translator question

"John W"

Here's the comment I saw that worried me:

"Sending large messages may be a problem if the CAV6.6 is timing sensitive in receiving commands since you'll need to concatenate multiple shorter ASCII Bobcat messages together. Adding to the challenge is that the command queue for the ASCII Bobcat doesn't appear to be buffered. If you send a second message to the Bobcat before the first has been sent, the second message will replace the first.

I've found that I need to wait two cycles through my C-Max code between sending ASCII Bobcat messages in order for each message to be reliably sent. You'll probably need to experiment to determine how many cycles are needed for your situation. Hooking the ASCII Bobcat up to a serial terminal can be an invaluable tool for determining the needed cycle count.

Hope this helps,

-Bill "

-- Bobby G.

I've never used the Bobcat. My experience was with the ASCII strings in the CPU-XA, Ocelot & Leopard. IIRC I was told they were added specifically to support HAI's ProLink ASCII protocol. They had a maximum length (32 chrs?) and were also useful with their modem.

I think the Bobcat has the same basic specs. The only added delay should be the time it takes to send the command (rather short) over ADNET (9600 bps) as the strings are stored in the Bobcat.

Two cycles through C-Max code is rather nebulous. It might be a few millisec>> "Robert Green" wrote:

The end of the thread indicated that the Bobcat suffered from the lost comamnd problem but the serial port built into the Ocelot was buffered and would not drop commands.

I may not be a programmer but I do remember lots of software that had to be reworked for faster PC's because of the liberal use of such timing loops. It was real problem with games and such that had worked on XTs and ATs but failed on much faster Pentium machines. I don't know enough about CMAX to know whether there's a better way to induce a delay, though, so it may be all they have.

-- Bobby G.

I have a circuit board layout that I think will work. I'll stare at it for a few more days. I usually find that causes traces and solder pads to move around mysteriously ending up in locations where I don't recall placing them. I'll send it off this weekend and should have it about 2-3 weeks after that. In the meantime I can start on the code using a BX24-AHT board

Mouser does not stock any through hole RTC chips so I will use SMT and will forget about using any of the supported serial devices for the RTC backup. This will save a bit of code since there will only be the one chip to interface with. I'll leave the choice of supercapacitor to the user. Mouser stocks 3 that will fit the space I've alloted. They will provide anywhere from 2-3 days to more than a month backup for the RTC chip.

I've settled on I²C for the RTC chip and will use the ZX-24's hardware I²C implementation which can run at 410kHz. The I²C bus can support another EEPROM chip should it prove necessary to increase the space for user data.

To use Rosetta with any of the controllers that normally use a TW523 (or PL513), the newly liberated TW523 (or PL513) will need to be attached to the other TTL port in order to supply a ZC signal for Rosetta and the controller (e.g. HAI, JDS, Ocelot). The controller, Rosetta and TW523 (or PL513) will share the ZC signal. I think the ZX-24 will be fast enough to have the controller receive pin follow the ZX-24 receive pin which will be listening to the TW523. This way, there will be no delay in the controller receiving what's on the powerline (or at least what the TW523 is willing to divulge on that topic). When the controller transmits, there will be a delay as Rosetta will need to see a complete code before deciding if it needs translation or merely needs to be sent to the powerline via the TW523 (or PL513). This should allow for sending/receiving a mixture of X-10, UPB and 2414S powerline commands

Rosetta will also be able to program delays (~2ms minimum, infinite maximum resolution) as needed between signals it sends to the powerline.

Sounds excellent! BTW, I received and email from smartproducts announcing their Zwave/X10 translator today:

From:Dave Houston snipped-for-privacy@whocares.com

I deleted the email already but IIRC it was about $150.

From:Dave Houston snipped-for-privacy@whocares.com

Absolutely!

From:Dave Houston snipped-for-privacy@whocares.com

I've finished the board layout and put together a bill of materials. It looks like my estimated cost was fairly close to the the mark.

I don't yet know what the board itself will cost so am estimating it based on a similar sized, low volume board (i.e. BX24-AHT).

For those buying the parts from Mouser and doing all of the assembly, it will cost about $35-40 plus shipping. This excludes the ZX-24 chip ($60 after March 1, $45 currently ), an enclosure (Polycase LP31F), the xPort Ethernet module (Mouser $49) and cables. This requires soldering of 4 SMT chips and several SMT resistors & capacitors (805 size) plus through hole connectors and capacitors.

If there's sufficient interest, fully assembled boards could probably sell for $50-60 including the enclosure but excluding the ZX-24, xPort & cables. Users can jusp plug-in the ZX-24 but the xPort cannot be socketed and will always require soldering - it needs the copper on the board for a heatsink.

So it will vary from about $80 to about $160 depending on whether you assemble it, when you buy the ZX-24 and whether you buy the xPort.

If you want X-10 RF input, it will require soldering a jumper on the main board (+5V power for the MR26) and replacing the serial cable on the MR26 which will also involve soldering. There is a replacement PIC I did a couple of years ago for the MR26 that allows it to handle X-10 security codes. This requires soldering an SOIC-8 SMT chip.

If you add the xPort, it web-enables anything connected to Rosetta.

Sounds terrific! My preference would be for a populated board as doing the SMT soldering can be tricky for a lot of people (including me).

From:Dave Houston snipped-for-privacy@whocares.com

And downright impossible for klutzes like me!

-- Bobby G.

Some enjoy DIY and some DON'T. ;)

Actually, it's not that tough to solder SMT by hand once you give it a try but I'm sure the majority will prefer the partially assembled route.

There has been a delay. When I went to submit the board design last night I learned the plant will be shut down for the next two weeks to celebrate the Chinese New Year. I'll try to get the code finished in the meantime.

I just enjoy it on a different scale.

The Winford breakout boards came today and they are sweet. They're also just about the smallest size of electrostuff I feel comfortable with. It should be a piece of cake to wire up an RS-232 line from the MUX to the PC. I got a DB25 and a DB9 and they'll both mount nicely in a small project box. Thanks again for the pointer. It's exactly what my ham hands need to work with electronic gear.

Isn't SMT work normally done by wave soldering?

Woof! It's the year of the dog, if anyone cares!

-- Bobby G.

I've ordered 2 boards, got ZBasic & the IDE working under Wine ( expect that the ZLoad will also work) and I've got my nice new smt iron waiting to be used. Now if I only had time! Grrr.

Anyway, SMT isn't too bad. My friend taught me that using the proper solder paste goes a long way towards doing it right. While I can't do the really fine stuff I can solder the JDEC (?) and SOIC quite easily. And that was with the wrong size iron (too big).

It's a VERY big deal here in Hawaii. My wife is Chinese so tomorrow we're going to the big celebration in Chinatown.

From:Robert Green ROBERT snipped-for-privacy@YAH00.COM

Wave soldering is used with through hole components; reflow soldering is used with SMT.

It's an art, just like woodworking or paper hanging. A professional wallpaper hanger will go through 100 or more razor blades per room, an amateur maybe three or four. A pro understands that $5 worth of blades may save $50 worth of time patching a bad cut from a dull blade. I do a fairly good job on plumbing and electrical work because they fit my big hands and I do a great job at cabinetmaking and woodworking because big hands are often a plus when working with hand planes and such and I've been doing it since I was 12.

When you get to SMT you're getting into watchmaking and diamond cutting skills. Risks are much higher, at least to the subject material. It's so easy to overheat and kill small components. With wood, the worst thing is to cut something too short so you "measure twice" and cut once and use a template or a jig whenever possible. With SMT no amount of measuring's going to save that teensy-tiny chip from too much heat. I'm envious of the guys who can do it because I know I am not one of them!

JEDEC: Joint Electronic Device Engineering Council SOIC: Small Outline Integrated Circuit

I've been doing almost as much reading trying to get my arms around the details of custom HA as I did for my Master's. I wake up in the middle of the night with some idea about how I am going to connect X to Y and have Z do something based on some state change. Like right now!

I guess if I really wanted to improve my skills I'd try *desoldering* some of the SMT's on all the boards in the junkbin. How do you guys vent your solder fumes? My most recent batch of solder came with a California Cancer warning that implied that if solder fumes didn't lead to sudden death, they would surely lead to an early grave.

Still, I *really* want one of Dave's BX-AHT's and someone's already bought Don's so I might just buy an unpopulated board or two from Jeff and have a go at it. I haven't added to Dr. Greenenstein's horror drawer of deformed devices destroyed by defective soldering. I assume Dave's board is somewhat more robust than the 4 layer laptop motherboards I've ruined!

:-)

-- Bobby G.