Modifying an EMCO Compact 5 CNC to control it via the open source software EMC on Linux

EMCO Programing Manual I have now scanned the Emco Compact-5/CNC programing manual into PDF format. YOu may download them freely, since EMCO no longer supports these machines in any way. This manufacturer follows the most modern and international standards and offers workshops and training in CNC machines, milling and turning.

October 16, 2008

This webpage will explain how two people modified an EMCO Compact 5 CNC tocontrol it with a PC running the open source software Enhanced Machine Controller (EMC) onLinux. We will describe exactly what changes we made to the EMCO, the custominterface we had to make to connect the PC to the EMCO and the softwareconfiguration used. We hope this will inspire other people who would like todo similar projects.

The project story

This project is the work of two friends: Jerry Spaulding and David Foster.They both took the Portland Adult Ed Machine Tool class in the fall of 2007; Jerry because he had plans to buildhis own 3-D CNC machine and David because he just likes those types ofclasses. The class was taught by Joseph Bolducand they learned how to use milling machines and lathes, but only brieflymentioned CNC machines. Jerry asked Joe about two EMCO Compact 5 CNC machinesthat were sitting idle in the shop. Joe said they were donated several yearsago, but never used because he didn't have the time or resources. (For thosewho don't know this machine, it was originally built in the early 1980's andused a tape control system and TV monitor, both of which greatly limit theusefulness of it today) Jerry asked if he could try to modify one of theEMCOs to control it by a PC. Joe said sure and thus the project was born. Itwas also during this time that David volunteered to help out. Neither Jerryor David have a background in machine tools, but they're both excellentprogrammers, have some experience with circuit design and love working onprojects like this!

The first step was figuring out if this was possible. We found a webpagewhere someone had already done something remarkably similar, but it containedvery little information describing the project. So we looked elsewhere andfound the Yahoo Emco CNC Users groupwhich has a wealth of information, manuals, files available. We obtained somegreat info there, including the indespensible maintenance manual. We also gota high resolution copy of the wiring diagram for the board interconnects whichproved extremely useful. During our research we found that the majority ofpeople who modified their EMCO Compact 5 CNC/PC machines had does it usingcommercially available controller boards and often replaced the stepper motorsalong with the controller board. That is one way to go, but being the cheapguys we are, we decided to try to do it ourselves.

The goal of the project is to control the stepper motors, which in turncontrol the X and Y axes of the cutting tool. The work piece is clamped intothe main spindle, which is not controlled by the PC. Using the maintenance manual we were able to determine which boards we could get rid of and which boards we needed. We removed the extraneous boards and started trying to figure out what to donext. One thing that needed to be addressed was being able to turn themachine on. These EMCOs came with keylocks and without the key you can'tturn it on. Unfortunately Joe did not have the keys for the machines, so wewere forced to remove the keylock and connect all the keylock I/O as if it wasalways turned on. This too a bit to figure out, but we labeled the wired andconnected them with wirenuts. Now the big red emergency-stop switch is usedto turn the machine on and off. Of course if we had the key, we could haveavoided this step.

Joe had obtained an old PC (1GHz) and I got an old 19' monitor. Jerryinstalled the EMC2 Live CD on the PC, which is just a custom version of UbuntuLinux with EMC. This was our PC control setup. Jerry also created a breadboardinterface that we could start using for testing.

The Interface Board

The initial breadboard we used was based on information Jerry had obtainedfrom the owner of this webpage.Jerry created a custom parallel port interface to the breadboard (cutting offone end of a parallel cable and sticking the wires we wanted in thebreadboard) and making another custom cable to connect it to the stepper controllerboard. It took us many weeks of debugging many different issues, but wefinally got a board where we could control the stepper motors from EMC. Someof the issues we had to deal with along the way were accidentily blowing thestepper controller board (luckily we had a backup), metalshavings in the case causing blown fuses and other board issues, encoders thatweren't aligned, figuring out the correct wiring for the stepper controlsignals, determining how limit switches would be integrated and lots ofgeneral circuit issues. A picture of the breadboard during this process isshown in one of the pictures below.

Once we had a working breadboard, the next step was to transfer thatcircuit to a PCB for more compactness and reliability. David had created andetched a few boards before using a homebrew method that worked pretty well.Long story short, it took several iterations of PCBs (and several weeks)before they finally got one that worked as required. The numerous boards werea factor of poor board design/etching, incorrect circuit and other things. Inone case, all of the part connections on the board were a few percent toosmall because of using an intermediate PDF step to print the board masks andthus resulted in parts (like the DB-25 connector) that couldn't fit the board.It should also be noted that some parts of the circuit were re-designedslightly during this period to increase the functionality of the circuit. TwoLEDs were also added for a visual check of PC and EMCO power.

Circuit Explaination

The circuit is basically just some optoisolators and inverters. Thestepper motors we were controlling would be working in normal full step mode.Full step mode energizes two phases at any time according to the sequence AB->!AB->!A!B->A!B->repeat, where A/!A and B/!B are the pairs of controlsignals that control the two stators in a motor. Since the control sequenceonly ever has one signal of the pair 'on' at a time, we can do something alittle sneaky where EMC will just output the A and B control signals for astepper motor and we'll create the inverse signals (!A and !B) in ourinterface circuit. Since we're controlling two stepper motors, EMC willactually output two pairs of control signals (AX/BX and AY/BY). These signalscome from the PC and are routed to the input of an optoisolator via a smallresistor. The purpose of the optoisolator is to electrically isolate the EMCOCNC from the PC. This can prevent damage from occuring in the PC if there areever any problems with the EMCO. The output of the optoisolator is on the EMCO power domain. The signals are then routed to inverters and thus we now haveboth (AX/!AX, BX/!BX) and (AY/!AY, BY/!BY); these are the 8 control signalsthat the stepper controller board requires.

We've decided for the time being not to try to control the spindle motor via EMCdue to complexity, but we did decide to feedback the spindle motor encodersignals to EMC so that it could use that information for threading and otheroperations. The Compact5 has two spindle encoders, located near the spindlebelts on the side of the machine. One encoder generates a signal once perrevolution. The other encoder has many slots in the encoder disk and willgenerate many more (I can't remember the exact number at the moment, but itwould allow EMC more resolution of the spindle angle at any moment). Theseencoder signals are pretty weak, so we routed them thru some free inverters inthe 7414 inverter chip to clean up the signal enough to drive theoptoisolators. These optoisolators go in the reverse direction and the outputof them are on the PC power domain where they are connected to some pull upresistors and the PC.

The other signals we're dealing with are limit switches. EMC can home thetoolpost using the limit switches. We added connections for up to 6 limitswitches (which will be multiplexed onto three PC inputs, two switches perinput). This was a limitation on the number of inputs on the parallel portand in EMC to use. The plan is to attach 4 limit switches to the Compact5,two for the X limits and two for the Y limits. We did some preliminary testswith limit switches and they appear to work nicely. We took the easy routeand kept the limit switches on the PC power domain, so we didn't need to dealwith optoisolators. The limit switches connect to the three input lines viasome pullup resistors.

Emco Compact 5 Cnc Manuals