POE Arduino with 12v output

237 views
Skip to first unread message

David Vandenberg

unread,
Dec 19, 2012, 9:05:25 PM12/19/12
to sydney-h...@googlegroups.com
Hey Guys,
As I have mentioned to some of you already, one of my projects is a cloud-based access control system.

I'd like to base it on Arduino and have it powered by POE.
I need it to supply power to an RFID reader and door strike. It needs internet access as the master access control database will be in teh cloud.
Just wondering what you guys think is the best way to use 802.11af to power it and have both 12v and 5v outputs.

AFAIK my options are:

1 - POE-enabled arduino, with a step-up transformer to 12v for supplying power to the external devices
2 - Plain ethernet arduino, with a separate 12v 802.11af splitter. Basic 5V regulator to power Arduino, and short cable to connect splitter to ethernet of arduino.

The main consideration here is that I want to keep it as simple and cheap as possible. With the ability to get this manufactured down the track.
Initially I think I'll be producing only 6-20 of these in the first year. But there is a possibility that I could ramp this up significantly and don't want to be manually building them and want to avoid support/maintenance issues.

Luke Weston

unread,
Dec 19, 2012, 10:14:05 PM12/19/12
to sydney-h...@googlegroups.com
The SilverTel Ag5100 used on the Freetronics 802.3af module has its voltage output set with feedback through an external voltage divider. On the Freetronics module it's set to 7.5V out to minimise power dissipation in the LDOs on most Arduino-like boards, but it could be changed to 12V by changing the resistor on the Freetronics 802.3af module. (At the expense of greater power dissipation in the Arduino's LDOs.)

Perhaps that's a good place for your R&D to get started, before you go to your own custom integrated board design suitable for scalable manufacturing at a lower cost.

Also, remember to keep in mind how much current the 12V door strike requires when it actuates, because the 48V-->12V 802.3af regulator will need to be able to supply that much current.

Standard, original 802.3af only guarantees something like 13-14W available to the device, after allowing for voltage drop in the cable, I can't remember the exact number. Assuming that the buck regulator is say 80% efficient, you're only going to have about 900mA at 12V available, maybe even a little bit less than that. Is that sufficient to actuate the door strike (and run the microcontroller and other stuff?)

David Lyon

unread,
Dec 19, 2012, 10:50:51 PM12/19/12
to sydney-h...@googlegroups.com
It's a nice idea !

One of the challenges is going to be the software
and how you do the cloud stuff.

As in what authentication you are going to provide
and what features you want (login/control etc). I'd
have a think about that.

I doubt the arduino could cram the logic in to be
ready for that.

The basic chip in the Arduino is an AVR and it's
not hard nor expensive to get custom boards
designed or made with the features that you want.

It takes around 2 weeks to get them back from
china using seeed-fusion.



Paul Szymkowiak

unread,
Dec 19, 2012, 10:08:28 PM12/19/12
to sydney-h...@googlegroups.com
A few thoughts:

a) There is a similar project currently underway at CCHS here in Melbourne, partially physically implemented, with active technical discussion. That project has leaned toward near field (NFC) rather than RFID. You can follow the discussion on the CCHS kent Lane list:

b) I don't know what budget you had in mind, but I know Jon Oxer has worked with this a lot in his own home:


which has resulted in a number of products on the freetronics site:



c) I think Lemming from the Brisbane hackerspace had a board he'd designed for RFID door access for hacker/ maker spaces. Can't see the links right now, but worth seeking out


Those links may at least provide you with good ideas for your own project.




--
You received this message because you are subscribed to the Google Groups "Robots & Dinosaurs" group.
To post to this group, send email to sydney-h...@googlegroups.com.
To unsubscribe from this group, send email to sydney-hackspa...@googlegroups.com.
Visit this group at http://groups.google.com/group/sydney-hackspace?hl=en.
 
 

Max Nippard

unread,
Dec 19, 2012, 11:35:46 PM12/19/12
to Robodino List
Here are Lemmings board details. 

---------- Forwarded message ----------
From: Lemming .
Date: Thursday, May 17, 2012
Subject: [ozhs-net] SNARC Simple NetworkAble RFID Controller.
To: hackerspace_brisbane <hackerspac...@googlegroups.com>, ozhs...@googlegroups.com, makers <mak...@makehackvoid.com>, artifact...@googlegroups.com, connected-commu...@googlegroups.com, robo...@googlegroups.com, sydney-ha...@googlegroups.com


Hi guys, 

As some of you know I have been working on this project for a while now. The SNARC is is a 99% Arduino compatible networked RFID controller with on board Ethernet, and a MoSFET for switching a relay or door strike or other large load. The board is designed to function semi-autonomously as an access controller for anything you want to use it for. We already have some example code going at HSBNE which does local authentication of 125KHz cards with network based updating and are working on adding new features to it all the time.


One of the planned features is a networked E-Stop system for use with machine tools. This means that if you have the machines hooked up with these for access control, you can also set it up in such a way that if you see Joe Bloggs on the other side of the room about to get sliced in half. You can hit the button in front of you and save critical seconds, rather than having to run across to where he is. Having these on the machines also means you can log how much the machines are getting used automagically and thus know when maintenance is due on them. 

The Edge at the state library here are working on an interesting project here with it where it will be used as a check in system for people to find other people with similar or meshing interests. Mark is doing this as part of his university studies and you can read his article about the board and his project here. 

The reason for this email is that I am now offering these for purchase in Kit form. This is just to try and recoup some of the time and effort I have spent developing, testing and refining this board. I will be opening up this board for everyone very soon, but I am currently unemployed and really need to see some reward for my hard work. 

Price is $60 per board, this is for a professionally made PCB, all components and headers and the Ethernet module. If you are interested in this please email me directly or reply to this email. I'll be placing the order for all the parts within the next week.

Regards,
Lawrence "Lemming" Dixon



David Lyon

unread,
Dec 19, 2012, 11:42:12 PM12/19/12
to sydney-h...@googlegroups.com
We could massively do better than that design by changing the colour of the LED's from green to blue. :-)

Nick Johnson

unread,
Dec 20, 2012, 3:08:43 AM12/20/12
to sydney-h...@googlegroups.com
On Thu, Dec 20, 2012 at 3:08 AM, Paul Szymkowiak <paul...@gmail.com> wrote:
A few thoughts:

a) There is a similar project currently underway at CCHS here in Melbourne, partially physically implemented, with active technical discussion. That project has leaned toward near field (NFC) rather than RFID. You can follow the discussion on the CCHS kent Lane list:

I don't mean to nitpick, but NFC _is_ RFID. RFID describes a broad range of technologies; NFC is one of those.

-Nick

David Vandenberg

unread,
Dec 20, 2012, 3:36:16 AM12/20/12
to sydney-h...@googlegroups.com
Thanks Guys, that's some good leads.

@DavidLyon - actually the cloud part is the easy bit for me. I've been doing web services since 1995, so I do that stuff in my sleep :)
But yeah I have no idea what the capabilities of the Arduino are for things like encryption, which at the very least I will probably be doing SSL for web calls.

@Luke - Looking at a few door strikes on Jaycar, it seems they draw 200-450mA, so max 5.4watts.

Iain Chalmers

unread,
Dec 20, 2012, 4:11:49 AM12/20/12
to sydney-h...@googlegroups.com
On Thu, Dec 20, 2012 at 7:36 PM, David Vandenberg <da...@thelab.com.au> wrote:
> Thanks Guys, that's some good leads.
>
> @DavidLyon - actually the cloud part is the easy bit for me. I've been doing
> web services since 1995, so I do that stuff in my sleep :)
> But yeah I have no idea what the capabilities of the Arduino are for things
> like encryption, which at the very least I will probably be doing SSL for
> web calls.

I'm 99.9% certain you _aren't_ going to be terminating an SSL/TLS
session on an Arduino.

If you want to run available-on-the-public-internet services, you're
probably going to need to use something with few more computrons and a
Arduino - perhaps a RaspberryPi or a WR703...

The other obvious alternative to to use a more capable machine t
listen to the internet (and terminate the SSL connection), and accept
the risk of non SSL/TLS secured connections between that machine and
the Arduino.

There is _some_ crypto libraries for Arduino here:
http://www.das-labor.org/wiki/AVR-Crypto-Lib/en but nothing as
plug-n-play as you might be used to with Linux/OpenSSL/Apache (you
might be able to build something using some of that crypto code and
the other end running the same crypto in Javascript. I strongly
recommend _not_ inventing your own crypto for anything except learning
purposes. I know I wouldn't trust my front door to crypto I'd
designed...)

big
>
> @Luke - Looking at a few door strikes on Jaycar, it seems they draw
> 200-450mA, so max 5.4watts.
>
>
>
>
> On Thursday, December 20, 2012 1:05:25 PM UTC+11, David Vandenberg wrote:
>>
>> Hey Guys,
>> As I have mentioned to some of you already, one of my projects is a
>> cloud-based access control system.
>>
>> I'd like to base it on Arduino and have it powered by POE.
>> I need it to supply power to an RFID reader and door strike. It needs
>> internet access as the master access control database will be in teh cloud.
>> Just wondering what you guys think is the best way to use 802.11af to
>> power it and have both 12v and 5v outputs.
>>
>> AFAIK my options are:
>>
>> 1 - POE-enabled arduino, with a step-up transformer to 12v for supplying
>> power to the external devices
>> 2 - Plain ethernet arduino, with a separate 12v 802.11af splitter. Basic
>> 5V regulator to power Arduino, and short cable to connect splitter to
>> ethernet of arduino.
>>
>> The main consideration here is that I want to keep it as simple and cheap
>> as possible. With the ability to get this manufactured down the track.
>> Initially I think I'll be producing only 6-20 of these in the first year.
>> But there is a possibility that I could ramp this up significantly and don't
>> want to be manually building them and want to avoid support/maintenance
>> issues.
>>
> --
> You received this message because you are subscribed to the Google Groups
> "Robots & Dinosaurs" group.
> To post to this group, send email to sydney-h...@googlegroups.com.
> To unsubscribe from this group, send email to
> sydney-hackspa...@googlegroups.com.
> Visit this group at http://groups.google.com/group/sydney-hackspace?hl=en.
>
>



--
"I say we take off and nuke the entire site from orbit.
That's the only way to be sure." - Ellen Ripley

David Lyon

unread,
Dec 20, 2012, 4:11:53 AM12/20/12
to sydney-h...@googlegroups.com
On Thu, Dec 20, 2012 at 7:36 PM, David Vandenberg <da...@thelab.com.au> wrote:
actually the cloud part is the easy bit for me. I've been doing web services since 1995, so I do that stuff in my sleep :)
But yeah I have no idea what the capabilities of the Arduino are for things like encryption, which at the very least I will probably be doing SSL for web calls.

AFAIK the AVR Arduino doesn't have enough memory to do SSL or HTTP encryption.

So initially, you'd be talking about transmitting door security information over the
network in clear text. I personally wouldn't be comfortable with that.

What you could do is find an encryption algorithm that will run on the arduino and
embed the private key into the eeprom of the arduino.

Then encrypt your data, and post that to your cloud server using simple http posts
maybe with some tweaks.

As for getting more power for your releases, maybe you could trickle charge some
18650 notebook batteries with the 5v on the POE rail. You only need ~ 4.2v do
that. Then run 3 in series to get your 12v.

Those batteries (ex-notebook batteries) are cheap and plentiful. The charging
circuits will cost about $4.

Luke Weston

unread,
Dec 20, 2012, 6:16:19 AM12/20/12
to sydney-h...@googlegroups.com


On Thursday, 20 December 2012 14:08:28 UTC+11, Paul Zee wrote:

c) I think Lemming from the Brisbane hackerspace had a board he'd designed for RFID door access for hacker/ maker spaces. Can't see the links right now, but worth seeking out


That would be SNARC, which is mentioned elsewhere in this thread. :)

There's also this one, which I built within a day just to see how hard it actually is to implement that basic level of functionality. Not that hard. And this might end up being taken and used by CCHS or Nothern Suburbs Makerspace or whatever in the future if it's wanted.

http://forum.freetronics.com/viewtopic.php?f=4&t=3082
 

David Vandenberg

unread,
Dec 20, 2012, 8:39:27 AM12/20/12
to sydney-h...@googlegroups.com
Hmm... Admittedly, I had just taken for granted that SSL would be possible on the Arduino or Netduino.
I might need to rethink things if they aren't even capable of securely consuming web services without some extra convoluted encryption.

Nick Johnson

unread,
Dec 20, 2012, 8:41:17 AM12/20/12
to sydney-h...@googlegroups.com
On Thu, Dec 20, 2012 at 1:39 PM, David Vandenberg <da...@thelab.com.au> wrote:
Hmm... Admittedly, I had just taken for granted that SSL would be possible on the Arduino or Netduino.
I might need to rethink things if they aren't even capable of securely consuming web services without some extra convoluted encryption.

The AVR is definitely not powerful enough to terminate an SSL connection. In fact, it can't even do the rest of the stack; the wiznet chips in the Arduino ethernet shield does all that for them.

You could implement your own crypto scheme; if you only need integrity and not secrecy, there are some secure hash functions you could implement fairly simply. Such a scheme is fraught with trouble, however, and very easy to get wrong. I'd suggest using a more powerful system and using proper SSL. How about the RPi as a prototyping platform?

-Nick

David Vandenberg

unread,
Dec 20, 2012, 8:53:15 AM12/20/12
to sydney-h...@googlegroups.com
Yeah sounds like the Rpi is the obvious choice if I want to do SSL.. I was really hoping to not have to run an OS though.
I'm definitely not a linux guru, and in my experience running an OS presents countless other problems with reliability and security.
I want this project to be as solid-state, set-and-forget as possible.
Need to restrategise I think.. Maybe I can some up with some encryption that is good enough and lightweight. Surely the .Net micro in Netduino would have at least one suitable encryption library. Will look into it more.

Nick Johnson

unread,
Dec 20, 2012, 9:07:28 AM12/20/12
to sydney-h...@googlegroups.com
On Thu, Dec 20, 2012 at 1:53 PM, David Vandenberg <da...@thelab.com.au> wrote:
Yeah sounds like the Rpi is the obvious choice if I want to do SSL.. I was really hoping to not have to run an OS though.
I'm definitely not a linux guru, and in my experience running an OS presents countless other problems with reliability and security.
I want this project to be as solid-state, set-and-forget as possible.
Need to restrategise I think.. Maybe I can some up with some encryption that is good enough and lightweight. Surely the .Net micro in Netduino would have at least one suitable encryption library. Will look into it more.

Well, there are more sophisticated processors that still avoid the need for an OS. I believe the ATXMega range has built in crypto support, as do some Arm Cortex M0 chips.

-Nick

Luke Weston

unread,
Dec 20, 2012, 9:36:19 AM12/20/12
to sydney-h...@googlegroups.com
If a door access control system is on the inside of a secure building, and it isn't directly accessible from the external Internet, does it really need to have strong crypto right up to the door strike?
How overengineered does it really need to be? In terms of malicious attack, deliberate security penetration, what are some plausible attack scenarios?


On Thursday, 20 December 2012 13:05:25 UTC+11, David Vandenberg wrote:

Madox

unread,
Dec 20, 2012, 3:04:04 PM12/20/12
to sydney-h...@googlegroups.com
I think he wants to connect it to the external internet, at least that is what I thought when he said cloud based... I had assumed that he meant the door strike itself is exposed as there was no mention of a control/interface server...

But then again, its probably just easiest to kill power to the door strike if I wanted to break in :)  Quite a few are fail open :)

Franc Carter

unread,
Dec 20, 2012, 3:10:09 PM12/20/12
to sydney-h...@googlegroups.com
On Fri, Dec 21, 2012 at 7:04 AM, Madox <mado...@gmail.com> wrote:
I think he wants to connect it to the external internet, at least that is what I thought when he said cloud based... I had assumed that he meant the door strike itself is exposed as there was no mention of a control/interface server...

But then again, its probably just easiest to kill power to the door strike if I wanted to break in :)  Quite a few are fail open :)

I'd be surprised if it was legal for them to fail in such a way that you could not open them from the inside
 


On Friday, December 21, 2012 1:36:19 AM UTC+11, Luke Weston wrote:
If a door access control system is on the inside of a secure building, and it isn't directly accessible from the external Internet, does it really need to have strong crypto right up to the door strike?
How overengineered does it really need to be? In terms of malicious attack, deliberate security penetration, what are some plausible attack scenarios?

On Thursday, 20 December 2012 13:05:25 UTC+11, David Vandenberg wrote:
Hey Guys,
As I have mentioned to some of you already, one of my projects is a cloud-based access control system.

I'd like to base it on Arduino and have it powered by POE.
I need it to supply power to an RFID reader and door strike. It needs internet access as the master access control database will be in teh cloud.
Just wondering what you guys think is the best way to use 802.11af to power it and have both 12v and 5v outputs.

AFAIK my options are:

1 - POE-enabled arduino, with a step-up transformer to 12v for supplying power to the external devices
2 - Plain ethernet arduino, with a separate 12v 802.11af splitter. Basic 5V regulator to power Arduino, and short cable to connect splitter to ethernet of arduino.

The main consideration here is that I want to keep it as simple and cheap as possible. With the ability to get this manufactured down the track.
Initially I think I'll be producing only 6-20 of these in the first year. But there is a possibility that I could ramp this up significantly and don't want to be manually building them and want to avoid support/maintenance issues.

--
You received this message because you are subscribed to the Google Groups "Robots & Dinosaurs" group.
To post to this group, send email to sydney-h...@googlegroups.com.
To unsubscribe from this group, send email to sydney-hackspa...@googlegroups.com.
Visit this group at http://groups.google.com/group/sydney-hackspace?hl=en.
 
 



--
Franc

David Lyon

unread,
Dec 20, 2012, 3:57:53 PM12/20/12
to sydney-h...@googlegroups.com
On Fri, Dec 21, 2012 at 1:36 AM, Luke Weston <reindeer...@gmail.com> wrote:
If a door access control system is on the inside of a secure building, and it isn't directly accessible from the external Internet, does it really need to have strong crypto right up to the door strike?
How overengineered does it really need to be? In terms of malicious attack, deliberate security penetration, what are some plausible attack scenarios?

People go to the guest room to use their wifi and have their wifi password. Somebody could use Ettercap to do a
man-in-the-middle attack, find all locks. Then return to the street at any time later and open all the locks via the
wifi at their own convenience.

See these videos for more information:

 - https://www.youtube.com/results?search_query=ettercap+man+in+the+middle

Anyway, there are measures that can be taken to protect against those particular attacks. I'm just answering
the question as it's posted assuming that the intended location is Fishburners. Which I could have wrongly guessed.



 

Kris

unread,
Dec 20, 2012, 7:22:55 PM12/20/12
to sydney-h...@googlegroups.com
Most normal door strikes are fail closed in my experience. Mainly because they are latch side not lock side. What I mean is they replace the strike plate I'm the door frame, so if you are inside you simply turn the door handle to open.
Most offices do not restrict you from leaving, only entering. Its a bugger for me because my access logs can never tell me if my office is empty or not.

Also on a related note, be aware that door strikes for external doors require a very different rating to normal internal doors. I really wanted to have my front door with rfid entry but found that my home insurer was not happy with it. Also they aren't ip65 rated, as in...they rust. (well the cheap ones anyway, blunderings do an all in one jobby but its like 300 dollars.
Sent from my Android phone with K-9 Mail. Please excuse my brevity.

Luke Weston

unread,
Dec 20, 2012, 8:07:10 PM12/20/12
to sydney-h...@googlegroups.com
Basically, keep in mind that your typical attacker / criminal isn't going to download the garbage file for the door locks off the Gibson, he's just going to go with the tried-and-true method that works on any platform, i.e. a crowbar.

The electronic door strike system doesn't have to be 100% impregnable to electronic attack (at any level of the hardware/software/network), it just has to be harder than breaking into the building conventionally.


On Thursday, 20 December 2012 13:05:25 UTC+11, David Vandenberg wrote:

David Lyon

unread,
Dec 23, 2012, 2:04:15 AM12/23/12
to sydney-h...@googlegroups.com
Hi Luke,

Nice effort. Obviously a lot of work went into that.

How much are these worth ? I have use for some if the price is ok. I'm thinking
I could use some as car-doorlocks. Having had my van stolen earlier this year, I will be
looking to cut the doorlocks out of my other old car and something like that might be cool.

David

Luke Weston

unread,
Dec 23, 2012, 5:25:27 AM12/23/12
to sydney-h...@googlegroups.com
I really like that board Luke.

It just seems to be a much more tightly integrated, more compact, lower profile sort of thing compared to SNARC, probably because it mainly uses SMD components and the original SNARC doesn't seem to use any SMD components, using relatively large, tall through-hole components and a plug-in pre-made daughterboard with the W5100 on it.

I wonder how much complexity would be involved with replacing the AG8100 daughterboard with something that is open right down to the ICs-with-detailed-datasheets level using a 802.3af controller chip such as the MAX5940? All the Ag8100 module seems to have on it is the 802.3af chip and a diode a couple of inductors and a couple of capacitors.

On Sunday, 23 December 2012 02:08:21 UTC+11, Luke Hovington wrote:
Hi everyone,
Im Hovo a Member at HSBNE..

I have been working secretly on a modified version of Lemmings SNARC,
as it's based on the same concept it has been called the SNARC+

Basically it has POE, RFID, Arduino compatible Micro and Ethernet back to base.
Im currently in a redesign phase as Atmel deprecated their chip I was going to use so I have to take that off the borad for a start.

If you want to check out my progress my github link is -> https://github.com/lhovo/SNARC_SMD
and some images of the first prototype are here -> https://groups.google.com/d/msg/hackerspace_brisbane/rc6EGhjuUoE/EIwAZBSGbLoJ

If you have any requests, comments or questions I would like to hear them.
I am planning at this stage to have the next version out by mid Jan and I think this is feasible at this stage.

Cheers
Hovo

Lemming

unread,
Dec 23, 2012, 8:45:47 AM12/23/12
to sydney-h...@googlegroups.com
There is an SMD version of the Normal SNARC in the works, it will have the ethernet integrated like SNARC+

The reason it uses PTH components is for ease of hand assembly. 

Luke Hovington

unread,
Dec 23, 2012, 11:11:03 AM12/23/12
to sydney-h...@googlegroups.com
Thanks, I haven't realt done a detailed costing as yet but the AG8112 is the most expensive part by far on the board at $8 each.
At present I have sunk a fair bit of my wages and time into this project as I want to see it realy succeed (plus we want to see them working in HSBNE soon as supplying power to each door is going to be a HUGE pain and expensive especially if you add a UPS for when there is no power).

As for moving away from the AG81xx and something on board would be awesome I would have to do some more reserch as there are lots of problems that arise when building these things. The module has 402 resistors and is really dense, plus the extra headache of having to deal with placing components on the board in strategic locations to reduce the noise going to things like the micro and Ethernet module.
Soooo for $8 I though I would skip this for the first release ;)

I would realy like to see the total board cost coming under $50 worth of parts. (not including my time to solder it all, yes I hand soldered the first board shown in the previous post - will use reflow next time or the option of getting it manufactured elsewhere)
This is of course one off prices, and the more interest the more the price can be reduced.

The price could also be reduced by placing pin headers instead of screw terminals (also save space on the board)
But I think it adds the the usability and ease of use and a better product. Thoughts?

For the RFID side I was initially going for 125Khz range as these are cheap and seem to do a good job. I have two chips in the back of my mind at present, Phillips HTRC110 and the Atmel U2270B, is there any others I should be looking at?

The NFC stuff looks interesting, though I think the board might have to be a bit larger to fit it all on.

At present the current design all fits on a board slightly bigger than Lemmings version 8x5cm but is still very compact compared to the arduino form factor.

Cheers
Hovo

Nick Johnson

unread,
Dec 23, 2012, 12:55:11 PM12/23/12
to sydney-h...@googlegroups.com
On Sun, Dec 23, 2012 at 4:11 PM, Luke Hovington <lu...@hovo.id.au> wrote:
Thanks, I haven't realt done a detailed costing as yet but the AG8112 is the most expensive part by far on the board at $8 each.
At present I have sunk a fair bit of my wages and time into this project as I want to see it realy succeed (plus we want to see them working in HSBNE soon as supplying power to each door is going to be a HUGE pain and expensive especially if you add a UPS for when there is no power).

As for moving away from the AG81xx and something on board would be awesome I would have to do some more reserch as there are lots of problems that arise when building these things. The module has 402 resistors and is really dense, plus the extra headache of having to deal with placing components on the board in strategic locations to reduce the noise going to things like the micro and Ethernet module.
Soooo for $8 I though I would skip this for the first release ;)

For prototyping and small runs that seems entirely reasonable. As soon as you start making your own boards, you can integrate this into them, and save yourself a lot.
 

I would realy like to see the total board cost coming under $50 worth of parts. (not including my time to solder it all, yes I hand soldered the first board shown in the previous post - will use reflow next time or the option of getting it manufactured elsewhere)
This is of course one off prices, and the more interest the more the price can be reduced.

$50 is quite a high raw materials cost. Going by the rule of 2.25, that's equivalent to a retail value of $112.50. What does it need to do besides scan RFID tags, interface over Ethernet, and open door strikes?


The price could also be reduced by placing pin headers instead of screw terminals (also save space on the board)
But I think it adds the the usability and ease of use and a better product. Thoughts?

Screw terminals are often a poor choice for permanent installations. If you can go with crimped headers, I would do so.
 

For the RFID side I was initially going for 125Khz range as these are cheap and seem to do a good job. I have two chips in the back of my mind at present, Phillips HTRC110 and the Atmel U2270B, is there any others I should be looking at?

This depends on your security requirements. Both 125kHz and Mifare Classic (basic NFC) tags are easily spoofed or forged. There are more sophisticated chips out there, like the Mifare DESFire, which operate on 13.56MHz.

For new applications, I'd go with NFC if you can.
 

The NFC stuff looks interesting, though I think the board might have to be a bit larger to fit it all on.

Why would the board have to be larger?
 

At present the current design all fits on a board slightly bigger than Lemmings version 8x5cm but is still very compact compared to the arduino form factor.

Cheers
Hovo

--

David Lyon

unread,
Dec 23, 2012, 8:51:04 PM12/23/12
to sydney-h...@googlegroups.com
On Mon, Dec 24, 2012 at 3:11 AM, Luke Hovington <lu...@hovo.id.au> wrote:
Thanks, I haven't realt done a detailed costing as yet but the AG8112 is the most expensive part by far on the board at $8 each.

Fair enough.
 
At present I have sunk a fair bit of my wages and time into this project as I want to see it realy succeed

Define 'succeed'.
 
(plus we want to see them working in HSBNE soon as supplying power to each door is going to be a HUGE pain and expensive especially if you add a UPS for when there is no power).

If you want I can donate you some Lithium-Ion batteries that will allow you to run for months.
 

I would realy like to see the total board cost coming under $50 worth of parts.

One way to do this (an idea from left field) is to remove the processor and the network chip and
offload all of that to a NetDuino or Raspberry-Pi.

It might be easier to sell as a 'shield' as well as costing you much less to produce.
 
This is of course one off prices, and the more interest the more the price can be reduced.

For the RFID side I was initially going for 125Khz range as these are cheap and seem to do a good job.

That will depend on what your market wants.

Purely just speaking for myself, I want it as a shield not with all the (expensive) microprocessor
and network stuff (that I already have somewhere). That I already have in my Raspberry Pi or Arduino.
 

At present the current design all fits on a board slightly bigger than Lemmings version 8x5cm but is still very compact compared to the arduino form factor.


Cool.

David



Lemming

unread,
Dec 24, 2012, 5:06:15 AM12/24/12
to sydney-h...@googlegroups.com
Sorry just wanted to respond to this bit.


One way to do this (an idea from left field) is to remove the processor and the network chip and
offload all of that to a NetDuino or Raspberry-Pi.

It might be easier to sell as a 'shield' as well as costing you much less to produce.

Cost comparison of an integrated board like the SNARC or SNARC+ vs seperate processor, network and IO systems is misleading.

It's all well and good if you only want one or two of them, but once you get into doing lots of them (like HSBNE)  the cost of the extra parts adds up quickly, especially once you add in assembly time.

For instance, your idea of using a Raspberry Pi. $44 for the board, say $1 for a MoSFET to switch the output with, $4 for a breadboard to mount external components too, $12 for an RFID reader, $10 for a power supply to run the whole thing and $1 for a 7805 to run the Pi. Add a few $ for a case to run it all. It would probably take you 45 minutes to an hour to assemble all of this. The finished product will also end up a lot larger than either of the SNARC's

With a SNARC, you are looking at ~$35 build cost, $10 for a PSU, $12 for an RFID reader. Add a few $ for a case. The psu socket is already there, the outputs already wired, it takes about 5 minutes to assemble.

Once we do start to get volume production going on these, which I'm looking to do, and pretty sure Luke is as well. The cost will drop FAST.

David Lyon

unread,
Dec 24, 2012, 7:37:39 AM12/24/12
to sydney-h...@googlegroups.com


On Mon, Dec 24, 2012 at 9:06 PM, Lemming <iner...@gmail.com> wrote:
Once we do start to get volume production going on these, which I'm looking to do, and pretty sure Luke is as well. The cost will drop FAST.

cool.

We'll remember to ask for "best-price" right from the very start. Thanks for the info.


Luke Weston

unread,
Dec 25, 2012, 10:30:02 PM12/25/12
to sydney-h...@googlegroups.com
Yeah, and it's a whole lot more open, and gives you more freedom to choose what sort of development toolchain you're using, compared to something closed like Netduino.

David Lyon

unread,
Dec 26, 2012, 6:18:16 AM12/26/12
to sydney-h...@googlegroups.com


On Wed, Dec 26, 2012 at 2:30 PM, Luke Weston <reindeer...@gmail.com> wrote:
Yeah, and it's a whole lot more open, and gives you more freedom to choose what sort of development toolchain you're using, compared to something closed like Netduino.

The Netduino is just one type of processor. In this application it has the neccessary processing power to
satisfy David's original request for basic security via TLS/SSL. Being at Fishburners, budgets are unlimited - haha.
I'm sure they can afford the tools.

Also, Arm chips are coming down in price.

The form factor on the NetDuino means that if the juicier security isn't used then the device could be used
as a regular arduino shield. Not saying I think they are great, but people know them.

In the meantime, I think I'm going to order something like this and get learning on RFID:

 - http://www.aliexpress.com/store/product/Free-shipping-New-Mifare113-56M-wiegand26-34-dual-Led-9V-12V-epoxy-packaged-high-cost-performance/314389_658740280.html



Lemming

unread,
Dec 27, 2012, 5:21:26 AM12/27/12
to sydney-h...@googlegroups.com
I would advise avoiding Wiegand readers like the plague, they are a pita to interface and code for.

Andrew

unread,
Dec 27, 2012, 4:44:13 PM12/27/12
to sydney-h...@googlegroups.com

http://www.plosone.org/article/info:doi/10.1371/journal.pone.0049365?imageURI=info:doi/

 

Carbon black mixed with polymorph, 15% by weight.

 

Seems to be within reach of mere mortals...

 

David Lyon

unread,
Dec 27, 2012, 5:54:37 PM12/27/12
to sydney-h...@googlegroups.com
Would be nice to see it working.
Reply all
Reply to author
Forward
0 new messages