How about a new AP Protection product then ?

"AP Sentinel"

It should Protect your AP and ensure as much Up Time as possible.

Post your Wish List in as much detail as you can, and i'll see if i can include them.

The Best suggestions will win a Free AP Sentinel, provided that it ever gets built or produced at all.

I'll be the Judge, Jury and Executioner, so no whining. I generally keep my word though, and runners-up can have a free R52, cos i got about 200 spare.

Features so far include :-

1. About 4x MT PoE ports with 12v, 24v or 48v PoE power options (each)

2. Programmable low-voltage shut off, with programmable re-tries in case the power comes back.

3. Scheduled rebooting.

4. Remote rebooting through a Web interface.

5. GSM option to remote reboot any of the power outputs

6. Multiple Power source options (e.g. solar, wind, normal PSU x2, nuclear

7. Big red button so Client can reboot without a brain

8. Automatically report stuff, like voltage/alarm

9. Relay outputs to control things like a Door Release mechanism

10. Camera option to see who's there with automatic FTP of images to somewhere

11. Digital/Alarm-style Inputs to detect if somebody opened a door, broke a window, stole something etc.

12. Digital keypad (to open door, box etc) with remote Code Change option

13. RFID card/keyfob access option

14. MT RS232 option, so it can report things using the AP it is plugged into

(Battery backup/solar/wind power is included under #6)

will it also make coffee and run to the store for me?

Coffee no problem.

As for store ... depends on whether the store accepts Mikrotik

well, i guess nobody wants one then.

I want one!

I would suggest to make it Modular...
One core with the basic stuff (PoE, power control, power cycle, web page)
And multiple pluggable modules to attach when needed.
It will make the product simple, cheaper for people that don't need some stuff (like door controls)
and you can invent new stuff if you have market.

Otherwise you will keep spending time in something that wont fly..

Edit: I would like to buy one for testing.

the thing is, it is already has a lot of very-very cool features, i just wish, the product will be cheap as possible(less than $100) and PoE passthrough feature.

I would suggest to make it Modular

That's a Good idea.

The only problem with 'Modular' is the production of lots of separate bits (modules).

How about having a Kit ?

You'd have 1 PCB with several sections that you could solder on the bits - or not.

E.G. if you never ever wanted alarm inputs, you would not have to Buy, or solder on the connector for the Alarm inputs.

Maybe the 'core' could be always populated, and have different versions of the same PCB - the only difference being which 'modules' were soldered on.

Of course, the less Components on the board, the less it will cost.

Is that any good ?

If < $100 (US i guess) is the Price Point, then i'd better get inventive.

Any more suggestions (and the chance to Win a Free one) before i start doing the PCB design ?

A Voltage / Current monitor would be nice, like a table/report for past 12 hours for each connected device and power source's at a 5 minute resolution?

Any more suggestions (and the chance to Win a Free one) before i start doing the PCB design ?

If you release only the schematic, very few of us will be able to use it.
At minimum, you should sell the PCB. If you sell a kit, that will be even better.

However, ideally, you should sell the finished product.

I suggest you start with a device capable of only the most requested features and then add additional functionality in the future in more advanced models or thru modules.

If you sell a kit, that will be even better.

Maybe i'll get somebody to sell the kit, or finished product.

You got any specific wishes to add to the mix ?

Something you've always wanted on an AP site but never had ?

A Voltage / Current monitor would be nice

I understand what you mean, and how it could be done, but cannot see a really good use for it.

If a Mikrotik suddenly starts consuming lots of amps, it will fry.
After that it will probably stop consuming any amps (well, after the fire burns itself out)

All the bits in between are kind of measurable ranges you can get off a production AP.

Why exactly would you want that feature ?

Maybe i'm being a bit thick.

One advantage of a Kit version is the Price.

The kit will definately be cheaper than a finished product.
... and the Price Point is < $100 US for a Finished Product.

At this moment in time, the Camera option is being dropped.

2 reasons :-

1. Price - it is gonna make it > $100, cos i'd have to use a Board with USB that runs Linux. (cheapest seems the UBNT RS board at 40 euros)

2. I bought a really good Pan/Tilt (no zoom) colour & night-vision IP Camera for 38 Euros off ebay.

The Pan/Tilt is great fun, and it does auto-ftp upload-on-change too.
This camera detected the best evidence for the existence of Fairies a few days ago.
It *might* have been a Moth, but *might* have been Oberon in person.

All it needs is a 5v 2A supply.
I got mine on a 12v car battery out in the Bush. (SolarTik controlled of course)

12v-5v Buck Circuit/code available for free if anyone wants it.

Here it is ....

The 'Fairy' is the mainly white thing.

The stuff at top right is my house.
I dunno why it looks like a Skutter off Red Dwarf from that angle.

One thing that amazes me (apart from the Fairy) is the striking similarity of the spiky plant on the left to some of the Voynich Manuscript illustrations .....

I have a few suggestions one being ditch the keypad and substitute wiegand its a simple protocol to implement and allows the use of just about any type of authentication device possible. Plus if you don't need a keypad it didn't cost you, and to keep it cheap for those who do make a cheap wiegand keypad or put a header for a matrix style keypad add on.

Unless I missed this in your features, I think the next most important thing to monitor next to voltage is temperature. I designed a temp. monitoring system over the weekend for my AP's using an arduino and ethernet shield and to allow expansion for more sensors I used the Dallas/Maxim 1-wire sensors. It's monitored through a web interface and every time I add a new 1-wire sensor it's device id and temp. pops up on the web page.

Another suggestion which comes from living in the southeast of the US is a humidity sensor. A great indicator of a problem with a gasket is rising humidity in the box.

My last suggestion is a LCD or at least an LCD interface. If you wanted to keep it user optional maybe throw in an option for the LCD size so users can pick their own to use.

Hope some of these ideas help and I must also say keep it as cheap as possible but don't cut it to low I've seen many other systems similar to what your designing go for a lot more than what you shooting for.

I think you could be rather successful with this and Good Luck.

Lyle

.... ditch the keypad and substitute wiegand ...

... important thing to monitor ... voltage ... temperature ... humidity sensor

... LCD [and buttons] interface.

Ditch the keypad ?
Wiegand is something i've never looked at despite hearing about it for years.
I will look at it now that you have suggested it.

Temperature & Humidity will add cost, but no reason *not* to include something like the SHT11 to give Temp/Humidity feedback/warnings, and the End User can choose to have it (at extra cost) or not. They could also add it later as cash allows.

[factoid for people in less warm climates: you can also calculate the 'dew point' from Temp/humidity, so you can tell when condensation will start to form]

Nice suggestions Lyle.

Your Dallas temperature monitor sounds cool. (no pun intended)

in a remote areea a solartik station will not need that stuff .. because eat amps ... no solartik instalation want this nearby . i don't like to replace an empty batery in the middle of the winter on the top of a hill far away from road access ...

so i guess you would not buy/install the LCD then.

simples.

because eat amps

This design takes the Amps in to account.

I will probably end up using a PIC chip, which is cheap and eats Nano-Watts.

Your battery will die of self-discharge faster than it will because of this design's Amps requirement.

As a clue, it might eat up to 5mA (0.005A). I will measure it and post the results once it is built.

There will be smaller/bigger current drawn for less/more features on the board.

The desire for a Web interface and <$100 is proving quite hard.

Anybody got any pointers on how to calculate the PPP 16-bit checksum/FCS without using a 256-word lookup table ?

PIC assembler solution would be nice

All the stuff i tried today didn't get the same checksum as a couple of examples i found, such as http://www.netfor2.com/ppp.htm

The desire for a Web interface and <$100 is proving quite hard.

Perhaps telnet/API interface (or SNMP?) would be more effective? Web interface could be built to interface with it or just use Dude...

Hmmm.

I like the idea of a really cheap device with Internet though.

E.G. if there is enough flash space, maybe it can have a Web interface and SNMP as well.

All i'm stuck on today is what f**k CRC PPP uses.
Tomorrow i'll pull the source for pppd apart and find out.

It may well be that the examples i found show bad CRCs - one certainly did.
Perhaps i need to capture some *real* PPP packets and make sure.

Well, i worked out the CCITT CRC-16 polynomial thing with a lot of Google help, but a recent post makes it all redundant now.

http://www.packetflux.com/index.php?mai ... ucts_id=21

It's already been done, and <$100.

Well, i worked out the CCITT CRC-16 polynomial thing with a lot of Google help, but a recent post makes it all redundant now.

http://www.packetflux.com/index.php?mai ... ucts_id=21

It's already been done, and <$100.

Nice link. I wonder why they are backordered.