RasPiComm v3
Posted by on August 10, 2012
I want to give you a short update about the RasPiComm progress. You’ll find the schematics of v3 below.
I reworked the whole board and rerouted it all. I tried to implement as much feature requests as I could while preserving the footprint.
I also decided to start working on a RasPiComm Plus board. It will be bigger and better! But it will not play in the same league pricewise.
But first things first, here are the changes I made:
- Moved the backup battery out of the way. Now it does not collide with the DSI (S2) plug on the Raspberry Pi.
- 5V tolerant inputs
- pluggable terminals for power, RS232 and RS485
- supressor diodes for the outputs. You can now directly connect relays to the outputs
- output is now a pinheader. The RasPiComm focus is on serial communication. Sacrifice I had to made to get the pluggable terminal headers in.
- Extra 2 pin header for 5V fan. In case you want active cooling on the Raspberry Pi
- Changed the through-hole joystick to a SMD variant. That resolves a little issue with the aluminum capacitor of the Raspberry Pi. No need for a tape or PCB support below to prevent accidental joystick activation
- Extra tantal capacitor for external 5V power supply through the power header
- SPI header
Shipment: Since my first post about the RasPiComm about 1 week ago I got a lot of positive feedback, so I really want to push v3 to production soon. Please give me a little time for finding the right partner for that. Eben Upton (Director of the Raspberry Pi foundation) is helping me to get this to Element14 (Farnell), this would be my preferred partner. Thanks Eben for your support! But as they are very busy with the Gertboard right now I don’t know about the timeframe yet. So I am working on it!
And thank you all for all the nice comments on my first blogpost about this! You all keep this going!
Downloads (Creative Commons Attribution-ShareAlike License)
Daniel Amesberger 
Pingback: RasPiComm – a Raspberry Pi piggyback board « Daniel Amesberger
Reblogged this on My Raspberry Pi and commented:
Nicely done, now take my money. 😛
Nice idea. I’d like to have one of these. Only suggestion O’d offer is to somehow allow the RPi GPIO connection to have a Male connector on your board so that I could plug in a GertBoard in addition to yours on the same RPi. Think this would be possible?
Hmm, I don’t think that it is viable to connect both boards to one RPi. The I/Os are configured and used differently on Gertboard and RasPiComm. The serial port for example is connected to the onboard microcontroller on the Gertboard, the SPI is used for A/D and D/A converter on the Gertboard, I use it for a rs485 interface chip. So this won’t really be possible, sorry.
Daniel, the element14 announcement is great news but I still feel you may want to investigate the earlier suggestion of involving Adafruit.
Having watched Limor Fried discuss the business’s aims on a video (Entrepeneur of the Year 2012) and having checked out the website it would seem likely that Adafruit may provide support through short-form and/or ready-to-assemble kits for those who can still see components that small 🙂
I already talked to Adafruit, they blogged about the RasPiComm but told me that they currently have 50+ new products in the pipeline and that they are currently not able to support it 😦
I also looked at the option with SeeedStudio, but I’m not happy with the pricing structure here yet. Want to do a low cost product, I don’t see yet how I can do this with SeeedStudio.
hi there
“GPIO voltage levels are 3.3 V and are not 5 V tolerant. There is no over-voltage protection on the board”
the above is a quote from http://elinux.org/RPi_Low-level_peripherals#Referring_to_pins_on_the_Expansion_header
But i see that u have interfaced the DS1307 directly to the gpio’s
i am confused, the SDA SCL lines in RPi board are already pulled up using 1k8 resistor
and one more thing read this PDF which says level converter is needed
Click to access an97055.pdf
Read page number 9,10
Thanks
Hi,
the DS1307 has an open-drain SDA pin, so connecting it to 3.3v via a pull-up does the trick (DS1307.pdf page 6). The SCL line needs 2.2v for logic 1 (page 2), so 3.3v is sufficient to have a reliable logic high. Since the DS1307 does not source any current we do not have a problem driving the DS1307 with 5v and having the I2C lines operate at 3.3v.
You are right about the internal 1.8k resistors, the 10k pull-ups are just in case you can reprogram the internal pull-ups in the RPi. They can be ommitted, but with 10k there is no potential problem, so I’ll leave them just to be sure.
Here is the datasheet: http://pdfserv.maxim-ic.com/en/ds/DS1307.pdf
If you still have concerns about the design let me know!
Daniel
Hi
thanks for the reply and for pointing out the logic levels.
I have to point out that the SCL and SDA are not pulled up internally but instead have a external pull up of 1K8 (page 2 of http://www.raspberrypi.org/wp-content/uploads/2012/04/Raspberry-Pi-Schematics-R1.0.pdf)
The 2 pullups (1K8) are named as R1, R2.
So there is no need for u to have the 10K there. I again verified this on the board.
Thanks
Prag
Thanks for looking that up, in that case the pull-ups can be omitted. It only helps in an extremly rare situation when you want to use the real-time clock with another i2c master and without the RPi. I think this special case can be safely ignored 🙂 I will remove the pull-ups in the next version. But they do not harm either, minimum pull-up in I2C sytems should be about 1k, the 1k8 and 10k in parallel will have a combined resistance of appox. 1k5. So no harm, but no use either 🙂
Thanks, Daniel
Is it possible to buy a prototype before series production is defined?
Regards,
Claus
It is theoretically possible, but the price will be quite high when we hand-place the components and use prototype PCBs. We would have to do at least 10 pieces, price per unit would be three times higher as in batch production (much higher part prices, higher assembling prices, much higher pcb prices), and shipping would be from Austria (which is not a big deal if you are in Switzerland 😉 ).
What will be the price of these units?
Never mind, found the answer in another post.
Reblogged this on Gigable – Tech Blog.
Hello,
It is a pity that you didn’t include the full duplex capabilites of the MAX (RS-422). what are your plans for the plus version? I am writing data acqusision software for the pi and I would love to see analog inputs (4/20ma?) as well as isolation of inputs/outputs/supply. Are you going down this road, or more digital communications?
Best Regards,
Alex
I thought about implementing RS-422, but it would need two different layouts to support RS-485 and RS-422.
I am currently working on a RasPiComm Plus. It will be available in the first quarter next year and it will support a lot of new things like digital and analog inputs and a wide range power supply (6-24V) all with real-time support through a dedicated ARM-chip. RS-422 and CAN is also on my list.
Opto-isolated I/Os will probably come with a seperate “RasPiComm Plus Industrial” version, because it is quite expensive. I would have to isolate all communication ports too which requires high-speed optocouplers.
If you have a wish-list I’m happy to hear about it, the design is not yet finished and – as with RasPiComm v3 – I will listen to community input and incorporate it into the design!
Daniel
From my experience with rs422/485 I had some adapters fry when used to connect to outdoors electronics during thunderstorms. I think there are many rs485 ICs with build-in isolation, although at a somewhat higher price.
Although isolation would skyrocket the price, I think for at least long line comms (rs485) it is essential for anything other than hobbyist circuits. SPI or I2C would probably be ok without any. I would love to see isolation on the power supply, shouldn’t be much pricier although a bit bulky.
Finally, at least in my case I would love to see MANY analog current/voltage sensors (10hz sampling rate max, I’m not greedy) and digital in/digital out (relays?) but that is more of a dreamlist than a wishlist, that could surely move up to 100eur per board so that wouldn’t sell that good. Maybe expandable modules that can be daisychained with SPI?
Keep up the good work, I was about to design my own board and I’m not that good with hardware design 🙂
Hmm, do you think that they have to be isolated? Of course, it is the best solution technically, no doubt. But it comes with other limitations like seperate power supply for the isolated part. Don’t you think that ESD protection (TVS diodes) would do the job in 99% of the cases? This is what I plan for the RasPiComm Plus. I think that ESD input protection does work very well and it comes with the benefit that strong power surges do not destroy the optocoupler, some TVS diodes can dissipate a lot of power.
Isolation of the power supply would have a big impact on price I think. But if you know cheap solutions I’m ready to listen! Our current design for the Plus board is using a buck converter circuit. This solution has a wide-range input of 6-24V and 5V 2A output at a quite low cost and PCB footprint, the digital inputs automatically adjust their threshold level to 1/2 of the input voltage. So if you supply 24V to the board, inputs below 12V are low and above 12V high. The outputs are high-side power MOSFETs driving up to 6A. I think this is a quite elegant solution. You can easily drive industrial components (valves, …) and connect industrial sensors directly.
If you know an isolated wide-range converter (6-24V, input, 5V 2A output) for a low price and small PCB footprint let me know!
Yes, the new RasPiComm Plus will have 2 SPI and at least 1 I2C. I thought about doing some sensors, actors and I/O boards which are connected via CAN bus. It is much more robust than using SPI and its done all over the automotive industry. I am not a big fan of SPI when it comes to connecting external boards. It is a perfect bus for communication between ICs onboard, but not so great when you need longer cables. I am still considering Gadgeteer-compatible connectors though. There are plenty of modules already available which could be connected via I2C and SPI.
I thought about making a website with possible features for the RasPiComm Plus and get the community to vote. So if you want you can follow me on twitter or this blog, I’ll keep you updated. We really putting focus on this (I have one software-developer on this full-time), I think the Raspberry Pi is a great platform and we really want to bring some cool boards out in the next months. And we also want to listen carefully what everyone wants. And there will be an industrial version of the board with optocouplers too, but this will take some time, we still have to see how the RasPiComm goes, then we will move on to the RasPiComm Plus and afterwards the RasPiComm Plus Industrial.
Disclaimer: I’m a software engineer so I might just don’t know what I’m talking about (although I have proper hardware training).
Diodes might protect from surges, but I am more afraid of ground loops in more elaborate installations than surges. In the industrial world there is a lot of analog sensing via ADCs so having a clean ground is essential. Of course if you want to combat ground loops you have to isolate absolutely everything and yes this will increase the price as well as the footprint. Nevertheless I was planning to develop a board with a budget of ~80-90eur (retail prices) for something like that which I think is pretty specialized. FYI for a power supply I was checking TRACOPOWER (not very cheap), although it might be cheaper to use a plain transformer for galvanic isolation.
Considering SPI & I2C, I also believe they aren’t too useful except for inter-IC communications, which could be useful only if you plan a more modular design with PCBs interconnecting very closely to each other. I think CAN is more than acceptable although I was considering rs485 or rs422 in a ring topology mainly because I am more oriented towards industrial standards (personal preference).
I would suggest for the evolution board, not to fear to cover the whole RPi with an extension board. I prefer using a heatsink to cool the cpu/mem than limiting the size of the addon board and any restrictions that this might impose to it. Also the coolest thing would be to have an enclosure (as you have already planned as I checked) that can house reliably the raspberry as well as the addon board and can withstand a bit of a harsh enviroment – No ventilation ports, maybe weatherproof, screw holes for wall mounting, maybe DIN rail mountable? This would be the perfect plug & play solution so that you don’t need to worry about the hardware.
FYI, you can check the project that I am currently working on here: http://www.raspberrypi.org/phpBB3/viewtopic.php?f=37&t=11955&p=175609, although I am also working on other (semi-commercial) monitoring projects in which the raspicomm would fit nicely.
I know Tracopower, and I use it for an industrial project, but they are far from cheap. I understand your concern, of course for industrial applications I always use optocouplers and I will bring them to the RasPiComm in an industrial version. But that will still take some time, firstly I have to sell these RasPiComms to keep the project going.
How urgent is your need with the opto-isolated version? I could speed things up a little bit if its urgent and build a RasPiComm with optocouplers for you. I like your project, we could probably work something out – you get free boards for development, and we can use your project as a reference implementation. I want to find some projects based on the RasPiComm to show what you can do with it. Your project seems to be a nice opportunity for that.
I think the correct answer is not urgent at all. The reason I want this is so I can slowly build a tightly integrated solution of industrial scale monitoring. Since I am not sure if people can trust something that is not a PLC, I don’t even know the marketability of such a product right now, and a lot has to be done on a software level anyway. Nevertheless it’s good to know that you know what I am talking about and have it under consideration.
Now on to the solar monitoring, since it’s geared more towards consumers, I think most people would prefer something cheaper than very reliable so your current implementation is exactly what I need, and the RTC is a very nice addition too. I would love to integrate tighter with the RasPiComm and for now I would love it if you could send me some info about how you plan to integrate the RasPiComm software-wise with the raspberry.