CM15A and MisterHouse Module Starting Point

Thanks for the link Neil.

How did you analyze the cm15a communication? Some kind of line analyzer? I'm thinking that there shouldn't be an arbitrary delay between addressing the devices and commanding them. Any ideas or can I email you privately?

X-10 still hasn't released any documentation on the protocol?

I like the Insteon concept, which you already know if you followed my rambling review, but still have reservations based on their track record. I'm also waiting to see what they do for wireless control.

I have designed two daughterboards for the CM15A (one RS232, one USB) which replace the MCU with a socketed 16F88 and an SMT 12F683. The socket allows the 12F683 to be programmed in circuit. The 16F88 can use a bootloader which allows updating its firmware. The 12F683 handles RF (TX & RX) and controls an ST485 which can interface with other RS485 nodes. I had planned additional RF receivers on the RS485 network to handle multiple frequencies and multiple protocols but that's open ended.

It's looking more and more unlikely that I'll have the energy to complete firmware for them. I'm willing to make the designs open source if you're interested in taking them forward.

The bare boards cost $2.50 each (plus postage) for QTY 1 and that's with soldermask and silkscreen on both sides. They would cost about $0.10 each (plus air mail from China) in any kind of quantity. I might be able to arrange a P/N with the board supplier so that individuals could order them at the $2.50 price.

The 16F88 has a USART and I use a triple inverter for a 0-5V RS232 link. It does require a conversion cable (USB to DB9) but that's pretty simple. The USB version uses Silicon Lab's CP2102 which is a better (and much smaller) single chip USB-serial bridge that only requires a couple of bypass caps.

The passives are 805 size which I think is the smallest that can be hand soldered (even by a lame, spastic geezer like me) but it's easier to do all of the SMT soldering in a toaster oven. There's probably no way the CP2102 can be hand soldered.

The completed boards should cost in the $12-17 range.

Good luck with it!

I really think the hardware is first rate. The software sucks. (At least the early versions - I haven't paid much attention to any updates.)

I'll send you schematics and will try to create a bill of materials and write an assembly procedure.

I'm going with Insteon. So far it

FWIW, I sent Bob Cusey (Smarthome's Insteon busines development guy) an email expressing concern about their licensing strictures. My primary concern was the freeware/shareware/open source issue. Bob and Smarthome's Business Development veep (didn't catch his name, alas) phoned me to get a more accurate handle on just what it was that bothered me. Apparently, they've seen some of the negative reactions here and on the various HA fora, but haven't been able to contact the people complaining in order to get to the crux of their issues.

They assured me that the license was not intended to preclude the freeware/shareware/open source options - the Smarthome validation of Insteon-related products is only intended to be required for commercial products that use the Insteon logo ( kind of like Microsoft's Windows-certified program). They agreed (this is my take on it) that the curent license is ambiguous and possibly hard to understand; they said that they would change it to make it clearer that amateur/small developers were not required to jump through the validation hoops.

I regret that I didn't raise Dave Houston's confidentiality/NDA concerns - the phone call caught me a bit by surprise (I had emailed Bob my phone number, at his request, about five minute before the phone rang). I understand Dave's issues, but they're not mine.

On the strength of Bob's and the veep's assurances, I ordered the Insteon SDK (and am about to order the hardware development module).

Again, this is all FWIW.

- Dennis Brothers

P.S. - I agree that Insteon looks like something you can hang your hat on.

I don't think the license is ambiguous. It clearly states that they have the sole right to decide whether or not you can sell your Insteon based product and clearly states that you cannot divulge any information of any nature once you've agreed to the license.

They may have intended it to only require their approval if you want to use the "Insteon Compatible" logo. If so, it's poorly written. I would expect an outfit as big as SmartHome to have good legal talent available.

From some messages here and from your phone conversation it may well be that they didn't intend it to mean what it clearly means. But neither anonymous messages here nor your phone conversation have the same legal status as the terms of the license. If they don't intend the license to mean what it says they should rewrite it so it says what they mean.

Actually, none of these are my issues, either. As I said in my review, I'm not planning to use Insteon as X-10 meets my needs and I understand how to handle the issues that frustrate others. I'm not even planning to write any software to support Insteon as, while the cost of the kit is very reasonable, it's still more than I would spend just to create a demo program for others. My junque boxe already has an Ocelot, a Leopard, a few Bobcats, a LynX-10, a LynX-10 PLC, etc. Back when Worthington had their fire sale on IBM logo X-10 kits, I bought a few ($5 for a SmartHome starter kit was hard to pass up.) so I have several CM11As and now a CM15A.

I hope nobody interprets my screed as a knock on the Insteon technology itself. I tried to make it clear that I think it's well thought out and promises the elusive robustness everybody wants at an attractive price and I sincerely hope it proves to be as robust as it looks. From a technical standpoint they seem to have addressed most of the issues that have been discussed here over the past few years. Only time will tell.

I'll still sleep better and be less cranky if they rewrite the license. If they want to contact me, they have my contact info in their files. They have contacted me in the past.

P.S. I still have a couple of questions. I hope you'll answer them after you get the SDK. ;)

Neil. I can understand that the device doesn't have the ability to buffer commands and that would be bad. After all I'm sending commands to the cm15a, not the X10 lines. But I cannot figure out 2 things.

1. Obviously the device can buffer 4 bytes as it can accept longer single commands than that...
2. If I have a response waiting from the cm15a, I write the first have of the message, read the queued response, and write immediatly. This works in less time than half a second, it's basically instantaneous. That's why I'm thinking there is something else that needs to be sent for the first half. Maybe a or something like that.

-- Paul

Neil,

I threw a preliminary web page together at...

with pictures of the two boards and schematics. I'll add bills of materials with Mouser & Digikey PNs later as well as some assembly tips and my suggestions on a communications protocol for both the main MCU and the RS485 network.

I still have 3 bare boards of each type that I will give to _qualified_ people who want to create firmware with the understanding that it will be open source. They will need minor modification as I originally used 2 pins on the PIC12F683 for RS485 TX & RX. Later, it dawned on me that I only needed one.

Later, I will either provide the Gerber files or (if the supplier will agree) a PN for order>> I have designed two daughterboards for the CM15A (one RS232, one USB) which

"Neil Cherry" wrote in

Congratulations! I'm sure you'll do fine. Sometimes having someone else lay out the structure is more of a blessing than you realize at first. I've been toying with setting up a Linux server for HA. I'd certainly buy a copy!

-- Bobby G.