Advice on Nd:YAG Flashlamp power supply

[zigunov]

Hi,

I was wondering if anybody could give some advice on my reverse engineering endeavors here... I found this (I believe, 600V) power supply made to charge/trigger capacitors for flashlamps used in Nd:YAG laser tattoo removal machines:

https://www.aliexpress.us/item/3256805411784901.html

I'm working on a custom Nd:YAG laser, which is why I really wanted to make this work. I got the supply in the hopes I may be able to figure out how to trigger it. I'm still somewhat puzzled so I figured maybe some of you may have this experience since there's literally no information about how this supply is supposed to be wired in the internet (at least, not in English...).

I attached a picture of the circuit board inside. The main terminals for AC input and 600V output for the capacitor charging (CAP+, GND) are well-labeled. Also, the flashlamp trigger (HV-J) is also labeled and is the screw terminal at the bottom right.

The control connector is a white header connector with 8 pins, but only 5 are connected (see figure at the bottom right). I labeled in the picture the terminals 12V, GND, TR, A and C. The TR terminal is clearly a trigger input, as it connects directly to the gate of a SCR through a small network (likely for overvoltage protection). From my understanding of how these supplies work, the control board requests a charging of the capacitor (for safety purposes, the capacitor is always discharged except just before the laser fires). When the capacitor needs to be charged, I believe something needs to happen to either A or C to tell the board to oscillate the CHARGING TRANSFORMER (blue ones on top right of picture, I'm not sure which one). After charging is done (which I believe is assessed by also, A or C as an output), then the trigger signal is sent to TR and the SCR shorts the TRIGGER CAPACITOR and the TRIGGER TRANSFORMER, generating a high-voltage spike (>20kV, I believe) at the HV-J terminal where the flashlamp is connected. This strikes an arc at the flash lamp and shorts the main capacitor connected to CAP+, firing the flash lamp with however much energy is stored in the capacitor.

This is as far as I was able to probe and figure out. I'm confused about what is the function of A and C and how that communication may happen. Input pin A seems to be connected to the SG3525, which seems to be a PWM regulator IC. Pin C, on the other hand, appears to be connected to the LM358 op-amp. Pin C measures 0.14V when I plug the board, which I reckon is an effective zero, and the voltage of the capacitor is indeed zero volts when I plug the board in. Pin A measures 3.3V when I plug the board in, which is rather strange.

I reckon pin A is supposed to start the charging process, while pin C is supposed to monitor the voltage of the capacitor, or the charging current. The resistor/capacitor network there is very dense and I couldn't probe it with sufficient accuracy to draw a circuit diagram there or even identify which pin exactly from each IC they are connected to. I tried putting a constant 5V on pin A but it doesn't seem to start charging the capacitor. When I ground pin A, it pulses the output voltage quickly to ~800V but it doesn't maintain it and in about <1s it returns to 0V (likely the RC discharge time).

I was just wondering if you had any ideas of maybe how this charging circuit is supposed to work. I am not familiar with the SG3525 IC, and from the datasheet I couldn't figure out whether this uses an enable/disable input or if I need to provide an actual oscillatory signal to it.

Any comments would be appreciated! I'll try to keep you updated if I find anything else. Thanks!!


============EDIT: This is how it works; I've done tests and was able to fire a flash lamp successfully! ==============

Alright, it works now! Took me longer than I care to admit, but I was able to figure it out:

Three connectors are functional in the 8-pin connector:
T = Trigger to flash the lamp (+12V = not triggering, 0V = triggered, falling edge transition);
A = Grounding signal (+0V) to start charging the supply
C = Analog input (+0 to +5V) to define charging voltage; output voltage = 140V/V of input voltage

This is the operation:

• Ground pin A to enable supply charging
• Set the desired charge voltage by setting an analog voltage on pin C (say, 5V=700V)
• Trigger pin T. Active is grounded through a PNP transistor. You need to provide +12V to maintain the pin T inactive. When triggered, an SCR discharges a small (470nF) capacitor charged with the set high voltage (700V) through the primary of a ~1:20 transformer that is in series with the flash lamp (series trigger; generates ~20kV at the flash lamp to strike an arc)
• Looks like there's a delay of 500us between pin T being grounded and the flashlamp actually firing. As I examine this better I'll provide more details.

[moffy]

It doesn't look like the PCB photo is attached.

[zigunov]

oh, dang it! Sorry! Now it is =)

[daisizhou]

Why do you need to reverse engineer this circuit board?
It looks to me like this is a PCB from China,In other words, you can get it at a very low price.
It could be from a medical device or a cosmetic laser,It looks like this device is not advanced or very simple

[zigunov]

If you read my post, I'm not trying to reverse engineer it to build another one; I just need to figure out how to operate it. There's no documentation about what are the expected signals at the inputs.

[daisizhou]

This type of power supply has a "pre-ignition" circuit,The flash lamp is lit by "pre-ignition", then the voltage decreases, the current increases, and enters the low voltage maintenance state.
You can search for an Israeli Phoenix laser. I remember there is a manual for it on the Internet. There is a block diagram on the last page.

[zigunov]

Apparently the manufacturer you mentioned makes CO2 lasers, but this supply is for a Nd:YAG pumped by a flash lamp. The specs are different, but most importantly, the control signals come from a digital touchscreen. An example of such a machine is this: (https://www.aliexpress.us/item/3256806988159318.html)

[Phil1977]

Many of these Laser PS somewhere have a connection for a safety key switch. Maybe that´s what you pulled down with your GND-connection so that the converter started.

[daisizhou]

Apparently the manufacturer you mentioned makes CO2 lasers, but this supply is for a Nd:YAG pumped by a flash lamp. The specs are different, but most importantly, the control signals come from a digital touchscreen. An example of such a machine is this: (https://www.aliexpress.us/item/3256806988159318.html)

Not CO2 laser, please see the last page of the link below
https://www.medwrench.com/documents/view/14195/alma-lasers-harmony-xl-new-service-manual-revc
Pages 125 and 126 have diagrams

[zigunov]

Hi, all,

Unfortunately the circuit pointed by daisizhou is not corresponding to the power supply I have here. I went through the circuit in more detail for the Input A and I think there may be some manufacturing mistake in the board I have; which may be why I may be struggling to make it work. Here's a little schematic that I was able to figure for the input train at position "A".

As far as I understand, the SG3525 PWM circuit will oscillate the transformer that powers the flashlamp capacitor to charge it to an intended working voltage. The part in D7 is labeled "KL3" in its package, which seems to refer to the BAT54 part. According to this (https://www.mouser.com/datasheet/2/149/BAT54C-192576.pdf), there are 4 versions of this part; which seem to be labeled with the same letters in the SMD package. I think the correct version for this part is the BAT54S, which has two Schottky diodes connected facing opposing directions (as I drew in my schematic below). This would make sense because then the mode of operation is straightforward: When the Input A is grounded, the (not) Shutdown Pin #10 of the SG3525 would see a low and start oscillation. When the input "A" is active (5V) or floating, the voltage seen in pin #10 is ~5V, which would shut down the oscillator, therefore being its default behavior (off).

HOWEVER, when probing the BAT54 part, it looks like the actual part installed is a BAT54C, where both diodes are pointing towards the common junction point; which makes no sense to me because the Input "A" would be unable to affect anything in such a configuration. What do you think? Maybe I'm seeing this incorrectly?

So, I grounded pin 10, and there we go, we get 2200V at the capacitor output (don't worry, I put a 6,000V capacitor there!). That's also somewhat concerning because it means that the output from "C" is actually used to control the charging voltage of the capacitor (i.e., the supply is not self-regulated). But again, that is also kind of a feature, because that also means the voltage can actually be controlled; which is nice.

I was thinking of just eliminating the BAT54 part. It looks like, apart from perhaps some protection, there's no real need for it. If you replace the diodes with short circuits, the circuit should work (to be tested!). Any thoughts?

Thanks!

[Buriedcode]

If in doubt, trace it out.

Thats the first thing I would do.  You might not need to know the exact values of the SMD caps to see how this all works together.  The SG3525 is common enough that you can compare the traced out circuit to others, and the opamp and comparator circuits can't be that complicated. I doubt its a 4 layer board given the voltages used, but you never know.

[zigunov]

Ok, I finally figured it out... Took me a while but now I get it. I also will update the original post in case others need this knowledge 

This circuit is the closest thing I found in the internet to the circuit for this power supply:
https://www.iplpowersupply.com/uploads/soft/200805/1-200P5104051.pdf
I also attached the pdf below in case the supplier website disappears in the future.

I was right about the original operation; but here's a summary:
Three connectors are functional in the 8-pin connector:
T = Trigger to flash the lamp (+12V = not triggering, 0V = triggered, falling edge transition);
A = Grounding signal (+0V) to start charging the supply
C = Analog input (+0 to +5V) to define charging voltage.

The BAT54C part is actually correct. I'm not sure how that can be, but it does work with it (I didn't need to make any changes there).

From my measurements, the charging voltage follows this somewhat linear curve, with a slope of ~140V/V:

• Vin=1V -> Vout=138V
• Vin=2V -> Vout=275V
• Vin=3V -> Vout=420V
• Vin=4V -> Vout=560V
• Vin=5V -> Vout=700V

[Phil1977]

A = Grounding signal (+0V) to start charging the supply

That´s what I had in mind regarding the safety key switch...

Thanks for the description! Can you tell where you buy the flashlamps and if you do some light pulse measurement?

[zigunov]

Hi, Phil, this is the flashlamp I got... They are reasonably cheap for their power. Each flash is ~20J of energy 

https://www.aliexpress.us/item/3256805302665250.html?spm=a2g0o.order_list.order_list_main.17.70391802w9TJZq&gatewayAdapt=glo2usa

I took a picture yesterday of my (very safe!) setup. The pulse width is about 150us as you can see in the scope. I used a low-capacitance photodiode for the measurement, its rise time is <1ns so we should be seeing only the flashlamp profile here.

[LaserSteve]

No Series Inductor?

Nd:YAGs storage time is about 256 uS, thus you want to shape the lamp pulse.

Your storage cap will live longer, as well.

This doesn't require magic ferrites, just a low ESR Air Cored Inductor.

STEVE

[zigunov]

Steve, if you see the series inductor is in the trigger transformer. It's a rather clever design, I remember I saw a paper from the 1980's discussing how it works. The ferrite couples a low energy trigger pulse from a 470nF capacitor, starting the lamp, but then when the current from the main capacitor runs through the inductor it quickly saturates the ferrite core and it then behaves as an air-core inductor. From the timing of the lamp flash, the inductance is approximately 100uH.