Problems with a factory-made Soarer's Converter

[Red_October]

Hello! I'm new here but I've been over on Geekhack for a few years. I've been having a problem with a Soarer's Converter and I thought I'd see if anyone here knows what's up.

I bought a Soarer's Converter for XT keyboards from Orihalcon on eBay a while ago, used it happily for some time, only to have it develop an intermittent fault. Seemed like the connection at the USB end was flaky. He had me send it back and sent me a new one. Now that one has developed a fault as well, at first it was intermittent too, but last week it just up and quit on me. I decided that, as I had already received a second one and that it had been some time, I would tackle this one myself, especially as my soldering skills have improved greatly and I have a lot more tools for electronics work.

I proceeded with disassembly and discovered that the circuit board inside the USB plug was potted in hot snot (da bigger da gob, da better da job, amIright? ...) which hadn't bonded to the board or components, or even the rubber shell, only the wires. I cut everything back and proceeded to solder up the USB and DIN-5 connections now inside an enclosure box with a panel-mount DIN-5 and a USB cable with a gland, salvaged from some other thing long ago. Unfortunately, when I was doing something, I managed to knock off a tiny little surface-mount component about 1/4 the size of a grain of rice. One of the two larger ones on the side without the microcontroller. Aided by a bright LED torch, I finally found the goddamn thing (on a filthy cement floor... fun!) and was able to put it back on the board with my needle tweezers. Testing it with a meter showed it seemed to be a capacitor, as with the probes in one orientation the meter counted upwards and in the other downwards (in Ohms mode, beautiful old HP 3455a). I put it back in what I thought was the correct orientation and tried the unit. Plugging it into a laptop for testing (Dell Precision M60 portable workstation) I got an error from Windows: "Power Surge on Hub Port". I checked all my work, found some little bits of wire from the USB side that may have been making contact with one-another and trimmed them back. Same problem. I swapped around the orientation of the mystery surface-mount device, still the same error. What's up here???

[fohat]

Orihalcon's cables are hardly "factory-made" ha ha.

I have a couple of them but have never cut one open.

While I don't know what kind of circuitry he puts in them, I do know that he takes care when he builds them and that they have a reputation for being absolutely sturdy and foolproof. Personally, I have always used Teensies in my builds and never had any problems at all unless I made my own mistake.

Without knowing any details, I would guess that your problem is likely to be anywhere but Orihalcon's cable. Too bad that you don't have any other way to test an XT keyboard to see if you could isolate the problem elsewhere.

Also, Orihalcon is open to learning what pitfalls are out there so he can continue to improve his products.

Just my 2 cents' worth

[orihalcon]

Have you tried it on other USB ports or with other computers? There's also the possibility that it could be the keyboard itself as Fohat mentioned. You could hand wire a converter with a teensy and see if you get similar issue to prove that keyboard isn't causing it.

If there's ever a faulty converter, I can always exchange, but it would need to be unmodified though. The ones I currently have are a newly manufactured batch, so perhaps there's some minor difference that might make those work where previous ones did not?

t's fairly rare that these get returned to me for not working, so the fact that you've had two do similar things makes me wonder if something else is causing it?

[Red_October]

Orihalcon's cables are hardly "factory-made" ha ha.

I have a couple of them but have never cut one open.

While I don't know what kind of circuitry he puts in them, I do know that he takes care when he builds them and that they have a reputation for being absolutely sturdy and foolproof. Personally, I have always used Teensies in my builds and never had any problems at all unless I made my own mistake.

Without knowing any details, I would guess that your problem is likely to be anywhere but Orihalcon's cable. Too bad that you don't have any other way to test an XT keyboard to see if you could isolate the problem elsewhere.

Also, Orihalcon is open to learning what pitfalls are out there so he can continue to improve his products.

Just my 2 cents' worth

I assumed they were factory-made because of the molded ends. They do have a slight seam, but they just didn't seem like a thing that someone would be churning out himself. I have an old 5155 that may or may not still work, but I don't recall if the Modular jack on the front comes out on the back as a DIN-5. I'll check. I have another XT keyboard and many, many other computers, so I will see what happens with combinations. Different USB ports on the same machine yielded either the same thing (other port on the cluster of two on the right side) or nothing at all (cluster of two on the back).

Have you tried it on other USB ports or with other computers? There's also the possibility that it could be the keyboard itself as Fohat mentioned. You could hand wire a converter with a teensy and see if you get similar issue to prove that keyboard isn't causing it.

If there's ever a faulty converter, I can always exchange, but it would need to be unmodified though. The ones I currently have are a newly manufactured batch, so perhaps there's some minor difference that might make those work where previous ones did not?

t's fairly rare that these get returned to me for not working, so the fact that you've had two do similar things makes me wonder if something else is causing it?

I'll tell you exactly what it is. It's a fundamental design flaw with the USB connector and its lack of mechanical strength. If any force is put on it, it flexes and breaks the solder joints to the board. I've been seeing this fault since the days of 64-Megabyte flash RAM sticks. This is why I wanted to put it in an enclosure with wiring and sockets. I don't think it's any fault of yours, I'm sure if you spec'd higher strength construction it would go right out of the nice price bracket it sits in now, assuming you actually CAN get meaningfully stronger bits. You have no idea how I wish they'd stuck with D-Subminiature plugs, or even full-size DIN, for computer interfaces.

I'm pleased with your products and your exceptional customer service when I did the one return, but I sell on eBay for work and have done lots of customer-facing sales, so I'm not going to be an asshole, especially the asshole who tries to return something he's fucked with himself. What you can maybe do for me is tell me what that little SMD I knocked off is, and what it's for. It had no markings. I can make photographs if needed.

[Blaise170]

I for one am glad that we are no longer using D-Sub connectors, those pins are impossible to fix if they get bent. Heavens forbid if a bin actually breaks off.

[andrewjoy]

How in the hell do you bend pins on a dsub?

I have seen it on VGA a few times but only when the whole connector and its metal shield itself is crushed never seen it on a DB25 or a DE9.

You have to be some sort of orangutan to kill d-sub as much as USB.

[Blaise170]

Apparently you've never worked in IT.

[andrewjoy]

Oh i work in IT.

But back in the days when d-sub was more common you had less trained lobsters thinking that they are IT professionals because they can reinstall outlook.

[Blaise170]

All it takes is someone trying to shove a VGA cable in backwards and continuing to do so until you eventually break pins.

[andrewjoy]

The ultimate test.

Give a toddler a USB cable and VGA cable and something to plug them into. See what breaks first. I bet you its the USB.

Its why VHS is superior to DVD a kid can break a DVD in seconds, but even if they get the tape out of the VHS you can just wind it back in , hell you can even tape it if it rips .

[Engicoder]

I have seen it on VGA a few times but only when the whole connector and its metal shield itself is crushed never seen it on a DB25 or a DE9.

Thank you for not saying DB9.
Edit...although I did actually see a true DB-9 connector once...so much connector, so few pins

[ag36]

I remember the days, studs of vga or dvi connector broke off from motherboard -_-

[Red_October]

Haha I work in perhaps the messiest computer shop on the Earth, and the only mangled D-sub I haven't been able to resurrect was one that was trodden on, repeatedly. Otherwise, with needlenose pliers and a stout tweezers, you can set them to rights. You can always replace them as well, any decent electronics shop will have the common sizes. I mean, yes, a gorilla with canned hams for hands can fuck the pins if they insist on getting, say, a DE9 to plug into a DE15HD, but if you have the sense to pour piss from a boot with directions printed on the heel, you'll notice that it's not mating rightly and re-orient it or make sure you have the right connector.

[Red_October]

Alright, sorry for the double-post but I've got some results of testing the thing out on another computer and with another keyboard: on another computer, through a hub, nothing. Also directly to the machine, nothing. With another keyboard (one of those Honeywell/Bull jobbies with all the lock lights and extra keys, previously determined to be an XT keyboard of some kind because it did work with a Soarer's Converter before) the power light lights, the lock lights toggle, but nothing makes it to the machine, and there is no USB connect/disconnect sound, nor does any new device appear in Device Manager anywhere. Will try this again, tomorrow, after re-orienting the stupid SMD again, as I had thought I had it right the first time.

[Red_October]

Aaand I'm gonna triple-post! Sorry!
I tried to reverse the component, but it fucked off from the grip of my needle tweezers, never to be seen again. A little work with the meter figured out where the traces go, turns out it's supposed to be a 10μf capacitor across USB line and neutral. Ho hum, dig up such a cap, install it, test it out. Same behavior as yesterday, keyboard's getting power but sends no signals to the PC, PC sees no USB device.

[Red_October]

Quadruple post! Sorry! I did some more fiddling around with it, several attempts at repair later, the board's fucked. Pads coming off, traces coming out, etc. Not really a board made to be worked on, at least for someone with my level of skill with components that small (I'm used to through-hole construction, not this surface-mount shit). I've ordered a new Teensy and I'll end up loading the programming on that.

[ag36]

Eh, there's edit button to edit your post.

[fohat]

Teensies have always worked perfectly for me straight out of the package with PJRC firmware and Soarer's code on top.

[Red_October]

Teensies have always worked perfectly for me straight out of the package with PJRC firmware and Soarer's code on top.

Excellent. I don't think I'll have any trouble wiring it up, nice through-holes on the board. I didn't spring for the one with pins, I wanted to make it socketed but I couldn't tell if it fit a standard DIP socket and the documentation seemed to suggest that the pins were for putting it onto prototyping boards, not into sockets. Since I'm not prototyping with it (with this one, at least, I've got something I'll post about later on that I might need help with) I didn't bother. No particular reason I wanted to make it socketed, just frustration at so many things being not socketed, and a sort of logic that states "whenever you service a directly-soldered component, make it socketed if space allows".

[ag36]

I recommend solder sockets with some ribbon cable, you can use any socket you want, any size any type even crimped terminal.

[Red_October]

Microcontroller came today, wired it up and it's working perfectly now! Just thought I'd follow up!