Micro controllers with UHF transceivers?

On Oct 25, 2:05 pm, larwe <zwsdot...@gmail.com> wrote:

> > I checked Atmel but among the available chip found only transmitters on
>
> Eh? Did you find your way to this page: <http://www.atmel.com/dyn/
> products/devices.asp?family_id=651>

Oh, wait a moment, never mind :)

larwe wrote:
> On Oct 25, 12:08 am, Jim Granville <no.s...@designtools.maps.co.nz>
> wrote:
> 
> 
>>There is no reason to re-spin proven die, to swap-out one core,
>>and put another in ?!.
> 
> 
> Suppose I told you that I /know/ it is being done, I'm not merely
> theorizing? The interest is that MSP430 is low-power-optimized and a
> lot of applications for the Chipcon standalone txcvr parts are battery-
> powered.

So you mean some NEW designs are being done with MSP430 ?

That makes sense, most likely they have MSP430 users demand,
(more than any technical reason) - if I was a MSP430 user,
I'd be asking the same thing.

That does not mean the existing devices are going to
vanish any time soon.
(Indeed, if Ti are smart, the new devices will be pin-compatible.)

We still have production design using ATF16V8BQL, and I see TI
STILL make 22V10's - those are both what could be described as
very much trailing-edge devices, but still in active production.

Atmel has done a similar 'common peripheral' swap between their
ARM7/ARM9/AVR32 cores, and all those devices can expect long design lives.


> I also believe those old parts are made on a fairly ancient process
> which TI would like to retire, but that latter comment is just
> hearsay.

Was TI the foundry for Chipcon, or did chipcon use a more
conventional foundry ?
Normally only when the volumes fall, do suppliers retire a part.

-jg

larwe wrote:
> On Oct 25, 11:46 am, Joerg <notthisjoerg...@removethispacbell.net>
> wrote:
> 
>> You think Chipcon might dry up after the aquisition? That wouldn't be
> 
> I think that the MSP430 with on-chip RF is going to replace the 51
> parts.
> 

But when? Waiting since over a year now ...

Thing is, my clients need designs now, not next year.


>> would be MSP430 with RF on board. But none so far. What's wrong with a
>> 51 core? Tons of products still use it, including almost anything
> 
> Sure, but the target markets for much of Chipcon's stuff would benefit
> from an MSP430 core.
> 

If it comes to market.


>> 433MHz band. Here, the FCC limits burst mode tricks to 20dB net power
>> advantage AFAIK.
> 
> Depends on the application. In my day job I work a little lower down
> in the band, ASK, low duty cycle.
> 
>> I checked Atmel but among the available chip found only transmitters on
> 
> Eh? Did you find your way to this page: <http://www.atmel.com/dyn/
> products/devices.asp?family_id=651>
> 

Says "Directory Listing Denied" :-(

-- 
Regards, Joerg

http://www.analogconsultants.com/

Jim Granville wrote:
> larwe wrote:
>> On Oct 25, 12:08 am, Jim Granville <no.s...@designtools.maps.co.nz>
>> wrote:
>>
>>
>>> There is no reason to re-spin proven die, to swap-out one core,
>>> and put another in ?!.
>>
>>
>> Suppose I told you that I /know/ it is being done, I'm not merely
>> theorizing? The interest is that MSP430 is low-power-optimized and a
>> lot of applications for the Chipcon standalone txcvr parts are battery-
>> powered.
> 
> So you mean some NEW designs are being done with MSP430 ?
> 

Not RF-on-chip designs. There ain't no chips :-(

[...]

-- 
Regards, Joerg

http://www.analogconsultants.com/

> So you mean some NEW designs are being done with MSP430 ?

Yes.

> That does not mean the existing devices are going to
> vanish any time soon.
> (Indeed, if Ti are smart, the new devices will be pin-compatible.)

They definitely won't based on the discussions I've had. The first
part will be an all-singing-all-dancing part [i.e. someone's ASIC sold
to the general market]. Other versions will follow. I'd expect to see
early silicon at the end of next year.

> Was TI the foundry for Chipcon, or did chipcon use a more
> conventional foundry ?

I'm not sure, didn't follow the relationship that closely.

On Wed, 24 Oct 2007 13:34:29 -0700, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

>Hello Folks,
>
>After some Google searching and perusing the sites of the usual 
>contenders I only found one uC family that has serious on-chip RF 
>transceiver capabilities, the Cypress CYWUSB6953 and its brethren. 
>rfPICs and others usually only have a transmitter.
>
>Anyhow, the Cypress will only serve 2.45GHz but I need the lower UHF 
>bands for range reasons. Is anything coming down the pike soon or will 
>that have to remain a two-chip solution?

It would help if you had specified the country in which it should
operate. The frequency bands, the power levels and duty cycles vary
with country or at least with continent.

You said that the range was too short on 2.45 GHz. What kind of
propagation environment do you have ? Free space, lot of thick trees,
indoor or what ? 

In a free space environment, the capture area of an omnidirectional
receiver antenna at a higher frequency can get quite small, thus the
receiver signal will be weaker. Also a lot of wet trees will attenuate
the 2.45 GHz signal. 

On the other hand, in typical indoor situations and at the streets of
a large city, there are going to be plenty of reflections and the
propagation is more or less independent of the frequency. The shorter
wavelength might even propagate more easily through narrow slits in
air conditioning ducts etc.

Paul

Paul Keinanen wrote:
> On Wed, 24 Oct 2007 13:34:29 -0700, Joerg
> <notthisjoergsch@removethispacbell.net> wrote:
> 
>> Hello Folks,
>>
>> After some Google searching and perusing the sites of the usual 
>> contenders I only found one uC family that has serious on-chip RF 
>> transceiver capabilities, the Cypress CYWUSB6953 and its brethren. 
>> rfPICs and others usually only have a transmitter.
>>
>> Anyhow, the Cypress will only serve 2.45GHz but I need the lower UHF 
>> bands for range reasons. Is anything coming down the pike soon or will 
>> that have to remain a two-chip solution?
> 
> It would help if you had specified the country in which it should
> operate. The frequency bands, the power levels and duty cycles vary
> with country or at least with continent.
> 

As many countries as possible. Definitely US, Canada, Europe and in and 
around the Gulf of Mexico, other cases also Asia. That leaves 433MHz and 
2.45GHz. Or frequencies below 100MHz but there aren't any integrated 
solutions for those.


> You said that the range was too short on 2.45 GHz. What kind of
> propagation environment do you have ? Free space, lot of thick trees,
> indoor or what ? 
> 

All of the above, usually comms between a pod inside the house to 
equipment outside. I guess our house would be the worst case, woodframe 
construction with aluminum backed insulation inside the walls.


> In a free space environment, the capture area of an omnidirectional
> receiver antenna at a higher frequency can get quite small, thus the
> receiver signal will be weaker. Also a lot of wet trees will attenuate
> the 2.45 GHz signal. 
> 
> On the other hand, in typical indoor situations and at the streets of
> a large city, there are going to be plenty of reflections and the
> propagation is more or less independent of the frequency. The shorter
> wavelength might even propagate more easily through narrow slits in
> air conditioning ducts etc.
> 

All I can say that around our house 2.45GHz does not work reliably at 
all. <200MHz works excellent.

-- 
Regards, Joerg

http://www.analogconsultants.com/

Joerg wrote:

> Jim Granville wrote:
> 
>> larwe wrote:
>>
>>> On Oct 25, 12:08 am, Jim Granville <no.s...@designtools.maps.co.nz>
>>> wrote:
>>>
>>>
>>>> There is no reason to re-spin proven die, to swap-out one core,
>>>> and put another in ?!.
>>>
>>>
>>>
>>> Suppose I told you that I /know/ it is being done, I'm not merely
>>> theorizing? The interest is that MSP430 is low-power-optimized and a
>>> lot of applications for the Chipcon standalone txcvr parts are battery-
>>> powered.
>>
>>
>> So you mean some NEW designs are being done with MSP430 ?
>>
> 
> Not RF-on-chip designs. There ain't no chips :-(

Always good to design with what you can get :)

It takes a while to move from "new designs being done" to have
actual production silicon in your hands.
In another post Lewin mentins end 2008 for early silicon, being
derived from someones ASIC.

A portion of the delay is engineering, and a portion is likely
commercial, as the ASIC customer negotiates a 'head start'.

Indeed 'someones ASIC' is a higher risk design path, as when that
large single customer moves, the volumes plummet, and that's bad news
for the smaller users. Better for smaller users to back parts that
have a wide user base, and moderate, but steady volumes.

Low power and fast wakeup, are not unique to the MSP430, and
Silabs have new '51 devices with very fast wakeups,
and very low powers.
Even the present Chipcon parts spec sub uA idle powers,
and operate is dominated by RF power, not core power.

So there is little technical rationale for a core change.

-jg

On Oct 25, 4:51 pm, Jim Granville <no.s...@designtools.maps.co.nz>
wrote:

> In another post Lewin mentins end 2008 for early silicon, being
> derived from someones ASIC.

That comment should be taken with the appropriate sample of ionic
compound :) It would probably be more accurate to say it's an ASSP
which is being designed to meet some specific customer's requirements
(not us), and it will be made generally available.

That end 2008 figure is also not an official figure from TI, it's just
the feeling I get from the last time I talked with them about it.

larwe wrote:
> On Oct 25, 4:51 pm, Jim Granville <no.s...@designtools.maps.co.nz>
> wrote:
> 
> 
>>In another post Lewin mentins end 2008 for early silicon, being
>>derived from someones ASIC.
> 
> 
> That comment should be taken with the appropriate sample of ionic
> compound :) It would probably be more accurate to say it's an ASSP
> which is being designed to meet some specific customer's requirements
> (not us), and it will be made generally available.
> 
> That end 2008 figure is also not an official figure from TI, it's just
> the feeling I get from the last time I talked with them about it.

Understood, that's a fairly common industry pathway. eg Atmel have a few 
devices like that.
The Plus is you can get a market-tester-part easily, the minus is they 
often lack general use features, and struggle for eco-system critical mass.

-jg