Hi,

I have a HeNe gas laser and high voltage driver. I've had it for about 10 years and it's now stopped working properly - the output has become pulsed rather than continuous. It's a rather fast pulse, you can't detect it just looking at it but if you move the beam it creates a series of spots. I think the capacitors in the driver are failing and it's not smoothing the input AC power any more.

I'm thinking of putting a capacitor in parallel with the driver to smooth the output. There is already a current limiting resistor in series with the laser tube so I hope the driver is still providing enough peak power to charge the cap and power the laser.

The driver's rated output is 2300V dc at 6.5mA (from a 220V AC input). I've worked out providing 6.5mA for 1/50th second at 2300V would require a 56.5nF capacitor. My local electronics supplier has 22nF, 3kV ceramic capacitors so three of these should provide enough smoothing.

My question: Do you think this is a resonable assumption of the cause of the problem and is the solution viable?

Should I dissasemble the HV driver and try to fix the components in it rather than working around the problem with an external solutiuon?
I don't know the working's of the driver, not even if it is switchmode or purely transformer based. It's just a sealed black tube.

ps, I don't have much experience of circuits at these voltages, but I am aware of the dangers, especially when combined with capacitors or DC supplies which probably contain caps.

 

Mine is sealed and potted, wasn't meant to be repaired, and kind of expensive to replace. I can understand wanting to fix it, rather than spending the extra cash. Ar you sure its the driver, and not the tube, they don't last for ever. 10 years seems like quite a long time...

 

No, I'm not sure it's the driver, kinda just hoping it's not the tube. The tube does fire up and produce laser output - it's just pulsed rather than continuous. If the mirrors had lost alignment or the gas was leaking I would expect there to be no laser output at all.

 

I'm reminded of fluorescent lights as they start to expire. They blink, but not at 60Hz (and certainly not at 120Hz full-wave).

Have you made an estimate of the frequency of the blinking? I would suspect it may be slower than what you would get from line frequency.

If the tube is not conducting, the voltage of the output would build up on the cap until it became higher than the normal operating voltage. High enough to fire even a failing tube. When it finally fired, the voltage would drop to the normal operating voltage, where the failing tube would no longer be able to maintain, so it would stop conducting, the current would stop, and the cycle would repeat. This would be at some unknown frequency, but I would bet it would be less than something based on line frequency.

If it randomly happens to be at or near line frequency, it doesn't exclude this being the problem, but if it is slower, it almost certainly identifies it. Worth a check.

 

When DC turns to AC a bad cap is a likely cause.

 

..........

 

Hi DonQ, thanks for your comments.

We've got a 50Hz supply here in Australia so I'd expect a 100Hz flash. Our TV's are also 50Hz. I'm estimating the laser pulses faster than 50Hz because when it's not moving it looks very steady - much more steady than a 50Hz CRT TV.

Also if you move it in a nice continuous motion the dots look very evenly spaced - this is what first made me think it was linked to the AC cycle rather than some random process.

I don't currently have access to a scope, is there some other way I could test the signal from the HV driver?

If there is not much chance it will break anything I think I'll spend a few dollars on some caps and try that idea out.

 

A lot of DVMs have build in freq counters. You could use a photodiode and feed it into a freq counter.

 

Sounds like your tube has started to go. Think of it as the active element in a neon relaxation oscillator.

 

Also if you move it in a nice continuous motion the dots look very evenly spaced - this is what first made me think it was linked to the AC cycle rather than some random process.

I don't currently have access to a scope, is there some other way I could test the signal from the HV driver?

What I was talking about would be considered a "relaxation oscillator", and can be very regular, not a random process.

I was just thinking about a gross visual measurement. If you move the spot on the wall at "x" meters per second, are there ...whatever... dots per meter? This could tell if you have 20 vs 100 Hz, but probably not 85 vs 100 Hz. Meant as an elimination test...

The eye is not reliable for detecting stationary blinking for anything over about 20 Hz. Film movies used to be (I think they still are) 24Hz, and NTSC TV is 30 Hz (interlaced at ~60Hz, but a new whole image at 30Hz.)

Any freq. meter, with a photosensitive sensor of some sort, would be so much better if you can somehow manage that.

 

I had exactly this problem, many years ago. Mine was due to one of the three transformers arcing internally, and the cure was a new transformer. The arcing was faintly audible, and in a darkened room tiny sparks could be seen in the windings.

 

Thanks very much for your help. I've Googled "HeNe laser relaxation oscillator" and found some very interesting reading about the electrical characteristics of the laser. The starting voltage of the laser is 8k to 12kvolts so my 3k capacitor idea would be bad. Around 2k volts is needed to maintain ionisation.

Looks like the oscillator problem is caused by either:
1) A failed ballast resister letting the circuit discharge too much.
2) A failed power supply unable to provide enough current.

I'm going to buy a new ballast resistor, and also a few spares at higher resistance. If the problem is #2 a higher resistance may reduce the circuit current to what the power supply can provide.

I'll also try to measure the frequency of the flashes.

My reading gives me optimism that it's not the laser tube that's the problem. The only real problem they can have is gas leakage which results in very slow starts (because the start voltage increases and takes time to build up), or no start whatsoever. My tube is instantly starting many times a second.

 

Here's an update on my progress:

I've bought a picaxe microcontroller and photoresistor and setup a light-pulse counting circuit. I wanted to use a photodiode or phototransistor because they react much faster but I could only find IR ones which would not pickup my laser.

Anyway, the laser started out pulsing at 75Hz, and as it warmed up it gradually crept up to 80Hz. After about 5 minutes there was a reading of 67Hz, and the next reading was then zero - The laser suddenly came on at full power continuous.

After turning it off, giving it a few minutes to cool down then starting it again it came on with the first start, the circuit didn't count a single pulse (It measures the number of transitions from on to off, so a clean turn on shouldn't register). I'll leave it overnight and see what happens in the morning.

I can feel a fair bit of heat coming through the case next to where the ballast resistor is, more than I've ever noticed before so I think this needs replacing. The only problem is I can't find anyone who sells a 75KΩ, 5W wire wound resistor. I've got 5 x 15KΩ 1W carbon resistors. I'll see how these go in series but I think the high starting voltage may burn them out.