I measured again and found the thinner side being 0.25-ish when pressed hard with the meter. The wider side is very close at 0.5 with all pins - plus or minus a few hundredths.
I took from the datasheet that this in fact is some generic part, the part name will give the specs.
We have Dn-mX, where n gives the number of pins (4 to 14), m the position of the common pin(s) , if any:
1 = Bussed with pin 1 in common.
2 = Bussed with centered common.
3 = Isolated.
5 = Dual termination.
and X is C for a common cathode type, A is common anode.
Those things seem rather expensive but make up for it by being out of stock. 
Did you consider using BAW56 or BAV70 or BAV99? Those are cheaper, more available, and have multiple manufacturers.
https://www.mouser.com/c/?m=TT%20Electronics&q=diode%20array
I never bought diodes more expensive up to now, but I have a very dense PCB where I can profit from the reduced traces. Even the DIP form factor is helping, as I can add another ālayerā on top of the regular copper to bridge a bus. I could do with a 16-diode variant, but since the pole ends at 14, I decided to take two 8-diode ones.
I can understand concern about density, but I would be worried about the fact that the entire series is out of stock, and even the quantities on order are low (such as 1000). Many of the Mouser quantities on the BAV-BAW series are over 1000 or 10,000. Such quantities give confidence of availability. With 0.1" lead spacing, those TT assemblies are not so tiny. How does it lay out compared to a bunch of SOT523s (for example) with two diodes per package? I donāt like to use those smaller packages such as SOT523 but I would do so if space were a problem. Normally I prefer SOT23.
Yeah, I would be worried too if this was a commercial project. But it is just for the fun of it for my own use, so the 32 items I needed were easily collected 
My older eyes and hands do not like SMD too much, so the DIP comes again as a benefit. I have a couple of SOIC24W chips to solder plus a handful of capacitors in 1206, so that seemed to be enough to torture me 
If you have the height available you could even stand up hand wired !N4148 in DO35 axial packages. But if you have what you needā¦go ahead.
I even have the 1N4148 in stock (SO123), but besides requiring a lot more soldering, these will require two pads each. That would give me another headache for my crammed PCB.
My idea was to stand axial leaded parts on end.
Do you mean SOD123?
BTW I much prefer SOD80 (mini-melf) over SOD123 for hand assembly. The SOD123 is happy to lay on its side or upside down; no such problem with the mini-melf. I can probably mount and solder 5 mini melfs in the time required for one SOD123.
Sorry, yes, SOD123. Chickenās foodā¦
My footprint for SOD123 has CrtYd rectangle 4.8x2.6=12.48 but my footprint for SOD523 has it 2.4x1.4=3.36 - 3.7 times less space.
Right, but times 16 for 16 diodes, plus the pad and keepout areas plus traces.
The arrays are narrower, are using 7 less connectors on the PCB and no traces at all. Suits me better.
Note that I was mentioning SOT523 and not SOD523.
Have you ever tried hand soldering mini-melf compared to SOD123? At least for me, much quicker and easier.
I didnāt liked mini-melf as it is round (worse to keep with tweezers). We tried also some diodes I would call rectangled mini melf (MCL4148 is an example I just found). After such tries we decided (previous century) to use transistors and diodes only in SOT323 (BC847W, BC857W, BAV99W, BAV70W, BAW56W). So even we needed one diode we used double. In my opinion it is extra protection against mistake with soldering diode the other way round (you canāt do it with SOT323). Recently (means 5ā¦7 years ago) I decided to get back to use single small signal diodes and because of it I have SOD123, SOD323 and SOD523 footprints but my designs are assembled by contract manufacturer and not manually. I stopped worrying about incorrect assembly - it is his worry 
Hi Piotr
It is surprising to me that one person (you) would find the mini-melfs to be more difficult and another person (me) would find the SOD123 more difficult.
- The cathode band is MUCH easier to see on the mini-melf
- You can (reasonably) lay down the SOD123 8 different ways on the pads (including on its side, upside down, reverse polarity) of which only one is correct. You can reasonably lay down the mini melf two different ways on the pads, of which one is correct. SOD123F may be a little easier to keep upright than ordinary SOD123.
- My hands are not so steady but I have no problem with the mini melf in tweezers. My hands are better for scrambling eggs than for doing surgery.
- The square mini melfs might be OK; but with one hand holding diode in tweezers and the other hand holding soldering iron (with solder already on one pad) the cylindrical mini melf diode will not roll away. The square mini melfs seem to be less common and more expensive than the cylindrical ones.
Reminds me of our decision when (with my employer) we started with SMT during 1990s. We used SOT23 transistors, 1206 resistors, and BAV99 diodes in SOT23 (even if using only one diode.) I do not remember what capacitors we used but probably mostly 1206s.
Fortunately we have a choice
Those time (about 1990) we (me and my brother) didnāt choose between diodes in mini-melf and SOD123 but between resistors in mini-melf and 1206. The flat side surfaces of the 1206 grip better with tweezers than round glass (larger contact area - mini-melf easier rotates vertically, and jumps out). Those time we used only resistors SMD. Other elements were hard to get. It was only 1 year after Poland left the socialist bloc (means lack of everything, including phone numbers - we, as a firm, had to wait 2 (or more) years to get a phone number, my parents were waiting 19.5 years until they got a number).
About 1995 we bought a semi-automatic machine for SMD assembly. The elements were taken by negative pressure with a flat metal tube (diameter of the order of 2mm). Flat top of element is better to get with it (the air does not escape). I suppose those rectangled mini melf were done because of it. I didnāt used that machine so Iām not sure but I can imagine that if you want to rotate element (and do it fast) then mini melf can not exactly follow this tube rotation because poor contact with it.
I have checked my footprints - mini-melf occupies 4.8x2.4=11.52 and SOT323 3.2x2.6=8.32 (we skip over using SOT23 very fast). So there were no reasons to use diodes in mini-melf as you had 2 in smaller package. End effect was - we used only Zener diodes in mini-melf as they could dissipate higher power and accept higher current pulses.
Regarding SOD123 - I wold never consider those extra 4 positions. It is practically not possible to not notice that legs are not at PCB.
Hi, Piotr
That mostly makes sense. I can imagine that vacuum pickup does not work as well with mini-melf (SOD80).
But just to confirm; I am unsure about your double negative:
Do you mean that you have no problem with SOD123 or that you find it difficult?
One difficulty is to see how the diode is rotated, but the greater difficulty:
When I put down the SOD123, it is probably laying on its side or upside down. Cylindrical mini melf cannot be on its side or upside down,
The difficulty is to rotate SOD123 correctly with tweezers even if the misplacement is obvious.
It is also difficult to see the cathode band.
Guys, funny to see you chatting along 
Just excuse me for a secondās interruption and let me leave my symbol and footprint for the D9-1C here:
Diode_Array_8_Common_Cathode.kicad_mod (2.6 KB)
d9-1c.dcm (105 Bytes)
d9-1c.lib (606 Bytes)
1 Like
My guess is that these are commercially obsolete.
Not much is made in through hole these days and these were obscure even in their day
Not possible to not notice means Iām sure I will notice if diode will be upside down. So I donāt expect any problem with it. Nowadays I am not soldering anything myself (only when I am looking for the source of problems in product) but when touching the sharp soldering tip you see that it goes under the diode leg instead of leaning against it.