GA144 one pin oscillator again

There has been debate on this forum about whether with the Green Arrays
GA144 chip, it is possible to make a crystal oscillator attached to one
pin and no further components.

The inventor of the chip demonstrates a working example, at least for a
resonator that has a lower Q. A high Q only makes it harder to start
the oscillation, but the principle is the same.

This is explained in a video lecture (showing actual, albeit
hard-to-read, code). Google for " fireside chat chuck Moore"

The video also demonstrates a colored character display generated on
the fly by a dozen of the 144 small computers on the chip.
There is no buffer for the 1000 characters or the character
definitions, it is all dynamic.

The computers are asynchronous, the resonator is needed for the
timing of the display only.

Groetjes Albert

--
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

I don't think there has been an debate about whether a crystal
oscillator is "possible" on the GA144.  I have not yet seen any proof
or even reasonable evidence that a robust, reliable software based,
crystal oscillator has been demonstrated on the GA144.  I will take a
look at the video you describe, but if it is anything like the other
app notes from GA it will be largely hand waving and leave the bulk of
the work to the viewer.  Demonstrating a reliable oscillator is very
different from producing an app note that shows one working in the
lab.

What type of resonator is demonstrated?  What frequency?  One of the
problems encountered before was how to start a high Q oscillator at a
higher frequency, I believe 10 MHz.  There is not enough resolution in
a 700 MIPS instruction stream to tune to the crystal rate to get it
going.

Rick


On Apr 28, 7:49 am, Albert van der Horst <alb...@spenarnc.xs4all.nl>
wrote:
> There has been debate on this forum about whether with the Green Arrays
> GA144 chip, it is possible to make a crystal oscillator attached to one
> pin and no further components.
>
> The inventor of the chip demonstrates a working example, at least for a
> resonator that has a lower Q. A high Q only makes it harder to start
> the oscillation, but the principle is the same.
>
> This is explained in a video lecture (showing actual, albeit
> hard-to-read, code). Google for " fireside chat chuck Moore"
>
> The video also demonstrates a colored character display generated on
> the fly by a dozen of the 144 small computers on the chip.
> There is no buffer for the 1000 characters or the character
> definitions, it is all dynamic.
>
> The computers are asynchronous, the resonator is needed for the
> timing of the display only.
>
> Groetjes Albert

rickman wrote:
> I don't think there has been an debate about whether a crystal
> oscillator is "possible" on the GA144.  I have not yet seen any proof
> or even reasonable evidence that a robust, reliable software based,
> crystal oscillator has been demonstrated on the GA144.  I will take a
> look at the video you describe, but if it is anything like the other
> app notes from GA it will be largely hand waving and leave the bulk of
> the work to the viewer.  Demonstrating a reliable oscillator is very
> different from producing an app note that shows one working in the
> lab.
>
> What type of resonator is demonstrated?  What frequency?  One of the
> problems encountered before was how to start a high Q oscillator at a
> higher frequency, I believe 10 MHz.  There is not enough resolution in
> a 700 MIPS instruction stream to tune to the crystal rate to get it
> going.
>
> Rick

I think I'm going to make some popcorn.
This should be entertaining.

-jim

In article <0c732237-196c-4393-ae5f-aa6736fb1bd6@c4g2000yqj.googlegroups.com>,
rickman  <gnuarm@gmail.com> wrote:
>I don't think there has been an debate about whether a crystal
>oscillator is "possible" on the GA144.  I have not yet seen any proof

I will not go into it whether that debate has taken place.

I'm talking about the debate whether even an experimental oscillator
using only a crystal, a single pin and no further components is possible
using the GA144. A demonstration is relevant to that discussion.

>or even reasonable evidence that a robust, reliable software based,
>crystal oscillator has been demonstrated on the GA144.  I will take a
>look at the video you describe, but if it is anything like the other
>app notes from GA it will be largely hand waving and leave the bulk of
>the work to the viewer.  Demonstrating a reliable oscillator is very
>different from producing an app note that shows one working in the
>lab.

Yes, this is a demonstration only, not giving enough information
to build the thing yourself. It will irk you. The source code is
shown, on a barely readable video screen, in the unfathomable
colorforth language.
Still *I* think it is an interesting video.

<SNIP>

>What type of resonator is demonstrated?  What frequency?  One of the
>problems encountered before was how to start a high Q oscillator at a
>higher frequency, I believe 10 MHz.  There is not enough resolution in
>a 700 MIPS instruction stream to tune to the crystal rate to get it
>going.

I take exception to this. It depends on the crystal frequency.
I intend to experiment with a 32kHz clock crystal.

>Rick

Groetjes Albert


--
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

On Sun, 13 May 2012 14:38:22 +0000, Albert van der Horst wrote:

> In article
> <0c732237-196c-4393-ae5f-aa6736fb1bd6@c4g2000yqj.googlegroups.com>,
> rickman  <gnuarm@gmail.com> wrote:
>>I don't think there has been an debate about whether a crystal
>>oscillator is "possible" on the GA144.  I have not yet seen any proof
> 
> I will not go into it whether that debate has taken place.
> 
> I'm talking about the debate whether even an experimental oscillator
> using only a crystal, a single pin and no further components is possible
> using the GA144. A demonstration is relevant to that discussion.
> 
>>or even reasonable evidence that a robust, reliable software based,
>>crystal oscillator has been demonstrated on the GA144.  I will take a
>>look at the video you describe, but if it is anything like the other app
>>notes from GA it will be largely hand waving and leave the bulk of the
>>work to the viewer.  Demonstrating a reliable oscillator is very
>>different from producing an app note that shows one working in the lab.
> 
> Yes, this is a demonstration only, not giving enough information to
> build the thing yourself. It will irk you. The source code is shown, on
> a barely readable video screen, in the unfathomable colorforth language.
> Still *I* think it is an interesting video.
> 
> <SNIP>
> 
>>What type of resonator is demonstrated?  What frequency?  One of the
>>problems encountered before was how to start a high Q oscillator at a
>>higher frequency, I believe 10 MHz.  There is not enough resolution in a
>>700 MIPS instruction stream to tune to the crystal rate to get it going.
> 
> I take exception to this. It depends on the crystal frequency. I intend
> to experiment with a 32kHz clock crystal.

Without commenting on the technical feasibility, the last time there was 
a discussion on this I looked at the white paper from the processor 
company and it seemed to be nearly all smoke and mirrors.  What little 
content there was could have been generated by a fresh college grad (one 
who got a low GPA) wandering into the lab and doing some hurried work at 
the last minute.

Either they have such a weak product that they feel they need the smoke 
and mirrors, or they have such little regard for their customers that 
they do not feel that decent applications support is worth paying for, or 
their top management is incompetent at selecting competent engineering 
staff, or money is so tight that they can't put together the competent 
staff they know they need.

None of these circumstances make me feel that designing in their 
processor would be a good risk.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

On Sun, 13 May 2012 14:38:22 +0000, Albert van der Horst wrote:

> In article
> <0c732237-196c-4393-ae5f-aa6736fb1bd6@c4g2000yqj.googlegroups.com>,
> rickman  <gnuarm@gmail.com> wrote:
>>I don't think there has been an debate about whether a crystal
>>oscillator is "possible" on the GA144.  I have not yet seen any proof
> 
> I will not go into it whether that debate has taken place.
> 
> I'm talking about the debate whether even an experimental oscillator
> using only a crystal, a single pin and no further components is possible
> using the GA144. A demonstration is relevant to that discussion.
> 
>>or even reasonable evidence that a robust, reliable software based,
>>crystal oscillator has been demonstrated on the GA144.  I will take a
>>look at the video you describe, but if it is anything like the other app
>>notes from GA it will be largely hand waving and leave the bulk of the
>>work to the viewer.  Demonstrating a reliable oscillator is very
>>different from producing an app note that shows one working in the lab.
> 
> Yes, this is a demonstration only, not giving enough information to
> build the thing yourself. It will irk you. The source code is shown, on
> a barely readable video screen, in the unfathomable colorforth language.
> Still *I* think it is an interesting video.
> 
> <SNIP>
> 
>>What type of resonator is demonstrated?  What frequency?  One of the
>>problems encountered before was how to start a high Q oscillator at a
>>higher frequency, I believe 10 MHz.  There is not enough resolution in a
>>700 MIPS instruction stream to tune to the crystal rate to get it going.
> 
> I take exception to this. It depends on the crystal frequency. I intend
> to experiment with a 32kHz clock crystal.

You may be able to make that work at 32kHz.  To verify that what you do 
will actually work correctly in the long run you not only need to be good 
at analog circuitry and at software, but you'll need the cooperation of 
the manufacturer -- and given their capabilities in that direction, I 
don't think you'll ever get there.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

In article <ge2dnSy0Xs4GdzLSnZ2dnUVZ_vEAAAAA@web-ster.com>,
Tim Wescott  <tim@seemywebsite.please> wrote:
>On Sun, 13 May 2012 14:38:22 +0000, Albert van der Horst wrote:
>
>> In article
>> <0c732237-196c-4393-ae5f-aa6736fb1bd6@c4g2000yqj.googlegroups.com>,
>> rickman  <gnuarm@gmail.com> wrote:
>>>I don't think there has been an debate about whether a crystal
>>>oscillator is "possible" on the GA144.  I have not yet seen any proof
>>
>> I will not go into it whether that debate has taken place.
>>
>> I'm talking about the debate whether even an experimental oscillator
>> using only a crystal, a single pin and no further components is possible
>> using the GA144. A demonstration is relevant to that discussion.
>>
>>>or even reasonable evidence that a robust, reliable software based,
>>>crystal oscillator has been demonstrated on the GA144.  I will take a
>>>look at the video you describe, but if it is anything like the other app
>>>notes from GA it will be largely hand waving and leave the bulk of the
>>>work to the viewer.  Demonstrating a reliable oscillator is very
>>>different from producing an app note that shows one working in the lab.
>>
>> Yes, this is a demonstration only, not giving enough information to
>> build the thing yourself. It will irk you. The source code is shown, on
>> a barely readable video screen, in the unfathomable colorforth language.
>> Still *I* think it is an interesting video.
>>
>> <SNIP>
>>
>>>What type of resonator is demonstrated?  What frequency?  One of the
>>>problems encountered before was how to start a high Q oscillator at a
>>>higher frequency, I believe 10 MHz.  There is not enough resolution in a
>>>700 MIPS instruction stream to tune to the crystal rate to get it going.
>>
>> I take exception to this. It depends on the crystal frequency. I intend
>> to experiment with a 32kHz clock crystal.
>
>You may be able to make that work at 32kHz.  To verify that what you do
>will actually work correctly in the long run you not only need to be good
>at analog circuitry and at software, but you'll need the cooperation of
>the manufacturer -- and given their capabilities in that direction, I
>don't think you'll ever get there.

Oh, I certainly agree with you there. Green Arrays consists of senior
citizens with very little in the direction of marketing potential.
I'm purely interested in the R&D aspects.
(Members of) the Dutch Forth chapter are experimenting to get a decent
development environment for the chip. We do get some cooperation from
Green Arrays.

>
>--
>Tim Wescott

Groetjes Albert

--
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

On May 13, 9:35=A0pm, Albert van der Horst <alb...@spenarnc.xs4all.nl>
wrote:
> In article <ge2dnSy0Xs4GdzLSnZ2dnUVZ_vEAA...@web-ster.com>,
> Tim Wescott =A0<t...@seemywebsite.please> wrote:
>
>
>
>
>
>
>
>
>
> >On Sun, 13 May 2012 14:38:22 +0000, Albert van der Horst wrote:
>
> >> In article
> >> <0c732237-196c-4393-ae5f-aa6736fb1...@c4g2000yqj.googlegroups.com>,
> >> rickman =A0<gnu...@gmail.com> wrote:
> >>>I don't think there has been an debate about whether a crystal
> >>>oscillator is "possible" on the GA144. =A0I have not yet seen any proo=
f
>
> >> I will not go into it whether that debate has taken place.
>
> >> I'm talking about the debate whether even an experimental oscillator
> >> using only a crystal, a single pin and no further components is possib=
le
> >> using the GA144. A demonstration is relevant to that discussion.
>
> >>>or even reasonable evidence that a robust, reliable software based,
> >>>crystal oscillator has been demonstrated on the GA144. =A0I will take =
a
> >>>look at the video you describe, but if it is anything like the other a=
pp
> >>>notes from GA it will be largely hand waving and leave the bulk of the
> >>>work to the viewer. =A0Demonstrating a reliable oscillator is very
> >>>different from producing an app note that shows one working in the lab=
.
>
> >> Yes, this is a demonstration only, not giving enough information to
> >> build the thing yourself. It will irk you. The source code is shown, o=
n
> >> a barely readable video screen, in the unfathomable colorforth languag=
e.
> >> Still *I* think it is an interesting video.
>
> >> <SNIP>
>
> >>>What type of resonator is demonstrated? =A0What frequency? =A0One of t=
he
> >>>problems encountered before was how to start a high Q oscillator at a
> >>>higher frequency, I believe 10 MHz. =A0There is not enough resolution =
in a
> >>>700 MIPS instruction stream to tune to the crystal rate to get it goin=
g.
>
> >> I take exception to this. It depends on the crystal frequency. I inten=
d
> >> to experiment with a 32kHz clock crystal.
>
> >You may be able to make that work at 32kHz. =A0To verify that what you d=
o
> >will actually work correctly in the long run you not only need to be goo=
d
> >at analog circuitry and at software, but you'll need the cooperation of
> >the manufacturer -- and given their capabilities in that direction, I
> >don't think you'll ever get there.
>
> Oh, I certainly agree with you there. Green Arrays consists of senior
> citizens with very little in the direction of marketing potential.
> I'm purely interested in the R&D aspects.
> (Members of) the Dutch Forth chapter are experimenting to get a decent
> development environment for the chip. We do get some cooperation from
> Green Arrays.
>
>
>
> >--
> >Tim Wescott
>
> Groetjes Albert
>
> --
> --
> Albert van der Horst, UTRECHT,THE NETHERLANDS
> Economic growth -- being exponential -- ultimately falters.
> albert@spe&ar&c.xs4all.nl &=3Dnhttp://home.hccnet.nl/a.w.m.van.der.horst


What exactly is "a decent development environment"?  Is the eval board
lacking in some way?  Is the software not good enough?

I am planning to build (not yet committed 100%) a beagle board
daughter card using a GA144.  I want to design it so it can be a
daughter card but it can also be a stand alone card or use the other
Beagle board daughter cards.  Would a board like this be of any use?
I do plan to add oscillators rather than try to get the GA144 to drive
a crystal using only one pin.

Rick

In article <112325d8-b554-4399-b0b2-a9824002ebce@ec4g2000vbb.googlegroups.com>,
rickman  <gnuarm@gmail.com> wrote:
>
>
>What exactly is "a decent development environment"?  Is the eval board
>lacking in some way?  Is the software not good enough?

I can't pin-point it exactly, but the operational definition would be
that it allows someone to switch from a Microchip environment to a
GA144 environment with a comparable amount of effort to switch to an
ARM. Now the colorforth hurdle is very steep.
An absolute requirement from my part would be the possibility to
use a code management system, source that can be communicated
as ascii file via a mail attachment.

>
>I am planning to build (not yet committed 100%) a beagle board
>daughter card using a GA144.  I want to design it so it can be a
>daughter card but it can also be a stand alone card or use the other
>Beagle board daughter cards.  Would a board like this be of any use?

What is a beagle board?

The target application for GA144 is in the realm
of a 3 by 3 cm solar cell powered MP3 player. A prototype for such
a beast can't be made with the current development environment.
(The wow would be the smallness and lack of even a backup battery,
light enough to bungle from your head phone. That would be
lost with the current development board.) A board accomodating that
more or less would be nice.

>I do plan to add oscillators rather than try to get the GA144 to drive
>a crystal using only one pin.

Probably wise, but you're moving away from ultra-low power then.

>
>Rick

Groetjes Albert

--
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

On May 17, 1:08=A0pm, Albert van der Horst <alb...@spenarnc.xs4all.nl>
wrote:
> In article <112325d8-b554-4399-b0b2-a9824002e...@ec4g2000vbb.googlegroups=
.com>,
>
> rickman =A0<gnu...@gmail.com> wrote:
>
> >What exactly is "a decent development environment"? =A0Is the eval board
> >lacking in some way? =A0Is the software not good enough?
>
> I can't pin-point it exactly, but the operational definition would be
> that it allows someone to switch from a Microchip environment to a
> GA144 environment with a comparable amount of effort to switch to an
> ARM. Now the colorforth hurdle is very steep.
> An absolute requirement from my part would be the possibility to
> use a code management system, source that can be communicated
> as ascii file via a mail attachment.

So you want ascii files.  What else is wrong with the development
system?


> >I am planning to build (not yet committed 100%) a beagle board
> >daughter card using a GA144. =A0I want to design it so it can be a
> >daughter card but it can also be a stand alone card or use the other
> >Beagle board daughter cards. =A0Would a board like this be of any use?
>
> What is a beagle board?

Beagle Board (BB) is an open source project using the TI high end ARM
processors on a small, low cost development platform.  There are
several other similar projects, but from what I see the BB is the most
popular.  They have a version which is a square board, around 3 inches
on a side (10 cm) and a newer version is the Beaglebone (3.4" x 2.15"
or 86 mm x 55 mm) which skips the video output and is more useful for
embedded apps.  Both have Ethernet and USB as well as an expansion
bus.  They are very similar to the raspberryPi boards which attracted
so much attention a month or two ago.  The difference is you can buy
the BB.

My goal is to leverage the popularity of the BB (as much as a GA144
board can) as well as leverage some of the peripherals.  So this board
will be form factor and I/O pin compatible with the Beaglebone most
likely.  They call the BB daughter boards "cape" boards, a throwback
to the Arduino shield boards.  So some BB users would buy this board
to add I/O capabilities to the BB,  To use the GA144 more fully they
can write forth code for it.  Others may be interested in using this
board standalone as a much lower power replacement for the BB itself.
Using it standalone the standard BB daughter boards can be
supported.


> The target application for GA144 is in the realm
> of a 3 by 3 cm solar cell powered MP3 player. A prototype for such
> a beast can't be made with the current development environment.
> (The wow would be the smallness and lack of even a backup battery,
> light enough to bungle from your head phone. That would be
> lost with the current development board.) A board accomodating that
> more or less would be nice.

It is highly unlikely that any eval board you find will be form factor
compatible with your project.  Schmartboards makes an adapter board
for the GA144 which you might be able to squeeze into your project
size.


> >I do plan to add oscillators rather than try to get the GA144 to drive
> >a crystal using only one pin.
>
> Probably wise, but you're moving away from ultra-low power then.

I looked into this and found any number of low power commercial
oscillators.  I've been away from the project for awhile (saving the
US) and don't recall any companies, but they are out there.  Since GA
never explained how to actually build a single pin oscillator I don't
see how you can say their solution would be any lower power than any
other.  That's the big problem with GA, to this day, nearly a year
after announcing their chips could be bought, they still don't have
any real info on how to use them on the hardware level.

Rick