What's new
What's new

Block Processing Speed On Older Controls? 2005 MiniMill


Nov 26, 2019
So I've been trying to do some single point forming work on some 3d prototypes. This of course means that since I'm not removing material, I should be able to run higher feeds rates without worrying about tool breakage :p

However, my 2005 Haas Mini Mill seems to start stuttering at around 10 blocks per second. Is this really the max processing speed of a 2005 generation control? The feedrate at this processing speed is about 20ipm, MUCH slower than I want to run.

I've gone through the settings, and everything seems be factory stock. Corner rounding is set to 0.025, finish is 'medium', etc. G103 (lookahead... if it exists?) is set to unlimited....or I at least call G103 without a parameter. If I do call a parameter with G103, I get a divide by zero error.

If I want to set my smoothing to .1 in CAM (with a regular tolerance of .0001), then I can of course up the feed rate a little bit, but it's still limited to around 10 blocks per second.... and of course a smoothing tolerance of .1 is an absolute joke.

In a few manuals from its era, Haas still claims 1000 blocks per second (found in a VF series 2004 manual).

Ultimately, if a new machine can process this code at a true 1000bps, or hell, 500bps, then I won't have an issue upgrading. I just want to be sure I don't throw $120k at something that won't solve my problem.

I contact my local HFO inquiring about the processing speeds of 2005 controls, but they had no idea (unsurprisingly).

So, at last, I have wound up here to see if anyone has any ideas... Thanks!
Pretty sure lookahead is good for at least 1000 blocks per second historically, but you'd need to be sure that high speed machining is active to run like you're expecting. Even the new machines sandbag a little bit on complex geometry without that option active.

It would help to see a snip of your code, too
Blocks per second does not equal lines per second. There is some sparse documentation on this if you look. The real problem is partly lookahead, which is real low without hsm.
Blocks per second does not equal lines per second. There is some sparse documentation on this if you look. The real problem is partly lookahead, which is real low without hsm.

The latest documentation uses blocks per second and lines per second synonymously. It might be different when you're talking about comments or blank lines, in which case BPS would be even higher.

"During machine motions, the control prepares future blocks (lines of code) ahead of time."

The second part of your statement is very true.
Haas has some old docs that explain it. I really think it means this: 1 ms per block means that for example, 100ipm = 100 / 60 / 1000 = .0016", which means the control has that resolution at that feed rate. @coyoinu, ask your mocon programmers if this isn't so, assuming you still work for haas.
Hey folks,

Thanks for the replies.

It seems I was under the impression that some level of lookahead was always on the machine, regardless of HSM being enabled or not (I thought I recall HSM enabling lookahead to 80 blocks, from 40).

This of course is a secondhand machine I've had for a couple years, but turns out the HSM trial was never activated. I turned it on, and sure enough the difference was night and day. It still does stutter here and there around 300ipm, but that makes much more sense than the previous stuttering at 40ipm.

So are the newer controls any 'faster' when it comes to processing speed compared to this 2005 control? Or would I expect the same level of performance if I were to buy a new VF3 from Haas today?

Also, an odd side note.... I get a divide by zero error on any program when I try to set setting 191 (default smoothness) to "rough" from "medium". Anyone else had this issue?