> I photoshopped that sucker, edited some key measurements, and that was what they used for the very expensive mold-making process.
I work in manufacturing and this sentence made me feel so many feelings. Surprise that this worked, envy that this worked, remorse on how hard we try to make things precise AND THIS WORKED. Just so, so many feelings.
One thing to keep in mind is that connectors are often parametric: You got some end bits, and a middle section you repeat N times. This means a lot of your dimensions are going to be "Y = AX + B", where X is just the pin spacing. If you know they're from the same family, you can extrapolate the dimensions without too much trouble.
Of course there's still a huge risk with DB-19 being a semi-custom format. Sure, it's almost certainly derived from the standard, but are you willing to bet the cash equivalent of a car that they didn't slightly tweak it for one reason or another?
It was a standard D-sub connector at one point in time, DA19P is going to be the first part of the real part number. Only issue is that no one makes this one anymore as the D-sub is pretty much abandoned by the consumer electronics industry. Serial ports, single density 9 pin, and VGA, double density 15 pin, being the only survivors and even those have been almost entirely removed. Defense contractors still use lots of D-subs as it's a simple and cheap way to pass a bunch of signals. The major connector manufacturers stopped making 19 pins because no one was asking for it. Defense contractors aren't going to keep that model alive with their low volume orders, it's easier to just use a 15 or 25 pin D-sub.
Since there are so many compatible drawings out there of D-subs, it's easy to find one that shows the breakdown of why the measurements are what they are. From there, you can easily backtrack and get the exact dimensions of a DA19P.
From what I gathered, they Photoshopped a CAD drawing and changed numbers around (and not far presumably) — I’m not sure why that is too surprising given that the manufacturer is building it off the measurements and not literally the scale of the actual drawing?
(And given that drawings used to be hand drawn with pencil and paper, you have way more ton of precision control in a non-linear tool like Photoshop that gives you precise measurements than some rulers and those types of tools.)
if he had molds made, the manufacturer actually sent the dimensions to a 3rd party mold maker. That company put them into some sort of CAD software to build a mold (which is basically a negative). Once that was completed, the mold was brought back to make the product
I don't get why you would do this in a world where prototyping is so extremely cheap. Even if you have no hardware, you can easily get e.g. a resin 3D printed model made for very little money, just as a sanity check. See if it mechanically fits, alignment looks right, etc.
except we don't even compile programs anymore... (in the traditional sense, anyway; I try to look the other way when build engineering does dark magic with webpacks and esbuilds)
At the point where JScrambler is implementing a VM in JavaScript to run the obfuscated code, I think it actually qualifies as being compiled. Which is unholy, but that's the world we live in.
Any measurement off by some small amount, and the connector won't insert, or be too tight or too loose.
It is the equivalent of code that compiles and runs successfully at the first attempt. It is nice, and with experience easier to achieve. But it is still part lucky.
We don't know and that is the problem. A lot of manufacturing is about process and we have rigid processes because it reduces risk. There are no checks in this, no qualification/pilot run, no one checking the design who understands the requirements. The only solace I can find is that the factory might have done all of this, they just want to cover their asses in case something goes wrong.
If you just want a one-off you can try something I did way back to make my own HP48-SX connector. Get some pins that match. Solder or crimp wires onto them. Put the pins in the mating connector. Careful use of epoxy to make the "body" of the connector. Viola- your own custom connector.
Specifically, a SCSI cdrom audio cable worked although I can't recall if ATA and SCSI cdrom drives has identical connectors or not. Perhaps the SCSI one tended to be white rectangle without a retention mechanism while the ATA one was one was black thinner, and longer, with a retention mechanism.
3 of the 4 wires (RX, TX, and GND) would then be soldered to a DB-9 female plug.
That's not a "DB"-19. The DB shell is much wider, sized to accommodate up to 25 pins. It's more likely a DA-19, or maybe a DE-19. Or, as the author specified their own dimensions, an unofficial variant.
It is the colloquial term for the connector (regardless of what the 'official' name may be):
> Several computers also used a non-standard 19-pin D-sub connector, sometimes called DB-19,[8] including Macintosh (external floppy drive), Atari ST (external hard drive), and NeXT (Megapixel Display monitor[9] and laser printer).
It might be the colloquial term, but if you ask someone for a DB-15 connector then theres a good chance you'll get the wrong one. I regularly use both DE-15 and DA-15 in my line of work (among others), and any document that refers to a DB-xx for anything that isn't the DB shell size will be corrected.
"The DB-19 isn’t really a B shell and probably shouldn’t be called “DB” anything if you’re being precise about naming. Apple called it DB-19 in their documentation, though, and the name stuck."
There actually WAS a DB9 connector, that is 9 pins in the B-side shell. They were used for manufacturing equipment. They carried 4 high-voltage differential pairs and a common signal for negotiation/detection. IIRC part of the specification was that the signal pins had to be milled solid parts rather than the much cheaper/more common rolled hollow pins.
RJ45 is/was the pinout/wiring of an 8P8C[1] (eight pin/eight contact) connector designed for the telephone system.[0] The wiring specification of Ethernet is ANSI/TIA-568[2].
The only thing RJ45 has in common with Ethernet is that it used the same 8P8C-sized connector. But even then, RJ45 had a key notch on the side and Ethernet does not.
IIRC, that was one of the reasons we had the center pair available with Ethernet for such a long time:
A given jack on a wall in a given office could be wired for a regular single-pair phone, or for Ethernet, or for both -- even at the same time (using the same cabling), as a design intent.
Of course this doesn't work with the gigabit spec we all wound up using, but it was never a popular option to begin with (and nowadays, we usually have VOIP if we even bother with desk phones at all).
The original idea was to have the first pair in the middle, the next pair outside of that, etc.. so that a smaller pair plug could fit in a larger sized jack.
However when it came time to specify the pinout when designing ethernet over utp they found the separation required to place the wires like this was too great and degraded the characteristics of the line so they placed a pair on each side keeping the center as is for some sort of compatibility. Like said you could have a phone line in the center of your ethernet. suck to be you it you did that then later wanted to upgrade to gigabit. It also suck if you had some idiot doing the wiring who had the bright idea "herp derp, two pairs of this are not used... I could run two ethernet lines over this cable"
Which is sort of my next complaint, tia-568, the intent was to have the A standard be used with the B standard for compatibility with existing phone lines . Every place I have ever been they just used the B standard. I mean it is not a huge deal, but it always bugged me.
The antiquated Registered Jack standard is RJ45S, which always gets butchered to RJ45 in common parlance [and this sans-S version has never actually existed].
For RJ standards, an S suffix signifies that it is a "single line" circuit.
Yes, and this is a common problem with language. There are many instances where colloquial knowledge is just wrong and its not clear how people even came up with their alternative interpretation. Maybe you have seen me previously ranting about such things.
It’s confusing. It’s a technical spec and the wrong use can be confusing, and can cause you to discover you have the wrong part at an inconvenient time. It’s a case where “size matters”.
It used to be quite common to see a serial port wired in a B shell, sometimes with just three pins, making a DB-3.
If you were making the point that I shouldn’t care if a bird were a raven or a crow in the background of a painting, sure, that would be pedantry (unless there were some important symbolism). But this is a simple case where a modicum of accuracy can make a big difference.
Apple themselves call the floppy connector a DB-19 (https://support.apple.com/en-me/112201). In this case, it would be better to use the "wrong" DB-19 term so that people seeking the referenced connector can find it.
I would never have guessed this configuration was so rare. D-sub connectors seem to have more active use in aerospace and defense that you might expect (Until I started working professionally, I always associated them with "old" computers, not high end stuff).
Stuff like this made me realize how I have no idea how big enthusiast markets are. If you asked me how many weirdo disk connectors for vintage macs and Apple II's could be sold today I would have guessed maybe one or two hundred at most, sold over many years.
It looks like they've sold tens of thousands of these.
If we guess 50,000 at $129, with say a 30% profit margin, which might be low for hobbyist markets, its $2m net. These "little" enthusiast markets are much larger and lucrative than I thought.
I've always been wondering, could you build connectors like this by CNCing the metal parts and 3d-printing the insulation? (Maybe add in some metal bending, but maybe that could be done manually using hand tools?) In this case, there actually seems to be a market for thousands of these connectors, but for many other connectors the market is... a couple dozen or so. Note: I have very little knowledge in this field.
Yes you definitely can. Obviously you should buy the pins instead of CNCing pins, but for the complex geometry of the main connector body you can definitely CNC it. Example of the pins being sold on Amazon: https://www.amazon.com/JRready-Terminal-Connector-205089-1-c...
If not needed under 10 I'd make it manually like you say.
For above 10 I'd make a jig to either cut and punch holes in 3mm plexiglass sheets, or a jig to quickly make cardboard moulds into which I'd place the pins and pour resin into.
To me, either of the above is preferable to ordering 10k of them.
In the past, I've seen examples of D-shell connectors with molded all-plastic connectors.
The most common example is perhaps the Atari 2600's molded female joystick connector, but I've also seen male iterations (usually on IDC connectors).
Other than lack durability and lack of built-in shielding, I don't see any problem with the idea of 3D printing such a thing.
With a decent resin printer, one should even be able to design it to use removable pins that just snap into place (which are things that are still made in factories every day).
> but then I found an old mechanical drawing of a DB-25. I photoshopped that sucker, edited some key measurements, and that was what they used for the very expensive mold-making process.
I'm glad that didn't result in five figures of group-buy money wasted.
Reading along with great enjoyment, until I got to the part where the author is photoshopping and editing a component diagram of the D19 to get more manufactured.
I think to myself, "how strange! even if the manufacturer isn't willing to pay for the standard, it's got to be worth it for the guy who's trying to get them built!"
A small bit of Wikipedia browsing followed, in which I realized that I've been laboring under a decades-long misapprehension that the D in D-sub connectors stands for DIN. Nope, just the shape, and not a DIN, or ANSI, or ISO standard. Just de facto.
Note that this isn't totally insane, there are in fact DIN connectors https://en.wikipedia.org/wiki/DIN_connector. Just so happens that D-sub connectors have nothing to do with them.
If you pay for a factory to make unique tooling like this, can you avoid them using the tooling for their own runs, to undercut you as the sole source for your target market (such as on Alibaba/Aliexpress)?
Yes, but this scenario is a bit different. If you ask a connector factory to custom-make you a connector it's common that they are allowed to add it to their catalogue, though you can of course stipulate something else in the contract.
You'll see this in the electronics markets in Shenzhen for example... they still have a lot of the connectors and switches from MP3 players, which I believe was one of the early boom industries in the region.
Did you work in china though? I don't want to make wrong assumptions but I think the intellectual property protections are taken a bit less serious there.
For a large order this might be a concern, but for a smaller order (as it sounds like this one was) I would think it's unlikely to be worth their time to compete with you.
You could ask them for the molds at the end or stipulate in the contract that they can't use them for anyone else, they might not abide by this or might not sign a contract requiring this but you can ask.
No there's still no new production of those, as mentioned in the comments. It's kind of a shame. You can adapt a DB-25M to fit a DB-19F by cutting some of the hood to size and removing the extra pins. It's not pretty but it works. That's not really possible to do with the female connector though.
There is a plastic part in the middle to hold the pins which would need a mold. But you are correct that the outside would probably be formed with a die
I miss how solid these connectors were once screwed in, compared to e.g. USB or HDMI* connections of today.
(*once had an HDMI soften over time due to all the heat spewing out of the graphics card it plugged into, took a few cables before I found one that didn't want to melt)
Now someone needs to do this for the 8 pin 'DIN' commodore connector. They seem impossible to find now other than from people who've already bought up all of them to make commodore video cables.
I believe it's the same issue with the DB-23 connector on the Amiga. It's would be somewhat surprising if no one has done something similar for that connector. Then again, to build an entirely new Amiga you'd still need to harvest the custom chips and then you may very well be able to get the DB-23 connector as well.
If I am not mistaken, it is the same connector as the one used by the Sega Megadrive 1, although with a different pinout. In principle, you should be able to get a Megadrive cable, cut it, and reorder the pins.
They seem readily available until you try to actually order them. Everywhere that sell electronic components has an entry for them still but has no actual stock.
.jpg){kind=link}
I work in manufacturing and this sentence made me feel so many feelings. Surprise that this worked, envy that this worked, remorse on how hard we try to make things precise AND THIS WORKED. Just so, so many feelings.