Hi,
in this forum we often talk about SSE, which I currently use on the pantograph with Mach3 (it has always worked great!).
I use Mach4 on the lathe with CSMIO IP S (again super satisfied).
Since I like Mach4 a lot, I decided to use the Mill version on the pantograph as well, as SSE is fully supported.
I make a brief consideration.
I use the same computer for both the lathe and the pantograph.
Both being ethernet, I added a PCI ethernet card.
Obviously if I use the lathe the pantograph is off and vice versa.
I have been using this system for 3 years without ever having any problems.
let's get to the point.
I don't really like the SSE + Mach4 combination.
Unresponsive jogging movements, with a slight delay.
Even worse when running gcode, if you increase the Feed slide (especially in 3D files) the machine shoots very abruptly (with Mach3 the same g-codes work without problems).
I immediately asked myself this question, where am I wrong ??
I point out that I have the exact same motor configurations, pulses and parameters.
Thanks in advance for your answers.

That is not my experience with ESS and Mach 4.  You should be able to control the buffer depth in the ESS plugin.  Contact Andy at Warp 9 and see what he has to say.  That would be the first place I would start. 

I have two machines in my basement using Mach4/ESS and they are both flawless, responsive, and NEVER give problems.  I keep an ESS on my desk to test with as well )Two thee currently.  And if you sneeze at the shop, you will have hit an ESS somewhere.  There wouldn't be that many ESS boards laying around our trenches if they didn't work well. 

Steve

Hi,
in fact I didn't say that the ESS works badly.
I want to understand where I could go wrong.
in the past I have made surveys for about 10 continuous hours of work without missing a beat (here I am referring to the use of Mach3).
let's start with a question.
what value do you use of "Look haead"?
Are there any videos where ESS works with a 3D relief? with good speed like F2000 / 2500? (I am Italian I use mm / min).
if i upload gcode to google drive or dropbox, would you like to upload it to the machine and then make a video?
specifying the parameters of speed and acceleration in order to put us in the same conditions?
regarding ESS frequencies I left the default ones.

I also tried the same gcode on the "Rosetta CNC" controller on the new cnc I'm building to see if he was the problem.
as you can see from the video I reached F6000 (mm / min) without batting an eye.
https://www.youtube.com/watch?v=ebd8xqGlsh8
Obviously I cannot compare Rosetta with ESS.
They are two completely different hardware, even in price, (Rosetta B costs more than 1000 euros) so the challenge is not the same.
Mach4 and Rosetta have different strengths (I'm talking about software).
Mach4 Mill is more suitable for mechanical parts, where it is possible to obtain synchronisms by wire.
With Rosetta (at the moment) this is not possible. Its strengths are rotary axes, supports RTPC, virtual machine, etc. Etc.
Returning to the purpose of the post, I am fascinated by Mach4! and I absolutely want to use it on the pantograph too !!

20 lines of lookahead mostly.  Yes, even for 3D work. But I mill steel and aluminum on a box way mill mostly.  So now your question has me wanting to ask questions.    What are you wanting out of the lookahead?  A smoothed profile that doesn't hold tight to the path?  Or do you want something that holds as tightly to the path as it can?

Either way, the lookahead will not have anything to do with your jogging issue.  That sounds like a buffer depth issue to me.  And when I say buffer depth, I mean the amount of time that is buffered in the motion controller.  If you have a large motion buffer, changes in the feed rate will be delayed.  And the start of the profile may be delayed as well, depending on how the motion controller implements the buffer. 

ESS ships with very conservative settings.  I can't remember what exactly off hand, but I think it is something like 122 or the like?  The cycle time for the ESS is 1 millisecond.  So that is over a tenth of a second in the motion buffer!

But for Mach 4 on a decent computer, you can get by with extremely small motion buffers.  Say like 30 milliseconds worth.  Say the motion loop is fired ever 10 milliseconds.  I like to stay "three mistakes high" (an old RC model airplane saying) so 3 x10 = 30 milliseconds.  With a small buffer like that, feed hold appears instantaneous.  Jogs start appear to start instantly.  The whole system seem to respond to human input "instantly", even though it really isn't.  Any buffer depth less than 100ms is hard for use humans to detect.  Personally, I like to stay below 60ms.  So 122ms of motion delay is going to be noticeable.  However, it will run on even the crappiest of PCs.   

So if you left the default setting for the motion buffer, I bet that is you jog issue.  Maybe it will solve you other issues as well. 

I have milling machines and lasers.  No routers.  I just like milling metals and catching things on fire.    So unfortunately, I can't run a part for you.  But I know there are plenty of people that do run 3D reliefs with Mach 4/ESS. 

I saw your Rosetta machine video.  ESS and Mach 4 can certainly do that on that hardware.  No problem at all.  I need to find a video of us running trochoidal tool paths on a machine at IMTS one year.  We kept up with the Hurcos!  It was amazing to watch.  But I have seen ESS do that plenty of times.   Now, will your pantograph machine move like that Rosetta?  I would like to see a video of that G code running on the pantograph machine.  It may shed a lot of light on the issues at hand. 

But check you buffer setting first!  it may be the answer. 

Steve

Waiting for some answers, this morning (I'm Italian, we have a different time zone, I also use a translator) I did further tests.
Look ahead value 200.
Buffer 0.18 (default value).
Frequency from 40 (default value) I went to 200.
I have loaded a 2D Feed 5000 roughing.
On linear lines, everything is fine!
When it finds the G2 or G3 bows, it sometimes takes very hard hits !!
I had to stop right away! the blows were so violent that the pantograph trembled (I don't want to run the risk of damaging screws or supports).
The computer is a gaming PC (played Tomb Raider in medium / high quality with no problems) so amply sufficient for Mach4.
In CV mode I did some tests, right now I don't remember if it was active during the tests.
I'll try making videos over the weekend, then uploading them to Google Drive.

Attention: at the first sudden click I stop the machine immediately.
If you have any videos, I will watch them very carefully!
I want to use Mach4 also on the pantograph!
I like it! and I want it!
note: I am using the Mill version on the pantograph, I don't think this is the cause.

I forgot to say that at low values of Feed (1500/1800) everything works fine.
automatically comes to think of two things:
1- wrong frequency setting.
2- poorly suited or insufficient hardware for Mach4.
From what I understand, Mach4 is powerful as software and requires high performance hardware.
I sincerely hope it is the first cause.
At the moment my finances are all busy completing the new CNC.

Hi,
I have used an ESS for six years and in the last year or so, daily.

The ESS settings for look-ahead and  buffer are at default, 180ms. You can notice, just, FeedHold delays and FeedRate delays but are so insignificant I don't
worry about them.

I run a dual core Atom single board PC with on-chip graphics. That must be about the lowest possible powered PC to run Mach4......and it runs seamlessly.
The only time I really notice the 'gutless' little PC is when I load a largish file....it can take seconds or even minutes for really big files. Once the file is loaded
it runs perfectly. I have had three 'Run Out Of Data' issues in six years, two of which were related to electrical interference by the spindle, and the last
one was over the summer when the temperature peaked at 38C....and the little Atom struggled,it has no fan at all.

From what I understand, Mach4 is powerful as software and requires high performance hardware.

That is contrary to my experience. The bottom line is that Mach4 IS NOT  a CPU power hungry application and even very low power PCs will do a good job with an ESS.

I'm building a new mill with servos and its capable of G0's of 25m/min and max G1's of 15m/min. The Atom/Mach4/ESS/BoB do the whole lot easy.
My steps per unit (mm) is 1000. Thus to get to 25m/min the max step rate is 416kHz, so I used differential signalling to the servos. The ESS is capable
of 4Mhz, ie eight times faster than I require!

Craig

@joeaverage
ok thanks for the info!
now I know the reasons!
I was convinced of the opposite! the site specifies the minimum PC performance, does not say anything about very fast PCs and possible contraindications!
In my case there is no point in replacing the PC, so I reluctantly stay at Mach3 Mill.
I will continue to study Mach4 Mill in parallel because I like it a lot! I will go slow to avoid problems!
Maybe later, when I get out of the costs of the new CNC, I recover an old PC to use only for the pantograph with Mach4.
So now that we have discovered the cause, I don't need to make videos anymore.
As always thanks for your availability !!
GOOO !! GOOO !! MACH4 !!!!

I searched and searched again!
I have not found any video of Mach4 + ESS working with a 3D relief!
Do you have any links?