USB-CAN adapter

[24c]

IHaving the DIP switch available through the top is practical but it also means it could be accidentally switched which may cause issues. If the switch is slightly recessed then accidental switching should be minimal.

I have a few USB/Wifi & Bluetooth adaptors that adopt this approach, but they have tiny dip switches, and you'll need your Mr Magoo glasses here as well.

And you're probably right that a short USB extension should have minimal impact.

I use these already on my MacBook flavours and a desktop set up, because the port centres are too compact on the side, and on the rear of the desktop. Even SD card readers cause problems, and it's not just laptops, it's USB hubs too. I have a short lead on one connection, so I always know which port ID I am using.

When they are ready, put me down for one of these please, thanks Jean.

[jbelanger]

When they are ready, put me down for one of these please, thanks Jean.

Will do.

I'm now looking at using a real USB enclosure. That would solve most issues about blocking another USB port. However, I'm not sure I can fit every component in the package I currently use and that would likely require having components on both sides of the board. I would also solder the cable directly to the board and would probably have the 3.5mm four contact plug mentioned before with a 6' cable (like this).

Jean

[Rod S]

Jean,

Is having a short flying lead soldered direct to the PCB on the USB side really such a problem ?

As mentioned above, there are many examples commonly used, my 4 in 1 memory card reader being typical.

You wouldn't have to try and reduce the size of your package any more and it's more friendly on the laptop port (minimal weight hanging off it and nothing rigid that could get knocked and damage the connector inside the laptop).

Anyway, assuming I can't convince you, I've ordered a couple of short USB extension leads (dead cheap over here, GB£1.73 each including postage!) in anticipation

Rod.

BTW, the other lead (4 core, 3.5mm plug) looks ideal.

[24c]

You wouldn't have to try and reduce the size of your package any more and it's more friendly on the laptop port (minimal weight hanging off it and nothing rigid that could get knocked and damage the connector inside the laptop).

+1 Bulky rigid USB adaptors on the side of laptops are a big issue for me, ever since I had to repair one.
I always use a cable when the laptop is not on a desk.

Personally I have no issues with the first design, because I would just add a short USB extension cable as I have these lying around, but have you seen these? This company do some interesting boxes.
Prototype USB thumb drive, comes in clear too, so you could have a flashing led (status?)

[jbelanger]

That's the exact enclosure I wanted to look into.

And one reason to go to such an enclosure is also to simplify my assembly process: if I have a case that needs making only one round hole for the CAN cable and not the rectangular hole for the USB connector then it's a lot easier and faster for me. Having a short flying lead for the USB connector would also simplify the case machining but would require more work to make the cable and having more wires in stock.

So I'll have a look at both options and see what makes more sense in terms of assembly time and overall cost. In the end, if I have to spend significantly more time on the assembly it will end up costing more.

Jean

[24c]

That's the exact enclosure I wanted to look into.

Excellent, as it's a snap together you can cover the mini dip switch, rather than make an access hole in it. Totally agree with the assembly process & your costs, and it's what I would have done (ex product designer).
Making a round hole for the cable is the right choice, if there isn't a suitable high density surface mount connector available. There is the common 5 pin mini USB that you mentioned, but Nikon cameras use another type 4 pin IIRC. There is also a four pin FireWire aka 1394 connector that might be useful, as these are used in video/hard drive connectors the cables are screened.This is a robust connector too.

[jbelanger]

I can see the advantage of using a common connector but I'm reluctant to use a connector for a different communication protocol. This could lead to people connecting it to the wrong place and either burning the adapter or a PC port. I don't see why someone would connect the adapter to a Firewire port or device but if it can be done easily there's a chance it will be done.

Jean

[Rod S]

Jean,

If you did use that case I would plug it into a short cable extension lead anyway as I wouldn't want anything that long sticking out of the side of my laptop whilst it's sitting on the passenger seat or the lap of a frightened passenger

The risk of damage to the laptop port is too great.

Rod.

[jbelanger]

Rod,

I understand and as mentioned the final decision will be based on ease of production. So it will most likely be either the small USB enclosure with the 6' CAN cable and male USB plug or the current case with a short pigtail (a few inches) with a male USB plug and the same 6' CAN cable.

Jean

[24c]

Personally, I'd like to see a thumb drive type enclosure, but for production simplicity, you then have a USB male at one end and another connector at the other end, That way all Jean has to do is drop a single pcb into a snap case, after he makes a hole in the casing. I suggested using a 4 pin FireWire connector because it was small, and to be honest, didn't think anybody would be daft enough to use a proprietary cable and stick the other end into a FireWire port of their computer or camcorder.

I realise now that Jean is right, so why not fit a 4 pole jackplug socket at the opposite end to the USB plug, that way there is no soldering, you just get a double ended 4 pole cable, for example. These are as cheap as a single ended version, that will need to be soldered. I also realise the female sockets are bulky, but Innovate use a similar set up (OK it's 2.5mm) for their serial link ups.

PS I would always use a stubby USB cable with a thumb drive on a laptop anyway, so the fact it could come like that doesn't worry me.

[aarc240]

Add another customer.
I too would use a short USB cable regardless of case style and would favor the 4 pole socket so a double ended cable did the job.

[jbelanger]

So there seems to be a consensus.

I have looked at the USB thumb drive case and there isn't enough clearance above the board on both sides and it does require a thinner PCB than standard. So I have ordered a new small prototype batch for the same case as the previous one but this one will allow connecting a cable on both the USB and CAN sides. So I'll put a short USB cable and a long CAN cable with the 3.5mm 4-pole plug.

But for the final version, I'll look into putting a jack on the board so th CAN cable can be removed. It's going to be tight and will require some layout adjustments but that will be worth it.

Jean

[keithg]

Jean,

I want one of these! I am unconcerned about the cabling, etc. Is this based on the open can adapter design referred to on the T5Suite pages? If so, it could double duty for me.

Let me know when you have some to sell.

Keith

[jbelanger]

Keith,

No this is not based on this since at least one of the devices mentioned is based on a proprietary chipset which would do the B&G protocol (at least in standard form). However, I definitely want to support at some point the standard protocol(s) used by OEMs and other automotive products. The issue is finding information that is easily digestible for implementing it.

If you have any link to help get more information, it would be really helpful. And I'll let you know when I have some available.

Jean

[keithg]

This is probably overload on info. Maybe useful, maybe not.

The one used by the T5/T7 guys is this one:
http://www.canusb.com/downloads.htm
This seems to be the only one that works with the software crafted by these guys.

This one is an atmel based device which also is purported to work.
http://www.mictronics.de/projects/usb-can-bus/

Saab's proprietary commands are in this manual on page 96:
http://trionic.mobixs.eu/Trionic%205.pdf

More information here:
http://code.google.com/p/reverse-that-t ... 2FT5CANLib

Keith

[jbelanger]

Keith,

Thank you for this. I'll have a closer look at them. But I've already seen something that looks useful.

And when/if I do implement this, would you be able to lend me something that would help me test and debug the code? That's always the sticky point because I don't have access to a lot of those devices and I can't afford to buy them for the small expected volume.

Jean

[Rod S]

Jean,

Any progress on this ???

Reason for asking, I want to keep this as an option on my implementation of your IOx-OEM board. My initial intention is to use the sparkfun serial to USB to connect to the IOx-OEM but it will tie up one of the serial ports on the CPU which I may later want. So I want to make my carrier board also have the option of accepting this CAN to USB.

If your final design is as suggested with two flying leads, that would either be by installing a suitable socket on my carrier board for your CAN cable - for which I need to know which pin is going to be which in advance - or, if you have now committed to flying leads, are you connecting the cables to through holes or to pads on your PCB and at what pitch - giving me the possible option of asking you for the bare (unpotted, no cables) board and plugging it into my carrier PCB with either through hole headers or SMD headers.

Hope that made sense......

Rod.

[jbelanger]

Rod,

I have made little progress recently because I need to look at the code in both the USC-CAN adapter and MS2/Extra. There are problems in the original code so that needs correcting.

The final design will use flying leads with the wires being solder to pads on the board. The plan, as mentioned before, is to have a 3.5mm 4-conductor plug on the CAN side and a corresponding jack on the CAN device side. I think that's better than integrating the board in your design since one of the goals of the adapter was to shorten as much as possible the USB side to keep interference to a minimum. The CAN cable could then have the needed length to connect to the device and since CAN is designed to handle a noisy environment, it should make communication more reliable (together with the use of a shielded cable, of course, and the board's isolated CAN transceiver).

The design I have now does have this but I have 5V going to the isolated CAN transceiver (to power the CAN side) which needs to come from the CAN device. The final design will incorporate a 5V power supply so that 12V will need to be supplied on the 3.5mm jack instead of 5V. This will allow the use of battery voltage directly and will not put any demand on the device power supply. Also, this is a more standard CAN arrangement since an optional 12V pin is available on almost all types of standard CAN connectors. However, this will require a few more components on the board and a slightly longer case.

Jean

[Rod S]

Jean,

Understood, so I'll put in a 4 way 3.5mm socket (it will save me some space anyway...) to future proof myself if I want to use this when it's ready.

But is there an industry standard for what the 4 pins should be ??? - I can't find one so presumably you will choose your own.

In which case it would help if you could say at this stage so I can match the connections, in advance, for the future, so to speak.....

I assume it will be 12V, GND, CANH and CANL but to which connections on a standard 4 pole jack.

Rod.

[jbelanger]

The tip is 12V, CANH is the first ring, CANL is the second ring and ground is the sleeve. And this is not an industry standard type of connector. I couldn't find one that was readily available and not too bulky (DB9 is a standard but way too big for this). This one is small, cheap, relatively available for the plug, jack and cable and you can get a Y adapter that goes in a jack if you want to chain multiple devices.

The CAN-EGT+ will have this jack together with DIP switches to enable the resistor but also to connect and disconnect the 12V and ground connection on the jack. This way if multiple CAN devices are put on the bus, it is possible to have only one device providing the power and ground on the bus so there is no power backfeeding or ground loops.

Jean