Machsupport Forum
*****VIDEOS***** => *****VIDEOS***** => Topic started by: chrisjh on October 09, 2013, 06:15:18 PM
-
I thought I’d share my experience with Mach3 Turn using Constant Surface Speed (CSS).
Attached is a short dissertation that could be used to bridge the missing information in the Mach3 Turn Manual.
I cover the use of G96 (CSS On) and G97 (Cancel CSS and return to RPM mode), as well as some tips on using G52 Offsets for gang tooling equipped CNC lathes.
Below is a video clearly demonstrating that CSS works well with Mach3 Turn, if you follow the rules outlined in my attachment.
The video also shows how I use Alodine 1200S to chromate convert aluminium parts. Please note that Alodine 1200S is not ROHS compliant as it is a hexavalent chrome. However, if you do not drink the stuff you will be OK.
http://youtu.be/2TsKji7VkYo
Enjoy
Chris Humphris
-
OK,
I've had a few inquiries regarding the chromate conversion process.
Attached is a procedure I have written which includes where to get Alodine 1200S in Queensland. Basically it is available from Henkel Chemicals worldwide.
Regards
Chris Humphris
-
Very useful information Chris, thank you for sharing.
Tweakie.
-
I had never tried CSS when using the parallel port but it has never worked correctly with any external controller. I asked many people who used the PP if it worked for them and at first all said yes it did but then when I asked a bit deeper then they discovered it didnt. Also I am almost certain (could dig and find it to be sure) Art has said it does not work correctly in G95 and it is impossible to make Mach3 work correctly with CSS/G95.
Can you definitely say that the feed per rev stays constant, ie whatever you have it set to. So for example going from large dia to smaller the axis sill speed up and keep the FPR constant? In all the tests I have done and all the tests others have done the axis will start off at its FPR but from that point on the FPR will actually be treated as the base point for FPM.
Hood
-
Hi Hood,
You raise some points I hadn’t considered. I have not tried CSS with a parallel port connection, as my lathe is controlled via USB Smoothstepper. So I can’t say for sure if a PP connection would work. However I can’t see why not, as my understanding of Mach3 (which is not great; my understanding not Mach3, that is) is that the Mach3 PWM Spindle Control stream is simply passed through the Smoothstepper and not manipulated.
I cannot say categorically that the feed per rev remains constant but it appears to do so, both for diameter changes along the X Axis and linear cuts along the Z Axis. I guess I could measure it but I am very happy with the way it behaves for me. In early days of development of my gang tooling setup, I found out that very large diameters slowed the spindle RPM to below the lower limit I set in Mach3, triggering the lower limit function of Mach3. The upper limit is set to 1500RPM, well above the 1200RPM limit set in my Danfoss VFD, so I never reach the upper limit.
FYI, I use Mach3 Version R3.043.056 exclusively for both my lathe and mill. The mill is PP controlled.
For the USB Smoothstepper, I use Plugin BETA Ver 0.015meg. I tried later versions of the Plugin, but reverted to BETA Ver 0.015meg when I found that they would not report the spindle speed to Mach3.
In summary, given that G95 and G96 is not documented in the Mach3 manual, I have always believed that I have been lucky to strike something that works, and that it was more luck than good management. I can’t recall why I even started down the CSS route, but I have vague recollections that I read somewhere that others had used it with success.
PS I searched the forum and found this:
http://www.machsupport.com/forum/index.php/topic,5503.0.html
Regards
Chrisjh
-
Well I would be doubly surprised if CSS is working correctly for you knowing you use the SS.
Should be relatively easy for you to see, if you start at a large dia and move in to centre you should see the FPR DRO decrease as the feed, when CSS is enabled, gets based on the FPR and RPM at that time.
Also if you start CSS on a large dia and then move in to centre and then without calling a further G96 just command the X to move back out to the starting Dia, I suspect you will see a hige difference between the time taken to move in and that of it moving out.
I tried searching the Yahoo group but couldnt hit on the exact post by Art (the search in Yahoo is really pathetic now since all the changes there, in fact the whole online Yahoo layout is now that way, it used to be difficult to follow things at times, now it is almost impossible). I did find some posts that mentioned Art had previously stated it was not possible to make Mach work correctly with CSS in G95 but that is all. will keep searching but look forward to your results if you can get some time.
Hood
-
Well managed to find some info on Yahoo, this was regarding a discussion where some said CSS worked, others said it didnt and basically it boiled dow to some saying it used to but not now, anyway heres what Art said on the matter.
Hood
here it is
---- Original Message -----
From: G_Mouer
To: mach1mach2cnc@yahoogroups.com
Sent: Monday, April 09, 2012 10:21 AM
Subject: [mach1mach2cnc] Re: CSS thread, Brian care to weigh in with your thoughts?
Is this one of those things that used to work way back when, and went funny over the different versions? Just curious.
Hi:
Ill kick in on this.. no , it was never different really. Due to the restrictions on the driver/interface doing proper CSS was
near impossible back then, so a spindle pwm contol doing CSS was the best I could shoehorn in at the time. Proper
CSS required things that a buffered trajectory path just wasnt compatible with. Lathes were not heavily used so dev time
was at a minimum. I know care is being taken to allow proper lathe control in future..
Thanks,
Art
www.gearotic.com
-
OK,
I had assumed that because it appeared to work for me, that CSS capability was an undocumented feature of Mach3. You have aroused my curiosity, so I performed some measurements.
This is what I did.
Setup some stock in the chuck and calibrated the X Axis.
Using the MDI, I issued the command G96 M3 S50
From the MDI, I requested G01 X50, X45, X40, X35, X30 , X25, X20, X15 in succession
At each of the diameters listed above, I recorded using an independent tachometer, the actual Spindle RPM, and the reported RPM from Mach3 DRO (S True)
Listed below are the results:
Ø (mm) Calculated RPM Measured RPM Mach3 DRO RPM Reading Tangential Surface Speed (Metres/min)
50 318.31 348 349 50
45 353.68 380 380
40 397.89 420 420
35 454.73 471 471
30 530.52 540 540
25 636.62 636 636
20 795.77 785 785
15 1061.03 1034 1034
Calculated RPM is from the formula:
RPM = Surface Speed (in Metres per Minute x1000) / π*Diameter
I repeated the measurements at a couple of random diameters and the results were totally repeatable. I also noted that the speed transition was very smooth, regardless of whether the diameter was increasing or decreasing.
I think the measurements prove that my combination of Mach3 and USB Smoothstepper is capable of at least approximating CSS. I am happy it works for me.
One thing that annoys and frustrates me is that at the first use of Mach3 Turn, CSS does not work as intended, and the spindle speed drops to below the Mach3 low speed limit. The workaround for this is to stop the program, rewind and start again. The lathe then behaves normally for the rest of the working day, even if I change part programs. I think this is a result of my initial conditioning code, and I haven’t solved this problem yet. I have simply learned to live with, and expect, this behaviour.
Regards
Chris Humphris
-
Excellent work Chris.
(Interesting about only working correctly after the second run).
Tweakie.
-
If I am understanding your above tests they are showing the RPM is close to what is required for a given surface speed. That really has never been in question, what is in question however is whether the feedrate adheres to the commanded feed per rev.
For example a feedrate of 0.3mm/rev at 100rpm should be also 0.3mm/rev at 1000rpm. In all tests I have done and others have done that is not the case and it seems Art has said that it can not be done in Mach3, certainly Brian has said that.
What actually happens is the moment of a feed being called in G95 and CSS the feedrate is then translated into a feed per minute and that is the rate that will ensue for the duration of the move and thus the axis will contnue to move at the same velocity. What should happen when using CSS with G95 is the axis should speed or slow to keep the 0.3mm/rev constant.
Hood
-
Hi All,
I haven't gone to sleep, rather been thinking long and hard, and performing some tests.
My tests (based on blind faith in Mach3's ability to report Feed per Revolution accurately in the F DRO in Mach3 Lathe) reveal that Constant Surface Speed is accurate but, as reported by Hood, Feed per Revolution appears to vary whenever the X Axis is in motion in CSS mode. I have concluded that, in my case, it doesn't really matter, but have proposed a couple of possible solutions to improve constant chip load.
My report is attached.
Regards
Chris Humphris
-
I’ve been thinking some more and decided to look very closely at some of my parted off material.
A true G95 motion should have a spiral with a pitch equal to the distance per revolution. (In this case 0.05mm)
The attached photo shows that the pitch decreases as the parting blade moves from the outside diameter of 38mm towards approx. Ø12.5, then reverts to a smooth 0.05mm per rev to the centre, when the VFD reaches maximum RPM.
Regards
Chris Humphris
-
Chris, I had hoped you had found the magic combo but really knew it would not be the case. Anyway until I junk the ESS from the lathe I cant really use CSS anyway as sadly the ESS and USB SS have issues with a step/dir servo for a spindle anyway.
Hood
-
Just curious but could you not just look at the Xdro as diameter and use that value as a percent of Spindle over ride.
When started it would see the X value as the base to start from then as you moved towards the center mach3 would adjust the Spindle overide to increase the speed same with moving out it would slow down.
Am I missing something ??
Just a thought, (;-) TP
-
Spindle override would be far too limiting, for example a very conservative surface speed, say 50m/min, so you start at 100mm Dia so RPM is 159rpm, now the max you could get with spindle override would be 250% or is it 300%? anyway we will say 300% so get to 10mm dia and you require 1592rpm but that 300% has limited you to 477rpm, so effectively the spindle override became ineffective at approx 33.35mm Dia.
Better option would likely be to use the Feed Override and compare to actual spindle speed (Mach does a decent job of controlling RPM in CSS) there is actually a guy that has done just this via a plugin, he posted recently on the General part of the forum,. It worked in my testing but I had some issues but more importantly for me the ESS has even greater issues in its spindle control which makes me not want to test further, roll on Mach4 so I can use the CSMIO on the big lathe and according to Brian CSS is done so this wont be a problem.
Hood
-
OK instead of using the SSO then just update the Sspeed dro then you have unlimited response up to max spindle speeds. I have a working model running in a macropump.
It can be started/programmed via a macro M96 as to on /off ,base diameter and Max spindle speed that it cuts OUT and does no further adjustments. Once started it follows as to the position of X being the control radius.
Interesting, (;-) TP
-
Hi all,
just for test purposes I have created a macro that calculates and set the speed and feedrate at the line where it is coded.
May be not free from issues
attached the macro and a sample G-code
#1 is surface speed in meters per minute
#2 is feedrate in mm per rpm
#7 is max rpm coming from the macro
alex