Power Control of the K40 LPS

This has been moved from a DM between @jkwilborn & @donkjr to the forum for everyone’s benefit. We will continue it here.

Hi, I’ve had a couple of questions about K40s that I thought you could point me to some information on them.

How does the basic power control, ‘knob on console type’ work, hardware wise…

The simple answer is that the pot sets the duty cycle on a TL494 PWM controller which in turn drives an H switch that drives the HV transformer which in turn charges a voltage doubler. The drive to the HVT is sensed and fed back to the PWM control for current control.
In addition, a 600V power source is derived from the AC line using a very complex multiplying circuit.

More info on my blog: Don’s Laser Things: Search results for LPS circuits

Is there really any difference from one of the lps in the ‘China Blue’ that I have and the K40 other than cost or how it’s labeled.?

I would have to see a photo of the two LPS connectors to be sure. Generally, these 40W supplies are the same. The 50-60W I have seen have an upgraded HVT but the PCB circuits are the same. That said, I have seen a few new LPS designs that are not the same as the typical K40. I have also seen 80W configurations with the same circuitry and 2x HVT’s.

If it sets the pwm, how is the controller able to change it for power control?

In my recommended configuration the LPS power is controlled by PWM from two places;

  • the controller connected to L with a PWM signal
  • the pot connected to IN controls the internal PWM signal

The resulting PWM for the system is the product of the two PWM values.

Does this function the same if they replace the K40 controller with a different one.?
I would assume they both have some kind of control via the PWM from the controller.

To my knowledge, the stock K40 controller only dithers and does not control power from the L pin. It turns the beam on and off with the L signal while the power level is controlled with the pot.
I am not sure how this changes if the stock controller is driven by the newer versions of compatible software.

More on my blog: Don's Laser Things: Engraving and PWM Control

I’ll have to look through your stuff and see if I can find a simple schematic… I’ve looked at the tl494 controller, so I’m somewhat aware of its functionality.

My schematic is posted on this forum but I did not create a simple or summary schematic. I can create one if you think it would help.

Aware of the hv lps and it’s basic operation, but haven’t really examined the current feedback area.

The current feedback mechanism has always been somewhat of a mystery to me. The current feedback seems open loop to me in that it is not actually measuring the current through the tube. Rather it indirectly measures the current in the output stage. When the tube (a negative resistance device) discharges the current waveform is a complex avalanche of current. I am not clear on how a closed-loop approach would be implemented or perhaps even necessary. I ended up figuring that since it works it must be sufficient.

Did you have a specific purpose for asking these questions? Just interested or are you wanting to use this information for making changes etc. Knowing the purpose would help me better answer your questions.


Thanks for moving the thread… Sorry I’ve been out for a while.

I’m just trying to get the relationship between how the current is controlled. I’ve watched my hv meter and have a slight handle on what it must be doing while I’m running certain types of jobs, more concisely, how I’m loading the tube. Having both meters let me relate, somewhat.

Right now I’m trying to add on a 600kOhm on the end of my hv meter resistor to tap about 1V to 1kv scale on the scope or in the ballpark… I’d like to see what its really doing. Would also give me a dependable number for a response time of the lps.

Thanks, take care

FYI: a generic discharge profile:

1 Like

Have you seen this with any kind of time reference?

I’m assuming the ‘townsend regime’ is were you don’t want to operate for any extended period…?

How much time is it from start to finish?


I do not have a time base for this in the K40 or for Co2 lasers in general.
I spent a lot of time when I did the initial work on K40 LPS to try and see this waveform.

My intention was to measure the response time of the marking engine to try and better understand engraving. This was promoted by the models I had built which suggested that conventional Co2 lasers do not have a response time that would result in good resolution. Note: the model did reveal better PWM values which are now used in most K40 setups.

I did not complete this quest for the following reasons:

  1. I did not like working on the HV end of this thing. I had multiple close calls with arcs and decided it was not worth being electrocuted and/or destroying equipment (scope).
  2. To get a meaningful response measurement you need to look at the optical output. I did not find a sensor that met the response time, power, and $$$ goals.
  3. Needed to sync the scope down to a single pixel and just did not get motivated to set this up since the K40 was obviously working … and # 1 & #2 were not resolved.
  4. Sarbarmultimedia did some work looking at the output with a scope and HV probe that showed the waveform and that satiated my curiosity.
  5. I decided to use printed patterns to measure the printed effect of response time on the image. Don's Laser Things: Measuring Laser Marking Quality

@43.51 of this video: RDWorks Learning Lab 97 Understand Your HV Power Supply via Florida - YouTube


I think I’ve schemed up a way to hook the scope to the hv tap… Hopefully I’ll get to it tomorrow.

If I get anything useful, I’ll post a screenshot of it, I think the scope can do that.

Wondering about number 2 or the response of the optical beam. Did you find any kind of sensor?

It might be worth frying a photo transistor to test this… cost wise…

Watched his supply video a while back, thanks for posting it.


1 Like

I did not find a sensor. Those I looked at in the right frequency range were slow and expensive.

1 Like

I have ‘broken’ the line going to the uA meter and inserted a 640K resistor.

Figured it would give me about a 1V to 1kV. I didn’t think I could watch it drain off… Kind of kills this idea… at least with a scope.

The pwm output is purple and the yellow is the voltage.

You can see where the line starts to go up that 1mS pwm pulse from the console. It’'s a single 1mS pulse at a low percentage.

Doesn’t really tell me much except I need to get at the power end :frowning:

Might waste a photo transistor and see what happens… never ‘lased’ a semiconductor.


1 Like

This was the test setup I built just before I abandoned the testing:


I’ve tried a few things and have my scope connected to the L-ON1 (laser enable), purple, which is also the trigger. Blue is across the 30mA meter and yellow is the pwm. PWM period set to 1mS.

Guess I’m going batty, it looks to me like the lps is ‘on’ when the laser enable goes low but the pwm doesn’t seem to change anything…

Does this make sense to you.?



1 Like

What does it look like if you expand the horizontal to see only one L-ON1 period?

Where are you getting PWM. I would have expected PWM and L-ON1 to be the same signal. I.e. the L connection on the LPS?

On the Ruida, the L-ON1 (laser one, enable) goes to the L input. LPWM1 (pwm) is wired to IN of the lps. P is just grounded as the Ruida has water protect inputs to stop the machine.

I had tried to capture something similar, but had and operator ‘oops’. I’ll get it up here asap.

I’m engraving a old Glowforge file I purchased quite a while back and run it in ‘passthrough’ mode.



Perhaps since the Lon and PWM are separate they are not in sync on the scope.
The laser appears to be cycling (like a PWM) during Lon just not in sync with the PWM trace?

I wonder if a grayscale line engraving wouldn’t be a better test design since you might see the pwm vary gradually on the scope depending on the time domain capabilities.

Very cool experiment since usually we see L PWM’ed and IN fixed at some max voltage.

I had similar experience with making sense of the current in the tube vs the power control.

I concluded that a known “start of horizontal” scope sync point for each scan is needed and then never pursued getting one.

I was thinking running a repeating vertical raster pattern … or something like that.

1 Like

might be able to trigger off the Y axis motor driver DIR pin.

The scope is triggered by the falling edge of the L-ON signal, so everything is related to that edge. The ‘T’ under the ‘STOP’ on the scope is where it was triggered.

I would have thought there would be some sort of synchronization, but appears that’s not the case.

I built my first scope, Heathkit, back in grade school. I have some time with them. Can’t figure out how I could have misconfigured it.

Suggestions, I can try…

@dougl I don’t know what triggering off the stepper would buy me. I don’t know of a direct relation to when the laser is firing and the stepper moving. However, I’m listening…

The first screenshot was not a gray scale. Here was the test image I was using in the first post.

‘Zoomed’ in the pattern is more clear.

However it must be working here is a 60 minute run of the finished item.

To simplify what was happening, I drew a line and set it to 300mm/s@50%… I’ve kept the period of the pwm to the slower 1mS.

This one is 100mm/s@50%… just to see if there is any difference. The scope horizontal scan rate is increased to 200uS.

This has the horizontal sweep to 5mS/div…

I was scrutinizing this area of the first image in this post.


A ‘negative resistance’ device …?

Not clear what’s going on… Any ideas?


I’ve found that when trying to find/see something on a scope, I look for ways to trigger on a known element relative to the signal desired to “see”. So instead of dithering an image which is just lots of on/off, I would shoot for grayscale power control where you’d expect a variation of the PWM power level( IN ) signal. Then, knowing the laser starts engraving the known pattern when the X axis is driven in a certain direction you might see the beginning of what your looking for.

I think that is what I was thinking. :wink:

Follow that… I know the that it’s either the PWM raising edge or the L-ON falling edge. It’s pretty clear from the image that lps ‘fires’ when L-ON goes low.

Why we don’t see any evidence of the effect of the pwm has me puzzled…


Since the L-on is just an enable do we really know its relation to the the PWM?

I don’t think the L-on is creating any useful information relative to the PWM?

Did you try syncing on the PWM.
Try syncing on the laser current and look at the pwm.

I would try drawing a set of vertical line patterns that repeat and are repeated for the full horizontal width that could be easily recognized. Make the pattern an integral of the horizontal pixel width.

Something like 5 lines, 10 spaces … repeat.
Then try; sync on the laser current trace and look at PWM. See if that gives a hint to the relationship.

1 Like