Von Neumann's popular IAS computer design (ILLIAC, JOHNNIAC, ORACLE, etc.) used an unusual analog adder design. It added two bits and carry-in as analog voltages, and then converted the resulting voltage to digital sum and carry bits. (Don't confuse this with an analog computer.) This seems convoluted to me, but apparently it was a viable alternative to a normal adder based on Boolean logic.
Edit: If you want full details on this adder, pages 60-63 of the manual [1] explains how the adder sums the voltages, getting 54V, 104V, 154V, or 204V from the sum. The digit resolver then turns this voltage into a sum bit. (The carry bit is easier, as it can be generated by a threshold.) You can look at the schematics [2] if you like vacuum tubes; the digit resolver is schematic #200.
It's just surprising they stopped at 3 bits. Weren't those parallel machines? I wonder why they didn't go for 5, 7, or 9. Too many comparators in the ADC? I suppose it was too early to use Gray's 10-bit(?) single-tube ADC—if it ever even saw mass production?
This took the place of a full adder, which has three input bits. (In other words, it was a single adder with three inputs, not a 3-bit-wide adder.) I don't see how 5 bits would help.
A carry in, two bits from each addend, and three bits out. Halves your carry-chain length in a ripple adder, and the DAC and adder parts of the circuit are a doddle; even an R–2R DAC is plenty, even at many MHz. But then you need a 3-bit ADC for the output. They didn't know about lookahead carry yet, did they? As I said below, the Chinese abacus uses it, but the relevance of that may not have been apparent at the time.
Toward the very end of the vacuum tube era, there were some "almost solid-state" tubes created, and mainly only used on televisions. You could find old TVs that had parts which were transistorized, and others that had these "tubes". Except they didn't quite look like tubes:
They were almost the size of then-contemporary transistors; just a tad larger. They were made in a special process (inside a vacuum chamber). Pretty much the zenith of the technology, and showed what it could be pushed to do.
Aha, I see National Union Electric Corporation is the same company as National Union Radio Corporation, having changed its name in 1954, five years before the Eureka merger. Thanks!
You'd probably need feedback to mix audio channels with it, since it wasn't designed to be linear.
It runs on 300 Volts likely from a supply that can source a couple of Amps. I'd rather take the flyback transformer, it will zap you but the voltage will collapse nicely. That 300V is there to stay.
> Hey, did you fill that due diligence position at your fund?
The response was nothing short of amazing, 30+ people responded, about half we made deals with an several of those are now regulars. Way beyond anything I imagined.
Oh, that's wonderful! Now I really wish I'd responded—at the time I thought it sounded like too much of a commitment. I hope things go really well there!
I'm curious, what other binary adder designs are out there? This adder-stage-chain design seems ubiquitous. I wonder if anything else can be done considering the current state of technology... or just for fun, like analog circuits or no clock, etc...
Carry-save adders (as used in Chinese 7-bead abacus) avoid the ripple-carry penalty mentioned in this article, as does the bit-serial design it mentions. Residue number systems are another option to limit carry chain length, which additionally save you loads on multipliers: https://web.stanford.edu/class/ee486/doc/chap2.pdf
When you don't need to add, just count and compare, you can use a LFSR, as the Atari video generator infamously did for sprite positions.
Modern computers, even back to the 1970s, usually instead use lookahead carry to get logarithmic carry delays instead of linear ones.
If you use a compound or otherwise non-binary base, carry logic is different, but the same basic options still exist. The 8008, 8080, 8086, i386, and amd64 instruction sets have “auxiliary carry” bits for use with BCD arithmetic; a separate DAA instruction adjusts the binary addition result to be a BCD result. The weirdest thing is that the 8086 added DAS, DAM (?), and AAA.
Edit: If you want full details on this adder, pages 60-63 of the manual [1] explains how the adder sums the voltages, getting 54V, 104V, 154V, or 204V from the sum. The digit resolver then turns this voltage into a sum bit. (The carry bit is easier, as it can be generated by a threshold.) You can look at the schematics [2] if you like vacuum tubes; the digit resolver is schematic #200.
[1] http://www.bitsavers.org/pdf/univOfIllinoisUrbana/ordvac/ORD... [2] http://www.bitsavers.org/pdf/univOfIllinoisUrbana/ordvac/ORD...