Deckel NC Milling Machine Web Site


Chapter one

Introduction by Web Site host:


What can be seen here below is the beginning of an ambitious project currently being undertaken by an individual in the UK. What is being done here can not be recommended for the average Hobbyist. As a matter of fact it must be strongly discouraged that anyone try to do something similar without significant experience or technical background. Way too many Deckels have become victims to botched conversion jobs, only to be reduced to scrap in the end.

So enjoy the following descriptions, see what can be done, but be strongly warned against misguided imitation.








Converting an FP4ATC to a Heidenhain Control, while eleminating the Tool Changer


5th August 2009


Machine: Deckel FP4ATC

Control upgrade to a Heidenhain 415, and conversion of the machine to single phase 240 volt input.

Machine’s original owner: Suffolk University (possibly a spin-of company)

Unfortunately the tool changer has to go due to space constraints but I never intend to run the machine in production conditions anyway…  so really the machine will be a FP4A 

I was introduced to the ‘Deckel’ family of machines by a friend of mine and I finally owned a manual FP3 after many years of searching, due to space I purchased a Bridgeport interact 1 Mk 2 machine for some time but was always on the look out for a suitable Deckel CNC.

On the 14th of July 2009 I located a machine on the internet in the UK (my birthday incidentally) and purchased it on the 15th with agreed delivery date of the 5th August.

With working on a project in Germany (machine tool related) I only had a handful of days to sell my Bridgeport…. Everything worked out fine after a enormous amount of time and effort spent organising… I even decided to relay my workshop floor with another 100mm of concrete and water proof membrane, so I booked 4 days of work and set too…….. 4 days later I was totally exhausted but had my deckel….. This was just the start!

 Picture 1: Deckel FP4 CNC (a deckel is a good choice of machine to leave out on the front drive as it bends in with the garden and surrounding countryside very well :-)


 Picture 2: Deckel FP4 CNC side view

Machine arrived; tool changer and panelling plus the electrical cabinet covered my entire driveway for three large cars!!! What Have I done?

Picture 3: Deckel FP4 CNC + all the other stuff that came with it (were do I park my cars now)

The machine was instantly covered and sheeted as it would have a 3 week wait until it gets moved across the dive into its final resting place. (we didn’t have one drop of rain until after the machine was moved!)

The panelling was stored in my shed just leaving me with the tool changer and machine control cabinate to contend with.

Fighting the English weather it was decided to pull the control cabinet in to my workshop in-case of a heavy downpour, once this was resolved. We attacked the tool changer and reduced it to nuts and bolts within about 3 hours.

Picture 4: Deckel FP4 CNC (Electrical Cabinet)

Tool changer comments

This unit is of a truly massive construction and the mechanical complexity but also the quality is truly remarkable, from the shell cam turret indexer with its 1KW motor through to the hydraulic tool clamping and loading arm. 

Picture 5: Deckel FP4 CNC (Hydraulic Power Pack in tool changer)

Inside the tool changer cabinet body lies the 2.2Kw hydraulic pump motor and valve assembly, + air filter regulators and distribution board for sensors solenoids etc.  all hoses were drained and stored away along with a other components, unfortunately the hydraulic power unit for the toolchanger also provides the pressure for the vertical and horizontal drawbars ….. dam it…

Picture 6: Deckel FP4 CNC (Removing the tool changers carousel)

Picture 7: Deckel FP4 CNC (Nearly stripped down)

Conventionally deckel like to submerge the pump in a casting pocket in the base casting of the machine. As I don’t have this pump unit I will have to contend with starting a 3HP motor every time I need to change a tool (I can feel my carbon foot print getting larger).

Martin to the rescue!! lucky for me he was breaking a FP4A so a spare hydraulic unit become available . good news for me and my electricity bill!

Picture 8: Deckel FP4 CNC (Hydraulic Pump)

With the tool changer in component parts I was able to turn my attention to stripping the electrical cabinate… again within a few hours everything was out…. Sorted into wire / relays/ contactors/ transformers / resistors / rectifyers / etc etc etc the empty cabinate was ejected from the work shop and I continued with relaying the floor in my workshop!! (I certainly know how to make like difficult for my self)

Picture 9: Before

Picture 10: After

Once the floor laying job was out the way I continued to prepare for the big move.

This involved what I call a first stage clean on the actual machine, removing all sludge from the sumps and old smelly oil + swarf particles.  Every cover was removed to give me unrestricted access to the hard to reach areas, the machine wasn’t that dirty considering its age and the more I looked at the machine I realise how little the university used it. 

Before going any further I rigged up the central lubrication unit ‘Willy Vogel’.... and relentlessly turned the pump unit on an off until the machine was dripping in oil…. I also released all the oil unions slightly to ensure each was getting oil to the outlet. The oil from the vogle unit charges small piston accumulators that once the Vogel unit gives it short burst of pressure, the accumulators can still provide a positive supply of oil until the next PLC command to run the lubrication unit 

With this done I though it would be a good time to move the axis around and get a feel for ball screws and any other issues. Connecting the 4 bare motor wires into pairs to my portable bench top DC power supply.  Everything appears to be in order with even / steady current consumption across the axis travel.

The X- axis did show 0.05mm backlash, after retightening the ballscrew nut (it was loose) this was rectified.

Picture 11: Machine moving

Picture 12: chain + Rope + rollers = moving machine

Picture 13: fitting table

Picture 14: My Manual FP3


Picture:15 Machine in position



The Heidenhain 415 Control


Picture 16: Heidenhain 415 Control

The heidenhain control for ‘me’ the best, and is perfect for what ‘I’ want to do with the machine. I am in no way putting any other control down if you’re happy with your Grundig / Fanuc / Siemens that’s great.

The conversational programming format and ‘free contour’ programming aids are excellent; combine this with Mathematical Parameter Programming and you can achieve in a few lines what would take ten thousand lines with a cam system… I do use CAD (SolidWorks) and many different CAM packages but the heidenhain can nearly always be used to create what you want.

Back to Basics

Now I own a set of ‘Deckel Castings’ with some bit’s of cable hanging from it…. Where do you start?

The answer is I don’t really know!

But getting some power on the Heidenhain was as good a point as any; a friend helped armed with some very basic PLC code to let us move about the control. A mock-up E-Stop relay and NC Start microswich was used to get things moving and try to get some parameters entered into the control.


Power and Drive

With the issues of not having 3 Phase the FP4A with its DC servo motors are an excellent option as you can’t get more single phase than a DC motor!

There are many PWM drives available some with a single phase input or 3 phase input but with a DC Bus input option also,  in this instance Control Techniques Maxi Maestro drives have been used in collaboration with a large DC power supply to provide the DC bus .

The Power supply currently comprises of a Large 5Kva transformer to step down the 240 volts (some times creeping to 250 volts in my neck of the woods) to 150volts AC, then through a beefy rectifier to charge two large capacitors, this forms the basis of the DC bus for the dives.

Initially the motors and Tacho’s were connected to the drive (being sure to leave the axis in the middle of the machine travel in case the tacho was not connected correctly, or the axis will run away at top speed!! Be warned) the drives were set up following well documented procedures and the aid of an oscilloscope.

Picture 17: if you look close though the Rats nest you can see a test DC power supply for the drives



Picture 18: Heidenhain and drives

Finally after fighting some gremlins and getting my head around the referencing system of this machine (only one limit switch on the x and y axis….. no reference dog’s or switches!) I managed to convince the heidenhain to handle this setup… with the machine referenced and the software limits set I was now able to move the machine round for the first time!..... Manually I kept the central lubrication system working to ensure I’m getting oil to the machine.

Axis testing

I have been lucky enough to purchase a Mahr Millitron probe system that can give full scale readings down to + - 0.3 Micron full scale deflection, I’m not saying its that accurate (Ie nanometres) but these LVDT probe are analogue and the readings analogue so the resolution is very high and I would say the repeatability is also very high.

I can now see how the machine dynamically responds to very small jog commands relative to the machine table… with the handwheel working I was jogging with 0.0001mm increments, and giving incremental moves of 10mm also and asking to come back to a 0 position. The machine responded very well, When given a position command (say 5mm move) the machine would over shoot by 1 micron then come back to the desired position… interesting to note that the Y axis becomes stiffer the further out it is (in terms of no load movements) due to the ballscrew distance between the nut and the clamped support gets shorter and produces less twist and stretch in the screw. This can be seen to its full extent when the Y is completely back (Y+ Max) and the Probe will oscillate 2 microns and settle to its final value after afew seconds.

Picture 19: Mahr Millitron Probe Readout with potential for 0.00001mm resolution (that’s 0.000000393” for you imperial guys)

This measuring exercise was conducted to give myself a good indication of how it performed in relation to my Bridgeport interact….. I would say it outperforms it by a factor of ten.  This conversion will be money well spent I think.

Lot’s has happened over the past  10 weeks but it will become quieter over the coming months due to the large amount of mechanical and electrical building work that needs to be done…. I still need to find an electrical enclosure!  And I have many auxiliary items to get working. Depending on the response to this work on the website will dictate the amount of detail I put into my descriptions and updates, if there is sufficient interest I will add more detail to what already up and increase the level as I go along




Chapter 2


The design of a Deckel machine


I would like to highlight some design points of the Deckel machine, which I think are interesting to note:

(they are a brief summary only and are not an attempt to be a vast technical definition and explanation)


Direct measurement: the ultimate place to reference the machine datum too is the physical tool tip or tool centre, however this is for practical reasons nearly impossible with current technology.

Some machines use rotary encoders on the back of servo motors or on the end of the ball screws to provide feedback to the control… this is a ‘assumed’ measurement based on the accuracy of the mechanical components linking the encoder to the machine table.


Consider a encoder mounted on the rear of a servo motor, the torque is then passed through a reduction timing belt drive to the ball screw… then to the nut that’s fixed to the table, let’s consider the real life situation

We ‘COULD’ have two pulleys that could be bored 0.005mm eccentric that will cause lobbing and create positioning errors, timing belts aren’t perfect and ware / profile inaccuracies occur and can alter their characteristics over time, the ball screw shaft has to be preloaded correctly to allow no axial play in the bearings without crowding the bearings.


This Torque is transmitted down the ballscrew (that can’t be described as perfect), as pitch errors have to be realized and preloaded ball nut’s also have issues, one big area for concern is heating of the ballscrew due to heat from rapid motions or external sources, (15 degree rise in temp over a 550mm screw could generate upto 0.090mm of expansion… does the encoder see this….. no… do you get a 0.090mm error creeping into the system ….. Yes


Ball screw stretch is also another reason for errors (and ballscrew sag on large axis machine) if you load steel with ‘X’ amount of force it will behave under hook’s law and extension can be seen, ball screws aren’t exempt from this Law…. Does an encoder mounted servo see this linear deflection…. Your correct NO.

Deckel like most high end machine tool manufactures will use turcite on the slide ways to reduce ‘stiction’.

ie the machine slide requires a larger force to get it moving then when moving at a constant velocity on its hydrodynamic film,


thus when the slide isn’t moving due to stiction but the ballscrew is rotating to produce torque a angular twist is generated in the ballscrew (so the ballscrew is now being twisted and put under tension or compression…. The motor will move the correct amount to generate your 0.001mm jog but the axis wont move at all ! once the stiction is overcome this extra force will then dissipate by suddenly catching up with its desired position this was 6 to 10 microns on my Bridgeport machine depending were you were on the bed.


If you measure the Real axis movement with a glass scale as close to the tool as possible you can eliminate most of these factors that will affect machine accuracy if optimum setting aren’t achieved in a motor mounted encoder solution, If you take a close look at a deckel ‘Y’ axis you can see the front of the scale is rigidly mounted to the casting were the rear of the scale is mounted on a bracket that can allow for thermal expansion of the ‘Y’ axis due to external heating or working heat etc etc.

deckel carefully designed the machines using laws of physics and not approximation this is why they perform well. Also I would say deckel is a firm believer in Evolution not revolution and there machine’s just got better as the designers gained more experience, to get 1 micron tolerance is unrealistic without temperature control rooms and coolant and high accuracy tooling and indeed the correct machine foundation etc etc.


Heidenhain Scales when purchased new will come with a inspection report that will show the accuracy of the scale in graph form along an axis, most of the scales I deal with are within 3 micron along its length but for the majority of the reading is within 0.5 micron along the scale, with a curl up or down at one end of the graph.


My Heidenhain 415 display can be set up to resolve 0.1 of a micron (0.0001mm) and I can give very small jog commands though the interpolation factor on the control that electronically gears the hand wheel or write a program with incremental moves of 0.001mm.



The heidenhain 415 was an easy choice…. A friend was selling one due to an upgrade to a later model, it was priced right …… a new 530 would have cost around £12,000 without any good software options
+ all the other things you need (probably would have been getting on for £15K- £16K, and for less than 10% of this cost I have a very able control! and with a slow machine like a deckel why do you need the latest high speed digital drive technology…..




Heidenhain TNC 415 Technical data



Axes                                     5 plus spindle

(NC axes and PLC axes can be defined as desired)


Program input                    In HEIDENHAIN Plain Language and to DIN/ISO

Memory for                         12 000 blocks approx.


part program

Positions                             Nominal positions in Cartesian or polar coordinates,

dimensions absolute and incremental


Input and display               0.1 μm




Linear interpolation 5 of 5 axes

Circular interpolation 3 of 5 axes 1)

Helix Yes

Rigid Tapping Yes


Block processing time    4 ms


Look Ahead                         · Defined rounding of discontinuous contour transitions (e.g. 3D surfaces)

· Collision viewing with the SL cycle for "open" contours

· Advance calculation of geometry for feed-rate adjustment


Free contour                       In HEIDENHAIN Plain Language with graphic support

programming FK


Coordinate                          Shift and/or rotate coordinate system, mirroring, reduce and enlarge -

transformations                also axis-specific


Tilting the                              Yes

working plane


Subprogram                       Program section repeats, subprograms, program calls



Fixed cycles                       Pecking, tapping, slot cutting, rectangular and circular pockets,

SL cycles (milling cycles whose contour descriptions are stored in

subprograms); the machine tool manufacturer can also integrate

customized macros


Q-Parameters                    Mathematical functions (=, +, -, x, ¸, sin a, cos a, angle a from sin a

and cos a, a, a2 + b2, tan a, arcsin, arctan, arccos, an, en, ln,

log, absolute value of a number, the constant p, negation, truncate places

before or after the decimal point)

logical comparisons (=,¹, >, <),

Parenthetical calculations


Program test                      By graphic simulation of the part program


Parallel operation              Yes, no graphics Yes, with graphics


File management              up to 100 files: programs in HEIDENHAIN and DIN/ISO format, also tool 1),

PLC datum shift, pallet tables1) and text files


Tool compensation           Tool length, tool radius in machining plane

– Three-dimensional tool

compensation with surface

normal vectors


Central tool file                  Various tool tables for 254 tools max. each, with flexible pocket coding,

tool life monitoring and sister tool organization1)


Data interfaces                  V.24/RS 232 C and V.11/RS 422

·"Blockwise transfer": programs that exceed the control capacity can be

downloaded block by block and simultaneously executed.

· Extended data interface with LSV/2 protocol for external TNC operation

across the data interface

Baud rate 38 400; 19 200; 9 600; 4 800; 2 400; 1 200; 600; 300; 150; 110

Keyboard TE 400 with integral QWERTY keyboard

Screen BC 110 14" colour monitor 640 x 490 pixels

Logic unit LE 407 LE 415 B / LE 425


Axis control                        Feed pre-control or operation with servo lag


TNC 415 :                             0.6 ms (contour)

Position control                   0.1 μm



Integral PLC

PLC inputs 56 + 1 "Control is ready" input; (Option: + 64*) per PL)

PLC outputs 31 + 1 "Control is ready" output "; (Option: + 31*) per PL)

.Two PL 410 max. can be connected

Option: analog inputs

± 10 V 4 per PL 4101) or PA

Option: Inputs for Thermistors 4 per PL 4101) or PA

PLC program memory Approx. 8 000 logic commands

PLC cycle time 24 ms 20 ms


Error compensation         · linear axis error compensation

· non-linear axis error compensation

· compensation of reversal spikes in circular movements

· compensation of thermal expansion

· backlash compensation

· stiction compensation

· offset compensation


Position encoders             HEIDENHAIN incremental linear and angle encoders (preferably with

distance-coded reference marks)

also HEIDENHAIN incremental rotary encoders


Reference mark                Following a power interruption, automatic reference value input if

evaluation                           reference marks are traversed


Max. traverse                     ± 100 000 mm

Max. traversing                 300 m/min



Feed-rate and                     0 to 150% with two potentiometers at the control panel

spindle override


Accessories                       Electronic handwheel 1 x HR 330 Portable handwheel

or 1 x HR 130 Integral handwheel

or up to 3 x HR 150 Integral handwheel with adapter HRA 110

Diskette unit FE 401

Touch trigger 3D probe TS 120/TS 511

Measuring 3D probe — TM 110

Touch probe for tool TT 110






For the machine conversion I intend to re-use as much of the original equipment as possible so far I can state my intentions as:


Controller : Replace with Heidenhain 415

Direct measurement scales : Remain

Contactors 24v : Reuse

Servo + Spindle motors : Remain

Servo Drives: changes to maxi maestro

Spindle motor : inverter driven

Hydraulic motor: Replaced with sump mounted unit and lower power

Transformers : replaced

PSU: switch mode and transformer rectifier replacement


Due to the fact I don’t know every single component I need in the electrical panel, hence I don’t know how big it needs to be.

So I have decided to Mock up the electrical panel by fixing two sheets of steel to a box section frame, this way I can arrive at the correct dimensions for a cabinet and remove the pressure to get it right first time!


Having a quick layout of some component and truncking  I marked out the panel and drilled it out on a FP5A (Deckels making Deckels)

Picture 20: marked out plate



Picture 21: Deckel making a Deckel…. Chicken and egg situation



Picture 22: first build up


The DC power supply for the axis drive would have considerable inrush current associated with it’s start up, also two large transformers are being used.

To enable a soft start of the system, the jog resistors for the spindle motor were used (with driving the motor though an inverter they become redundant)


Picture 23: start up resistors

When the main system contactor pulls in the two transformers are initially fed through the resistors and a small electronics unit looks at the DC voltage on the main DC bus capacitor when this reaches 190volts the resistors are shorted out by another contactor


Picture 24: Initial buildup of main contactor bank




  Picture 25: Some of the build up         

Now I can start the machine and run the X and y axis without any issues, with a relatively tidy mock up electrical panel.


Next time I will talk about gear changing (this is already working but will upload in a week or so)

The main items left are the only the tool release and the Z brake and axis drive….. Then build everything into a nice new panel ….. Neatly ….. The mount the monitor and keyboard….. Then arrange some coolant supply…… replace the glass…… the list goes on and on and on.. we will see J








Chapter 3

Gear Changing

The Deckel is described as having 21 speeds, but in fact it is only a 18 speed gearbox. The other speeds are achieved by running the motor slower (4 pole, 1400 rpm) instead of a 2 pole at 2,800rpm.

The ultimate result of changing gear is :

The gear shifting is achieved with a bank of 3 motors and 9 micro switches, the DC motors drive cylinders with plastic detents located around the circumference corresponding to various gear positions, when a particular gear position is reached the corresponding micro switch is pressed. This can be seen in the following short clip.

Looking at the table below it can be seen there is a very regular pattern to the progression of spindle speed against the gear position.



Unfortunately the Heidenhain can only deal with an 8 speed gearbox so everything will have to be done through the PLC.

The essence of the PLC gear change program is a series of questions and actions looking at each bank individually, working from gear selector bank 1 to bank 3.

The original layout of the gear selector bank had around 20 wires required  to perform the actions of motor power and switch outputs all requiring valuable PLC inputs and outputs, the below method has 7 wires totalling 4 outputs and 3 inputs on the PLC.

Looking at the Gear number table it can be seen that each gear bank essentially has a Low/medium and high gear within it, and if we deal with each bank individually to arrive at the desired gear speed we are not required to look at all the current switch / gear positions before we start the gear change routine.



 When selecting the gear range motor or gear bank it is possible to energise the common of the switch bank without starting the motor,  the motor is only driven when the start transistor is switched on at the base, the below diagram should cast some light over this.




The below image depicts the circuit used on my machine and the relay connections to the motor,

Note how each relay shorts out the motor when in the off position, this will bring the motor to s sudden stop when the relay drops out, if the motor is allowed to freewheel to a stop the gear position could over run and miss the gear. The left hand relay just changes the polarity on the motors so the direction can be changed so the gear can be reached faster.

The relays can be seen operating in the below link.





The basic flow chart below depicts the fundamental questions asked within the PLC program, as the first gear bank decides if it is high or low range the first questions must be asked….

“Is the requested speed higher or lower than 568”?

“Is the high range switch already in position” ?

“if the switch needs to move, what direction should the drum rotate to find the gear the fastest”

Once the back gear has been selected it can move through the other banks using a similar method to select the finer range banks.


All the time the gear selection is in progress the inverter is driving the spindle motor very slowly backwards and forwards to ensure the gears mesh as they are selected.

An interesting and quite useful feature of the setup I have is the ability to have variable speed control of the spindle motor though the potentiometer (POT) on the Heidenhain keyboard,  The inverter can accept an analogue demand from the Heidenhain control (0 to 10 volts) so after selecting a gear i can fine tune the cutting speed from the POT on the control and vary the frequency of the inverters output.