HP 15C CE ships with mediocre batteries

[rprosperi]

(Yesterday 08:01 PM)mfleming Wrote:  Here's something to ponder. The HP15C CE I received months ago worked fine, no key bounce problems at all. Then more recently I began to see certain keys repeating when pressed, often enough that I was going to pull out my cable and update the calculator software.

Remembering this thread, I thought, why not change the batteries? Replaced the 2032's that came with the calculator with some Amazon Basics in their adult-proof packaging. Yep, no more bounce. Puzzling...

~Mark

Thanks for posting this Mark!

I've been wondering if there was any meaningful correlation between battery state and presence of bounce. I got one of the very first units and have never experienced any bounce issues at all.

And for several months there were almost no such reports (perhaps 1 or 2?) of bounce issues. Then as time went by, there suddenly were a notable number of reports, and then it seemed to quiet back down again.

No reports, then some, then none again might correlate with batches of batteries used in manufacturing, if there is some cause and effect, but I've never heard of such a case, and frankly don't have enough EE skills to even consider if this is feasible.

Any experts out there?

[RPNerd]

(Today 03:44 AM)rprosperi Wrote:  No reports, then some, then none again might correlate with batches of batteries used in manufacturing, if there is some cause and effect, but I've never heard of such a case, and frankly don't have enough EE skills to even consider if this is feasible.

Any experts out there?

It is plausible.

Whenever a key is pressed, a contact is made, which means that a small current is flowing somewhere. That current, however small, inevitably has to come from the batteries. This means that the voltage from the batteries is going to drop a small amount, but it could be just enough for a gate somewhere in the calculator's guts to no longer register the key press, leading to another condition that draws less power from the batteries, in turn causing the output voltage to go back up and the key press to be registered once again.

If this is the case it could probably be avoided with better decoupling of the +3V rail.

[mfleming]

(Today 11:55 AM)RPNerd Wrote:  If this is the case it could probably be avoided with better decoupling of the +3V rail.

Reasonable conjecture. Lack of a capacitive reservoir might also account for memory loss when replacing depleted batteries one at a time. Anyone willing to slap a big cap across the input rails?

[AnnoyedOne]

(Yesterday 08:01 PM)mfleming Wrote:  I began to see certain keys repeating when pressed...

As far as I know the HP-15C CE uses the SoC internal RC oscillator and not an external crystal (to reduce cost I assume).

I suspect that as the CR2032 batteries drain the system clock slows until it eventually stops altogether and the calculator "freezes". Perhaps that causes "key bounce" issues as well. I wouldn't expect that (the key scan rate should slow) but...

(Yesterday 08:01 PM)mfleming Wrote:  ...Amazon Basics in their adult-proof packaging.

BTW Amazon aren't the only ones to put CR2032's in adult-proof packaging! A few days ago I had to cut some Maxell ones, around the cell edges with scissors, to get them out! Normally you push on the front clear plastic and the back foil breaks. Not with these ones.

Interestingly CR2032 batteries are nominally 3V0 (a bit higher when new) and the ATSAM4LC2CA SoC is a 3V3 part. The ATSAM4LC2CA datasheet states (p1121 of the PDF) that the minimum operating voltage as either 1V68 ("Linear mode" <= 12MHz), 1V8 ("Linear mode" > 12MHz) or 2V0 ("Switching mode"). The choice is hardware selected (see p45 of the PDF). I don't know what the HP-15C CE is configured for.

A1

PS: I don't think that the HP-15C CE/12C+ have any large capacitor across the supply rail. Batteries removed = data loss every time. That saves the cost of a component! Maximise profit and your users can go...

[RPNerd]

(Today 01:29 PM)AnnoyedOne Wrote:  PS: I don't think that the HP-15C CE/12C+ have any large capacitor across the supply rail. Batteries removed = data loss every time. That saves the cost of a component! Maximise profit and your users can go...

In defence of manufacturers, there's also an element of wanting to avoid problems. Some caps can be leaky, even ceramic caps of a large enough capacity. Back in the early days of the DM42, a batch went out with leaky surface mount caps decoupling one of the 3V rails causing unnecessarily high battery drain. I had to replace that part in one of mine and did so for a few others' machines too.

[AnnoyedOne]

(Today 02:00 PM)RPNerd Wrote:  Some caps can be leaky...

True. Electrolytics can be problematic especially over time. I used tantalums but they cost slightly more. Usually lower ESR though.

A1