Well, there’s min power in machine config, and min power in the layer settings.
RF lasers produce output basically proportional to the duty all the way down. So min power in config is 0% (well IIRC I think it may require 1% be entered).
It does scale. As it goes through corners, the short area where it’s at like 5% of the commanded speed needs 5% of the commanded power to deliver consistent joules/mm.
If you used 100% min power, there’s no scaling, and it uses the max power all the time. This overburns in corners and vector start/stop and tight curves. Mostly no one notices overburn unless they’re only trying to score the surface. If doing through cuts, you might see the edge a little more blackened or slightly wider than it could be.
But, if you’re doing a lot of cutting detail and all, long term, then you’re using more net tube power over time than needed. This does add wear on HVDC excited tubes. RF tubes have little usage wear and it doesn’t make much difference.
Start speed IS kind of weird and annoying on the Ruida. IIRC I had to set it to 3mm/s. I wish it could be zero. If command 10mm/s, when it’s at 0.5mm, then it needs 5% of the max power, ideally.
When I set it to 3mm/s, that is usually fine. But in rare cases people have tried to cut really thick stuff super slow and that creates a VERY unexpected result. The machine runs up to the commanded speed almost instantly (since the target speed is very slow), but is then stuck at min power (basically no output) indefinitely because the target speed is below the start speed. That’s pretty undesirable nonsense behavior, and there’s no good machine config for that case. Even if I entered 50% for min power in machine config, that’ll be cutting at only 50% with 2mm/s in the layer, and then every other cut will have overburn too. The workaround is machine config has 3mm/s start speed 0% min power and if you cut at 2mm/s you put 100% min power in the layer settings. The times where power scaling is even used basically goes to zero because it takes almost no time or distance to reach just 2mm/s.
It does come up. It’s a powerful beam with excellent shaping and good air assist- it actually does very straight cuts all the way through 2x4 and other thick lumber. But it may need to go slow. It’s hard to predict how lumber cuts. Moisture content matters, and some wood types just won’t ablate into a long channel and just char, regardless of speed or attempting multiple passes.
IIRC, when I tried to set the start speed lower, that’s one of the things that made it do something very glitchy. Like, I do recall one config case that led to pressing “start” and the head would, every time, start moving very slow with no output, waste like 5 sec of time, then jump back and do the real cut. I don’t recall for sure if that was the glitch you’d see from setting start speed too low but I have seen a wide range of inexplicable, pointless behavior when playing with the Ruida config.
I’ll write a formal tutorial up on linearizing the Ruida on HVDC machines. It’s pretty simple to do at its basic core, actually. But without a full tech explanation with good evidence pics I suspect the thread would break down into people arguing over it.
Once you do some simple tuning, you can get power from “the low end” much further down, the output is basically linear, and the correct min power in the layer settings is usually 0%, as long as you are under a top transition speed. I used 400mm/s and almost no one ever vectors faster than that.
One of the tests to optimize it involves engraving copier paper with long runs and tight detail and corners, but NOT burn through or fail to mark on the long straightaways, and give a reasonably consistent brown shading along the vectors. With min power set to zero in the layer settings, even if max power is scaled back too.