Everybody sings the praises of the Tomlinson Dellorto and Weber manuals. Your experience may vary, but I've found them to be borderline useless to actually tune a set of duals. They're written with all the precision of a 1979 Car Craft article (they read like a Hot VWs feature). They talk about stuff and give "rule of thumb" settings for mixture screws, air-correction jets, etc., but don't tell you much of anything specific. They have 2 or 3 nice exploded views, but I have an internet connection and can get those views pretty much anywhere, so the books were read once and set aside.
I've read more about this than most people. I think I'm the only man alive that tried to read all 10 bazillion pages of the AF thread over on theSamba. Internet opinions are like belly buttons-- everybody's got one, but some of them are just wrong.
I've had an Innovate wideband for years (purchased before I knew I could get the same thing for 1/4 the money), and have used it to set the idles and mains. I was never very comfortable knowing what to do with the air jets, so I just left them. Carbs are pretty much fancy air-leaks, and I figured I had them as good as they could get without going to EFI.
My car always ran way better than 90% of the other speedsters I've driven, and always returned me at least 25 mpg, no matter what engine I was running. Some engines ran better than others, but there's always the interplay between the actual engine, the ignition curve, and the fuel mixtures.
There are known mechanical combinations that work well (the 2110 with an W120 and 1.25s, the 2332 with an FK8 and 1.4s, etc.). Most of us have one of them, and they're great. But the thing is, no matter what engine is back there-- it can be made way better by tuning it. That's ignition and fuel.
Most of us have a distributor we drop in and set to 30* advance at 3000 RPM. Idle advance is what it is, and we live with the curve. I'm convinced this is 75% of tuning-- getting the ignition curve right, and we just live with whatever we've got for a distributor. 009s are notoriously suboptimal, but most of us don't even have that-- we're trying to do this with a Chinese 009 (or 034) copy. I ran a CB Black Box on the twin-cam motor, and was blown away how much better a stable and customizable ignition curve was. People have no idea. I figured this was about as good as it was going to get.
Then I ran with Danny's Spyder. It was a revelation.
He's got the reputation as a "carb whisperer", and he's made a lot of guy's cars run a lot better by just working with one of the 3 (mechanical, spark, fuel) systems. But I'm convinced that his crank-fire setup was at least 75% of the reason his car ran as good as it did (with carbs, not EFI). His engine pulled hard everywhere, even with worn rings (in 2018). I'm not running his car down at all (far from it), I'm just pointing out that even when the engine was getting tired and needing a rebuild-- it ran well enough that 99.99% of the people out there wouldn't have touched a thing on it.
He played with his ignition curve for years, and optimized it with 50+ different maps. When I drove it last fall, it had a fresh rebuild but still had carbs. There were no holes anywhere-- none. There was no discernable transitions from idles to mains, no hesitation off idle, no stumble on an accelerator stomp. It was as close to perfect as I've ever driven. A 2332 has more outright power, but his car spooled up and took off in a way that most bigger engines I've driven have not. The power the engine has was everywhere. Total output is not the only measure of performance, and not even the one that matters the most.
I came home resolved to go to crank-fire, and I'm deeply "in process". The current plan is to eventually get to EFI, but to set up a Microsquirt system with ignition only to start, then go to injection once I'm completely comfortable with TunerStudio. Baby steps. The Black Box was a gateway drug, but it relies on a distributor pickup. I'm excited to dive into the world of dead accurate and dead stable crank position ignition pulses.
I really do think that we're all starting in the wrong place when we try to tune. Fuel really can't be set up optimally unless spark is worlds better than most people settle for with a fake 009. I'd encourage everybody to get a Black Box at a minimum.
Anyhow, I reread the John Connelly article Ed linked and it helped me get excited about playing with my carbs again once I'm up and rolling with 50% of the Microsquirt (the ignition).
John is a smart guy, and provides way more information than Tomlinson ever did with "the books", but he's not a writer either. It's going to take a season or two to optimize the carbs even with a better ignition. I've got every jet I could possibly need, and have the O2 sensor to get it done.
I don't even think I'm going to put the new engine in the Speedster this year. For once, I think I'm going to stand pat and stick with 'ol trusty (the 2110), and use it for the test-bed for the Microsquirt/crank-fire thing. I can dial in the carbs using John's article as a guide, and see how well I can make this run (and how comfortable I can get with the software) before I introduce a whole different set of variables with an engine needing a proper break-in, etc. The EFI part of the program will be last, and is probably several years away.
Anyway, read the article Ed linked. There's a huge amount of information-- and I can guarantee that if you use half of it, your car will run better than it ever has.