NewRPL announciator question

[nickapos]

Hello to everyone, sorry for the simple question but i have not been able to find a clear answer in the documentation.
On my 39gs NewRPL calculator I can see that the announciator ((*)) is constantly on.

According to github this should be the busy announciator, but why is it always on even when there is no computation happening.

Am I reading this right?
I wonder if I have something on that is constantly running and consume energy or if this is normal.

Thank you very much

[SammysHP]

AFAIK that annunciator is a low battery warning. I have the same problem with my calculator and already reported it to the developers a long time ago. The ADC values have wrong thresholds, but for some reason I've never continued to fix them. Maybe because I got lost in the code somewhere. IIRC I asked the developer how I can read the ADC value, but didn't get a helpful response.

Edit:

Quote:2022-01-11

Hi Claudio,

I've just noticed that the low battery annunciator on my HP 39gs with
newRPL is on, but the batteries (NiMH) are still at 1.25 V each (5 V
total). IMHO that's still a very acceptable level.

From looking at the code it doesn't seem there is a command to print
the current voltage or ADC value and also nothing for runtime
calibration. Do you think it would be feasible to add both:

1. a command to put the current voltage or ADC value on the stack
2. a hidden variable that can be used to calibrate the voltage reading
(offset to the hardcoded thresholds)

Got a reply that it works as intended. Under load the battery voltage should drop. I was also told to wait until the annunciator starts blinking. My response:

Quote:I've just tested it under load (NUMINT) using my oscilloscope. There was
almost no drop, only 0.01 V. Eneloop have a pretty good discharge curve.
A method to get the current voltage or raw value would be great
nevertheless.

No further messages are recorded.

[nickapos]

(05-05-2023 01:39 PM)SammysHP Wrote:  AFAIK that annunciator is a low battery warning. I have the same problem with my calculator and already reported it to the developers a long time ago. The ADC values have wrong thresholds, but for some reason I've never continued to fix them. Maybe because I got lost in the code somewhere. IIRC I asked the developer how I can read the ADC value, but didn't get a helpful response.

Edit:

Quote:2022-01-11

Hi Claudio,

I've just noticed that the low battery annunciator on my HP 39gs with
newRPL is on, but the batteries (NiMH) are still at 1.25 V each (5 V
total). IMHO that's still a very acceptable level.

From looking at the code it doesn't seem there is a command to print
the current voltage or ADC value and also nothing for runtime
calibration. Do you think it would be feasible to add both:

1. a command to put the current voltage or ADC value on the stack
2. a hidden variable that can be used to calibrate the voltage reading
(offset to the hardcoded thresholds)

Got a reply that it works as intended. Under load the battery voltage should drop. I was also told to wait until the annunciator starts blinking. My response:

Quote:I've just tested it under load (NUMINT) using my oscilloscope. There was
almost no drop, only 0.01 V. Eneloop have a pretty good discharge curve.
A method to get the current voltage or raw value would be great
nevertheless.

No further messages are recorded.

Thank you very much for the info. I measured the batteries with my multi-meter and replaced the ones that were depleted. I would expect this annunciator to appear at a lot lower voltage than the one I have right now. I think i will ignore it for now.

[Claudio L.]

Let me explain:

newRPL has 2 thresholds for battery level:

a) The first level is intended to be a long warning, you know your batteries are getting low but there's no need for you to run and change them yet. This is when the indicator is constant, either all the time or just when the high clock kicks in (at the beginning you'll see it pop up whenever you do an extended calculation that takes more than a couple of seconds and the 192 MHz mode kicks in, later you'll see it all the time, even in idle mode).
b) The critical level is when the indicator is blinking. Once it's blinking you know you have critically low voltage and batteries need to be replaced.

Ant that's it. newRPL does NOT turn itself off like the normal operating system. It will try to run until the battery dies completely (that's also why you are thinking the warnings come out too early)

Why 2 levels? Because it depends on how you use the calculator and how good or bad your batteries are (and the type, alkalines are 1.5V, while NiMH are only 1.2V). When you run programs that do "heavy" calculations that make the high clock kick in, some battery brands will drop the voltage too much and the CPU may crash. That's when the "critical" warning is blinking.

So, when the indicator is constant you have low voltage but still OK to keep the batteries. When you tax the CPU you might see it start blinking temporarily and back to steady on when it's idle. You should replace your batteries if you need to keep using the high clock.
When the indicator is blinking in idle, it's very late in the game, don't try to run in high clock unless you have very good brand batteries (for example, NiMH high-demand batteries can deliver high currents with little voltage drop, even when they are close to depleted and you'll likely be OK, but a non-brand standard alkaline might crash the CPU if you push it).

Summarizing:
If the indicator blinks in idle = replace me now, better safe than sorry.
Solid indicator in idle = keep a set of batteries handy, especially if you need to push the calculator to its limits. If it turns to blinking while the calc is being taxed... go ahead and replace, or stop making it work too hard if you want to keep the batteries.

I use NiMH and because of the low voltage the steady indicator comes in early, but I know these are rock solid and can deliver a punch without dropping the voltage too much, so I pretty much ignore the solid one. However, when the voltage drops and it starts blinking, I need to replace the same day because this type of batteries hold the voltage very well, then they drop dead quickly with little warning.
Regular alkalines have higher voltage, but also drop the voltage much more when pushed, so when you see the steady warning on idle you need to keep an eye when you tax it to see if it doesn't start blinking. If it blinks in high load, it's time to replace them, if you wait for blinking in idle you risk crashing the CPU under high load (some brands might still work well though, your mileage might vary).
If you want to squeeze the last bit of your batteries, then buy always the same brand and type, and run some tests to see how long it lasts when it's blinking, and whether you can run the CPU in high clock when blinking or if it makes it crash. Know your battery to know your limits.

[Claudio L.]

I'll add a little more hard data for those who care:

Originally I had code displaying the voltage. Made no sense other than cause anxiety... I calibrated it with my unit, but other members of the team reported their unit had different calibration, so there's that: each calc reports slightly different voltage. Voltage also doesn't tell you much because what really matters is the voltage drop while under load. And that load also is different from calc to calc, especially if you have a an SD card. The critical operation is a program that calculates something and writes results to an SD card: CPU kicks in the turbo while the SD card is writing to flash. But power consumption on SD cards varies wildly too, so does the voltage drop.

So I ignored the voltage and went with crash: calc crashed at about 4.5V when under load, so the blinking indicator I calibrated slightly above around 4.7V. The solid indicator if I recall correctly is calibrated around 4.9V. Think that if you are using alkalines you should be around 6V with 4 new batteries. NiMH are 1.2V but the brands I tested typically have around 1.3V to 1.37V when charged and the voltage drop isn't that significant.

I understand people might prefer to custom-calibrate the indicator but really, it will apply to only one calc, with one type of batteries, and you'd have to sit down with your multimeter to read the voltages with the calc under load, then try several SD cards to make sure... it's a pain, not worth anybody's time I think.

[nickapos]

Thank you very much Claudio for the info, it clarifies everything.
In my case the annunciator is constantly on, just checked the voltage again and it is 5.3V.
In this case should the annunciator be on?

[Claudio L.]

(05-07-2023 03:10 AM)nickapos Wrote:  Thank you very much Claudio for the info, it clarifies everything.
In my case the annunciator is constantly on, just checked the voltage again and it is 5.3V.
In this case should the annunciator be on?

As I said, on my machines I calibrated for it to come up around 4.9V, but we realized that there's variability between calcs, I guess yours gives for 5.3V the same reading I got for 4.9V. I'm not having a lot of free time right now but I'll take a look and see if we can add some user settings so you can calibrate yourself from within newRPL. Give me a couple of months or so, I'm really tied up with "real" work.

[nickapos]

(05-09-2023 01:13 PM)Claudio L. Wrote:  

(05-07-2023 03:10 AM)nickapos Wrote:  Thank you very much Claudio for the info, it clarifies everything.
In my case the annunciator is constantly on, just checked the voltage again and it is 5.3V.
In this case should the annunciator be on?

As I said, on my machines I calibrated for it to come up around 4.9V, but we realized that there's variability between calcs, I guess yours gives for 5.3V the same reading I got for 4.9V. I'm not having a lot of free time right now but I'll take a look and see if we can add some user settings so you can calibrate yourself from within newRPL. Give me a couple of months or so, I'm really tied up with "real" work.

Hi Claudio, thanks for the clarification.
For the record this is not a request for fix, i am just trying to understand if this is normal or not. It does not prevent me from using my calculator anyway so it is not a big deal.
It is a minor issue, and quite possibly this might be something that depends on quality control of the hardware itself. My calculator is a device bought from AliExpress last year so I would not be surprised if it was not completely up to HP spec.