Need help with tricky AC power noise.

[bcro]

Hello. I work in an optics lab and am seeing evidence of instability in a laser's wavelength due to noise from the power line (120V, 60Hz). In a (successful) effort to clean up both the power and the ground lines, I have things hooked up as follows:

Outlet-->UPS-->Transformer-->Laser, (Config. 1)

where:
UPS = Tripplite SU750XL (on-line, double-conversion UPS)
Transformer = Tripplite IS250 (Isolation Transformer)

This cleans up the power nicely, but the instability in the laser only goes away when I unplug the UPS from the outlet:

XXX-->UPS-->Transformer-->Laser. (Config. 2)

I'm afraid I don't know details about the electronic internals of the laser system, but could any of you enlighten me on the difference between Config's 1 and 2? In either case, the 120V sine wave is just that generated by the UPS, and the Transformer is still there cleaning up the ground line. What's changing when I unplug the UPS?

Thanks a bunch.

 

[keith1200rs]

Have you tried it straight into the outlet without the UPS? Are you sure it is 60Hz noise? The thing you have lost by unplugging the UPS is the battery charging inverter in the UPS. I guess that could be the problem.

What type of laser is it?

Keith

 

[bcro]

Straight into the outlet gives the same noise. In fact, the UPS/Transformer were only installed in an attempt to fix the problem (on the assumption we simply had dirty power).

The laser is a tunable, diode laser. It's wavelength is tuned with a piezoelectric element. Hence, voltage fluctuations into the laser end up as fluctuations in the output wavelength.

I am not set up to perform a Fourier Transform on the signal trace, but I have confirmed at least that a component of the noise is 60Hz, just by triggering the scope at the line frequency.

It's a sure enough head-scratcher, but unplugging the UPS eliminates the problem every time, regardless of what else is or isn't in the loop.

 

[keith1200rs]

I wonder if it is somehow noise relative to ground e.g. And earth loop. The UPS might effectively float both of the mains wires. Have you tried an isolation transformer straight from the mains?

Keith

 

[bcro]

I've attached a diagram showing all of the device configurations I've tested, and the results. I definitely see the Transformer making a difference when I view the ground line on a scope, but it doesn't seem to affect the laser's output.

I'm waiting to hear back from the laser manufacturer on this issue. Tripplite's tech support has proven useless.

If floating the mains lines is a possible solution, is there a device out there that would do this without having to run on batteries? At this point I'd be willing to buy one and try it out.

 

[keith1200rs]

Assuming that somewhere the laser power supply connects to ground, maybe breaking that connection and replacing it with a resistor or inductor may help.

Keith

 

[bcro]

Worth a shot I suppose. What would be a reasonable value (or range of values) for a resistor to put inline with the ground?

On another note, if running off of the unplugged UPS works because it breaks the connection to the building's ground line, how bad would it be to simply disconnect the ground line from the laser and run hot/neutral only? If I kept the UPS inline so as to guarantee proper voltage, would I have anything to worry about there?

 

[keith1200rs]

I have seen 1M ohm used to link ground to -ve DC. Sometimes with a capacitor I parallel. I am not sure if it creates any safety issues.

I didn't suggest pulling the ground wire out of the mains because it is usually dangerous! I have occasionally done it though!

Keith

 

[FvM]

I don't know your equipment, but all usual isolation transformers are disconnecting the protective earth. If so, the ground loop hypothesis would be refuted. If the the protective earth is still connected through the isolation transformer, you can interrupt it without danger (at least assuming a recognized class II transformer that provides safe iolation).

One simple explanation would be, that short time mains voltage fluctuation simply pass the UPS charge controller and the battery. Only a sensitive measurement could verify this assumption. The direct mains connection can also carry conducted electromagnetical interferences, e.g. sourced from a near by AM radio station or VFD motor controllers. But it's difficult to know without determining the susceptible part of the instrument, unfortunately there are several option.

In any case, the instrument seems to be badly designed. If the manufacturer's support is acting lazy, you should tell them about a planned scientific paper where you report detailed experiences with the performance of different diode laser controllers...

 

[bcro]

Okay, so I took both the UPS and Transformer and put an ohmmeter across the incoming and outgoing ground terminals. It showed no resistance, which I assume means that neither device disconnects the protective earth. Perhaps what I want is an isolation transformer that does disconnect the protective earth, and does so safely. If I go shopping for one, how do I know which ones disconnect the earth line? Are there any that you guys would recommend?

By the way, I certainly agree that the big disappointment here is the laser system's susceptibility. These systems cost $15k a piece, and we have 4 of them. I am still waiting to hear from the manufacturer.

 

[keith1200rs]

I am not sure where you are, but as said by FvM, an isolation transformer should not connect ground - that is the point of them. The yellow ones used on building sites should 'float'

Keith

 

[bcro]

The solution for now seems to be to run off of that Tripplite Isolation Transformer with the safety ground disconnected between the laser and the transformer. This removes a large part of the instability, and doesn't seem to give any adverse side-effects.

I recently did some reading about "balanced power" transformers:

http://www.b-p-t.com/balanced-power-ultra-isolators.html

Maybe that's worth a shot.

Thanks again for the help.