Fine tuning Ramping

[sheldon@dingwallguitars.com]

Ron at Think and Tinker has suggested some ramping values to extend the life of our small diameter cutters. The confusion lies in translating Ron's suggestions into Shopbot values.

He is suggesting setting the acceleration rate to 1"/second/second. This to me sounds like a combination of ramp speed and ramp distance. Is that correct and if so, how would I calculate that?

The starting ramp speed and move speeds suggested are: .05"/sec ramp and .63"/sec move.

[gerald_d]

Sheldon, the 1"/second/second sounds like a very, very slow ramp. It means that it will take a full second for the machine to reach a speed of 1"/second - or 2 seconds to reach a speed of 2"/second etc.

But yes, I share your problem of converting figures like this to "ShopBot values". I have not yet come across an explanation of what the ShopBot "ramp" really is. Some ShopBotters swear that the ramp should be changed often (according to cutter, material, material, speed, etc) but we have left ours at the factory default and will only fiddle with it once we know what that that little figure really means in terms of distance/second/second (rate-of-speed-change). Maybe the SB-value only represents time-to-reach-full-speed?

An explanation from ShopBot would be appreciated.

[sheldon@dingwallguitars.com]

Thanks Gerald. My understanding from the manual is that the ramp distance is measured in steps. 150 if memory serves me correct. To access the setting, you need to open the ramp values setting, then close the dialog box, but before hitting enter type 6 commas then the desired number of steps. I've played with settings from 100 to 1000 doing aircuts just to see how it affects vibration.

Based on what you are saying, it sounds like the answer my question would be to adjust the ramp distance to accelerate from 0 to .63" in .63 seconds. Is that correct?

[gerald_d]

A "ramp" is a straight line incline/decline. When used in the terminology of a "ramped speed" it tells us that the speed increases/decreases with a uniform rate - i.e. a straight line ramp. In physics terms, the rate of speed change is constant, or the acceleration/deceleration value is constant. From my very rusty school physics:

t = time (seconds)

d = distance (inches or number of motor steps)

v = speed or velocity (inches per second or steps per second)

a = acceleration (inches/second/second or steps/second/second)

Where we have a constant a value (i.e. a straight line ramp) the following formulae apply:

v = a x t ............................(1)

d = 1/2 x a x t2........................(2)

v2 = 2 x a x d ...........................(3)

In your example, we have v = 0.63 and a=1.0. From formula (1), t= v/a = .63/1 = .63 seconds, as you has already worked out above. But, if we calculate the distance, using formula (2), we can see that it is 1/2 x 1 x 0.632 = 0.2 inches. (If you want to go at a higher speed, and still use the suppliers's recommended a value, you can see that the distance gets very, very long)

And, if the ShopBot cuts two line segments that are in line with each other, immediately after each other, there is no need to ramp down to zero speed before starting the next segment - and the SB does not do this, it goes smoothly at constant speed from the one segment to the next. I mention this so that we realise there is more to this ramping function than meets the eye.

[gerald_d]

When all else fails, read the manual - should have done this before writing most of the above! Have found the manual here at home and discovered the following:

ShopBot has explained how the ramps work, but the terminology confused me for a while. The VR command requires, firstly, that you select a speed from which the ShopBot will increase the speed. To my mind, this base speed must surely always be zero???? I mean, what happens below the base speed - does the SB jump instantly from zero to the base speed and then gently increase up the set move speed? This is not logical.

Secondly, the VR command wants a distance, measured in pulses, over which to go from base speed to full speed. If one is trying to maintain a fixed acceleration value (as recommended by a supplier) one must change the VR settings every time one changes the final speed settings.

Acceleration in terms of inches/second/second is perfectly logical way to define a ramp, that is not sensitive to any other settings - I can see why why a bit supplier uses that method, and cannot see why SB doesn't.

(Another thought - it is meaningless to have a ramp to protect the bit, if the bit has not entered the material yet, or has to cross a groove already in the material.)

[bob buttons]

i once in a while use a 1/32 bit for fine detail work for various projects. i usually cut at
.70 MoveSpeed . they do break with the shopbot factory ramp speeds alot. but my bot runs smoothly. i notice with ramp speeds altered from the factory settings my machine doesnt run as smoothly. especially when jogging around the table. there is a very good chance i put in bad ramp speeds in my trial and errors creating the less than smooth performance but i would like to add some ramping to help when i run the small bit for detailing. what would be a good ramping speed setup?

[gerald_d]

This morning at the shop, I tried to test my theory that the "ramp speed" (i.e. the speed from which the ramp starts) should always be zero. Set everything to zero and the software crashed completely - the SB will not start at all! We are now reloading the SB software.........

Would really appreciate someone from ShopBot telling us how the ramps really work. What happens before the "ramp speed" is reached?

[gerald_d]

Our crash may have been a "normal blue Monday" disk crash - coincidence that it happened with the ramp experimenting. But this did make us go to the SB website for the latest software to discover that version 2.35 for PRT's and 2.34 for older SB's was freshly put up on Friday.

[gerald_d]

The reported "crash" was not as severe as I thought. Putting zeroes into the ramps puts a line into SHOPBOT.INI that trashes the .INI file and stops the SB from running. This line in SHOPBOT.INI needed to be restored with a text editor:

VR, 30, 30, 45.001, 60, 45.001, 45.001, 150, 250, 100, 12

But here we see 10 parameters after the VR, while our quick ref card only allows for 8.

And, if we reduce the ramp speed values to something close to zero, we get a very uneven ramp-up (judging by the sound).

Wouldn't it make a lot more sense to enter on;y the 0.5's and 0.6's that everyone talks about and that the SB does the rest of the calc itself?

[sheldon@dingwallguitars.com]

Bob, what rpm are you using for your 1/32" bit? .7"/sec sounds pretty high for a feed rate. I would contact the bit supplier to find the ideal feed rate for the material, rpm and cutter design.

Most information I've read recommends setting the move ramp speed at 60% of the move speed. I find 40% seems a little smoother, but I'm re-examining all my ramping values right now.

I don't know how others do it, but I insert move and jog speeds and ramps in every part file.

e.g.
VS 1,,,3
VR .4,,,1.2

This way you can drop your jog speeds down if you're using small diameter bits with short jogs between cuts to reduce the cutter head shake at the end of the jog, then speed the jog back up to jog home.

By the way, Think and Tinker is now selling precision machined 1/8" collets for Porter Cable laminate trimmers. I've got one on order, I'll post if it lives up to it's ad claims.

[Ted Hall, ShopBot Tools]

Well you guys have certainly gotten me confused!

Here's the deal:

1) ShopBot (DOS) ramps are linear changes in speed from the start ("Ramp") speed to the final move speed [Yes, this is just what it says in the manual]. This is not the best way to do it, but 4 years ago it made sense, given what we had to work with.

2) When the profile is linear, you do not get pleasing ramps when starting at zero speed for anything over about .25in/sec move speed because too much time, relatively, is spent on the early part of the ramp. As you get to higher speeds, having a start speed above zero approximates a curved acceleration profile (which is better). That means, the faster you are going, the higher you want to set the starting ("Ramp") speed up to about .9in/sec, the point at which you would really be trying to get moving too quickly.

3) Remember that we are working with 'steps', not smooth continuous movement, especially at lower speeds. This is one of the reasons the ramp distance is specified in steps, not duration. Let's assume for simplicity that we had an X unit value of 250 (meaning 1000 steps inch) and that we are using a move speed of .6, a ramp of .2, and a ramp distance of 100 (the first parameter after the ramp values in the command). The average speed for the ramp distance will be .4 and the time it takes to cover this distance will be 1/4sec (100steps at 400steps/sec). This would be an acceleration of 1.6"/second/second as I see it. By doubling the length of the ramp, we would double the time and thus cut the rate in half.

4) Now that I've confused even me. Here's the rule of thumb that seems to work. For speed below about 1.4, set the start ("Ramp") at about 60% of the speed. For speeds above, set the ramp and .8 and don't worry about it. The distance can be lengthened a little to soften the ramp, or shortened to tighten up performance in certain busy 3D files. It's often best just to play around a little with the values to get the smoothest motion for the type file you are running.

5) The new software has ramping that has a curved profile. This makes a smoother start, and is distinctly smoother at the peak of the ramp in the transition to the final speed.

[gerald_d]

Hi Ted, your example in point 3) is a good reference to illustrate my confusion. If I want to double the move speed to 1.2, I have learnt that I should also double the "ramp" to 0.4 - i.e. if I want to keep a fixed ratio between move speed and ramp as per your rule of thumb. Plugging those figures into your calculation, the time taken to get up to full speed is only 1/8sec and the acceleration is a huge 6.4"/sec2. On paper, the ramps get much "rougher" as the speed settings increase!

However, what happens on paper, and what happens in practice are often two different things, and then it is best to discard the paper.
You seem to have discovered a simple formula which works for SB's and that is fine (your point 4), but can I suggest that this is pre-programmed into your end of the software? Now we have the situation where the SB operators are supposed to enter a new "ramp" value every time they change the speed value. The ramp is a bit of a fuzzy concept and it is tedious to keep fiddling with it.

If your "rule of thumb" is pre-programmed, you can give us experimenters an off-screen facility to fiddle with things like the 1.4, 60%(.6), 0.8(80%) and distance. (There would then be no need for a permanent display of the "ramp" on the screen). I reckon this will reduce your service calls and save you from having to make long posts on the Forum
.

[gerald_d]

Maybe this little sketch helps:

Ted took a move speed of 0.6"/s with a ramp of 0.2"/s (being 1/3 of the move speed). I played devil's advocate and doubled the move speed to 1.2"/s and also used 1/3 of that as the ramp - 0.4"/s. So, we have a ABC profile and a DEF profile. Both work smoothly if we use the rule of thumb.

But, if DEF worked smoothly, then DGE was okay and therefore DG was okay. Why can't we use DG to get up to the speed of 0.6"/s?




4898

[mikejohn]

I have been studying this thread for days, and I am confused.
I am a very newby, yet to take delivery of the SB.
Assume I am using a small bit,say 1.5mm (1/32) and I plunge it 4mm into MDF, then I wish to make a groove at this depth. It doesn't matter what my move speed is, surely I have to start from zero?
I have to accelerate to my cutting speed at the start of the cut. I cannot instantly go from 0 to 30% of my final move speed, the 30% being my ramp start. I'm obviously missing something.
But what?
Mike

[gerald_d]

Mike, just pretend you havn't seen this thread. We managed for two years without changing the ramps - until Sheldon came along and started this thread.

[rgbrown@itexas.net]

"Assume I am using a small bit,say 1.5mm (1/32) and I plunge it 4mm into MDF, then I wish to make a groove at this depth"

If you are using a "1.5mm" bit it will be close to 1/16" and you might break it with a cutting plunge of 2.7 X bit diameter. If you are cutting a grove 5X bit diameter, and are successful, I want to know your bit supplier......

[gerald_d]

Ron, how did the 2.7x become 5x ?

[papadaveinwy]

Gerald I THINK Ron was meaning the depth of plunge was 2.7x and the length was 5x Hypothetical. David in Wyoming

[mikejohn]

As I say, I'm a newbee!
I cant even convert metric to imperial in my head!!
I will want to use a 1.5mm bit to cut grooves, but only 2mm deep. Am I going to have problems? Will I need to make more than 1 pass?
Is there a rule of thumb for cutting depth against bit diameter?
Sorry about going off thread (I would put a sad face here, but I still can't work out how to do it

[mikejohn]

I did what Gerald told me, and it worked
:o

[rgbrown@itexas.net]

In "English"
1.5mm = 1/16"
.75mm = 1/32"

""Assume I am using a small bit,say 1.5mm (1/32)"

4/1.5 = 2.66 or ~ "2.7"
4/.75 = 5.33 or ~ "5"

Strength of bit will be related to the relative "squares" of the diameter while stiffness will be a "cube" factor.

What are the numbers? Is it 1.5mm or is it 1/32"???????????????

[gerald_d]

Hi Mike, you have caused what is called "thread drift" and we have gone way off the topic of "Fine Tuning Ramping". Actually, I welcome that in this case, because I think that my logic and graphs were causing more heat than light.

On the millimeter bits, I see now that the 1/32 from your first post caused some of the confusion - and at that stage you wanted to go 4mm deep. More recently you are talking of a 1.5mm bit going 2mm deep in MDF. My rough rule of thumb is that the plunge depth should be no more than 2x bit diameter in softwood and 1x in hardwood. This is just a personal choice and the "limits" can always be stretched.

(you were looking for the colon and the capital O to give you
)

[mikejohn]

The conversion between metric and imperial was a mistake on my part.Sorry!
One further question concerning ramping.
If I cut a groove 4mm deep with a 20mm bit, then cross this groove with a 1.5mm bit, 2mm deep, what will happen to my bit when it meets the mdf on the other side of the groove? Is this where ramping comes in?
Mike

(good fun this)

[gerald_d]

Mike, the programs are not intelligent enough to know that the bit is crossing a groove. (see the last line of my Sunday, March 16, 2003 - 12:11 pm post). I dont think that anyone bothers to program in a ramp for this situation.

Because the bit is circular, it does enter a straight edge progressively - over a distance equal to the radius. But now we are really getting down to nit-picking.

Ramping is not such a big issue in the daily life of a SB'er. If you really want something to look at while you eagerly await your SB, figure out how you are going to spin those tiny bits at a decent speed.......
:-) :^) <---- older-style, sideways smiles.

[mikejohn]

Er! Um! Spinning tiny bits?
Do I have to reconsider my Router purchase?(actually, I still haven't worked out which router to buy!)
Now we are way of thread

[gerald_d]

Go here (http://www.talkshopbot.com/forum/messages/29/852.html?) for an intro on driving small bits. And here (http://www.talkshopbot.com/forum/messages/312/1256.html?) for some banter on routers.

[rgbrown@itexas.net]

Speaking of Off thread, I believe Mike also has confused "APPROACH" with "RAMP".

RAMP is the time/distance it takes for the machine to go from "stopped" to "Cutting speed".

APPROACH is the entrance procedure for plunging a bit.

[mikejohn]

OK! chaps.
Where do I find a glossary of CNC/routing terms?

??

[pcampbell86@charter.net]

Hi Mike Paul Campbell here
i just cut my first sheet of parts with the shopbot and hey its a very big deal getting your Z
zero correct in the use of tabs and much more difficult to get a totally flat table than people seem to say aroun here
...and thats critical if you are using tabs for small parts and still want to cut right through the work piece without eating into your bleeder board....as for what router to use for 1/16 bits i really would consider and i mean this...a dremel rotary tool for such tiny parts..and no dust collection system as you dont want your parts sucked out of space
..like i did first try..hehe..well just thought i better say somthing..here

[designer@motorhome-world.com]

Hi everyone.I am running a PRTalpha 120x60 and seem to have problems running it at speeds exceeding 4.5inches/sec. Does anyone have a suggestion for a possible solution?

[gerald_d]

Hi Gakiem, my neighbour from Cape Town,

I think that you will have a much better response in this thread (http://www.talkshopbot.com/forum/messages/26/4760.html), so I took the liberty of moving the discussion there.