CW arc lamp PSU design, now the rat has moved out.

[LesioQ]

As it goes to a capacitor, I'd vote for position B, being likely a path to trigger pulse. However if the trigger circuit was to receive a series of trigger pulses, this might as well be a resonant circuit for driving a repetitve flashlamp-based laser. Sure this was designed for an arc lamp, not a flashlamp ? This triggering 555 is also a tell-tale ...

My other thoughts are around the inductors, especially the buck one, being too similiar to a toroid transformer. Also it's inductance seems too high to be a ferrite-core-wound. It's insulation ... prevents cooling, so welcome in CW arc lamp circuit. It's quite a goal to get a 3.5milli Henrys at >20Amps.

I cannot locate the LEM in the pictures ... could You point it ?
How big was the heatsink for IGBTs ? Was it fan-equipped ?

Could You post some pictures of the front and back of the PS housing ?

Piotr.K

[weartronics]

Hi Piotr

As it goes to a capacitor, I'd vote for position B, being likely a path to trigger pulse. However if the trigger circuit was to receive a series of trigger pulses, this might as well be a resonant circuit for driving a repetitve flashlamp-based laser. Sure this was designed for an arc lamp, not a flashlamp ? This triggering 555 is also a tell-tale ...

I am pretty sure the power supply is for CW arc lamp. The front panel says "CW LAMP DRIVER", and I think the lamp electrodes are intended for CW operation. So I think you and Steve are correct, it is probably a "start boost" capacitor to provide high current required for a short time at the lamp start, separate from the HV trigger pulse.

I am trying to reverse engineer the 556 board now, but it is very hard work. At this stage, I suspect one 555 is used to generate the IGBT gate drive waveform, and the other 555 is used to generate a series of trigger pulses automatically until the lamp is lit (then it stops), this is my guess.

My other thoughts are around the inductors, especially the buck one, being too similiar to a toroid transformer. Also it's inductance seems too high to be a ferrite-core-wound. It's insulation ... prevents cooling, so welcome in CW arc lamp circuit. It's quite a goal to get a 3.5milli Henrys at >20Amps.

I agree it looks like a toroid transformer, but it only has a single winding, and it is definitely wired in series with the lamp ... I hope it is sufficient for CW use.

I cannot locate the LEM in the pictures ... could You point it ?
How big was the heatsink for IGBTs ? Was it fan-equipped ?

The LEM is very well hidden, I will point it out in the photos below. The 3 phase AC to DC brige, IGBT and huge series diode are mounted to an aluminium block which is water cooled from outside the power supply, there is no fan.

Could You post some pictures of the front and back of the PS housing ?

Coming right up! Since I think it is certainly a CW power supply, I will connect the large mystery capacitor in position B so it can provide the "start boost" current. If the photos provide any other information, I welcome any other comments you have!

Best regards,

weartronics









Here are the logic parts that I removed from the power supply. The PLC does not start up after the rat incident, and I don't know anything about the program or the connections to fix it, this is why I need to redesign the logic parts.



This is the cooling pump (heat exchanger has been removed for cleaning). The logic for the cooling pump is more complicated than the logic for the power supply! I will remove all the undocumented microprocessor boards (I think they need to interface with the PLC, which is broken anyway). Then I will simplify this to just use the flow switch, thermistor, solenoid, etc...



Here is the LASER resonator. How can I identify which end is the HR and which end is the OC? The Q-switch is Crystal Technologies model 5100 type 45, but I have no way to know whether this is 24MHz or 27MHz. Unfortunately the suitable Q-switch driver was not included, but I found a 27MHz driver on eBay which has a 50/50 chance of being suitable.

[mixedgas]

usually there are giveaways in the numbers and markings on the optics, and 9 times out of 10 the HR is at the qswitch end.

Take a picture of the chips next to the LEM would you? I think thats a giveaway as to where the control board is.... And make sure we know what the part numbers are.

Steve

[mixedgas]

email crystal tech, they no longer make them, but they still know what they are from the part numbers. It usually takes their sales guy 2-3 days to get back to people who contact him.

Steve

[weartronics]

Hi Steve,

The 16 pin DIP next to the LEM is a 556 (dual 555 timer), the 8 pin DIP below it is a CA3240 dual op-amp, and the 8 pin DIP further to the right is a LM311 comparator.

The DB-9 connector was connected to the PLC, it must have some signal to indicate the desired current, and some interlocks etc... The red connector at the top has a twisted pair running to the SCR board (the SCR triggering signal no doubt) and a twisted pair running to the over-temp switch on the power electronics cooling block.

The little grey alien near the red connector is a fiber optic TX (i.e. an LED), which ran through a plastic fiber to a fiber optic RX (i.e. a photodiode), which drives a transistor totem pole which drives the IGBT gate. The rat liked the cladding and he ate quite a lot of it, but it seems the core was not so tasty, so it will probably still work fine .



I am working to reverse engineer the schematic and function of this board today. If I can work out the function, this board may be useable.

Edit: I really think this board is too complicated to reverse engineer, I think it will be much easier to design something to replace it.

Best regards,

weartronics

[weartronics]

email crystal tech, they no longer make them, but they still know what they are from the part numbers. It usually takes their sales guy 2-3 days to get back to people who contact him.

I actually emailed Crystal Tech (now "EPCOS") about this numerous times in March last year. I got an initial reply from a Mr George Sadilek, but he could not give my information about any specific types. He said he remembered they were making them in 24MHz and 27MHz versions, but I got no reply when I asked if there was any way he could find specs for the specific crystal. I have just emailed him again, you never know!

[mixedgas]

That LEM board screams PWM, based on its layout ....

Steve

[weartronics]

That LEM board screams PWM, based on its layout ....

The signal driving the IGBT gate seems to originate from a 555 output, then a transistor, then the fiber TX, then the fiber RX, then a totem pole, then then IGBT gate. So like you, I am pretty certain that it is a digital PWM signal through the fiber. I think I will simply pop the LEM and the fiber TX off this board, and measure the LEM and drive the fiber TX myself with a 25kHz PWM, modulating the duty cycle for happy lamp currents.

As for the trigger pulse, I guess I can sense when the lamp current is zero, and pulse the trigger repeatedly (say every couple of seconds) until the lamp current jumps to the setpoint. The way I understand it, it should almost always light first time, but occassionally may need a few bites to get going. My only concern is the polarity of the trigger pulse, I suppose the trigger pulse must be the opposite polarity to the normal lamp drive current, so that it will be blocked by the jumbo diode.

Best regards,

weartronics

[mixedgas]

the 60 page PE lighting guide pretty much does it

http://optoelectronics.perkinelmer.c...Flash+Lighting

my lamp is a QCW, second one down in the column.

Steve

[LesioQ]

That LEM board screams PWM, based on its layout ....

Steve

Yeah, though PWM done the hard way via timers due to some reason ...

Well, seems like You're just a PCB away from testing it with a dummy load!

Piotr.K