Working toward a nanoPEB case

[iKarith]

Okay all, I was just about all set to post a nice thread about building a case for the nanoPEB with pictures and measurements and all, but apparently from tapatalk on my iPad where all the pictures I took were, you CAN'T. Having an account with the forum wasn't enough, I had to create a tapatalk account. And I did that, and now I don't have permission to post a new topic. I have permission to reply though, so here's the stupid [CENSORED] [CENSORED CENSORED CENSORED] [CENSORED]ing [CENSORED CENSORED] post in a web browser so I have something to reply to.

 

Have I mentioned I hate tapatalk? And I hate forums on mobile devices? And that I don't use forums often because I tend to use mobile devices a lot, and I can't SEE what I'm doing without the stupid app which doesn't work right and sure doesn't act like a forum? Grrr...

[iKarith]

Okay, so I took lots of pictures, let's see if I can attach one.

 

 

Okay, so there's the nanoPEB itself. (This is clumsy.)

 

 

The first problem is that the CF adapter doesn't have the CF card stick out the back. Well, that's one of a couple problems actually. The other is that the CF IDE board is just flapping around there. That won't hurt anything, but it makes me nervous.

 

 

You can see in this picture that it can and will touch the top of the piggybacked chips. When it's not doing that, there's about 1/8" between the top of the chip and the bottom of the CF IDE board. Note there's a couple SMT chips that hang down lower that that, but there are no components between the chip on the nanoPEB and the PCB of the CF IDE board.

 

 

This last pic shows the other problem: the CF IDE board sticks down below the nanoPEB a hair. Just about 1/64" as it turns out. That may be within tolerance of a 3" square enclosure. Of course I'd recommend larger than that because the main PCB itself is 3" square and the DE-9 completely protrudes beyond that, as does the power jack.

 

The real issue here is: Should the CF IDE board be fit into the enclosure the main PCB fits into? If so, it's got to be thick enough for both boards and is going to have to be open in some way. If not, there's room for a layer of plastic between the two boards that can provide a solid shell around the main board and be drilled/cut for the switch, edge and IDE connectors, and the power and De/DB plugs at the back. That leaves the CF IDE board exposed, but a cover can be attached over that.

 

Or as I have seen a couple of folks do, the enclosure for the whole thing needs to be much shorter on one end with a notch so you can get the CF card out. (That or something to extend the IDE connector to push the second board back—doable, but it makes mechanical stability a much bigger deal.

 

Thoughts? Y'know, aside from that Tapatalk is annoying and overly complicates things.

 

 

Sent from my iPad using Tapatalk

[Stuart]

You should be able to separate the CF adaptor from the NanoPEB board using a bit of IDE ribbon cable with appropriate connectors (might need a different CF adaptor with a different connector type, depending on what connectors you're trying to use). So shouldn't be a problem moving the CF adaptor relative to the NanoPEB board, or changing its orientation completely.

[iKarith]

The use of the ribbon cable was what I was considering. The board hasn't got any mounting holes however as it is meant to mount into a motherboard's IDE connector directly. You'd want a male to male 40 pin IDC ribbon cable and you'd have to then mount the board with retention clips. Doable, but annoying. A stacking 40 pin header (wire wrap socket essentially) would probably push it back enough to have the card coming out the back next to the D-sub, but that'd make the whole thing mechanically pretty precarious.

 

You could also replace the CF board with SD. I'd seriously consider doing so if the filesystem were FAT32. I happen to know that at least newer Sandisk SD cards will, if you leave some space unpartitioned, use that as extra unused blocks for wear leveling and bad block remapping. Using the memory device raw as we are, it can't do that.

[dphirschler]

Are we talking about fitting the nanoPEB inside a 3" project case from Radio Shack?

 

Oh, I see iKarith's response. I'd love to see the nanoPEB modified to use SD cards. Is that possible?

 

Darryl

Edited June 4, 2016 by dphirschler

[iKarith]

Yeah, THIS should do what you want. Finagling the IDE connection the right way around is a thing you'd have to cope with, and the SD card would have to be in the raw format just like the CF currently needs to be. But it ought to work. The thing that's there now is a pretty standard Syba CF to IDE adapter.

[Shift838]

clearance will also be an issue, at least with both of my Nano PEB's it looks like the IDE to CF card adapter has been extended a bit to gain enough clearance to go over the EPROM.

[iKarith]

Using that device, you'd probably use a short IDE cable rather than the set of adapters that are plugged in now. I probably should note that there is the possibility the SD adapter won't work actually, since the nanoPEB/CF7+ manuals make it very clear that compatibility is ... limited.

 

I was just off trying to see if I could do something useful with the thing prior to some form of XB being available with it. Not really--can't even seem to use the catalog program. That might be that I actually just don't know what I'm doing, but I get some kind of I/O error in 50. I suspect it's trying to do something TI BASIC doesn't support. I need to ping Greg about an ETA.

[Shift838]

Using that device, you'd probably use a short IDE cable rather than the set of adapters that are plugged in now. I probably should note that there is the possibility the SD adapter won't work actually, since the nanoPEB/CF7+ manuals make it very clear that compatibility is ... limited.

 

I was just off trying to see if I could do something useful with the thing prior to some form of XB being available with it. Not really--can't even seem to use the catalog program. That might be that I actually just don't know what I'm doing, but I get some kind of I/O error in 50. I suspect it's trying to do something TI BASIC doesn't support. I need to ping Greg about an ETA.

 

do you have a speech plugged in on the console. I have seen this error if a speech is plugged in. I pulled mine off and i had no issues.

[iKarith]

 

do you have a speech plugged in on the console. I have seen this error if a speech is plugged in. I pulled mine off and i had no issues.

 

I do ... is that going to be an issue? Or was there a workaround?

[+hloberg]

 

do you have a speech plugged in on the console. I have seen this error if a speech is plugged in. I pulled mine off and i had no issues.

I have my CF7 (NanoPEB with prn port) attached to my speech without issue. you just got to make sure the speech contacts are really, really, really clean and tight to the console. I removed the clips from the speech and the console to help with tightness of the connectors. Only problem I have ever had was I had to swap my original CF7 to an updated model due to incompatibilities with the F18a video. the creator of the device said the f18a took too long to come up sometimes which messed with the timing of the memory on the CF7.

[PeBo]

I leave my nano permanently connected to a speech synth...I found the nanoPEB to feel flimsy, so I feel more secure disconnecting both when switching to the PEB (which I run thru a separate speech unit).

 

I've never had an issue, so have to agree with hloberg that the problem is in the contacts.

[iKarith]

That didn't appear to be the issue. I assume there's a problem using TI BASIC with the catalog program since the error is exactly the same, in line 50. Or old dsk1.catalog is not how you load it. It didn't like run dsk1.catalog at all.

[+Ksarul]

Note: most of the time, when looking for programs on a disk, you need to be using CAPITAL letters, as the TI is case sensitive, unlike a lot of machines from that period. The other thing to note is that the error you are getting would fit the fact that it failed to find a program named "catalog" on the disk. Try "OLD DSK1.LOAD" and see if that gives you a program (which probably won't work, but a lot of it may list, depending on which additional Extended BASIC commands are in it).

[iKarith]

I can now catalog the disk by doing "OLD DSK1.CATALOG" and then RUN. Weirdly, OLD and even DSK1 seem to work lowercase, which is why I didn't think it mattered. The CATALOG program fails on DSK3 with a bad subscript somewhere. I wonder if that means there's something not right with the volume? Either way, it seems there's not a lot I can do with it until I have a better BASIC.

 

The lack of a backspace on that keyboard is really annoying, and the fact that there is no function that serves the purpose in TI BASIC ... Ugh. Ctrl-H doesn't even work. I have discovered that left does, if followed by DEL. Is this corrected in other BASICs?

[+mizapf]

Device names are case-sensitive; many people already failed at that by using "dsk1". The lower-case letters were rarely used, maybe because they were not really lower-case but small capitals. For that reason, the Alpha Lock key remained depressed almost all the time, except when you used the Joysticks - so there was the common advice to release Alpha Lock now.

 

(Case insensitivity in the PC area is yet another of those haunting plagues brought to us by good old Bill G.; every reasonable OS has case sensitivity (ah, feeling better now).)

 

The TI keyboard processing does not use any CTRL combinations in the usual ANSI way. Accordingly, CTRL-H won't work, neither does CTRL-M, CTRL-G or whatever. All CTRL combinations are mapped to character codes above 128. This is also true for Extended Basic, and I suppose for other BASIC variants as well. There is a special keyboard mode, however, where you get a different mapping (I think it is number 4).

 

I never missed Backspace too much; you have Arrow Left and DEL, and ERASE for the whole line.

[iKarith]

Okay, with that solved, back on topic!

 

What I find is that you really want some solid thing to grab hold of when disconnecting the nanoPEB from your expansion port. If you were to create a vaguely 3 inch square tray and cut holes for power and serial in the back and the card edge connector in the bottom and just have that much extra attached to the nanoPEB to give you something to hold on to when disconnecting the thing, that'd be something significantly useful all on its own.

 

If you made that "tray" a little bigger than it needed to be at the top and bottom, you could create a U-shaped shelf under the CF board that would serve as a bed for the CF board to lay on. There's a couple SMT parts in the way of such a bed being quite flat, but it's easy enough to leave some space for them. This keeps the CF board from flapping inward toward the main nanoPEB board.

 

The top cover will provide some mechanical retention to keep the CF board from flapping away from the nanoPEB's main board, from being pulled out of the IDE connector, and give you something to resist your effort to extract the CF card.

 

Normally I'd prototype something like this using foam core, but the nanoPEB is reported to be somewhat static sensitive (among other nuisances), so, yeah that's a problem. I then got on to the idea of using coroplast. And actually, that makes sense as more than a prototype material! If the pieces are cut with some thought to aesthetics, the channels won't be too annoying. A box to hold a 3 inch square board whose edges are solidly glued will be tougher than most 3D prints and a helluva lot cheaper to make too. And finally, I assume most of you guys can manage a sheet of coroplast, a razor knife and good straight edge, a hand drill, a hot snot dispenser, a tube of silicone, and a plasticard scraper. Besides that,

 

THIS! IS! HOMEBREW!

 

Somehow I figure 3M Command picture hanger technically-not-velcro can actually hold the thing together. That stuff's tougher than you expect.

 

Thoughts?