EEVblog #867 - The Search For The First TTL Chip

[coppice]

attached are photos of 5400 parts in a TO-86 ceramic flatpak (surface mount) package, along with the plastic carriers that they were shipped in. 

Thanks for sharing, very interesting, never knew/saw that they were packaged individually. Are those plastick carriers ESD safe? They don't look like it.

I think they should be ESD safe. The earliest plastic tubes for DIP packages were, so the semiconductor makers took that issue seriously from the early days. People assembling boards, on the other hand.......

[Brumby]

Don't know about other chip families, but the 7400 series had a fairly low input impedance.  IIRC an output only had a fan-out of 10.

ESD didn't really become a big thing until CMOS came to town.

[richfiles]

I don't recall the exact logic type, but the Victor 3900 calculator used around 30-ish very early LSI chips (about 250-300 P-channel MOS transistors per chip). They were slated for a release in the summer of 1965, but were delayed to october when they discovered ESD was blowing their calculators up. They mitigated a lot of the risk by bodging resistors into the design, but the chips themselves were inherently vulnerable, as they simply didn't have internal ESD protection.

I have to imagine the industry had become WELL aware of ESD risks around that time period!

[MLXXXp]

Although a number of defence contractors eventually formed the legs of flatpack devices, and surface mounted them, they were not intended to be surface mount devices.

I've only seen a few photos of flatpacks actually soldered to circuit boards. As best as I can recall, they had the leads bent down for surface mounting, as described in these articles:
http://firetrucksandequipment.tpub.com/TM-9-254/css/TM-9-254_135.htm
http://www.tpub.com/neets/book14/59f.htm (scroll down to near the end)

How were they originally intended to be mounted?

[Svuppe]

Week 5 of likely 1980, as it has the Nat semi logo they were using around then.

Oh, so the 3-digit number is a date code. Never would have guessed that myself.

[rrinker]

Yeah, you see the older 3 digit date code a lot in older equipment teardowns. I guess in the early days they weren't thinking about the same part continuing across decades. Not just a Y2K issue.

[MLXXXp]

Are those plastick carriers ESD safe? They don't look like it.

I can't say for sure. I have no way to test them. My guess is that they aren't ESD safe. At least in my experience, there wasn't much concern about ESD with TTL devices back then, mostly only for FET transistors and CMOS chips. However, as has been pointed out by coppice, other ESD safe packaging was being used at the time, so these carriers could be.

[coppice]

Although a number of defence contractors eventually formed the legs of flatpack devices, and surface mounted them, they were not intended to be surface mount devices.

I've only seen a few photos of flatpacks actually soldered to circuit boards. As best as I can recall, they had the leads bent down for surface mounting, as described in these articles:
http://firetrucksandequipment.tpub.com/TM-9-254/css/TM-9-254_135.htm
http://www.tpub.com/neets/book14/59f.htm (scroll down to near the end)

We did that for airborne electronics, but you can't do that for missiles. The G forces are too high, and components would rip off the board.

How were they originally intended to be mounted?

I believe flatpacks were originally developed for the minuteman missile. Try looking up the electronics in that, designed to operate at enormous G forces.

[SeanB]

Flatpacks were intended to be spot welded to the PCB, and then conformal coated to hold them down physically.  They were formed and soldered as well, and with this and a conformal coat ( a wide range of them, both solder through or remove totally before soldering) they were very popular as a way to get the size of a board down to fit a box envelope.

Beware that the pinouts of the flatpacks are NOT the same as the DIP versions, the power pins are in the middle, and not the common pin 8/16 or 7/14 as on the rest of the TTL family.

There were a few missile designs where the flatpack devices were bonded to a flexible PCB and then selectively conformal coated to protect the pads ( using a flexible coat) and then the whole lot was rolled up in a jig to fit the space envelope, and placed in a mould and then further potted in a solid resin, making a final structural part that had electronics, capacitors and such all in a cylinder that contained other missile structure. Low failure rate there, mostly from actual physical damage breaking the structure.

Still have the special Weller Flatpack tip I bought, and it is in a Solomon handle now. Made changing the failed flatpacks a 5 minute job, instead of the 30 minutes otherwise using a stainless steel wire and stripping each lead off the pad individually.  Yes the plastic holder, you get 3 types, the ones illustrated, another with a 2 part clip together package and the third with the leads rolled around a bar to lock the pack in place. They are all a static dissipative plastic, though the insulation was measurable in Mohms, as you had a test jig you could simply place the whole package in as is, and it broke the leads out to a larger DIP pattern for insertion into a breadboard or an ATE system, so you could do a parameter check on the IC you were installing. Very fiddly trying to get 16 grabbers onto the leads otherwise without twisting them, and a twisted lead was a pain, as you could easily pop off the lid in trying to bend it straight, especially those in the top section. They did come in both ceramic hermetic with lids, ceramic with a glass frit seal, ceramic with side brazes ( only the larger number of pin packages though for MSI, which were almost a PLCC part with leads) and plastic moulded packages.

Strangely enough you get a modern adaptor that takes the flatpack, or a modern SMD part, and converts it to the DIP form, but with a sealed package. This is a compatibility part to replace old obsolete DIP parts with more modern SMD or old production flatpack parts, where you do not want to respin the board ( or certify it really ) to keep old things in military and aerospace running.  There are even 24/28/40 pin adaptors with the inside being a FPGA and memory or an ASIC to replace obsolete DIP parts. Where else will you get a pin compatible new part to replace an obsolete 2708 OTP Eprom in equipment, and still have the same part interchangeable in the equipment. They were often used not as a memory but as an 8 input 8 output combinational logic block to replace a board full of TTL.

[MLXXXp]

Flatpacks were intended to be spot welded to the PCB, and then conformal coated to hold them down physically.

Still falls under the definition of "surface mount", as far as I'm concerned  . Thanks for the detailed explanation of the use of flatpacks.

The ones I have were from a cheap "grab bag" mail order from Poly Paks or somewhere similar. I thought they might be useful for experimenting, even if I had to bodge them into something, since prime parts were expensive for me at the time.

I never actually encountered any flatpacks in "real life".

[rrinker]

 Somewhere I have a bunch of 7 segment LEDs that are in the flatpack format. I have no idea where I got them from, but now and then as I root through some junk boxes I come across them.

 My favorite (because it was my first computer) retro processor, the RCA 1802, has an unobtanium support chip, the 1861 'Pixie' graphics chip, a very rudimentary low res b&w graphics chip. A few different people have made substitute versions using a GAL or PAL (and one I think with a micro - though that may be a whole 1802 CPU) pin-compatible with the original 1861, so you can drop it in anywhere there was an 1861 and it works.

 This is getting a bit afield of the early 7400 series - well, I have some older TTL logic and some micro stuff (8085 I think) that I snagged from a forgotten Teletype terminal that was in the lab we used to build our senior projects in college. I'll have to check the date codes. Pretty sure most of it is no later than mid-70's.

[Ampere]

Dave, one thing that would be fun to me to make part of a wayback wednesday segment would be if you have copies of those issues, is going through YOUR old articles and projects that were in EA, and Silicon Chip. 

No one could explain the project in depth like you could and it would be extra fun for those projects which you still have the original prototypes to show as well.

Just a thought.

I'm all for this idea. I always appreciate the videos where Dave explains how something works. Fundamentals Friday and the power supply design series were my favorites.

[Alex Eisenhut]

Although a number of defence contractors eventually formed the legs of flatpack devices, and surface mounted them, they were not intended to be surface mount devices.

I've only seen a few photos of flatpacks actually soldered to circuit boards. As best as I can recall, they had the leads bent down for surface mounting, as described in these articles:
http://firetrucksandequipment.tpub.com/TM-9-254/css/TM-9-254_135.htm
http://www.tpub.com/neets/book14/59f.htm (scroll down to near the end)

We did that for airborne electronics, but you can't do that for missiles. The G forces are too high, and components would rip off the board.

How were they originally intended to be mounted?

I believe flatpacks were originally developed for the minuteman missile. Try looking up the electronics in that, designed to operate at enormous G forces.

Heh, even the Sprint's 100G acceleration looks sedate compared to an artillery shell and they were able to cram a tube radar/metal detector in there in WWII.



Oh, and missiles don't spin at 20000RPM...

[Brumby]

[chris_leyson]

Thanks Alex, nice picture  Looks like they replaced the battery electrodes with a piece of wood, I guess it's easier than trying to section dozens of thin zinc plates. Here's a close up photo of a Mk45 oscillator, the oscillator triode is mounted inside the coil former.

Triode close up, you can just see one side of the grid winding and the "mouse trap" spring, right hand side above the anode, that keeps the filliament under tension when undergoing 100's of G acceleration.

[chris_leyson]

Oscillator section

[SeanB]

Also used to detonate the first atomic bombs dropped, though there they also had back up altimetric fuses and conventional timers as well. The proximity fuses were only armed after a certain time interval and once the bomb had dropped low enough, so as not to go off too close to the bombers dropping them.

[bitwelder]

More thumbs up for the Wayback Wednesday! 
I think it would make an interesting segment for both young players (who haven't seen that stuff before) and old geezers (who are nostalgic about those times)

[jh15]

I repaired I think it was the airspeed indicator for the 747s that had a flatpack 555 timer

It also had a unijunction transistor in it for a "clock" or maybe watchdog timer.
Only other unijunctions I worked with were in Bell System oem equipment.

I'm sure I have some from dumpster diving (not really, production girls knew I was a nerd and brought floor sweepings before hitting the floor. Example being parts with leads cut wrong, etc).

[GeoffreyF]

I was 15 by the end of 1966, living in New York City, USA.  I remember those days the discussion in US hobby magazines turned on RTL mostly.  Don Lancaster had a book on that with some projects.  TTL was very expensive then, beyond most hobbyists.  One might buy a chip to play, but that was it.  I was working with TTL professionally about 1973.  Later working at Digital Equipment Corp there was a corporate legend that the JK Flip flop was so called because it was invented by Jeffrey C Kalb, then a VP at DEC.  What else to name the inputs but his initials?  Indeed he holds patent US 3591856 for the TTL "J-K Master-slave flip-flop".   That patent was awarded in 11/7 1967.  The PDP 11/20 (the first model of the line) was all TTL, with some discrete transistors.  The 11/70 the peak of that line was also TTL. The Vax 11/780 was mostly TTL still. 

I worked for the US Distributor of the EMS Putney Synthesizer as my first paid work with TTL.   They had a capacitance touch piano keyboard and sequencer for their AKS Synthi which fit in an attache case.  That device was all TTL with 1k Shift register memory.   The Touch keys were done by allowing the inputs of 74150 chips to float, with some bias to keep them consistent.  It was an out of spec use.  Later versions of 74150 were "Too good" for that trick to work.

[ornea]

Another thumbs up for the Wayback Wednesday! 

Thanks Dave for the uploading what you had, much appreciated.

I spent hours thumbing thru my old magazines until the termites got them.