Ttl 7400 Series

There has been a resurgence in retro technologies in recent years, such as projects built using 7400 series TTL.

Why I loathe 7400 series TTL

I always explain that I loathe 7400 series TTL and have finally decided to explain my reasons, so you be the judge ?

The Roll Form Batch Controller

In the early 1970’s I worked for a company designing a ‘roll form controller’. This innovative machine used a rotary encoder, made of reed relays and magnets, (but that’s another story) which measured the length of steel sheet that was shaped by the roll form machine steel rollers into various profiles such as roofing iron or beams used in building construction.

Length and batch numbers were preset via large front panel thumb=wheel switches and large orange Nixie tubes displayed the cut length and quantity set. It used heaps of TTL logic as this was the current technology, and Microprocessors were still a few years away from revolutionizing the industry.

I was present at factories where it was installed and I watched as it failed miserly despite every effort to fix the one big problem that made it totally worthless.

This was my first insight into why TTL (the early transistor-Transistor-Logic) was unusable.

What Was The Problem ?

The machine worked perfectly, the display would show the distance measured as the huge roll form machine fed steel sheet from a massive roll thru the forming dies, and at the right preset length a huge ‘flying guillotine’ would smash down, slicing off a section which would fall onto the previous units on a palette ready to be shipped to the customer. This continued as the ‘batch’ counter counted down to zero when the machine would stop.

But being in a huge factory, there was always someone using an electric arc welder somewhere, and this caused the problem that stopped this product being successful!

The machine might be working, and display a length preset of 4 Meters, and a batch of 100 units when someone in the huge factory started welding and suddenly the length preset would change to 1000 meters and a batch quantity of a million units !!!

This was a disaster! left unattended the roll former would rapidly unspool the entire massive steel roll all over the factory. Fortunately we never let this happen, but the possible costs and danger were massive and prohibitive. Could they have re-rolled the un-spooled steel roll afterwards ? I doubt it as it would have been kinked and too heavy to handle.

The engineers tried everything, filters, redesigns, everything they could think of, but they couldn’t even see the electrical spikes that caused the problem, let alone fix it. TTL was just too fast and was responding to electrical spikes they couldn’t see.

Industrial Alarms

In the late 1970’s I was building digital industrial alarm systems at work every day, they were 19” racks full of 7400 series TTL cards which latched alarms into ‘annunciator panels’. The panels went into large industrial plant ‘control rooms’ which looked like scifi films, i.e. large boards consisting of square switches with lights inside them. The button itself was engraved with the name of the alarm, i.e. “motor-1”

Hundreds, perhaps thousands of these buttons would be in the plant control room alarm boards, all wired to the TTL racks, which were in turn wired to switches inside plant machinery. Each alarm board ‘alarm’ had a choice of a NO or NC switch input, all TTL inputs used 220 ohm resistors for pull up or down etc.

When a alarm was activated, the light in the switch would flash, and pressing the switch would reset the flip flop that had latched the alarm. The control room operators spent their days pressing those switches because of the huge number of false alarms that always plagued those TTL only systems.

The standard procedure was ‘if a alarm is flashing, press the button to reset the alarm, and if it lights again, it’s a real alarm, not a glitch.

But why were they plagued by false alarms ?

The CACC

At work we had about 10 electronic engineers, all working on their own projects. One was named ‘Ray’, and his first project after graduating as a EE from university was what he called the ‘CACC’ it was a a ‘calculator assisted coordinate controller’ and the idea was the calculator like control pad hung from a massive overhead interior crane and one entered the coords to automatically send the crane to any location in a large factory. Ray was a great guy, smart, friendly and took the time to explain what he was doing.

The PCB he made was about a foot square and full of 7400 series TTL as that’s all we had back then, (as it was the latest technology) apart from a National PACE 16 bit cpu development system, but that was in use and anyway would have been too expensive for the CACC project.

The CACC all worked nicely, but had one fatal flaw, intermittently it would forget the coordinate! which Ray put down to the TTL being reset by a super fast ‘runt’ pulse, probably caused by electrical mains spikes caused by the hundreds of welders and motors in the huge factory that also housed our small R&D department.

Ray couldn’t see the pulse, he spent months trying to find it and work around it, but it persisted.

In those days a foot square PCB was a big deal, especially one jammed with 7400 chips, so he didn’t design another due to the time and cost, and in the end the poor guy was sacked for the failure of his project. What a blow to a really friendly and smart EE on his first project, I really felt this was most unfair as he had put all his best efforts into this project.

Those are the main reasons that I loathed 7400 series TTL in the 1970’s. I can’t really think of better reasons to loathe any technology than unreliability that you can’t fix because you can’t diagnose it.

CMOS TTL Saves The Day

Then 4000 series CMOS logic came out and all those issues I’ve described above just went away!

Everyone was pulling 7400 series chips and replacing them with 4000 series ones, and all the buggy projects were now working 100% reliably. The 4000 series were running from 12v and had a huge noise margin, and I rekon you could hear the massive sighs of relief from EE’s worldwide who’s jobs had just been saved in the nick of time.

Sadly it was too late for lots of EE’s like ‘Ray’ who had lost their first jobs because of 7400 series TTL :(

The Iron Ore Flow Guage

By comparison the radio-isotope iron ore flow gauge (https://patents.google.com/patent/US4205230) project I was assigned to, worked flawlessly with a NMOS National PACE 16 bit cpu and some small number of TTL such as quad tristate latches etc. I traveled to Port Hedland in Western Australia to install the prototype and watch the testing results. Nothing interfered with its operation, no nasty spikes affected it at all.

Life Was Hard For a EE In Those Days

Life was hard in those days, the brilliant EE I was assigned to (who designed the ore flow guage ) was also sacked when he wouldn’t release the project for billing as it ‘wasn’t quite finished’ …

After he was gone, I lost interest in that company and resigned because too many outstanding EE’s I worked with had been sacked and I no longer enjoyed working for that company.