TOS FN20 Upgrade to CNC

[Ark1]

When setting up the Gecko 320x installing a jumper between the Error Reset terminal 5 to the + Terminal 7 allows the servo to be checked and tuned. When commisioning the drive and Servo the manual requests that the armature connections not be made and that the servo motor should be turned by hand with the encoder connected and energize to put the servo drive into alarm and fault. Alarm is any encoder count of less than 128, at greater than 128 the drive goes into fault. Seems simple enough and it worked flawlessly on the 31st of December 2010. The day here in Canada was +60 degrees and my shop was quite warm with everything working and tested I decieded to run a small facing program written by Mach3 addons and checked the speed of the axis and all worked very well. The drives did what the manual said they would with the base setting of PID at 11 oclock. A few days had passed, the temperature in the shop now about 40 degree I set up to run the facing progam, chucked a part in the vise, loaded the GCode program and hit cycle start. A few stops, starts and then the table headed for the stops as the cutter head was turning and the part was ruined. I lowered the table, and tried again to run the program and again the table ran to the stops.What had changed, temperature of the machine and friction. The TOS FN20 uses gibbs and ways with a film of oil between them. My shop was colder than when I set up the program to run.

What I learned. If the lost steps occure and are greater than 128 the Gecko 320x goes into fault. The drive now signals Mach3 to reset the axis to zero and then the program continues to run and accepts the new zero. Well if the machine has alread moved 8 inches on a 13" table and is told to move 8 more inches the table is heading for the stops at the feed rate written in the GCode for the part to machined program. Not good.

With this found and it took me a few hours to figure out what the hell was happening, I started this post.

Mach3 Fault when Drive faults. - CNCzone.com-The Largest Machinist Community on the net!

Many thanks for the responses and I was prepared to purchase a Gerc board. A quick email to Marcus Friemens at Gecko informed me that the drives I had could have all the Error Reset terminals connected together and this would hault the drives, I would then have to manually reset the drives to continue. Let us face a fact here that 129 steps out of 61,000 will not produce an error in the part that I will see. I could live with that. However more research led me to read about this product.

Page Title

An encoder board that will hault the machine and signal the offending drive.

I have now gone full circle as I had installed the Rodgers Board to use as DRO manual machine with Mach3 displaying the axis posistion when the need arises.
Currently that board is not connected to the encoders or the software loaded for it's use. It is on my to do list, which will require more hard wiring and print updates and more testing.

For now I will be completing the run of parts that I have on back order and then when time permits make the chances to use the Rodgers Board that will allow me to use the hand wheels to posistion the axis and keep the DRO in Mach3 alive and counting.

With only 1 Milling Machine I still have the need to turn a hand wheel from time to time.

I supose the Z axis could by reworked with a larger servo, power supply and drive to increase cold shop performance. For now I will work with what I have being mindful that a cold machine can cause problems with increased friction and loading of the servos and drives.


Ark1

TonyK.

[Ark1]

The trouble with building all of this is that at times an problem will crop up that was missed in the thought of design.

I was milling an 8" x 1/8" slot for an O ring seal to make a vacuum chamber for degasifing a mixed rubber compound. Using an 1/8" end mill and my spindle speed at 2500 rpm and 3" a minute feed and the slot was progressing well.

Then the spindle stopped but the X axis continued to travel and broke my new end mill. Missed that one. Interlock the spindle with Mach to stop travel. After a bit of trouble shooting it was apparent that I had burned out the 1.8Kw spindle motor. The shop was cold, the belts tight, the oil in the gear box thick and the overloads set just a bit on the high side to compensate for the 600 Volt dynamic Phase Converter inbalance. Where to go from here. I did not want to get the motor rewound, so I opted to purchase a new 3 phase 240 volt motor and VFD. The unit I purchased was from Transwave Converters in the UK. One of the problems was the limited amount of space for the Spindle motor. The frame was metric 90L. The VDF is a Jaguar Cub and I bumped the motor up to 3 Hp or 2.2Kw.

The motor is a universal foot mount unit that allows the feet on the motor to be placed on any side of the motor to allow for clearance of the Junction Box. I tried installing the motor with the JB on the top but was forced to remove the motor and remount on the mounting plate due to clearance issues.

When I ordered the VFD I also ordered a remote start stop station. I have incorperated this in the Mill Control station as I still want to be able to use the mill in manual mode for small jobs. By doing this the remote mill station comes with a relay on the Jog selector button that will allow me to use the dry contact on the Sound Logic Board and have Mach3 turn the spindle on and off. In run mode I can use the VFD station to start the spindle when I use the mill manually.

I have followed the IMO drive manual and have installed the VFD in the control cabinet of my Mill as instructed. I am hoping noise will not be an issue. Time will tell when I start the mill and run through the comissioning steps. The VFD also has a fault system that has a set of dry contacts that I will put in the Mach3 EPO circuit. If the drive faults it will signal Mach3 to stop the axis feed and hopfully not destroy a part in the process or break a tool.

I am sure glad I made prints as I wired up the mill as this change is simple now to do. Just review the prints, change the drawings, implement the changes and proceed to commissioning.

Here are the current pictures of where I am.

Just a mention here. Pete Moss of Transwave Converters was most helpful with this purchase and sent me links to review as well. Shipping from the UK to Canada was only a day! and cost $140 by DHL. Shipping weight was 60 pounds. DHL does not charge brokerage on items less the $1600 only a $7 handling fee to clear the package into Canada. My package left the UK at 3:00pm was in Cincinati by 6:00 AM the next morning and at my door in Grimsby Ontario Canada by 4:00 pm the same day. The drive package with remote station and motor was $637 CND.

Ark1

[diyengineer]

At least now you have a new motor One less thing to worry about failing later on haha. Loooks goood!

The trouble with building all of this is that at times an problem will crop up that was missed in the thought of design.

I was milling an 8" x 1/8" slot for an O ring seal to make a vacuum chamber for degasifing a mixed rubber compound. Using an 1/8" end mill and my spindle speed at 2500 rpm and 3" a minute feed and the slot was progressing well.

Then the spindle stopped but the X axis continued to travel and broke my new end mill. Missed that one. Interlock the spindle with Mach to stop travel. After a bit of trouble shooting it was apparent that I had burned out the 1.8Kw spindle motor. The shop was cold, the belts tight, the oil in the gear box thick and the overloads set just a bit on the high side to compensate for the 600 Volt dynamic Phase Converter inbalance. Where to go from here. I did not want to get the motor rewound, so I opted to purchase a new 3 phase 240 volt motor and VFD. The unit I purchased was from Transwave Converters in the UK. One of the problems was the limited amount of space for the Spindle motor. The frame was metric 90L. The VDF is a Jaguar Cub and I bumped the motor up to 3 Hp or 2.2Kw.

The motor is a universal foot mount unit that allows the feet on the motor to be placed on any side of the motor to allow for clearance of the Junction Box. I tried installing the motor with the JB on the top but was forced to remove the motor and remount on the mounting plate due to clearance issues.

When I ordered the VFD I also ordered a remote start stop station. I have incorperated this in the Mill Control station as I still want to be able to use the mill in manual mode for small jobs. By doing this the remote mill station comes with a relay on the Jog selector button that will allow me to use the dry contact on the Sound Logic Board and have Mach3 turn the spindle on and off. In run mode I can use the VFD station to start the spindle when I use the mill manually.

I have followed the IMO drive manual and have installed the VFD in the control cabinet of my Mill as instructed. I am hoping noise will not be an issue. Time will tell when I start the mill and run through the comissioning steps. The VFD also has a fault system that has a set of dry contacts that I will put in the Mach3 EPO circuit. If the drive faults it will signal Mach3 to stop the axis feed and hopfully not destroy a part in the process or break a tool.

I am sure glad I made prints as I wired up the mill as this change is simple now to do. Just review the prints, change the drawings, implement the changes and proceed to commissioning.

Here are the current pictures of where I am.

Just a mention here. Pete Moss of Transwave Converters was most helpful with this purchase and sent me links to review as well. Shipping from the UK to Canada was only a day! and cost $140 by DHL. DHL does not charge brokerage on items less the $1600 only a $7 handling fee to clear the package into Canada. My package left the UK at 3:00pm was in Cincinati by 6:00 AM the next morning and at my door in Grimsby Ontario Canada by 4:00 pm the same day. The drive package with remote station and motor was $637 CND.

Ark1

[Ark1]

Just a short post if you are looking for a manual for the TOS FN20 or other older machines check out these links.

http://www.lathes.co.uk/manuals

http://www.lathes.co.uk/catalogues

I stumbled upon it yesterday and have found a manual for my machine the TOS FN20. This vendor has hundreds of manuals for older machines and if you have a manual for an older machine contact Tony Giffiths as he is interested in obtaining manuals and blue prints for older machines.

Ark1.

[GM1G]

dude your mill and build are about a million times more awesome than mine

im totally jelly nice work sir

[Ark1]

Thanks for the kind comments.

I am not sure if I will ever master the true extents of G code and programing. All of this is just one big lesson that have been a delight to try.


Ark1

[Ark1]

I shot a video of the spindle working now on the VFD and how I installed the controls for manual use of the mill or auto under Mach3 control.

In auto Mach3 closes a dry set of contacts that are across the jog button.

Since jog does not use a seal in contact as long as Mach3 keeps the contact closed the motor will run. Simple yet very functional.

See this video Link.

[nomedia="http://www.youtube.com/watch?v=E5TstpkDdIQ"]YouTube - IMO Jaguar Cub VFD[/nomedia]



Ark1

[Zygoat]

looks nice, seems to work well.

[Ark1]

Now that I have my machine working I have taken my notes and scratch drawings and have made the following video on how I made the drawings.

[nomedia="http://www.youtube.com/watch?v=KYkcIJb0g04"]YouTube - EWD Drawings[/nomedia]



Ark1.

[Ark1]

I have read a bit about Cam software.

Where to start and what to do with it, then there is the price of the software.

Reading in the forms I down loaded the trial for CamBam and as programs go and lack of knowledge present's it'self, I set about making a simple manifold flange. Following through the tutorials and moving back and forth to my project manifold flange I finally came up with a finished project and created a G code file and watched the tool path in Mach3. 750 lines of code.

The intent of all of this is to look for problems and of course I found them.

The ultimate test would be a flange in steel or stainless steel so the cut is only .015" per pass. The test metal is aluminium, dry cut, feed rate is at 30" per minute and the rapids about 45" per min.

Although CamBam suggests to set the rapid speed at the beginning of the drawing and at a faster feed rate, this proved to be too much for my machine mass and the servos installed. The Zero was lost several times once on each axis at a different time and each time during a rapid. In the end I slowed the feed down to 10" a minute.

One of the other parameters in the set up for a new drawing with CamBam is setting of the machining stock. Using 3/8" stock only allows milling to 3/8" which now requires that the stock be perfectly set up in the mill to mill through the 3/8" blank. I should have set the stock limit for .400" to over come this. Learning is the solution to all of this and that too takes time.

The video is a bit long, over 10 minutes. Spindle feed on the .250" 2 flute cutter is 1000 rpm. A few times I have zoomed in too close for my camera to focus and that is out a bit as well.

The lost steps and the reset in the Gecko drives are what caused the bulk of the problems. An error of 129 steps or about .0005" is causing the lost zero. That amount of error is not significant, but the software sees it then causes the problem.

The next step in developement is to to install the Rodgers DRO board to over come this and hault the machine to prevent the mess from happening.

Take a look if you have the chance at the video.

I suppose now after being at this for a while, NEMA 42 servos would have been a better choice. The Gecko's limit this build to 80 volts motors. I have not seen any Nema 42 80 volt brush servos on ebay yet, but I am looking.

Still working on it.

Ark1.

[nomedia="http://www.youtube.com/watch?v=UNecW98NSRc"]YouTube - Flange milling[/nomedia]

[RotarySMP]

Nema 42 would be huge motors for a machine this size.

[diyengineer]

I tried cambam. It was really easy and nice to use!

I have read a bit about Cam software.

Where to start and what to do with it, then there is the price of the software.

Reading in the forms I down loaded the trial for CamBam and as programs go and lack of knowledge present's it'self, I set about making a simple manifold flange. Following through the tutorials and moving back and forth to my project manifold flange I finally came up with a finished project and created a G code file and watched the tool path in Mach3. 750 lines of code.

The intent of all of this is to look for problems and of course I found them.

The ultimate test would be a flange in steel or stainless steel so the cut is only .015" per pass. The test metal is aluminium, dry cut, feed rate is at 30" per minute and the rapids about 45" per min.

Although CamBam suggests to set the rapid speed at the beginning of the drawing and at a faster feed rate, this proved to be too much for my machine mass and the servos installed. The Zero was lost several times once on each axis at a different time and each time during a rapid. In the end I slowed the feed down to 10" a minute.

One of the other parameters in the set up for a new drawing with CamBam is setting of the machining stock. Using 3/8" stock only allows milling to 3/8" which now requires that the stock be perfectly set up in the mill to mill through the 3/8" blank. I should have set the stock limit for .400" to over come this. Learning is the solution to all of this and that too takes time.

The video is a bit long, over 10 minutes. Spindle feed on the .250" 2 flute cutter is 1000 rpm. A few times I have zoomed in too close for my camera to focus and that is out a bit as well.

The lost steps and the reset in the Gecko drives are what caused the bulk of the problems. An error of 129 steps or about .0005" is causing the lost zero. That amount of error is not significant, but the software sees it then causes the problem.

The next step in developement is to to install the Rodgers DRO board to over come this and hault the machine to prevent the mess from happening.

Take a look if you have the chance at the video.

I suppose now after being at this for a while, NEMA 42 servos would have been a better choice. The Gecko's limit this build to 80 volts motors. I have not seen any Nema 42 80 volt brush servos on ebay yet, but I am looking.

Still working on it.

Ark1.

[Ark1]

Nema 42 would be huge motors for a machine this size.

The TOS FN20 has it's plus and it's minus. Looking at commerial CNC machines the Z is normally done on the spindle head. Little mass in comparison has to be moved. The Z on my machine is geared down quite a bit and counts 60,000 steps to the inch. It is too bad that the Gecko drive has a fixed erro/fault count of only 129 steps as in this application as the error only equates to .00000775" certainly something that could be over looked in this application. The solution could be just as simple as slower rapids or should I say slower acceleration of the rapid to compinsate for this occuring. The next step incorperating the Rodgers board would certainly over come this issue of loosing the zero and keep milling.

The Nema 42 motor would only be required on the Z axis, but yes you are correct stating that it would be huge and a bit of rebuilding to adapt to this application.

Thanks for the comments.

Ark1.

[GM1G]

you could always find an old weight set and start looking at it funny like i did...

[Ark1]

Did cross my mind.

I could hide a gas lift shock behind the lower way cover too.

Years ago when flying 60 size model helicopters a ground or grass strike of the tail rotor would shear off the .080" metal drive shaft. I then set out to find a better material for the drive shaft that had a higher carbon and a larger diameter to prevent the shearing of the shaft.

Problem solved, only to now present another problem... that was stripping the teeth of of the two miter gears in the tail rotor gear box. A $2 shaft now would out last the $45 gear set. I in short order went back to using soft .080" wire drive shafts.

Ark1.

[diyengineer]

Viper Servo Motor control products

You could always swap your Z axis for one of these drivers. The Viper 100 (what i use) can be used with a 25volt-100volt power supply. You can program the error count from 64-2000. Just a thought. I don't know how much the rodgers board is, might save money? The Driver is $229. <=== kinda spendy

That sucks about the error. Even if you did use a viper driver and have the error set to 2000, i believe with a quick estimate:

2000 counts/129 counts= 15.5

15.5 x .00000775" = .00012"

SO even with an error 15 and a half times greater, set to the vipers max, you would only get a .00012" error.

What is a rogers board?

The TOS FN20 has it's plus and it's minus. Looking at commerial CNC machines the Z is normally done on the spindle head. Little mass in comparison has to be moved. The Z on my machine is geared down quite a bit and counts 60,000 steps to the inch. It is too bad that the Gecko drive has a fixed erro/fault count of only 129 steps as in this application as the error only equates to .00000775" certainly something that could be over looked in this application. The solution could be just as simple as slower rapids or should I say slower acceleration of the rapid to compinsate for this occuring. The next step incorperating the Rodgers board would certainly over come this issue of loosing the zero and keep milling.

The Nema 42 motor would only be required on the Z axis, but yes you are correct stating that it would be huge and a bit of rebuilding to adapt to this application.

Thanks for the comments.

Ark1.

[Ark1]

See this link.

Page Title

I think it is doing the same as you have mentioned with the Viper, but if you get into the description it states "

No special
screensets for Mach3 are required. Making operation almost seamless with Mach3. A following error of any size can be set for each axis within the Rogers plug-in."


This happens all on one board for all 3 axes, which I already have installed but not connected to Mach3 or the encoders yet. A bit of wiring change is required but I have the sheet already drawn. My fear as I am new to all of this was taking on too much at one time. I have proven the machine and Mach3 control. The Rodgers DRO board should now allow me to use the machine in manual with the reset tripped and the DRO showing machine position in Mach3 being used manually. Even with my X axis with a metric ball screw. NICE! Also I can set the amount of error per axis and be rid of the lost steps and Mach3 re-zero the offending axis.

Like I say, I'm working on it.

Ark1.

[zsolt832]

Hi Ark1!

WOW, I just found this forum and I'm amazed. There is not a lot of information available about old machines, especially FN20. I realize that you posted 12 years ago, still it was great to see the pictures. May I ask about the quality of the manuals? My manuals are French, Hungarian, and German, but the images are low quality in some cases (except for the French), and I'm looking for better quality (and hopefully English) manuals.

Zsolt

[Ark1]

Zsolt, Sadly I did not find an English Manual. I was in contact with someone years ago in Great Britain but they thought they had a manual, I paid for it and then they got back to me as the Manual they had was only in Hungarian and was for an FN25, so they sent me a refund. If you come across one please let me know as I too would like to purchase a manual for this machine.

Thanks.

Ark1 ( TonyK.)