Anyone wondering about "Standard" TO220 Pinout?

The only contact most people will have had with any of the TTL families was the ZX81/BBC Micro era up to the IBM PC AT
After that there was an almost total replacement of TTL glue logic by ASICs
I still have an IDE bus development card with fixed pads for the bus transceivers

Great trip down memory lane.
Iā€™m pretty much at the point where the word ā€œstandardā€ means nothing to me.
Back in 1972 I bought a used PDP-8 that was retired from the assembly line at DEC in Maynard, Ma, USA. Talk about crazy logicā€¦ but learning how to program and fix it set me up for a nice transition into microprocessors later (the MC6800).

A lot of these medium power parts are gone for power supplies. I really like the new switchers with all the switches and snubbers on the same die. The new TI parts that also have the coil are great, just hard to get. I have 50 of the TPSM53604RDAR 36V 4A step down regulators on order. Ordered them a year ago, will get them about this time next year. 100 week lead times ! But for a lot of microprocessor projects it solves the power supply problem and most of the EMI problem. So you donā€™t see TO-220s and DPACs as often now.

While I may still use a TO-220 (I have a lot them), Iā€™ve moved to D-PAK. Hereā€™s a recent board I did with KiCad and the 3D printed chassis for it. It is basically a big ā€œORā€ gate with two inverted inputs (for pull to ground). It works for 3 to 20 volts. I have lots of PNP, NPN, and N-FETS in DPACs so thatā€™s what I used. Hard to find a ā€œgateā€ like that with that voltage range than can handle 2 amps. Itā€™s all ā€œdiode-transistor-FETā€ logic.

I find if ā€œfunnyā€ that my favorite D-PAK FET for use with interfacing microprocessor logic is the old MTD3055EL, an N channel FET with logic level gate thresholds, 3A capability. Iā€™ve still got about 1200 left over from a production run. If a smaller SOT-23 or its smaller variants canā€™t hack the current, this is great part. The number being a 3055 is just good fun.

Key Take-Aways:

  1. the word ā€œStandardā€ does work like it used to.
  2. there are less medium power semis in use today.
  3. Gate-level design and chips are declining in use, but the history is fun
  4. there are less people making PCBs and more people making software.
  5. Designing circuits and PCBs is an addiction.
  6. It is hard to express how cool KiCad and 3D printing are, and how well they work together.


1 Like

One of my first jobs involved a PDP-8a. Ferrite-core memory!! The bootlader for programs had to be keyed in by frontpanel switches. Each bit of the accumulator was an own PCB module IIRC.

1 Like

The first processor I implemented was the INTEL 4004.

Whopping 4 bits. sigh Getting all emotional. Everything was new and sometimes a first back in the day. Today everything is there already. Just rearranging known stuffā€¦

Full rack unit sized Classic 8 for me. I made some PCBs to convert the bus to TTL. First peripheral I made for it was a dual tone generator. Here it is in a room in my parents house, I was in my senior year of high school. You can see the paper tapes. The paper tape reader/punch was more expensive than the PDP-8.
Learning electronics and software, old school style. And getting a new Heathkit for both Christmas and my birthday for years.
Electronics as a hobby is making a come-back, complex chips and cheap processors (PI and others) are helping this. Love the journey I had, envious of the journey that lies ahead for the young pups.
Very much like the KiCAd-OSHPark connection. I used to expose PCBs using a positive etch system in sunlight (for UV light) for 30 minutes, etched in Ferric Chloride, drilled with a Dremel, ā€œplated throughā€ with wire pieces left over from the Heath Kits. Making PCBs is so much easier now, less of a barrier for people to play and learn.

1 Like

Yeah, thatā€™s her ( not it: her ). Complete with teletype, punched tapes and all. Do you still have it?

Making tape and sticky symbols on transparent sheets or Edding on copper?

Still have the machine, but, it developed a bad core in the stack around 1982. Iā€™m sure the huge caps on the ferroresonant supply are shot. 250#, 750 watts, .3 MHz with the equivalent of 6 KB of core, just in a 12-bit 4K format.

I used blank sheets that were for the overhead projectors and ordered the tape from Bishop Graphics. Didnā€™t use a computer for layout until 1997 starting with Protel. Built a ā€œrealā€ light table around 1978 when I was at Purdue. So glad things are easier today. People can put their energy into creation. Of course, in some ways things were easier when just about all logic chips were in a ā€œstandardā€ 14 or 16 pin DIP and everything could be connected up with wire-wrap. Progress means change.

Thatā€™s true. Tooling is not that much of an issue nowadays. Just finished a prototype with 2 PCBs and an enclosure. One PCB attached to the bottom half, one PCB (the carrier for Nixie tubes) to the top half.
Tight fit to keep it small. Boards in KiCad - enclosure in Fusion360 - Kicad-Step export to check - 3D print ā†’ perfect fit on the first attempt. All free.
I could go on forever, which I will not do :slight_smile: Nice chat though.

Heckuva nice photo!!

I donā€™t know what parts your Dpaks are. But my soldering abilities are only fair, and I can hand solder the 3 x 3 mm MOSFETs such as Infineon BSZ series without too much trouble. (The package designations are too confusing and I will not even attempt that.) They pack an amazing amount of power into a package which is much more space efficient than a Dpak. But a Dpak is inexpensive and very easy to work with.

Thatā€™s exactly what Iā€™m doing! KiCad, Fusion360 (the software I love to hate), export step file, make chassis, print on Prusa 3D. Also use Fusion 360 to make 3D models when I canā€™t find something on line.
Had to make/test ā€œrule setsā€ for different screws and their insets into the chassis, how to get mating ridges that always work, best thickness, etc. This also affects the PCB layout as I donā€™t use spacers in the chassis, I have the chassis clamp the board down. So I have to leave some 6 to 10 mm diameter room open on the board around the mounting holes. Might be a good topic for a thread someday, ā€œHow 3D printed chassis affect your PCB designā€ā€¦ we should share fusion 360 stories and strategies sometimeā€¦

BobZ: the DPAKs are mostly because there are a core set of my goto parts. I have the smaller sets, but if space is not critical or I just want a conservative design I use them. On the main PCB for this project, itā€™s all 0402 and the SC70 (I think) FETs and internally biased transistors.

Thatā€™s how I mount PCBs in enclosures in most cases:

Basically yes, maybe creating a new category.

It makes me feel all warm and fuzzy inside, seeing current technology used to resurrect a sixty year old icon.

All right, Iā€™ll take the bait:
It would have to be a display for a 10 GHz frequency counter or a clock. :slightly_smiling_face: :slightly_smiling_face: :slightly_smiling_face:

Far to few tubes for a 10 GHz counter, but enough for a clock. It is a Nixie clock, project name YANC (Yet Another Nixie Clock). IN-16 tubes, ESP32 processor and stuffā€¦

Iā€™m at the stage where I no longer regard the seconds digits as needed for a clock, although my clocks can be synchronised to the top of the minute. :wink:

Sure. Btw, I own a watch with only an hour hand to slow down :-). I just liked the proportions of the 6 digits.
And watching the warm glow as the seconds tick along is mesmerizing.
But then, one IN-16 now is around $10 for NOS, soā€¦

I use NTP for synchronization.

I use my stomach. :stuck_out_tongue_winking_eye:

But seriously i have an old Sharp desk calculator with nixies that I rescued from a flea market for a couple of bucks decades ago I should do something with. The nixies from the RTL based iron-constantan thermometer mentioned earlier are already glowing again.

I feel like I should nixie this discussion and send it down the tubes.

OMGā€¦:roll_eyes:
(Some characters for discourse)