dbr@kbrx.com wrote:> Check www.rfm.com for "trc102 range test at 433.92mhz". It's an article > discussing low power, ie no certification required, operation. If I > remember correctly, the trc102 is no longer available, but the text is > educational. >I'll second that (the notion to look at RFM). My experience with them has been good so far, and they offer pre-certified solutions. Here's one that is still available: http://www.rfm.com/products/spec_sheet.php?fhssmulti-functionconfigurablemodulesforwirelesssensornetworking_&record=DNT90C I don't know this one personally as we used its higher powered brother because of range requirements but it's probably quite similar. Good documentation, works right out of the box. Oh, and I put s.e.design back in the address field so the thread doesn't break. [...] -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Wireless that's "Fall off a log" easy?
Started by ●October 11, 2010
Reply by ●October 14, 20102010-10-14
Reply by ●October 16, 20102010-10-16
On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote:> Are there any wireless sub-systems that are just "fall off a log" easy > to integrate into an embedded system? > > The ideal gizmo would go into an OEM product, wouldn't require the OEM > to mess with any certification issues, would make the product appear as > a serial device to a host computer via BlueTooth, would talk to it's > local embedded processor via serial, and would be plug and play. > > The higher the data rate the better, up to about 100kbps or so. > > Any such things out there? Any that you have personal mileage with and > like or particularly dislike? I'm being lazy and haven't even looked at > ads or done any web searches.Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html I haven't used those modules myself. I've used some of their AM stuff in the past (albeit for experimentation rather than in anger) but that is only good for 2400 baud. This is good for 115200, SPI interface and $7.35 per unit from Digi-key. No Bluetooth though. -- Andrew Smallshaw andrews@sdf.lonestar.org
Reply by ●October 16, 20102010-10-16
Andrew Smallshaw wrote:> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >> Are there any wireless sub-systems that are just "fall off a log" easy >> to integrate into an embedded system? >> >> The ideal gizmo would go into an OEM product, wouldn't require the OEM >> to mess with any certification issues, would make the product appear as >> a serial device to a host computer via BlueTooth, would talk to it's >> local embedded processor via serial, and would be plug and play. >> >> The higher the data rate the better, up to about 100kbps or so. >> >> Any such things out there? Any that you have personal mileage with and >> like or particularly dislike? I'm being lazy and haven't even looked at >> ads or done any web searches. > > Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html > > I haven't used those modules myself. I've used some of their AM > stuff in the past (albeit for experimentation rather than in anger) > but that is only good for 2400 baud. This is good for 115200, SPI > interface and $7.35 per unit from Digi-key. No Bluetooth though. >You mean this one? http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf Couldn't find the words FCC or "pre-certified" in there, so be careful. Also, I found that any schemes that aren't frequency-agile are not all that reliable. Especially with rather vulnerable AM protocols. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Reply by ●October 16, 20102010-10-16
On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote:>Andrew Smallshaw wrote: >> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>> Are there any wireless sub-systems that are just "fall off a log" easy >>> to integrate into an embedded system? >>> >>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>> to mess with any certification issues, would make the product appear as >>> a serial device to a host computer via BlueTooth, would talk to it's >>> local embedded processor via serial, and would be plug and play. >>> >>> The higher the data rate the better, up to about 100kbps or so. >>> >>> Any such things out there? Any that you have personal mileage with and >>> like or particularly dislike? I'm being lazy and haven't even looked at >>> ads or done any web searches. >> >> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >> >> I haven't used those modules myself. I've used some of their AM >> stuff in the past (albeit for experimentation rather than in anger) >> but that is only good for 2400 baud. This is good for 115200, SPI >> interface and $7.35 per unit from Digi-key. No Bluetooth though. >> > >You mean this one? > >http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf > >Couldn't find the words FCC or "pre-certified" in there, so be careful. >Also, I found that any schemes that aren't frequency-agile are not all >that reliable. Especially with rather vulnerable AM protocols.Just curious... What's the worry about certification? Certifying an intentional radiator costs a little, but not all that much more than certifying an unintentional radiator (containing a certified module). If our lab's interpretation of the (FCC) regs is correct, the intentional radiator covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but it is an interesting kink in the regs.
Reply by ●October 16, 20102010-10-16
krw@att.bizzzzzzzzzzzz wrote:> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: > >> Andrew Smallshaw wrote: >>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>> to integrate into an embedded system? >>>> >>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>> to mess with any certification issues, would make the product appear as >>>> a serial device to a host computer via BlueTooth, would talk to it's >>>> local embedded processor via serial, and would be plug and play. >>>> >>>> The higher the data rate the better, up to about 100kbps or so. >>>> >>>> Any such things out there? Any that you have personal mileage with and >>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>> ads or done any web searches. >>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>> >>> I haven't used those modules myself. I've used some of their AM >>> stuff in the past (albeit for experimentation rather than in anger) >>> but that is only good for 2400 baud. This is good for 115200, SPI >>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>> >> You mean this one? >> >> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >> >> Couldn't find the words FCC or "pre-certified" in there, so be careful. >> Also, I found that any schemes that aren't frequency-agile are not all >> that reliable. Especially with rather vulnerable AM protocols. > > Just curious... What's the worry about certification? Certifying an > intentional radiator costs a little, but not all that much more than > certifying an unintentional radiator (containing a certified module). If our > lab's interpretation of the (FCC) regs is correct, the intentional radiator > covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but > it is an interesting kink in the regs. >The FCC has made the process fairly friendly lately but last time the cost was another $5k, AFAIR. Plus you must use a TCB certified lab which may mean a longer drive, and possibly higher fees for the regular EMC job. Some companies just ship the stuff, without one of their engineer witnessing the testing (I strongly discourage clients from doing that). If you build thousands of units a year it certainly makes sense. Of course, you also have to factor in and amortize your own design work on the RF portions of the project that would otherwise already have been handled by the module vendor. That often gets overlooked. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Reply by ●October 16, 20102010-10-16
On Sat, 16 Oct 2010 12:57:04 -0700, Joerg <invalid@invalid.invalid> wrote:>krw@att.bizzzzzzzzzzzz wrote: >> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: >> >>> Andrew Smallshaw wrote: >>>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>>> to integrate into an embedded system? >>>>> >>>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>>> to mess with any certification issues, would make the product appear as >>>>> a serial device to a host computer via BlueTooth, would talk to it's >>>>> local embedded processor via serial, and would be plug and play. >>>>> >>>>> The higher the data rate the better, up to about 100kbps or so. >>>>> >>>>> Any such things out there? Any that you have personal mileage with and >>>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>>> ads or done any web searches. >>>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>>> >>>> I haven't used those modules myself. I've used some of their AM >>>> stuff in the past (albeit for experimentation rather than in anger) >>>> but that is only good for 2400 baud. This is good for 115200, SPI >>>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>>> >>> You mean this one? >>> >>> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >>> >>> Couldn't find the words FCC or "pre-certified" in there, so be careful. >>> Also, I found that any schemes that aren't frequency-agile are not all >>> that reliable. Especially with rather vulnerable AM protocols. >> >> Just curious... What's the worry about certification? Certifying an >> intentional radiator costs a little, but not all that much more than >> certifying an unintentional radiator (containing a certified module). If our >> lab's interpretation of the (FCC) regs is correct, the intentional radiator >> covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but >> it is an interesting kink in the regs. >> > >The FCC has made the process fairly friendly lately but last time the >cost was another $5k, AFAIR. Plus you must use a TCB certified lab which >may mean a longer drive, and possibly higher fees for the regular EMC >job. Some companies just ship the stuff, without one of their engineer >witnessing the testing (I strongly discourage clients from doing that).It costs us about $2K (maybe a little less) to have the scans run for unintentional radiation, assuming it passes the first shot. We had a nasty experience were it took many attempts and several weeks. $5K would have been a bargain. That disaster was at least half the fault of the testing lab, though, so we try to not go there anymore. It is silly to just ship stuff off to the lab. If anything goes wrong (it always does) there is no one there to fix it. I'm the one who generally goes. I kinda like Atlanta. ;-)>If you build thousands of units a year it certainly makes sense. Of >course, you also have to factor in and amortize your own design work on >the RF portions of the project that would otherwise already have been >handled by the module vendor. That often gets overlooked.I'm not talking about designing the RF section, rather only the certification part. specifically, whether to use the RF module's certification and then certify your box as an unintentional radiator or certifying the entire box as an intentional radiator.
Reply by ●October 17, 20102010-10-17
krw@att.bizzzzzzzzzzzz wrote:> On Sat, 16 Oct 2010 12:57:04 -0700, Joerg <invalid@invalid.invalid> wrote: > >> krw@att.bizzzzzzzzzzzz wrote: >>> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: >>> >>>> Andrew Smallshaw wrote: >>>>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>>>> to integrate into an embedded system? >>>>>> >>>>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>>>> to mess with any certification issues, would make the product appear as >>>>>> a serial device to a host computer via BlueTooth, would talk to it's >>>>>> local embedded processor via serial, and would be plug and play. >>>>>> >>>>>> The higher the data rate the better, up to about 100kbps or so. >>>>>> >>>>>> Any such things out there? Any that you have personal mileage with and >>>>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>>>> ads or done any web searches. >>>>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>>>> >>>>> I haven't used those modules myself. I've used some of their AM >>>>> stuff in the past (albeit for experimentation rather than in anger) >>>>> but that is only good for 2400 baud. This is good for 115200, SPI >>>>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>>>> >>>> You mean this one? >>>> >>>> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >>>> >>>> Couldn't find the words FCC or "pre-certified" in there, so be careful. >>>> Also, I found that any schemes that aren't frequency-agile are not all >>>> that reliable. Especially with rather vulnerable AM protocols. >>> Just curious... What's the worry about certification? Certifying an >>> intentional radiator costs a little, but not all that much more than >>> certifying an unintentional radiator (containing a certified module). If our >>> lab's interpretation of the (FCC) regs is correct, the intentional radiator >>> covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but >>> it is an interesting kink in the regs. >>> >> The FCC has made the process fairly friendly lately but last time the >> cost was another $5k, AFAIR. Plus you must use a TCB certified lab which >> may mean a longer drive, and possibly higher fees for the regular EMC >> job. Some companies just ship the stuff, without one of their engineer >> witnessing the testing (I strongly discourage clients from doing that). > > It costs us about $2K (maybe a little less) to have the scans run for > unintentional radiation, assuming it passes the first shot. ...Well, but that's in Jaw-jah, not Taxafornia :-)> ... We had a nasty > experience were it took many attempts and several weeks. $5K would have been > a bargain. That disaster was at least half the fault of the testing lab, > though, so we try to not go there anymore. >That's why I urge clients to send an experienced engineer along _and_ do pre-compliance checks inhouse.> It is silly to just ship stuff off to the lab. If anything goes wrong (it > always does) there is no one there to fix it. I'm the one who generally goes. > I kinda like Atlanta. ;-) > >> If you build thousands of units a year it certainly makes sense. Of >> course, you also have to factor in and amortize your own design work on >> the RF portions of the project that would otherwise already have been >> handled by the module vendor. That often gets overlooked. > > I'm not talking about designing the RF section, rather only the certification > part. specifically, whether to use the RF module's certification and then > certify your box as an unintentional radiator or certifying the entire box as > an intentional radiator.For small qty stuff (meaning not millions/year) it's almost always better to use the module's cert and slap that extra sticker from its mfg onto the client's box. Only one more sticker, next to all the other mandatory ones like "Do not put in mouth" and all that :-) -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Reply by ●October 17, 20102010-10-17
On Sun, 17 Oct 2010 17:14:42 -0700, Joerg <invalid@invalid.invalid> wrote:>krw@att.bizzzzzzzzzzzz wrote: >> On Sat, 16 Oct 2010 12:57:04 -0700, Joerg <invalid@invalid.invalid> wrote: >> >>> krw@att.bizzzzzzzzzzzz wrote: >>>> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: >>>> >>>>> Andrew Smallshaw wrote: >>>>>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>>>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>>>>> to integrate into an embedded system? >>>>>>> >>>>>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>>>>> to mess with any certification issues, would make the product appear as >>>>>>> a serial device to a host computer via BlueTooth, would talk to it's >>>>>>> local embedded processor via serial, and would be plug and play. >>>>>>> >>>>>>> The higher the data rate the better, up to about 100kbps or so. >>>>>>> >>>>>>> Any such things out there? Any that you have personal mileage with and >>>>>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>>>>> ads or done any web searches. >>>>>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>>>>> >>>>>> I haven't used those modules myself. I've used some of their AM >>>>>> stuff in the past (albeit for experimentation rather than in anger) >>>>>> but that is only good for 2400 baud. This is good for 115200, SPI >>>>>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>>>>> >>>>> You mean this one? >>>>> >>>>> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >>>>> >>>>> Couldn't find the words FCC or "pre-certified" in there, so be careful. >>>>> Also, I found that any schemes that aren't frequency-agile are not all >>>>> that reliable. Especially with rather vulnerable AM protocols. >>>> Just curious... What's the worry about certification? Certifying an >>>> intentional radiator costs a little, but not all that much more than >>>> certifying an unintentional radiator (containing a certified module). If our >>>> lab's interpretation of the (FCC) regs is correct, the intentional radiator >>>> covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but >>>> it is an interesting kink in the regs. >>>> >>> The FCC has made the process fairly friendly lately but last time the >>> cost was another $5k, AFAIR. Plus you must use a TCB certified lab which >>> may mean a longer drive, and possibly higher fees for the regular EMC >>> job. Some companies just ship the stuff, without one of their engineer >>> witnessing the testing (I strongly discourage clients from doing that). >> >> It costs us about $2K (maybe a little less) to have the scans run for >> unintentional radiation, assuming it passes the first shot. ... > > >Well, but that's in Jaw-jah, not Taxafornia :-)Same companies. The unintentional scans (line and radiated) take about four hours, which translates to $1600 in the 10m chamber (*much* more reliable than a slightly less expensive 3m chamber).>> ... We had a nasty >> experience were it took many attempts and several weeks. $5K would have been >> a bargain. That disaster was at least half the fault of the testing lab, >> though, so we try to not go there anymore. >> > >That's why I urge clients to send an experienced engineer along _and_ do >pre-compliance checks inhouse.In our case, these were useless. With our equipment I can't even get a decent sniff.>> It is silly to just ship stuff off to the lab. If anything goes wrong (it >> always does) there is no one there to fix it. I'm the one who generally goes. >> I kinda like Atlanta. ;-) >> >>> If you build thousands of units a year it certainly makes sense. Of >>> course, you also have to factor in and amortize your own design work on >>> the RF portions of the project that would otherwise already have been >>> handled by the module vendor. That often gets overlooked. >> >> I'm not talking about designing the RF section, rather only the certification >> part. specifically, whether to use the RF module's certification and then >> certify your box as an unintentional radiator or certifying the entire box as >> an intentional radiator. > > >For small qty stuff (meaning not millions/year) it's almost always >better to use the module's cert and slap that extra sticker from its mfg >onto the client's box. Only one more sticker, next to all the other >mandatory ones like "Do not put in mouth" and all that :-)But the module's cert doesn't certify the whole box, unless there isn't anything else in the box.
Reply by ●October 18, 20102010-10-18
krw@att.bizzzzzzzzzzzz wrote:> On Sun, 17 Oct 2010 17:14:42 -0700, Joerg <invalid@invalid.invalid> wrote: > >> krw@att.bizzzzzzzzzzzz wrote: >>> On Sat, 16 Oct 2010 12:57:04 -0700, Joerg <invalid@invalid.invalid> wrote: >>> >>>> krw@att.bizzzzzzzzzzzz wrote: >>>>> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: >>>>> >>>>>> Andrew Smallshaw wrote: >>>>>>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>>>>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>>>>>> to integrate into an embedded system? >>>>>>>> >>>>>>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>>>>>> to mess with any certification issues, would make the product appear as >>>>>>>> a serial device to a host computer via BlueTooth, would talk to it's >>>>>>>> local embedded processor via serial, and would be plug and play. >>>>>>>> >>>>>>>> The higher the data rate the better, up to about 100kbps or so. >>>>>>>> >>>>>>>> Any such things out there? Any that you have personal mileage with and >>>>>>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>>>>>> ads or done any web searches. >>>>>>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>>>>>> >>>>>>> I haven't used those modules myself. I've used some of their AM >>>>>>> stuff in the past (albeit for experimentation rather than in anger) >>>>>>> but that is only good for 2400 baud. This is good for 115200, SPI >>>>>>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>>>>>> >>>>>> You mean this one? >>>>>> >>>>>> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >>>>>> >>>>>> Couldn't find the words FCC or "pre-certified" in there, so be careful. >>>>>> Also, I found that any schemes that aren't frequency-agile are not all >>>>>> that reliable. Especially with rather vulnerable AM protocols. >>>>> Just curious... What's the worry about certification? Certifying an >>>>> intentional radiator costs a little, but not all that much more than >>>>> certifying an unintentional radiator (containing a certified module). If our >>>>> lab's interpretation of the (FCC) regs is correct, the intentional radiator >>>>> covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but >>>>> it is an interesting kink in the regs. >>>>> >>>> The FCC has made the process fairly friendly lately but last time the >>>> cost was another $5k, AFAIR. Plus you must use a TCB certified lab which >>>> may mean a longer drive, and possibly higher fees for the regular EMC >>>> job. Some companies just ship the stuff, without one of their engineer >>>> witnessing the testing (I strongly discourage clients from doing that). >>> It costs us about $2K (maybe a little less) to have the scans run for >>> unintentional radiation, assuming it passes the first shot. ... >> >> Well, but that's in Jaw-jah, not Taxafornia :-) > > Same companies. The unintentional scans (line and radiated) take about four > hours, which translates to $1600 in the 10m chamber (*much* more reliable than > a slightly less expensive 3m chamber). >But then all the high ones need to be worked of out in the range, mostly using a receiver. Which adds hours. That's why the winter rule is BYOJ (bring your own jacket) :-)>>> ... We had a nasty >>> experience were it took many attempts and several weeks. $5K would have been >>> a bargain. That disaster was at least half the fault of the testing lab, >>> though, so we try to not go there anymore. >>> >> That's why I urge clients to send an experienced engineer along _and_ do >> pre-compliance checks inhouse. > > In our case, these were useless. With our equipment I can't even get a decent > sniff. >Mostly I just use a computer-controlled scanner. If there is a nasty emission somewhere it'll find it. Sure, I can't quantify it here in the lab but after a couple of decades on the beat you know what'll cause problems, just like an older cop can look a guy in the eyes and tell that "he dunnit".>>> It is silly to just ship stuff off to the lab. If anything goes wrong (it >>> always does) there is no one there to fix it. I'm the one who generally goes. >>> I kinda like Atlanta. ;-) >>> >>>> If you build thousands of units a year it certainly makes sense. Of >>>> course, you also have to factor in and amortize your own design work on >>>> the RF portions of the project that would otherwise already have been >>>> handled by the module vendor. That often gets overlooked. >>> I'm not talking about designing the RF section, rather only the certification >>> part. specifically, whether to use the RF module's certification and then >>> certify your box as an unintentional radiator or certifying the entire box as >>> an intentional radiator. >> >> For small qty stuff (meaning not millions/year) it's almost always >> better to use the module's cert and slap that extra sticker from its mfg >> onto the client's box. Only one more sticker, next to all the other >> mandatory ones like "Do not put in mouth" and all that :-) > > But the module's cert doesn't certify the whole box, unless there isn't > anything else in the box.For the intentional part it usually does, the mfg would tell you that (and supply the cert/sticker). Unless you do some unauthorized mods. The unintentional part needs to be tested just like it is with non-RF gear. If the mfg waffles on the stickers and cert copies, run. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Reply by ●October 18, 20102010-10-18
On Mon, 18 Oct 2010 11:24:11 -0700, Joerg <invalid@invalid.invalid> wrote:>krw@att.bizzzzzzzzzzzz wrote: >> On Sun, 17 Oct 2010 17:14:42 -0700, Joerg <invalid@invalid.invalid> wrote: >> >>> krw@att.bizzzzzzzzzzzz wrote: >>>> On Sat, 16 Oct 2010 12:57:04 -0700, Joerg <invalid@invalid.invalid> wrote: >>>> >>>>> krw@att.bizzzzzzzzzzzz wrote: >>>>>> On Sat, 16 Oct 2010 08:14:50 -0700, Joerg <invalid@invalid.invalid> wrote: >>>>>> >>>>>>> Andrew Smallshaw wrote: >>>>>>>> On 2010-10-11, Tim Wescott <tim@seemywebsite.com> wrote: >>>>>>>>> Are there any wireless sub-systems that are just "fall off a log" easy >>>>>>>>> to integrate into an embedded system? >>>>>>>>> >>>>>>>>> The ideal gizmo would go into an OEM product, wouldn't require the OEM >>>>>>>>> to mess with any certification issues, would make the product appear as >>>>>>>>> a serial device to a host computer via BlueTooth, would talk to it's >>>>>>>>> local embedded processor via serial, and would be plug and play. >>>>>>>>> >>>>>>>>> The higher the data rate the better, up to about 100kbps or so. >>>>>>>>> >>>>>>>>> Any such things out there? Any that you have personal mileage with and >>>>>>>>> like or particularly dislike? I'm being lazy and haven't even looked at >>>>>>>>> ads or done any web searches. >>>>>>>> Have a look at http://www.quasaruk.co.uk/acatalog/ALPHA_TRX_Module.html >>>>>>>> >>>>>>>> I haven't used those modules myself. I've used some of their AM >>>>>>>> stuff in the past (albeit for experimentation rather than in anger) >>>>>>>> but that is only good for 2400 baud. This is good for 115200, SPI >>>>>>>> interface and $7.35 per unit from Digi-key. No Bluetooth though. >>>>>>>> >>>>>>> You mean this one? >>>>>>> >>>>>>> http://www.rfsolutions.co.uk/acatalog/DS-ALPHA-TRX-5.pdf >>>>>>> >>>>>>> Couldn't find the words FCC or "pre-certified" in there, so be careful. >>>>>>> Also, I found that any schemes that aren't frequency-agile are not all >>>>>>> that reliable. Especially with rather vulnerable AM protocols. >>>>>> Just curious... What's the worry about certification? Certifying an >>>>>> intentional radiator costs a little, but not all that much more than >>>>>> certifying an unintentional radiator (containing a certified module). If our >>>>>> lab's interpretation of the (FCC) regs is correct, the intentional radiator >>>>>> covers a *lot* of the sins of the rest of the box. BTW, we don't do this, but >>>>>> it is an interesting kink in the regs. >>>>>> >>>>> The FCC has made the process fairly friendly lately but last time the >>>>> cost was another $5k, AFAIR. Plus you must use a TCB certified lab which >>>>> may mean a longer drive, and possibly higher fees for the regular EMC >>>>> job. Some companies just ship the stuff, without one of their engineer >>>>> witnessing the testing (I strongly discourage clients from doing that). >>>> It costs us about $2K (maybe a little less) to have the scans run for >>>> unintentional radiation, assuming it passes the first shot. ... >>> >>> Well, but that's in Jaw-jah, not Taxafornia :-) >> >> Same companies. The unintentional scans (line and radiated) take about four >> hours, which translates to $1600 in the 10m chamber (*much* more reliable than >> a slightly less expensive 3m chamber). >> > >But then all the high ones need to be worked of out in the range, mostly >using a receiver. Which adds hours. That's why the winter rule is BYOJ >(bring your own jacket) :-)I don't follow.>>>> ... We had a nasty >>>> experience were it took many attempts and several weeks. $5K would have been >>>> a bargain. That disaster was at least half the fault of the testing lab, >>>> though, so we try to not go there anymore. >>>> >>> That's why I urge clients to send an experienced engineer along _and_ do >>> pre-compliance checks inhouse. >> >> In our case, these were useless. With our equipment I can't even get a decent >> sniff. >> > >Mostly I just use a computer-controlled scanner. If there is a nasty >emission somewhere it'll find it. Sure, I can't quantify it here in the >lab but after a couple of decades on the beat you know what'll cause >problems, just like an older cop can look a guy in the eyes and tell >that "he dunnit".Scanner? Does it cover the entire range (30mHz to a few GHz)?>>>> It is silly to just ship stuff off to the lab. If anything goes wrong (it >>>> always does) there is no one there to fix it. I'm the one who generally goes. >>>> I kinda like Atlanta. ;-) >>>> >>>>> If you build thousands of units a year it certainly makes sense. Of >>>>> course, you also have to factor in and amortize your own design work on >>>>> the RF portions of the project that would otherwise already have been >>>>> handled by the module vendor. That often gets overlooked. >>>> I'm not talking about designing the RF section, rather only the certification >>>> part. specifically, whether to use the RF module's certification and then >>>> certify your box as an unintentional radiator or certifying the entire box as >>>> an intentional radiator. >>> >>> For small qty stuff (meaning not millions/year) it's almost always >>> better to use the module's cert and slap that extra sticker from its mfg >>> onto the client's box. Only one more sticker, next to all the other >>> mandatory ones like "Do not put in mouth" and all that :-) >> >> But the module's cert doesn't certify the whole box, unless there isn't >> anything else in the box. > > >For the intentional part it usually does, the mfg would tell you that >(and supply the cert/sticker). Unless you do some unauthorized mods. The >unintentional part needs to be tested just like it is with non-RF gear.Understood, but my point is that the intentional limits are considerably higher than the unintentional, outside some forbidden zones, so it's far easier to pass. Actually, they're so lax that you'd have to *try* to fail them. ;-)>If the mfg waffles on the stickers and cert copies, run.That's not my point.
