OpenGD77

Correction

We do leave it in Data mode

Perhaps it should be left in Command mode.

I guess that's worth a try

Na.

Doesn't make any difference.

Also it doesn't make any difference even if I stop updating the display when the backlight gets turned off.

I think that probably rules out the display as a way of saving power.

Yes.

I need to take some measurements again when the official firmware is running in its power saving mode.

Because I think the current taken when the Rx is turned off, is only 25mA.

But the MK22 takes between 25mA and 35mA on its own, depending on what "peripheral" clocks are running.

And I've also measured about 6mA of current even if I halt the MCU before anything is enabled.

So, I'm currently unsure how the official firmware achieves 25mA, unless perhaps they just reduce the clock rate, rather than going the whole way and entering Very Low Power run mode.

I don't know if its possible to change the master clock PLL on the fly. I may give it a try and see what happens, but I expect the radio will hard-fault and reboot

I'm not sure if I'm really seeing this.

But it looks like the first time I read the Flash memory, the current consumption goes up by 7mA, and does not go down again.

In theory the Flash memory chip is supposed to return to its "standby" state when its not actively in use, but possibly something about the software SPI interface I wrote, is causing the IC to remain in its active mode.

There is some cryptic text in the data sheet about the Data Out pin needing to be "High Impedance" as well as the Chip Select pin being High (deselected), for the Flash chip to enter its "standby" state.

I've tried leaving the Data Out pin , from the MK22 CPU, in both High and Low states, and the overall current consumption by the radio didn't go down after reading the Flash memory chip.

I've yet to try de-initialising the GPIO pins connected to the chip, to make the Data Out, and Data In pins become high impedance, but I guess thats worth a try

I've also ordered a Arduino style Flash memory chip module, so I can do some direct current measurements on an individual chip, however I can't by exactly the same chip on a module, and I'll have to hope that all the Winbond Flash chips behave roughly the same.

I should receive the module by the end of the week if I'm lucky, but the postal service is a bit slow at the moment because of the pandemic.

OK.

After a bit more experimentation I think I've found how to put the flash chip into its standby mode, which it should always return to.

Well, apart from perhaps 2mA, but I think the 2 mA difference is possibly not the Flash chip and its something inside the MK22 microcontroller

I'll now need to work out how to integrate this change into the firmware, so save about 7mA...

It doesn't sound much, but this is around 10% of the battery consumption when the display backlight is turned off.

I've now made a version with modified interface to the Flash memory chip, which appears to save around 8 or 9mA on the GD77, but only around 3mA on the DM1801.
I don't know how much current it saves on the RD5R, but at least 3mA based on the other 2 radio types

9mA on the GD77 equates to around 10% of the constant current drain on the battery, so is definitely a worthwhile change.

I initially had problems with the changes to the Flash memory access, which prevented some radios being able to read the Flash memory and those radios hung on with CAL DATA ERROR and would not work at all.

About 10 people have now tested a bug fix version, which no longer seems to have this problem.

This version should be considered as highly experimental, and should only be loaded, if you know that there is chance that it won't work, and you may even need to remove the battery to reset the radio.


https://drive.google.com/drive/folders/ ... sp=sharing


But I need to get this version tested by more people before it even makes it to being a Beta, hence why I've posted the files to my google drive.

Note.
The JA versions are Japanese only, and have the Japanese font instead of the normal "latin" / western font

(This is the same as the last version I emailed to Beta testers, and I seems to now work for all of them)

VK3KYY wrote: ↑

Fri Jan 22, 2021 9:23 pm

But I need to get this version tested by more people before it even makes it to being a Beta, hence why I've posted the files to my google drive.

Hi Roger!
I just flashed this version to my GD-77 and did some basic tests. It starts up and shuts down normally, receives and transmits normally and UI operates normally as well.
My battery percentage is now at 37%, so I'll check whether the lower power consumption is observable.
I'll write back soon but for now, it looks well.

73, Pavel

OK.

Thanks

I think either the Flash memory works or does not work from the beginning.

I don't know if 9mA / 90mA is very noticeable. Some people thing the battery lasts longer, but I have not done any tests myself to measure the increase in battery life, and it will depend on whether the radio is idle all the time or if some time its either receiving or transmitting.

The most difference will just be when the radio is idle and the backlight is off (for radios with a display)

I am testing some more changes to reduce the current.

I think I can reduce the current by another 23mA, so the total reduction is 30mA

Current Beta takes 91mA on Rx idle, and I now have a version which takes 61mA.

Tx current is almost the same, but during Tx, the majority of the current is the RF / PA, and this is unavoidable.

I will post this in a few days when it had some more testing by a few people.