To find K40 conversion information use the "INDEX", "SEARCH" or "LABELS" sections in the sidebar.

Monday, November 28, 2016

K40 Gantry Cable Management

Reliable Cable Management

Sooner or later you have to create a reliable way to manage the LED cabling and air assist hose that connects to the movable gantry.
This post outlines the use of a drag chain and the associated parts and build procedure.
This drag chain is purchased and installed horizontally from the K40 head bracket to the back right corner of the main compartment.
Two brackets are cut and bent to affix one end of the chain to the head and the other to the side wall of the main chamber.

For other information on the K40-S build use the  K40-S BUILD INDEX with schematics


Please consider donating (button to the right of this post).

Your donations help fund additional research, tools and parts that I will return to the community as information.

The Drag Chain Assembly

The bracket design(s) can be found in the Sketchup Warehouse. 
Note there are multiple brackets in this file collection only two apply to this build.
I did not cut them on the K40 (it is being built) but I designed these parts to be laser cut.
The brackets are fabricated using .090 acrylic that I got from HD.
I bent these on my shop made acrylic bender but you can bend it in a vice with a heat gun.

Drag Chain

The drag chain I used came from amazon:
Black Plastic Drag Chain Cable Carrier 10 x 15mm for CNC Router Mill.
Best I can tell the dimensions refer to the inner opening in the chain not the OD. However they are not precisely what I measured. Your results may vary. Different size chains are available but if you use other than this chain you will have to modify the brackets design.

Head Chain Bracket

This bracket mounts on the head using two of the base plates mounting screws.
The screw on the right and back are used, the left plates screw is exposed through a large hole in the bracket. This allows the bracket to be installed without removing or loosening all the base plate screws (just two) and thus messing up the alignment. My future plan is to print or cut a new base plate with all these features integrated. 
To install the head chain bracket, insure the left front screw is tight then unscrew the right and back screw and add this bracket. Then add back and tighten the right and back screws through the head chain bracket.

Cabinet right side mounting bracket

This bracket is mounted to the drag chain and then to the main cabinets right side wall in the back right corner. The wiring and the air hose are routed from the electronics cabinet through the sidewall and into the drag chain.

Cutting the sidewall

Mounted right side bracket

Removed piece

The right sidewall cutout

Drilling the mounting holes in the side wall

Later there will be a template for drilling the holes in the side wall in the design file.
However you mount the chain to the sidewall, insure that the chain is mounted at the same level on the side wall as it is on the head. I moved the gantry into the back corner and traced the top of the drag chain on the wall. Then position the bracket mark the holes and drill. Hopefully the template will do this for you.

Cabinet Installation

The location  and direction of the head bracket and the rear wall mounting is critical in getting the chain to clear all the optics as well as have smooth operation. The head bracket has fixed hole and mounts to the head plate so its easy to get in the right place but the rear bracket must be put in the right location manually.
Note: assembling the end links to the brackets can be annoying especially when you have big fingers like me. I used # 4 heat set inserts for plastic so that I did not have to hold a nut while bolting the bracket. You can install these using a soldering iron.
In my setup I used 24 links not including the end mounting links. More links will create binding at the rear of the machine and it gets in the way of the beam when it is in the left rear of the cabinet..
The below pictures show the endpoint brackets in the 4 corner positions. Admittedly the head bracket can be optimized for strength and moved to the right to gain additional clearance away from the optical paths. The idea for the chain to be long enough  to reach the corners but not so long that the arc of the chains wrap ends up in the 1st mirrors path when in the upper left or its trailing loop in the path of mirrors 2 - 3 when at the extreme right .
The distance from the back of the cabinet to the furthest forward mounting point on the head plate, when it is fully against the back wall, is about 73 mm and the width of a this drag chain when it is fully coiled back on itself is 68 mm. This means that when the drag chain is against the wall there is approximately 5 mm of gap between the chain and the wall. In practical application probably 1/2 that.  

The chain must clear the optical path when rolled and at the extreme left rear position.

Check upper right for clearance
Back left showing clearance from laser beam

Good at lower left, not to tight but not to long

Lower right looks good. Probably could move the bracket right a bit.

New Head Design

I completed a new head design and brackets that integrated this same drag chain.

Others approaches

I wondered why some mount the chain like mine (hinges vertical) vs rolling the chain horizontally across the x axis and vertically down the right side of the main bay (hinges horizontal). Like this design from +Ashley M. Kirchner.

The reported problem with running the chain with the hinges vertical vs horizontal is that the chain wears and sags and then can interfere with the gantry, sounds logical will see how these wear!

+Anthony Bolgar K40 has the same basic arraignment but winds the chain along the back wall from left to right. This clever approach eliminates any interference with the beam path. For some reason I could not get that method to work without binding, perhaps it is because my drag chain was bigger.

Other designs and parts

The operating drag chain

Note: the motion is jerky because I was manually jogging it.


Like most K40 conversions there is more than one way to solve the gantry cable management problem. I have used the above method because it was simplest for me to implement and met my minimal needs.
+Ashley M. Kirchner design:
I like this design approach best but my goal was to have a K40 conversion whose parts can be laser cut so those without a 3D printer are set up. His bracket is in the SU design files linked above. I think this approach probably wears the chain better so it may last longer and it clearly is a setup that by design is out of the beam path.

To Do:

  • Design a version of the +Ashley M. Kirchner setup that can be laser cut using a 7mm x 7 mm chain.
  • Add drilling and cutting templates
  • Design an integrated "laser cut" head plate that has all these features:
    • Drag chain mount
    • Laser finder mount
Enjoy and leave comments if you have suggestions or questions.

Maker Don

Thursday, November 17, 2016

K40-S Lift Table Integration


The K40 bed is fixed and one quickly finds out that being able to vertically position the bed enables a much easier focus process when cutting and engraving objects of various thickness.
I tinkered with building my own lift table and finally realized that that Light Objects lift table was less expensive than I could build myself. This table is used without these build instructions.
Two approaches are described.

  1. The lift function can be made to be stand alone as this post instructs:
  2. Integrate the lift table with the Smoothie which this post instructs.
This subsystem is added to the K40-S BUILD INDEX with schematics

Thanks to +Robert Rossi  who collaborated and tested this build on his K40.


Please consider donating (button to the right of this post).

Your donations help fund additional research, tools and parts that I will return to the community as information and how-to's. 


Smoothie Integration

Integrating Smoothie with the Light Objects lift table is pretty straight forward.

Limit Switches

You will need to mechanically affix limit switches to the Light Objects lift table. Ideas for mounting the limit switches are included in the Standalone Lift table post. There is a link in this post to a bracket that can be laser cut.

Wiring the limit switches

  • Connect the lift tables limit switches to the Zmin and Zmax end-stop connections
  • Configure the z axis in the configuration file
If you are using the GLCD you are done with the wiring. The "JOG" Z axis function on the GLCD will server to move the Lift Table up and down.

Adding UP/DOWN buttons

If you are not using the GLCD and/or want discrete control buttons
  • Connect UP/Down buttons to spare input pins on the Smoothie. I used P1.22, P.23. They can be found on the smoothie board next to the Ethernet connector on JP32. The buttons must switch to gnd.
  • Program the configuration file to use the input switches

Example switch configuration modules:

Up switch
switch.zplus.enable      true                     
switch.zplus.input_pin      1.22^

switch.zplus.output_on_command G91_G0Z10_G90

Down Switch

switch.zminus.enable      true                      
switch.zminus.input_pin      1.23^
switch.zminus.output_on_command G91_G0Z-10_G90


UP/Down buttons: any NO momentary button will work.

Lift Table Power

I don't know exactly what the power requirements for the lift table is but I do know that it is more than 1 ampere @24VDC. The lift mechanics put quite a load on that stepper.
If you are using the LPS to provide 24VDC to your Smoothie its capacity is only 1 ampere. 

It is therefore possible that you could be overloading the 24VDC supply. Measure the current and voltage when running your Z table configuration to verify you are not overloading the supply. There are no fuses on the 24VDC  from the LPS so if you exceed the load it will damage the internal regulator and you will need to repair the damage or purchase a new supply. (I added DC fuses in my power design)

To minimize power consumption insure that the table mechanics are aligned properly and moving smoothly. Mine had a bent shaft (LO replaced) that cause increased current draw.

Wiring diagram

Go the K40-S BUILD INDEX above to get the latest schematic. 

Tuesday, November 15, 2016

K40-S Interlocks & Breakout Board


Interlocks are an important feature of your K40. Although simple in principle I found them annoying to wire.
I created a breakout board to make this process easier.

The K40-S BUILD INDEX with schematics


Please consider donating (button to the right of this post).

Your donations help fund additional research, tools and parts that I will return to the community as information and how-to's. 

How to add interlocks to your K40

Interlocks cut from the main schematics

To see how this board interconnects with my build look at the INTERLOCK BOARD tab on the schematics link above. You may need to modify this build for your configuration.

Note: J11 in the schematic above connects to the ARMED LED (on the control panel) through a resistor to 5V. See the main schematic.

WARNING: in a previous version there was an error caused by the Interlock Open LED that allows the laser to fire without the "Laser Switch" being activated. Make sure that your LED is wired like above and not directly across the "+P, Gnd like it was before..

Below is a sketch of the interlock circuit. You should not put anything but switch contacts in series with the interlock loop. The P+ or WP connection is connected to the cathode of an opto-couplers LED and that pin must be ground to allow proper current to the op-ocouplers input diode.

If you want to add an optional ALARM led you can add it outside the circuit as shown. The LED must be connected into the circuit exactly as the sketch below shows.
In normal operation the LED reminds me that laser is enabled. When the LPS won't fire I find it useful to see if that ALARM led is lit. If not I know the interlock circuit is open.

Equivalent interlock circuit showing optional ALARM led

The interlock LED now indicates that the laser is "ARMED", I like that better anyway :).


The board parts are simple:
Mine is 5 ports. I suggest adding as many ports as you can to use for further expansion. You can plug a shorting wire into any ports that you are not using.


This board through J75 connects to the LPS "Laser Switch" loop for your laser supply type.
J11 goes to a led + resistor on the front panel. It tells you if any interlocks are open. This was added after the following pictures were taken.

To add an interlock you create a simple two-wire assy's that comes from each interlock switch to the breakout which uses screw terminals.

Note: You can ignore the middleman board I mounted it there for convenience and then later removed it. If you are not mounting the MM board your board can be smaller.


After it was all wired I realized that the Middleman board and interlock board have no connections with each other and no functional relationship.
Moving the Middleman board off the interlock board and into the Smoothie cabinet, cleaned up the wiring.

Interlock board with Middleman removed

Middleman moved to left wall of cabinet

Interlock Switches

I found some switch assemblies that make the mounting of interlocks and there interconnect easier and with no soldering requirements.

Purchase:Interlock switch assy.

Although this assy was designed for endstops it also works great for interlock switches.

Switch Assembly Schematic

In your interlock circuit you don't want any active devices, even LED's. Therefore you only need contact closures for this application. You want the NO and Gnd connections to be connected into the interlock circuit above. That is, pin 3 and 4 are connected to the interlock breakout. Pin 1 one is unused in this application.

Main Compartment Interlock Mounting

The main compartment can easily be protected by adding the interlock switch to the lower right of the middle partitions surface, no bracket is needed. An interposing block is affixed just above it in the cover. When the cover closes it energizes the switch and enables the laser. 
I used a piece of wood that was 26 mm x 26 mm x 25.4 mm but you may have to custom cut one depending on where your switch is mounted on the partition. I drilled holes for the sheet metal screws right were the partition bends.
Make sure to put insulating spacers under the board so that the lands do not short against the case.

Note; that's my old soldered version hanging in the bay waiting to be removed.

Laser Compartment Switch Brackets

I designed the bracket for my laser tube compartment out of acrylic so it can be laser cut. Normally I bend brackets but in this case due to the small size of the bracket I glued it. The designs slotted holes allows for adjustment of the switch PCB and mounts on the inside of the outer wall of the laser tube compartment. I mounted it near a hole about that is about 11" from the left edge of the cabinet. This gave me easy wiring access to the electronics compartment and kept the switch well away from the Anode of the laser avoiding any electrical noise or arc problems. The bracket is mounted with two sheet metal screws after drilling through holes in the face of the cabinets sheet metal. The sheet metal screws were threaded into the plastic before assembly and I found that one hole held the bracket solid, two holes are provided. The tricky bit is to get the bracket holes in the right place on the cabinet. I held the bracket up to the surface and marked them. You will find at the place there is two layers of sheet metal to drill through. Be careful and don't poke through and hit the tube.

The cover has an interposing block, double back taped (you could screw it in if you wanted) into the cover that closes the switch as the cover closes.

This arrangement also allows a magnet to be placed on the ledge to defeat the switch.
I don't recommend defeating interlocks, make sure you wear eye protection if you do!
Defeat magnet

Cover interpose block

The SU Design

The design file

Enjoy, and please comment
Make Don

Sunday, November 13, 2016

K40-S AC Power Systems Design


This post outlines the AC power design for my K40-S configuration. 

The K40-S BUILD INDEX with schematics


Please consider donating (button to the right of this post).

Your donations help fund additional research, tools and parts that I will return to the community as information and how-to's. 

K40-S AC design

See the K40-S schematic on the main page for details.

AC wiring:

AC wiring will be distributed with terminal blocks like these:

Auxiliary AC Outlets:

There will be four external plugs. The blower and air assist are controlled from panel switches and digital controls. These are the AC plugs:


Main and emergency power control will be implemented using: 

Friday, November 11, 2016

Click Here for the Index To K40 Conversion

Converting and Improving a K40

I have invested an inordinate amount of time converting my K40 to a better cutting and engraving tool. This post concatenates my K40 conversions posts into a build index. I am doing this with the intention of helping others reduce or eliminate their research, design and build time for K40 conversions.

There are many RIGHT ways to convert a K40 but the most often frustration I see from K40 owners is having a clear and as-simple-as-possible, but not simpler,  set of build instructions that result in a working configuration.
This build log should have everything you need to know to clone my design including parts. If it doesn't or something is missing let me know in the comments or G+ me at:

+Don Kleinschnitz

My approach in this conversion was to start with a clean sheet of paper and engineer the needed modifications. That means that I am analyzing each subsystem, understanding its operation, specifying a conversion design followed by build and then finally verification of performance.

This post pulls together detailed posts for each subsystem with the expectation that in the end it will give the community a specific configuration that works. 

My ultimate intention is for this information to be simple to use for most makers yet detailed enough to satiate an engineer.

Some of this content was designed by myself along with the associated testing and research. However a large amount of content linked in here came from the hard work of others. I just linked things together in a way that made sequential sense to me.

I thank and acknowledge the work of many other members of these communities that contributed in a significant way.

This is a work in progress so expect changes. I will keep this post updated as my analysis/design/build/test and usage evolves.


Please consider donating (button to the right of this post).
Your donations help fund additional research, tools and parts that I will return to the community as build and design information and how-to's.


Machine in operation testing minor jobs and making repairs and improvements

Advantages of this conversion design?
This design adds the following improvements to the standard K40 experience:
  • Protocol: GCODE enables a better part design tool chain
  • Increased safety: interlocks, mains shutdown, grounding verification
  • Documentation: includes ENGLISH schematics and drawings for everything.
  • Cutting/engraving performance: Air & vacuum assist 
  • Robust DC power supplies
  • Open source software tool chain: such as: LaserWeb, SketchUp, Inkscape
  • Support: a rich G+ community with engineers and experienced laser builders and users willing to help.








If you want to get right to the core of the electronics here are the full Build Schematics and below is a simplified sketch of the Smoothie wiring. Pretty one coming soon. The lift table integration is not included on this sketch


CAUTION! Unmodified K40's are dangerous.

K40's do not come with interlocks however every K40 should have interlocks installed!

The K40 laser is dangerous and can blind you.

Interlocks should be installed on the main cover and the rear laser compartment.

Don't defeat them when your inside the machine. Its not that hard to make an adjustment and close the lid to check the results. 

Wear your safety glasses especially if you have defeated interlocks which I just suggested you do not do!


  • Coming soon, keyed Laser enable switch
  • Coming soon, mains shutdown
  • How to  verify K40 grounds



I have elected to use two control panels on my K40-S. The stock panel will server as the "K40-S Operations Panel" and the GLCD will serve as the "Controllers Operations Panel. My reasoning was as follows:
  • These two panels have no interconnections with each other
  • I needed more real estate than the stock panel could provide
  • I wanted to make the smoothie and its interconnects modular so I could reuse the packaging for other CNC projects and test it on the bench.

Controller interface panel 

Operations Panel


DC Power 

AC Power

The Laser Power Supply & PWM

Interfacing to the the Laser Power Supply

Laser Power Supply Test, Repair & Installation


Interfacing to the gantry electronics

Cable management on the gantry

K40 Motors

The finder LED

I used a simple approach to implement the finder because I wanted to keep the head bracket simple. The disadvantage of this approach is that the pointer is offset from its true position when the table goes out of the focal range. To date I do not use the finder to locate anything accurately so I am using a single diode vs two with dot or line output. If I do need more accuracy I plan to move the finder so that it enters and exits the light path through the objective lens and therefore totally remove it from the head altogether (more on this idea later).
  • The finder diode is mounted on the head.
  • The finder uses a Laser diode that is mounted in a holder. The holder is adjustable with a screw.

Gantry parts


Controller packaging

This packaging design mounts the controller outside the K40 and the DC supplies inside.




CO2 Laser Theory and Operation

Basic Principles of Operation
How do CO2 lasers work

Laser Specifications

This is the best reference I have found for laser operating specifications.

Laser tube mounting brackets

Mirror mounts and assy. changes

K40 Optical Systems Modelling for Better Alignment

Aligning, Cooling, Operating and Protecting The Laser





Coming soon...
  • AC power and grounding
  • DC power and grounding
  • Interlocks
  • PWM
  • Endstops
  • Steppers


Sunday, November 6, 2016

K40-S Laser Power Supply Control "Take 2"

K40 Laser Power Supply (LPS) Control

During my journey to convert my K40 to Smoothie control I and many others have been frustrated trying to find a reliable, predictable and understandable means to digitally controlling the K40's laser power. For history see my previous post which I am superseding with this one.







Please consider donating (button to the right of this post).
Your donations help fund additional research, tools and parts that I will return to the community as information and how-to's. 

Simple Smoothie PWM Control 

A simple diagram showing how to connect to the LPS PWM from a smoothie 5x. You only need 2x wires.


In general, driver MOSFET's outputs on Smoothie compatible boards are labeled [+ or VBB] and [- or gnd]. 
The +/VBB is connected internally on the board to a [+] power source it is not a signal. 
The [-/gnd] is connected to the DRAIN of the MOSFET (not ground). 
For open drain connections you want to be connected to what may be marked [-/gnd] pin.  Pick up an actual ground for that signals cable from another pin (like the picture above).

You can remove the pot and install the red jumper which allows the PWM to have full control. I do not recommend this configuration. Although you will have full software control of power you will have to regularly calibrate your machine for max power. See Setting " Power Calibration"

Leave the Pot in

I recommend this approach!

You can elect to keep the "Current Regulation" pot installed in its stock configuration. If you do the actual laser power will be the product of the pot setting AND the PWM setting.

I.E. if the pot is set for 50% power and the PWM is asking for 50% power then the actual power will be:

Actual Power = Current Calibration % * PWM duty cyle

Example: Pot set a 50% and LaserWeb asking for .25%
Power = .5 *.25 = .125 or 12.5%

If you leave the pot in you can allow the software to control the power within its min - max range (0-100%) that the pot is set to. The Pot becomes an overall intensity control and you do not have to calibrate the machine for MAX power.

See this developing post for more detail regarding the PWM configuration effect on engraving:

Actual photo of how my machine is wired

Power calibration

I do NOT recommend taking the "Current Regulation" POT out.

Its my expectation that you remove the pot and pull the IN to 5vdc or set the "Current Regulation" (IN) to full on you will need to calibrate the power settings in your K40-S to keep the power range below the 18ma max operating range.

Power settings and verification

I have not completed the testing on this process but figured those looking for precision control of power would value having a preview. This is less important when cutting than when engraving or carving.

I recommend running a grey scale test and if that images but does not look good I would suggest something is wrong with your settings not your connection to "L".

When using the L line for normal operation with a pot installed you should have it set to a position that limits the current to 18ma and let the controller vary the power within that range. If you have the pot removed and the IN line jumper-ed to 5VDC you have to insure that the driving software never exceedes the max 18ma current level.

Setting the max power

If you are set up right the controller will not exceed the limits set in the configuration and you will not exceed where you want the tube to operate i.e less than 18 ma.

To calibrate:

  1. Verify your max power by using the test button to find out what the max current is with the pot full on. Example: your current meter at max pot setting = 24ma
  2. Then calculate what % you want the max limit of be and set that in the configuration file. Example: 18/24 = .75. You want the max power that the controller ever asks for to be less than 75%.
I advise testing by sending G codes for various power levels and verifying that it stays within limits on the current meter ("Current Regulation Pot"). Make sure the Gcode keeps the laser on long enough to get a good reading on the analog meter.

[later I will post a test for this]

After this is set up the controller should have full control of the power while keeping the power below the 18ma limit. 

You may need to re-calibrate as your laser wears and power levels drop.

Note: this approach to Max power setting may reduce your software's dynamic range for engraving. That is why I do not recommend setting the max power this way. 

See: for further development and research.


The remaining sections provide design and engineering information

Thanks for help from these folks .....

Thanks to +Paul de Groot and +Kim Stroman  for providing a schematic and broken power supply, respectively, as a source of information.

Multiple versions of LPS

We now know that this confusing situation is exacerbated by the existence of multiple versions of LPS's that are in use. Many look the same externally but are different internally.

You cannot tell what type of supply you have by any external means ... that I have found.

Below are pictures of  three few supplies. We have a schematic of Pauls and Kims. Mine I can only poke from outside because it is in service.

Caution: I know there are other versions of LPS out there and this analysis applies to the ones shown in this post. I will add the others as I get samples to trace.

Donate your dead LPS to research

If you have a dead one that you want to donate to this research email me at:

PS designation

For this post I am creating a designation for these supplies. To date I have found 3 types and I designated them like this:
  • First letter: the color of the AC power connector
  • Second letter: the color of the control connector
  • Third letterL the color of the DC power connector
  • Third letter: the color of the Power-on LED

Green-Green connectors with Red power LED

Designated G-G-G-R in this post
+Paul de Groot's supply

Green-Green with Green power LED

Designated G-G-G-G

+Kim Stroman's supply. Note: flyback disconnected

Green-White connectors with Green power LED

Designated G-W-W-G
My K40 LPS

More LPS style and interconnect references


For G-W-W-G style supply: 

LPS schematic PDF

LPS schematic download (.sch)

For GGGG style:

A share-able schematic that is a work in progress...

PWM circuit sketch not in main schematic yet
PWM circuit

For the G-G-G style supply here are some simplified schematics of the input controls:

Note that for HV isolation the "ENABLE" and "L" inputs to the LPS are routed through opto-coupler's. The external control signal simply provides a ground to the cathode of the opto-couplers transmitting LED to operate.
In the case of Smoothies open drain control (see PWM Control Via L, below) the opto-couplers LED is connected to the FET's drain and when the FET switches on current flows in the led illuminating it, correspondingly turning on the receiver transistor.

LPS control theory of operation

Although we are not yet finished understanding how external controls are implemented internally on all LPS versions we suspect that from an external perspective the fundamental behavior of these controls are the same.

In general the LPS is switched mode AC-DC supply that controls current with Pulse Width Modulation that drives a flyback arrangement and voltage multiplier. This supply creates a regulated and very high voltage.


So as to minimize confusion; in a K40 conversion there are two PWM functions at play. The one that is asserted by the digital controller (like a smoothie) and the another that is employed inside the LPS using a PWM IC. These LPS's are designed to operate both stand alone and with a remote means of controlling the level of power.

IMO: When a digital controller is providing PWM control the internal PWM is redundant.  

Basic control function behavior

There are three fundamental control functions on these supplies that are sometimes referred to by different names. I will give them a generic name for this post and also associate them with their real names where I can.

The "ENABLE" signal controls the internal PWM controllers output. If Enable is not asserted the laser will not fire because the internal PWM's Duty Factor (DF) is either 0 or in alternative implementations its output is disabled.
  • Ports that behave this way are: TH,TL, K.
At least two ways have been found for how this control is accomplished.
  1. Enable is connected through an opto-coupler that when asserted* enables the output of the internal PWM generator. If enable is not asserted the PWM generator is disabled.
  2. Enable is connected to an opto-coupler who's output transistor is connected to a differential amplifier.  When enable is NOT asserted* it biases one leg of  the amplifier insuring that the IN voltage will not generate a PWM DF > 0. If enable is asserted* the PWM DF will proportionally follow the IN voltage. 
*In all cases opto-couplers in these LPS are digitally asserted by grounding them. These signals are internally connected to the cathode of the opto-couplers transmitting LED.


The FIRE signal enables the output of the internal PWM to run if asserted and remain off if not.
  • Ports that behave this way are: L and Enable #1 above. 
In all the supplies above we have found that FIRE is always accomplished this way:
  1. Connected through an opto-coupler that when asserted* enables the output of the internal PWM generator. If FIRE is not asserted* the PWM generators output is disabled. 
Note 1: The internal "Test" button on the LPS motherboard is connected to this signal through an isolation diode.
Note 2: You will notice that this configuration of FIRE is exactly the same as #1 Enable behavior above. In some supplies it seems that there is an AND function of Enable and FIRE that controls the output of the PWM generator. [more verification needed here].

Power Control (PC)

The PC signal is the means by which the LPS power can be controlled with an analog signal. This signal can be in the form of a variable voltage from a pot or an analog voltage from a remote controller. 
To add to the confusion PC can be overloaded by a digital signal of the right voltage making the IN a power ON-OFF function. More on this further down. 

IMO: this control was not intended to be a digitally controlled input. That is in spite of the fact you will find many configurations and vendors that promote using it as PWM control. 
  • Ports that behave this way: IN
At least two ways have been found for how this control is accomplished.
  1. The IN signal adjusts the current through an opto-couplers input LED whose receiver provides a corresponding and proportional voltage to the PWM generators DF control. This IN signal is isolated and connected to the cathode of the coupler's input diode. In this configuration the IN signal expects a resistance to ground that changes the current in the opto-couplers LED transmitter and correspondingly changes the current in the output transistor. Its note-able that the opto-couplers used in this analog mode and the ones used in a digital mode* are the same component part. The only operational difference is how the current through the diode is provided.
  2. The IN signal is directly connected to one leg of  the internal PWM controllers differential amplifier. When enable is asserted this amplifiers output follows the IN signal adjusting the internal PWM's DF and in turn the output power. When enable is not asserted the other leg of the amplifier is at large enough voltage to prevent the amplifier from outputting a voltage proportional to IN effectively creating a PWM DF of 0. In this configuration the IN signal is not opto-isolated from the supply.

Putting LPS PWM control into practice on a K40 conversion

The current investigation and its corresponding espoused theory provide a basis for effectively using PWM control in a K40 conversion irrespective of a specific supply's actual internal operation.

The challenge with defining the RIGHT and WRONG way to configure PWM control of a K40 LPS is that you can make multiple approaches work. The fact that there are multiple unrecognizable versions of these supplies does not help matters. 

Therefore to start with I am going to define how I plan to use the information provided above. Then when I complete building and testing it I will update this post. Other configurations than the one outlined below may work. Its my judgement that the one below is the simplest and most reliable.

Interlock control: 

Leave the stock wiring for the "Laser Switch" and insert in series with this circuit any additional interlocking functions including cover switches and temperature monitors. 

PWM control via 'L":

Connect the controllers PWM function through an open drain (OD) or open collector (OC) to the L pin. Choose a transistor that is connected the controllers PWM function. Insure that in your controller you configure the input to this transistor to assert in such a way that the transistor provides a ground. In other words, when PWM from the processor is TRUE the transistor should be turned on. This transistor is connected to the L pin without any form of pull-up or level shift-er. This pin from the controller will be isolated from the LPS.
In the supplies we tested this signal can be found on the rightmost pin in the LPS DC connector.

A typical Open Drain configuration

The equivalent circuit of the K40-S PWM connection

Since we are providing digital power control to the LPS through "L" we do not need another form of power adjustment for PWM control but we do for max power setting.
As the laser wears the current required to get the same power changes so the power ranges changes .
Certainly we can change the max-min ranges in smoothies config file but why go through that annoyance.
Leave the pot in and it functions as an intensity control (see Power Settings and Verification above) 
However recognize that the controlled power is the product of the pot setting and the Smoothie PEM setting.
Therefore if the controller is set at 50% and the pot is set at 50% the actual power may be (.5*.5) or .25%. This is because the controller is turning ON the LPS whose power is fixed at 1/2,  1/2 the time .... is that sufficiently confusing :).

More detail at these posts:

Level shift-er use:

It has become common practice to use a level shift-er connected between a 3.3vdc PWM controller signal and either the IN or L signal on the LPS.
I believe that this configuration can be made to work but creates a level of complexity that is unnecessary. Most controllers have Open Drain or Open Collector drivers available. These are less complex to configure and wire and use less parts.

When an OC/OD is not available:

If there are no OD/OC outputs available I would use a discrete circuit that connects to a processors 3.3vdc signal to (R3) and connect "L" to the collector of Q2. Remove R4.

To test this setup. Before you connect to the LPS leave R4 in and connect this circuit to the controller and look at Q2 collector with a scope.  Insure that the collector of Q2 is being pulled to ground with the assertion of a PWM pulse.

I do not plan to test this configuration.

A level shifter circuit I found on the web (untested).

PWM on "IN" with pot installed:

I do not recommend connecting a PWM  control to the IN with a control pot installed. Unless the pot is left in the full power position you will create a complex condition where the incoming PWM is combined with the bias the pots wiper provides. This configuration creates IN bias values that change with the pot position.

PWM on "IN" with no pot installed:

A 0-5vdc PWM signal can be connected to the IN pin without a pot and operate properly. I don't see the value in using this approach as the configuration I recommend gives you the option to keep or discard the pot while providing opto-isolation. This configuration will also require constant max power calibration and configuration settings changes. 


The first step was to install the "Simple PWM control" and test its function.

Test the following:

  1. With interlocks open
    1. The laser does not fire with "Test" and/or without the PWM present.
    2. The laser does not fire when the machine is powered down or powered up
    3. The laser does not fire when the smoothie (controller) is reset
  2. With interlocks enabled:
    1. The laser fires when PWM present
    2. The laser does not fire when PWM is not asserted.
    3. The laser does not fire when the machine is powered down or powered up with PWM not present.
    4. The laser does not fire when the smoothie (controller) is reset
  3. The PWM duty factor is controllable from the SMOOTHIE

Signal quality and polarity

Note tests done on a G-W-G supply

Top: PWM from Smoothie
Bottom: PWM at the "L" pin

Top: PWM from Smoothie
Bottom: current in LPS
A table of measures

Above are some scope traces verifying the configuration:

  • Pict#1: The signal looks good and the "L" pin is being driven to the correct levels with the correct polarity.
  • Pict#2: My first try at correlating laser output with PWM. At first glance I am not convinced that the laser is responding properly to this freq of PWM. More testing needed to draw any conclusions.
  • Pict#3: Smoothie PWM with measurements table.
Note: these tests were run using the GLCD "LASER" functions to set and run PWM values.

Index of LPS types

For reference I am including pictures of these PS layouts and packaging.



LPS Album

Laser Power Sources

Theory of laser power sources
Enjoy and comment,