Oscilloscope, mains, isolation transformer, double check

[cthree]

Hi guys,

I tend to stick to 12 volts and less but I want to design and build a sine wave light dimmer which will obviously run at mains voltage. I don't care to die or suffer pain so I'm planning to take the recommended safety precautions:

Unplug before touching
One hand in my pocket
Static probing only, no probing of the live circuit

In addition I am planning to buy an isolation transformer because I must use my scope to probe and characterize the AC waveform (kind of necessary for a dimmer application) and I also don't want to inject any nasty noise back up the spout.

Can I get a quick confirmation just for my piece of mind that I have this right? Is it correct and normal practice to have a grounded scope (earth reference) probing a DUT connected to an isolation transformer? I know it has probably been said before many times, I just want to hear it again just to be sure I haven't missed anything.

I am planning to buy a BK Precision Single Output Isolation Transformer here http://www.bkprecision.com/products/power-supplies/1604A-single-output-isolation-transformer.html It's not very expensive ($115 plus shipping) and I think (hope) the 1.25A capacity is enough to provide adequate current for a dimmable LED bulb.

I am also considering the Dale/Fluke MI/IT400 "Medical" Isolation Transformer http://www.ebay.ca/itm/Dale-Fluke-IT400-Medical-Grade-Isolation-Transformer-110V-NEW-ORIGINAL-BOX-/161010666194 Which actually seems a much better unit with higher output (2.8A) and 3 outlets instead of just one and is half the price at $50 plus shipping.

I do understand that if I get the Dale/Fluke IT400 I need to take is apart and remove the ground jumper to make it an actual isolation transformer but I'm ok with that.

Any final words of wisdom?

Thanks.

[nctnico]

Buy a differential probe because this is the safest way.
Mains are usually protected by a GFI. If you use an isolation transformer the GFI is no longer effective. Worse, if you connect one leg of the transformer to ground through the oscilloscope you have lethal voltages on several points. And don't be tempted to disconnect the scope from ground!

[perfect_disturbance]

For a DC DUT does a good lab power supply provide adequate isolation? Does an ungrounded isolation transformer relieve concerns about ground loops when using a scope on AC powered devices?

[KedasProbe]

Keep a flashlight close anyway

[cthree]

Buy a differential probe because this is the safest way.
Mains are usually protected by a GFI. If you use an isolation transformer the GFI is no longer effective. Worse, if you connect one leg of the transformer to ground through the oscilloscope you have lethal voltages on several points. And don't be tempted to disconnect the scope from ground!

Mains in Canada are not protected by a GFCI unless the business end come within 2m of plumbing or the outdoors. My bench is connected through an AFCI because it is in a bedroom.

Can someone clarify the comments regarding GFCI/RCD. I will look into a differential probe for my scope but the grounding issue is still a concern.

[cthree]

I'm not sure the RCD/GFCI advice is good advice. I'm sitting here thinking about it and I don't see how it applies to my application. I'm going to be working on an ungrounded, two wire lighting dimmer connected to a lightbulb. If I'm going to to get a shock its going to be a short circuit from hot to neutral rather than from hot to ground is it not? I'm not even sure the isolation transformer is any practical value in this circuit since the DUT isn't connected to ground.

I will be using my scope to examine and test the mains waveforms on both input and output to the load. I will also use it for the low voltage DC circuits of the device.

What is the rationale for the GFCI?

[IanB]

A GFCI is a magic charm that protects from all harm 

But seriously, you can get all carried away with this stuff and start doing things without good understanding of when they do or don't help. So it's good to be skeptical as you are being.

Fundamentally, you don't absolutely need even an isolation transformer, as long as you keep track of where your grounds are and be aware of what's live in your circuit.

You need to know that ground on the oscilloscope is tied to mains ground. So you must either attach the ground lead of the probe to something that is already connected to mains ground, or attach it to something floating and isolated. But as soon as you attach that ground probe to a floating circuit it won't be floating any more, it will be ground referenced--you have just neutralized your isolation transformer 

There's a saying in the workshop, "measure twice, cut once". Similarly in the electronics lab, "think twice, before you attach that probe".

Also, it is often a good idea to clip probes to test points with the device turned off, before turning it on. That way, there is no danger of your hands slipping and touching anything live, or shorting something out.

[IanB]

The advantage of the isolation transformer is you are not limited in where you can attach the scope ground probe. You can attach it to part of your circuit under test that would be "neutral" or "live" in normal operation, or anywhere in between. You can't do that if your circuit is powered from the real mains.

The isolation transformer is not so much to protect you, but to enable you to probe the circuit under test without restriction on where you can place the ground probe.

[madires]

You need to know that ground on the oscilloscope is tied to mains ground. So you must either attach the ground lead of the probe to something that is already connected to mains ground, or attach it to something floating and isolated. But as soon as you attach that ground probe to a floating circuit it won't be floating any more, it will be ground referenced--you have just neutralized your isolation transformer 

He could create a differential scope probe by adding two channels (one inverted) as long as the probes and the scope support the voltage. Of course he should remove the ground clips first. That way he doesn't ground the DUT and the isolation transformer protects him in case he touches anything with mains voltage. We shouldn't rely on RCDs only, they can fail. He could also wear isolation gloves to add another layer of protection.

[ftransform]

I recommend soldering wires to places you want to test, hooking up your oscope to that and then turning it on.

Soldering a little extension wire is little effort, it will provide a secure place for your scope to grip and it will ensure that you connect to the right place and do not short anything out.

Unless you are doing stuff that does not like extra wires... in that case use your mad surgeon skills.

[nctnico]

There's a saying in the workshop, "measure twice, cut once". Similarly in the electronics lab, "think twice, before you attach that probe".

Thats absolutely true! A few weeks ago I was at a customer to do some measurements with a scope on a big industrial battery charger. It turned out the output of the charger is connected to the mains directly from a rectifier! I had to use two probes and let the scope substract the signals to get some kind of differential probe.

[lewis]

Use a differential probe. End of! Much safer and less susceptible to measurement errors. You don't need to worry about isolation transformers, which ground potential is where, and you can measure voltage across any component in your mains earth referenced application.

[cthree]

I like the differential probe idea accept for the prices. It is obviously the right way to go 

Having digested it, a GFCI and differential probes are what I seem to need. An isolation transformer would be beneficial if the DUT is grounded, but isn't in this case.

The GFCI won't help with a short between hot and neutral but will trip if the hot should come into contact with some grounded thing on the bench which is extra good if I'm in the middle.

Does that sound sane?

Thanks everyone for their advice and input. Are there high voltage differential probes that are the bees knees, fairly priced and universal bnc? 25 MHz is way more than enough, this is 60 Hz stuff. My scope is an analog 50Mhz tek 2225. I want to get a Rigol DS2102 but I promised my wife I'd sell some crap in the basement to pay for it and well.....  O0