The socket that is used on the current bot board is either way too narrow, or the Atoms are way too wide to fit them. It is just too difficult to get them into the socket without bending pins, no matter how careful I am.
I just messed up my second Atom because of this. I just bent one whole side of an Atom PRO trying to carefully get it to go into a socket it is much too wide for.
I won’t be buying any more Atom microcontrollers.
PLEASE use a different socket that is easier to get these Atoms into on the next bot board.
I agree, but I think this is a general “error”. Well thats a bad word for it. Bassically there not ment to fit perfect, and yes bending the pins is the only way to install the atom onto the board. I hate bending them myself, thats why once its in its not coming out!
Maybe you should post some pictures in a thread so we can see the extended damage. You might be able to squeeze the atom into the baord if none of the pins broke off.
IDT there is really anything they can do to make the Atom fit better, besides making them a bit smaller. What I found was that the socket was much too big for my atom.
I don’t have a camera good enough to show what has happened to this Atom PRO. I know very well how to get chips into sockets, but the Atom PRO does not match up with the socket on the bot board. The Atom PRO is simply too wide to fit this socket easily enough to not cause some damage to the chip. One whole side of the chip has bent under legs now. I’m not even going to bother trying to get the pins straightened again.
I just can’t afford to pay $59.95 for a chip that will probably end up damaged in some way due to having to try fitting it into a socket that is too narrow for it. I now have two damaged Atoms - a Basic Atom, and an Atom PRO, and neither are usable now. This essentially means I am no longer able to even work with my robots until I choose another microcontroller and/or get a PIC programmer. I’m out of robotics until I can get another controller or a PIC programmer.
What Atom chip were you working with that was too small for the socket?
Ive used both Atom and Atom Pro and they both seemed too BIG for the board. Ive had to bend the pins inward for them to fit. Now you say you ahd to bend them out? Than this means the sizes are not the same and are not accurately made…
Both my Atom and my Atom Pro were just a shade too wide to fit comfortably into the socket as they came - by maybe 0.05" or less. A little bit of careful ‘group bending’ of the legs and a slow and careful insertion were able to get them into place without any real issues.
There may be a bit of a mismatch, but I think that’s more of a design issue of the module, than of the board. The ABB uses a standard 0.60" DIP socket. The Atoms use a bit of circuitboard 0.60" wide, that is designed to take pin legs that are edge-mounted to the outside of the module, meaning that they’re just plain going to be slightly wider than a standard socket unless the legs are bent inward.
Three obvious solutions:
A) Bend the legs of the module inward slightly, and coax it into place. This is what most people do, and while it makes moving the module in and out of the carrier board, neither the module nor the board are really designed for this. Socketing a ZIF socket into the ABB might make insertion and removal a bit easier, though the increased size would most likely overlap jumpers or components on the board (I’ve got an old BASIC Stamp board with a 24-pin ZIF socket, which I used when I had one module that was shuffling around between a few testbeds and applications).
Replace the standard DIP socket on the ABB with two single-row SIP sockets, spaced 0.65" apart. This is utterly non-standard spacing, and just seems strange.
iii) Redesign the ATOM modules to use a bit of circuitboard that’s 0.55" wide, to compensate for the edge-mounted pins.
Since this is a thread about redesigning the ABB, the second option would seem the more likely, but since it’s designed around the spec for the Atom modules, should they decide to change the width the other way (as they may have done between revisions, with reports of both too narrow and too wide modules), that would make a wide-socket ABB doubly incompatible, possibly to the point of not being workable.
The width issue - both in terms of being too wide, and of possible variations - is a BasicMicro issue, not an ABB issue. (I wonder if asking about this width thing in the forums on their website might yield some answers…)
Yeah that’s kind of wierd as one would tend to think the entire point of making the modules with what appeared to be a 0.600" spacing is that they would fit a standard 0.600" DIP socket. Maybe they have a tolerance issue on how the things are assembled or something.
The socket that was used, although very high quality, was not supposed to be used. It’s a machined socket. The next batch will have the normal low profile socket. I agree that it can be difficult to get the pins lined up. I just place the Atom on it’s side and gently roll the pins all at the same time to bend them inward slightly. This had to be done with the normal sockets too, but the round holes on the machined socket make them harder to align. Even though I admit it’s not as easy as it should be, I have never damaged anything trying to install the Atom into the Bot Board.
Send me the two Bot Boards and the two Atoms. I will personally see to it that they are repaired or replaced and installed into the Bot Boards for your robotic enjoyment. All you will be out is shipping one way. All you have to do is ask for help…
You are too kind, Jim! I can’t tell ya how much I appreciate this. The Atoms just don’t seem to like me, and I do know how to handle chips. I’ll pack everything up and get it all back to you in a couple weeks.
Until then, I’ll just go back into design mode and see what new designs I can come up with in… There are some things I want to look at changing for Walk 'n Roll, and I need to design some SES compatible motor mounts to use on WALTER and Walk 'n Roll. Both WALTER and Walk 'n Roll will be using motors for their wheeled drives rather than CR servos. WALTER has this way of gaining weight…
Quite likely I am WAY too late on this for the new spin of ABB but my recent purchase of an ABB + BAP has led to the following discovery, which seems loosely mentioned on the BAP forums… if you go searching for it after spending several hours trying to figure out why a 5 line program can’t enablehserial, sethserial, and hserout a character from a variable to a hardwired com port from a desktop pc… the same one used to program the thing to begin with.
The DTR serial signal is used to reset (or otherwise grab the attention of… whatever) the BAP so it can be programmed. However if you try to use any normal kind of communications application, like hyperterminal for instance, the normal thing for them to do is drive DTR (and RTS) high even if you don’t have hardware handshaking enabled. The result appears to be your BAP sits in reset or some mode where it won’t talk to its hardware serial port. This was gleaned from a couple posts on the basicmicro forum. I’ll test this more thoroughly tonight after I hack a 3-wire cable together at lunch.
Based on that info though, having a jumper shunt between pin 4 of the DB-9 on the ABB and the ATN pin of the BAP socket would make it a lot easier to flip between programming and testing of a serial communication based program. As it is you apparently need to swap from a real 9 pin serial cable (with all 9 pins connected) to a 3-wire cable in order to test the code you’ve just written with most normal terminal programs.
Yes if you have the handshake lines in the cable and the connected device drives DTR active. If you are writing a custom application and KNOW about this it is less of an issue. I just put this out as a nice to have on the ABB if it were possible because not only would it make this easy to work around but just having the jumper would make users aware of it being something that “might” need to be considered.
Well due to time constraints I was a bit conservative. The board is pretty much as it was with these changes. The PS2 game controller port had an inverter on the data line that was required for the Basic Stamp II. However the limitations in the Basic Stamp II code prevented it from working with many 3rd party controllers, and the inverter was too slow for the Atom and Atom Pro processors. Furthermore the Atom Pro has a 3.3vdc I/O pin that happened to fall on the PS2 port. Note, the Mini-ABB came before the Atom Pro. So we moved the pins used by the PS2 port from 4-7 to 12-15 to fix the Atom Pro problem, and we removed the inverter so you can use the port instead of using the I/O pins on the main I/O row. The new Bot Board II no longer supports the Basic Stamp 2 for Play Station 2 game controllers. The PS2 will now work with the Atom Pro which will be the processor of choice for all future development. There just wasn’t time to take on some of the more ambitious design changes. Thanks to SN96 for making the changes to the board! By the way your parts are shipping today, plus a little extra you weren’t expecting. Thanks!
