: "Listen to me telling you how
important I am", but that doesn't go a long way toward helping the OP
solve his problem.

As far as I know, what I posted was adequate, and probably
minimalistic, for the OP's requirement.

Do you have a quarrel with that?

--
Indeed, but, in the end, I believe all the errors were worked out.