EmbeddedRelated.com
Forums

Battery for RCM6700

Started by seecwriter August 16, 2012
All,

I have a RCM6700 Deluxe Develop Kit. The 120uA current drain for the bbram and rtc on the rcm6700 is not friendly. I have not found the source of the 120uA specification except in this forum. The docs I have say TBD. I ran some tests that show that the actual value is quite a bit lower; around 25uA @ room temperature. What is the original source for this number? Has anyone else run any empirical tests?

My test for reference;

* connect a 3.6V SAFT LS14250 battery thru a 1N4005 diode to a 120Kohm resistor. The negative battery terminal was connected to J4-2 on the proto board with the resistor connected to J4-1.

* modified the samples\RTC_TEST.c program to also include a bbram variable. The bbram location was printed to the stdio window, then initialized and printed again.

* 1st time running the program the bbram variable was un-initialized garbage. After init, it would retain the init value as long as the battery ciruit was in place. If battery circuit was removed, 1st print was garbage again. RTC values were retained/erased in same way.

* The 120K resistor limited the current as follows:
NOTE: voltage drop across diode was 0.4V

3.2V/120,000 = 27uA

* memory & RTC operated correctly until resistance was raise to 150Kohm

Dave

--- In r..., "seecwriter" wrote:
>
> Mario,
>
> We just did the same thing, we added a separate RTC chip to our
> motherboard that should last over 10-years on the battery. On boot-up
> I set the Rabbit clock from the off-board clock so I can continue to
> use the DC built-in features for time.
>
> Thanks,
> Steve
>
> --- In r..., "mario_wtbbh" wrote:
> >
> > Hey Steve,
> >
> > I just noticed this post now.
> >
> > We raised the high RCM6700 VBat problem with Digi back in March and this was confirmed as a hardware problem. Initially the fix (new modules) was expected around July, but now they are saying "Christmas".
> >
> > As a back-stop, we added a separate SPI RTC chip that draws about 2uA and are using a 1000mAh battery (Renata CR2477) (with a jumper to VBat) on sites that are usually powered, with short periods on battery.
> >
> > I've had to convert all our data storage to flash, with ring buffers to avoid "wearing out" the flash.
> >
> > This itself has lots of side-issues (e.g. Needing to mount/dismount the flash to commit the flash table, which lives in battery-backed RAM).
> >
> > Regards,
> > Mario
> >
> >
> > --- In r..., "seecwriter" wrote:
> > >
> > > That's not a bad choice. But it's half the capacity I need for a
> > > 5-year life, unless I'm over thinking this.
> > >
> > > We may have to go to a separate time-of-day clock device that uses
> > > far less current than the RCM6700's 120ua in order to get the life
> > > span we want with a reasonable size battery.
> > >
> > > Steve
> > >
> > > --- In r..., "Moore, Robert" wrote:
> > > >
> > > > On the OP7200 I used a AA 3.6 VDC lithium battery (Saft LS14500CBA or
> > > > equivalent) accessible from the outside of the instrument case using a
> > > > Bulgin 122-BX0011/1 . . . I'm assuming the RCM6700 has the same pinout
> > > > for this. My biggest concern was to keep the end-user from opening the
> > > > instrument.
> > > >
> > > > bob
> > > > --
> > > > Robert K Moore, PhD
> > > > Designer, Macular Metrics
> > > > Robert_Moore@
> > > > 401-226-6888
> > > >
> > > >
> > > >
> > > >
> > > > On Thu, Aug 16, 2012 at 12:31 PM, seecwriter wrote:
> > > >
> > > > > The battery current draw for the RCM6700 seems rather high, requiring a
> > > > > large battery to maintain the clock for 5-years.
> > > > > Can anything be done to reduce the battery current? I'm just rying to
> > > > > maintain the time-of-clock.
> > > > > What have others done for a battery?
> > > > >
> > > > > Steve
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
There is a known bug in the die of the R6000 cpu that causes the excessive current for
the time-of-day clock. While 25ua is better, it will still kill your battery too soon for my
taste.
 
We remedied the problem by adding our own TOD clock chip to the motherboard.
A PCF8563 from NXP. It draws about 650na over temperature. On bootup I read the
PCF8563 and set the R6k's TOD clock to match it. Then I just use the built-in clock
functions.
 
Steve
>________________________________
> From: morton_street
>To: r...
>Sent: Tuesday, August 13, 2013 11:03 AM
>Subject: [rabbit-semi] Re: Battery for RCM6700
>

>
>All,
>
>I have a RCM6700 Deluxe Develop Kit. The 120uA current drain for the bbram and rtc on the rcm6700 is not friendly. I have not found the source of the 120uA specification except in this forum. The docs I have say TBD. I ran some tests that show that the actual value is quite a bit lower; around 25uA @ room temperature. What is the original source for this number? Has anyone else run any empirical tests?
>
>My test for reference;
>
>* connect a 3.6V SAFT LS14250 battery thru a 1N4005 diode to a 120Kohm resistor. The negative battery terminal was connected to J4-2 on the proto board with the resistor connected to J4-1.
>
>* modified the samples\RTC_TEST.c program to also include a bbram variable. The bbram location was printed to the stdio window, then initialized and printed again.
>
>* 1st time running the program the bbram variable was un-initialized garbage. After init, it would retain the init value as long as the battery ciruit was in place. If battery circuit was removed, 1st print was garbage again. RTC values were retained/erased in same way.
>
>* The 120K resistor limited the current as follows:
>NOTE: voltage drop across diode was 0.4V
>
>3.2V/120,000 = 27uA
>
>* memory & RTC operated correctly until resistance was raise to 150Kohm
>
>Dave
>
>--- In mailto:rabbit-semi%40yahoogroups.com, "seecwriter" wrote:
>>
>> Mario,
>>
>> We just did the same thing, we added a separate RTC chip to our
>> motherboard that should last over 10-years on the battery. On boot-up
>> I set the Rabbit clock from the off-board clock so I can continue to
>> use the DC built-in features for time.
>>
>> Thanks,
>> Steve
>>
>> --- In mailto:rabbit-semi%40yahoogroups.com, "mario_wtbbh" wrote:
>> >
>> > Hey Steve,
>> >
>> > I just noticed this post now.
>> >
>> > We raised the high RCM6700 VBat problem with Digi back in March and this was confirmed as a hardware problem. Initially the fix (new modules) was expected around July, but now they are saying "Christmas".
>> >
>> > As a back-stop, we added a separate SPI RTC chip that draws about 2uA and are using a 1000mAh battery (Renata CR2477) (with a jumper to VBat) on sites that are usually powered, with short periods on battery.
>> >
>> > I've had to convert all our data storage to flash, with ring buffers to avoid "wearing out" the flash.
>> >
>> > This itself has lots of side-issues (e.g. Needing to mount/dismount the flash to commit the flash table, which lives in battery-backed RAM).
>> >
>> > Regards,
>> > Mario
>> >
>> >
>> > --- In mailto:rabbit-semi%40yahoogroups.com, "seecwriter" wrote:
>> > >
>> > > That's not a bad choice. But it's half the capacity I need for a
>> > > 5-year life, unless I'm over thinking this.
>> > >
>> > > We may have to go to a separate time-of-day clock device that uses
>> > > far less current than the RCM6700's 120ua in order to get the life
>> > > span we want with a reasonable size battery.
>> > >
>> > > Steve
>> > >
>> > > --- In mailto:rabbit-semi%40yahoogroups.com, "Moore, Robert" wrote:
>> > > >
>> > > > On the OP7200 I used a AA 3.6 VDC lithium battery (Saft LS14500CBA or
>> > > > equivalent) accessible from the outside of the instrument case using a
>> > > > Bulgin 122-BX0011/1 . . . I'm assuming the RCM6700 has the same pinout
>> > > > for this. My biggest concern was to keep the end-user from opening the
>> > > > instrument.
>> > > >
>> > > > bob
>> > > > --
>> > > > Robert K Moore, PhD
>> > > > Designer, Macular Metrics
>> > > > Robert_Moore@
>> > > > 401-226-6888
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > On Thu, Aug 16, 2012 at 12:31 PM, seecwriter wrote:
>> > > >
>> > > > > The battery current draw for the RCM6700 seems rather high, requiring a
>> > > > > large battery to maintain the clock for 5-years.
>> > > > > Can anything be done to reduce the battery current? I'm just rying to
>> > > > > maintain the time-of-clock.
>> > > > > What have others done for a battery?
>> > > > >
>> > > > > Steve
>> > > > >
>> > > > >
>> > > > >
>> > > > >
>> > > > >
>> > > > > Yahoo! Groups Links
>> > > > >
>> > > > >
>> > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
>
Hi

I'm also using the RCM6700 so this issue interests me.

I have a few thoughts....

1) There is an BAT54 Schottky diode inline with the Ext Bat signal on the RCM6700 board, so need to add your own external diode. This will give you some extra run time.

2) The Battery should only be supplying current when the main board has no power, so the battery life is based on how long the unit is unpowered (in my case, it will be powered most of the time)

3) There is an additional 100K series resistor on the RCM6700 for the 6000's VBATIO input. So you've really got 220K for the VBATIO. This may be too much.

4) The Ext Bat supply powers a 1.2V regulator for the RCM6000's VRAM input. So, any resistance you add is reducing the available supply to this chip. You really have 2V overhead in your scenario (3.6 - 0.4 - 1.2), so with a 150K resistor any current draw over 13.3uA will shut down the regulator.

I'd consider a smaller resistance, and then actually measure the current draw.

However, your tests have made me go back an re-evaluate my own battery supply :)

Phil.


--- In r..., Steve Trigero wrote:
>
> There is a known bug in the die of the R6000 cpu that causes the excessive current for
> the time-of-day clock. While 25ua is better, it will still kill your battery too soon for my
> taste.
>  
> We remedied the problem by adding our own TOD clock chip to the motherboard.
> A PCF8563 from NXP. It draws about 650na over temperature. On bootup I read the
> PCF8563 and set the R6k's TOD clock to match it. Then I just use the built-in clock
> functions.
>  
> Steve
>
>
> >________________________________
> > From: morton_street
> >To: r...
> >Sent: Tuesday, August 13, 2013 11:03 AM
> >Subject: [rabbit-semi] Re: Battery for RCM6700
> >
> > 
> >
> >
> >
> >All,
> >
> >I have a RCM6700 Deluxe Develop Kit. The 120uA current drain for the bbram and rtc on the rcm6700 is not friendly. I have not found the source of the 120uA specification except in this forum. The docs I have say TBD. I ran some tests that show that the actual value is quite a bit lower; around 25uA @ room temperature. What is the original source for this number? Has anyone else run any empirical tests?
> >
> >My test for reference;
> >
> >* connect a 3.6V SAFT LS14250 battery thru a 1N4005 diode to a 120Kohm resistor. The negative battery terminal was connected to J4-2 on the proto board with the resistor connected to J4-1.
> >
> >* modified the samples\RTC_TEST.c program to also include a bbram variable. The bbram location was printed to the stdio window, then initialized and printed again.
> >
> >* 1st time running the program the bbram variable was un-initialized garbage. After init, it would retain the init value as long as the battery ciruit was in place. If battery circuit was removed, 1st print was garbage again. RTC values were retained/erased in same way.
> >
> >* The 120K resistor limited the current as follows:
> >NOTE: voltage drop across diode was 0.4V
> >
> >3.2V/120,000 = 27uA
> >
> >* memory & RTC operated correctly until resistance was raise to 150Kohm
> >
> >Dave
> >
> >--- In mailto:rabbit-semi%40yahoogroups.com, "seecwriter" wrote:
> >>
> >> Mario,
> >>
> >> We just did the same thing, we added a separate RTC chip to our
> >> motherboard that should last over 10-years on the battery. On boot-up
> >> I set the Rabbit clock from the off-board clock so I can continue to
> >> use the DC built-in features for time.
> >>
> >> Thanks,
> >> Steve
> >>
> >> --- In mailto:rabbit-semi%40yahoogroups.com, "mario_wtbbh" wrote:
> >> >
> >> > Hey Steve,
> >> >
> >> > I just noticed this post now.
> >> >
> >> > We raised the high RCM6700 VBat problem with Digi back in March and this was confirmed as a hardware problem. Initially the fix (new modules) was expected around July, but now they are saying "Christmas".
> >> >
> >> > As a back-stop, we added a separate SPI RTC chip that draws about 2uA and are using a 1000mAh battery (Renata CR2477) (with a jumper to VBat) on sites that are usually powered, with short periods on battery.
> >> >
> >> > I've had to convert all our data storage to flash, with ring buffers to avoid "wearing out" the flash.
> >> >
> >> > This itself has lots of side-issues (e.g. Needing to mount/dismount the flash to commit the flash table, which lives in battery-backed RAM).
> >> >
> >> > Regards,
> >> > Mario
> >> >
> >> >
> >> > --- In mailto:rabbit-semi%40yahoogroups.com, "seecwriter" wrote:
> >> > >
> >> > > That's not a bad choice. But it's half the capacity I need for a
> >> > > 5-year life, unless I'm over thinking this.
> >> > >
> >> > > We may have to go to a separate time-of-day clock device that uses
> >> > > far less current than the RCM6700's 120ua in order to get the life
> >> > > span we want with a reasonable size battery.
> >> > >
> >> > > Steve
> >> > >
> >> > > --- In mailto:rabbit-semi%40yahoogroups.com, "Moore, Robert" wrote:
> >> > > >
> >> > > > On the OP7200 I used a AA 3.6 VDC lithium battery (Saft LS14500CBA or
> >> > > > equivalent) accessible from the outside of the instrument case using a
> >> > > > Bulgin 122-BX0011/1 . . . I'm assuming the RCM6700 has the same pinout
> >> > > > for this. My biggest concern was to keep the end-user from opening the
> >> > > > instrument.
> >> > > >
> >> > > > bob
> >> > > > --
> >> > > > Robert K Moore, PhD
> >> > > > Designer, Macular Metrics
> >> > > > Robert_Moore@
> >> > > > 401-226-6888
> >> > > >
> >> > > >
> >> > > >
> >> > > >
> >> > > > On Thu, Aug 16, 2012 at 12:31 PM, seecwriter wrote:
> >> > > >
> >> > > > > The battery current draw for the RCM6700 seems rather high, requiring a
> >> > > > > large battery to maintain the clock for 5-years.
> >> > > > > Can anything be done to reduce the battery current? I'm just rying to
> >> > > > > maintain the time-of-clock.
> >> > > > > What have others done for a battery?
> >> > > > >
> >> > > > > Steve
> >> > > > >
> >> > > > >
> >> > > > >
> >> > > > >
> >> > > > >
> >> > > > >