Reset supervisor - revisited

[eyewonder300]

I am wanting to use an MCP111 reset supervisor chip with a 'F877A, with the board set up for ICSP programming. The ICSP will be from a Microchip ICD-2.

I have read the only other thread I could find on 'reset supervisor', but it does not seem clear enough for me to use. I have also printed out & read (several times) AN820, TB087, and TB013. Same confusion.

I am looking at MCP111/112 data sheet, page 12, figure 4-4. This is the same as TB087, page 4, figure 3. This figure shows the MCP111 output going through RS (for current limiting) & then into the MCLR pin of the PIC. Also connected to the MCLR pin is Rpu with its other end labeled 'Vdd/Vpp'. Is the top of this Rpu where the programming Vpp from the ICD-2 comes in? If it is, then when the ICD-2 is disconnected, there is no voltage on that resistor to pull-up the MCLR pin to its Vdd level to allow normal operation (the MCP111 output is open-drain).

It looks to me like the above will not work.

Looking at AN820 page 2 figure 2, it shows an MCP120 as the supervisor, but with a quite different hookup. We still see a current limit resistor (Rcl) from the supervisor output to the PIC MCLR pin. What's now different is that there is a pull-up resistor (Rpu) directly from the supervisor output to Vdd, which after normal reset (non-ICSP operation) pulls-up the open-drain RST/ pin, which sets MCLR pin high enough to siganl normal operation. There is also a Rpd pull-down resistor from the MCLR pin to ground. And, the actual Vpp signal from the ICD-2 appears to be connected directly to the MCLR pin. This setup makes sense to me.

Then going back to TB087 page 6 figure 5, its the same wire-up as AN820, expect a Schottky diode has replaced the Rcl current limiting resistor.

Help in understanding all this is appreciated.

Steve

[Mark]

Why are you wanting to use MCP111? Are you having problems with a design?

[eyewonder300]

Mostly to be able to use the MCP111 for reliable operation. This is a one-off personal project, not a production item.

Steve

[Ttelmah]

I resume you are referring to the awful diagrams in the Technical brief TB087?. This has to rate as one of the documents with the highest error rate around!. Rpu, is shown as going to Vdd/Vpp, with the comment then made, that some PIC's don't need the resistor, since theyhave one internally (which with the diagram as shown, would leave no programming connection...).
The correct connection, is Rpu to Vdd, and Vpp to Mclr. Figure 5, gives the 'best' configuration really, but it is a more expensive solution.

Best Wishes

[libor]

..a related question

In what applications is it advised to use a separate reset chip ?
PICs have power-on-timers, brown-out-resets, low-voltage-detection, etc. Also a separate analog pin can be used to measure incoming voltage if needed. A RC circuit can also be added to the reset pin to solve the slow powering up problem.
What can a reset chip do what a PIC alone can not ?

[Ttelmah]

Many of the older chips, don't have brownout detection. Also on some of the latter chips, it is unuseable, when operating off low supply rails (on a couple of models the brownout threshold is fixed, and stays the same on the 3.3v versions, giving almost no margin). Generally on the more modern PICs, an external reset circuit is not needed for monitoring of the local supply, but may still be needed if the rise time of the supply rail is very slow, or to monitor another rail.

Best Wishes

[eyewonder300]

Ttelmah,

Thanks for the help. Figure 5 (with the Schottky) was the only one that made sense to me, so I will go with that.

Thanks,

Steve

[eyewonder300]

Tried out the figure 5 arrangement, using 10k & 100k for Rpu & Rpd, Schottky diode, and all seems to be working on my project.

Thanks,

Steve