Brown out protection for S08SH8

[RobG]

I read somewhere that there is a possibility of an HC08 MCU failing to restart after brown out unless Vdd dropped below a particular level before recovering. I'm having trouble locating that document but I think it was an HC08 app note. I have the picture in mind of a graph that showed that if Vdd dropped below a certain level (but not low enough for POR) that if it subsequently recovered the MCU would not get a new POR.

I wish to change to HCS08SH8. My question is; Does the HCS08 (specifically the SHx) also have this same characteristic?

Rob

[Ake]

Hi,

The 9S088SH has got a Low-voltage detect which is controlled by the SPMSC1/SPMSC2 registers. The warning voltage can be set at 2.74, 2.92, 4.3 or 4.6 V and the trip point at 2.56 or 4.0 V.

In addition to that, you should use the COP, so that if your program gets stuck somewhere, it gets a restart.

On the earlier HC908GP32, it had a POR rearm voltage of only 0.1 V which could cause some problems.

On the 9S08SH this has been risen to 0.9 V - 2.0 V which has removed the problems.

Regards,

Ake

[RobG]

Hi Ake,

Thanks for your response. I do understand all the low voltage protection features. If it was as simple as that I have no problems. In my servo motor application, if excess motor load is the cause of brown out then it's not an issue because Vdd won't go below LVI warning level (let alone LVI trip level) simply because on warning level the motor is disabled and therefore Vdd will recover from that point.

The problem is other devices in the user defined system causing system wide brownout. Or just bad batteries/wiring etc. I have no control over that. That's why it's possible for the voltage to go very low and why to date I have been using an external low voltage reset IC.

Are you saying that the HC08 problem of not rearming POR is now cured in SH8 simply by lifting the rearm voltage?

Rob

[bigmac]

Hello Rob,

Note that neither an external reset nor the LVD module actually generates a POR, however they will have very similar effects to each other, should they occur.  I can see no advantage in using an external reset device, especially for the SH8.  To enable the external reset pin on the SH8 will mean losing a pin otherwise available for GPIO.

The depth of the "brownout" will determine whether a POR preceeds the release of the LVD reset condition.  What aspects of the POR do you rely on, that do not occur for other methods of reset?  Whether or not a POR has occured for a particular reset is easily determined by reading the SRS register.

Regards,

Mac

[RobG]

Hi Mac,

To answer your question, "What aspects of the POR do you rely on..?" It's simply about ensuring an orderly restart after brown-out. On earlier MCU's it was possible to come out of brownout with MCU in limbo and posibily even damaged. The document that discussed this is AN2158. The gist is that if Vdd goes below 1V but recovers before going under the POR re-arm level of 0.1V (JL/JK devices) "it" could hit the fan.

I'm inclined to agree with you. Having just read HCS08 Unleashed section 5.1 Reset since my last post it hit me that the external LVI  is of no use. My thinking is simply that the external LVI only holds the reset pin low. So what! That has nothing to do with POR. Even if what I was concerned about is a real possibility (not restarting correctly after brown out due to not going as low as the POR re-arm voltage) then releasing the reset pin is of no use. For two reasons: A, because reset pin is not active until SOPT1:RSTPE bit is set and B, because the MCU did not get a POR anyway! Or am I wrong on point A because reset pin is still active since SOPT1:RSTPE bit was set before the brownout?

You also make a good point about using any external reset circuitry at all. Perhaps the fact that there is not a dedicated reset pin is a big enough hint that external reset is not essential. So for brownout protection we'll just use the LVD system.

Rob

[bigmac]

Hello Rob,

With respect to the stated power down conditions, AN2158 seems to require that a set of conditions must be observed that can never be guaranteed in practice, since the duration and depth of a "brownout" is indeterminate.  The application note alludes to dire consequences in the absence of a POR, but is otherwise very vague.  I suspect there might have been a certain amout of "backside covering" within the document.

Keep in mind that the JK/JL series was a very early HC908 offering, with separate versions for crystal oscillator and RC oscillator.  The POR process allowed extra clock startup cycles, that would be important to allow the crystal to stabilize.  This appears the only difference between POR and other forms of reset, that I am aware.

However, for more recent offerings that contain an internal oscillator, the start-up time of the internal oscillator will be of much shorter duration, and any form of reset will generally revert to using the internal oscillator, with use of an external crystal specifically selected during initialisation.

This would leave me to believe that a LV reset should be sufficient to ensure correct operation, with POR occurring only when the brownout was sufficiently deep to meet the POR voltage threshold.

Regards,

Mac

[RobG]

Mac, good reasoning and evaluation on your part. The fact that reset pin is only gpio during POR suggests external RC is a waist of time. So why this note on page 25 of the SH data sheet;

RC filter on RESET pin recommended for noisy environments

I understood external RC is only of value at POR and not for power down when internal LVD can be used.

Thanks for your help.

Rob

[peg]

Hi Rob,

The recommendation to put a RC filter on the reset pin has been more recently added to other datasheets as well.
I presume its purpose is more of a noise filter during runtime rather than a "keep it in reset for a little longer at power on".
So it is still relevant if you have the reset function enabled.

[RobG]

Hi Peg,

Thanks. But how is the RC filter on enabled reset function of benefit during run time? Perhaps it gives debounce to push button reset?

Rob

[bigmac]

Hello Rob,

The external pullup resistor and shunt capacitor on the reset pin is a little bit of inexpensive insurance against a spurious reset occurring, should a noise burst or interference level be induced at the reset pin.  Even if I was using an external LV reset device, I would still provide the filter.

The reason for using the external pullup resistor, say 10k, even though the device may have an internal one, is that the external resistance will be a much lower value, requiring a significantly greater induced current for a spurious reset to occurr.

Regards,

Mac

[RobG]

Thanks Mac, Peg and others.

Excellent help as always.

Rob