I have an LPC3180 header board that needs 10MHz+ crystals to run per spec. I plugged in a ~2MHz crystal and it ran, as slow as expected. I've underclocked PIC parts down to being static, even using a button as a clock. Can this be done with ARM MCUs? Can this be done with ARM MCUs interfaced with SRAM chips?
Underclocking ARM and SRAM
Started by ●August 11, 2006
Reply by ●August 11, 20062006-08-11
On 11 Aug 2006 13:23:01 -0700, "ghazanhaider" <ghazan.haider@gmail.com> wrote:>I have an LPC3180 header board that needs 10MHz+ crystals to run per >spec. I plugged in a ~2MHz crystal and it ran, as slow as expected. > >I've underclocked PIC parts down to being static, even using a button >as a clock. Can this be done with ARM MCUs? Can this be done with ARM >MCUs interfaced with SRAM chips?Read the datasheet - there should be a min clock freq spec - if this is 0 Hz, it will work! SRAM certainly won't have a problem.
Reply by ●August 12, 20062006-08-12
In article <us1qd21qhijfiphc5ed1u362j9ibrqv7e3@4ax.com>, mike@whitewing.co.uk says...> On 11 Aug 2006 13:23:01 -0700, "ghazanhaider" <ghazan.haider@gmail.com> wrote: > > >I have an LPC3180 header board that needs 10MHz+ crystals to run per > >spec. I plugged in a ~2MHz crystal and it ran, as slow as expected. > > > >I've underclocked PIC parts down to being static, even using a button > >as a clock. Can this be done with ARM MCUs? Can this be done with ARM > >MCUs interfaced with SRAM chips? > > Read the datasheet - there should be a min clock freq spec - if this is 0 Hz, it will work! > SRAM certainly won't have a problem.For an external crystal, the LPC3180 accepts from 1 to 20 MHz. It also has an internal 32.768KHz oscillator, which can drive the system clock at 13 MHz via a PLL. --Gene
Reply by ●August 12, 20062006-08-12
Gene S. Berkowitz wrote:> In article <us1qd21qhijfiphc5ed1u362j9ibrqv7e3@4ax.com>, > mike@whitewing.co.uk says... > > On 11 Aug 2006 13:23:01 -0700, "ghazanhaider" <ghazan.haider@gmail.com> wrote: > > > > >I have an LPC3180 header board that needs 10MHz+ crystals to run per > > >spec. I plugged in a ~2MHz crystal and it ran, as slow as expected. > > > > > >I've underclocked PIC parts down to being static, even using a button > > >as a clock. Can this be done with ARM MCUs? Can this be done with ARM > > >MCUs interfaced with SRAM chips? > > > > Read the datasheet - there should be a min clock freq spec - if this is 0 Hz, it will work! > > SRAM certainly won't have a problem. > > For an external crystal, the LPC3180 accepts from 1 to 20 MHz. > It also has an internal 32.768KHz oscillator, which can drive the system > clock at 13 MHz via a PLL. > > --GeneOops! I meant an LPC2138. Its minimum per user manual and datasheet is 10MHz. I've been keeping an eye on the lpc3180 and its been on my mind. Samples arent out yet.
Reply by ●August 13, 20062006-08-13
ghazanhaider wrote:> Gene S. Berkowitz wrote: > > In article <us1qd21qhijfiphc5ed1u362j9ibrqv7e3@4ax.com>, > > mike@whitewing.co.uk says...> > For an external crystal, the LPC3180 accepts from 1 to 20 MHz. > > It also has an internal 32.768KHz oscillator, which can drive the system > > clock at 13 MHz via a PLL. > > > > --Gene > > Oops! I meant an LPC2138. Its minimum per user manual and datasheet is > 10MHz.Interesting. I wonder why it's minimum is higher than other members of the family (such as the 2106)? Robert
Reply by ●August 14, 20062006-08-14
> > Interesting. I wonder why it's minimum is higher than other members of > the family (such as the 2106)? > > RobertI dont know but I intend to play with much lower frequencies. Maybe its for the DAC and UARTs, I wont need em. Maybe ISP wont work. But last I tried a blinky prog 2MHz worked great. Crystal I mean.
Reply by ●August 14, 20062006-08-14
ghazanhaider wrote:> > > > Interesting. I wonder why it's minimum is higher than other members of > > the family (such as the 2106)? > > > > Robert > > I dont know but I intend to play with much lower frequencies. Maybe its > for the DAC and UARTs, I wont need em. Maybe ISP wont work. But last I > tried a blinky prog 2MHz worked great. Crystal I mean.As I recall, on the other members, the UARTs worked fine (not surprising) but ISP expected 10MHz or better. The big issue was the PLL which required a minimum of 10MHz input. Maybe the ADC has some timing limitations (Maybe hold time on the S/H?) but I would have thought the DAC to be all right. Robert
Reply by ●August 14, 20062006-08-14
sub2@aeolusdevelopment.com wrote:> As I recall, on the other members, the UARTs worked fine (not > surprising) but ISP expected 10MHz or better. The big issue was the > PLL which required a minimum of 10MHz input. > > Maybe the ADC has some timing limitations (Maybe hold time on the S/H?) > but I would have thought the DAC to be all right. > > RobertWell, I'm aiming for a really low power system, and will start testing underclocking the lpc2138 first. The pll divider only takes me to about 1MHz or a little less... but if the 2MHz crystal test worked so should possibly a 32KHz. In the longer run I have the lpc2888 in mind, but thats a bigger testing rig at 1.8V plus the low power sram and flash. That thing has a 10MHz limit too and its timing with sram and flash will be tricky. Gotta make it work. Gotta be the first one to make a solar powered linux computer that you can put in your pocket! Currently I need a square foot of panel in a bright indoor room as a minimum. Aint got pockets that big.
Reply by ●August 15, 20062006-08-15
ghazanhaider wrote:> sub2@aeolusdevelopment.com wrote: > > > As I recall, on the other members, the UARTs worked fine (not > > surprising) but ISP expected 10MHz or better. The big issue was the > > PLL which required a minimum of 10MHz input. > > > > Maybe the ADC has some timing limitations (Maybe hold time on the S/H?) > > but I would have thought the DAC to be all right. > > > > Robert > > Well, I'm aiming for a really low power system, and will start testing > underclocking the lpc2138 first. The pll divider only takes me to about > 1MHz or a little less..The datasheet has power consumption numbers down to 0 Hz (on the graph), maybe there is another clock divide down somewhere like on the Atmel ARM's.
Reply by ●August 18, 20062006-08-18
OK heres an update. A 2MHz and a 1.8432MHz rectangular crystal both worked, but a cylindrical 32.76kHz crystal failed. With the 2MHz and PLL divider at 16 it will run at 125kHz. I think I'll go xtal shopping till I find the slowest and lowest power at which it'll work, then I'll do it with the lpc 2888 or lpc3180 (whichever is at 1.8V) for the lowest power.