OK, first I searched and nothing on this so time to ask.

I've seen allot of topics on isolation transformers, and no topic about connecting to mains without it will fly.

My question is, with my first year of labs behind me I don't remember it being a requirement of the set up, but that being said, the closest we got to mains power is a variac, that I plugged a transformer into it, then banana plug into the bread board. I'm looking for my lab notes, but i remember the Vpeak at the diode full wave rectifier being in the 40V range.

The transformer was certainly no where near the size of these isolation transformers I've seen, on youtube and very close in size to the transformer we put in the regulated power supply project which was a 120/26 center tap transformer.

At what point (voltage wise I guess) do I need to transition to an isolation transformer?

I don't (yet) have a variac, and plan to for now use the 5V regulated 0-15V power supply built in one of the project labs.

 

Every transformer that consists of a primary winding and a secondary winding and that are not electrically connected is an isolation transformer regardless of voltage.
Usually when we talk of isolation transformers we are referring to 120VAC input and 120VAC output. But a step-down transformer such as 120VAC input and 15VAC output also provides isolation from the mains.

In general we always use an isolation transformer. PC switching supplies tend to be not isolated. They remove the transformer to reduce weight and cost.

 

Does it matter if the center tap is to ground?

 

No problem if center tap is on secondary. A " Variac " does not supply isolation.

 

From the perspective of what voltage is safe to work on live - there is always robust argument about this. There was a related thread recently running on this site. One can make an argument for any voltage to be unsafe under the "right" conditions.

You might adopt the typical ELV definition - e.g <25Vrms [35V peak] or <60V DC ripple free. Depends on which electrical standard applies for a particular country. These might be considered low risk values - but definitely not risk free.

The point is best made that one should always endeavor to work (where practical - practicable??) on electrical equipment with the power off. In the absence of that option being available you take every reasonable step to reduce the risk to an acceptable level.

I was always very nervous about using a variable auto-transformer without an isolation transformer. I think Variac is a trade name.

If you earthed the secondary center-tap of a higher voltage ratio transformer (let's say 120V:120-0-120V) you would confront the same hazards applicable to a direct 120V ac mains connection.

 

....PC switching supplies tend to be not isolated. They remove the transformer to reduce weight and cost.

PC switching supplies do indeed have isolation transformers! But, because of the high frequency used in the switching circuit, the step-down/isolation transformers have a much smaller physical size than transformers used for the same power rating at 50 or 60Hz.

Ken

 

True, there is high frequency transformer in a switching supply but all the electronics on the switcher side are HOT! There is zero isolation.

Where would you use an isolation transformer? Here is an example.
Many years ago we used an inexpensive TV as a video monitor. This did not have a built-in transformer but it came with a polarized AC power cord. Depending on which way the user plugged the power into the wall outlet there was a 50-50 chance that the chassis of the TV would be LIVE (as in HOT). So we used a 120/120VAC isolation transformer to prevent this from happening. I suppose this made our inexpensive solution not so inexpensive after all but it was still cheaper than buying an industrial video monitor.

 

Here is another example of where you could use an isolation transformer.
Most if not all oscilloscopes are grounded and hence voltage measurements are taken with reference to GND.

Now suppose you wanted to measure the voltage across two terminals using the probe and the ground clip. That would not be a good idea because you would be grounding your circuit through the ground clip.

You use an isolation transformer to take the scope off ground, i.e. your scope is now "floating". Now you can use the ground clip as your reference since it is now a floating reference.

Alternatively, you could power your test circuit (power supply and everything connected to it) via the isolation transformer and achieve the same effect.

 

Does it matter if the center tap is to ground?

"Ground" is a catch-all term without a definite meaning.

Your circuit "ground" is typically the point you reference all other measurements and signals to.

On the AC line, there is a "neutral" and sometimes a "ground." Here "ground" means literally the Earth's potential and is sometimes achieved by connecting to plumbing lines. The neutral is connected to this ground somewhere. A two prong AC plug uses these two voltages.

As power normally flows from the "hot" to neutral the voltage at the plug on neutral may be slightly above ground, hence the additional third prong is there for additional safety.

On lab power supplies it is common to see plus and minus outputs, and an additional ground output. The additional ground output is the AC ground or third prong on the plug.

"It depends" so you sometimes may (but usually not) connect a lab supply output to earth ground.

Here is another example of where you could use an isolation transformer.
Most if not all oscilloscopes are grounded and hence voltage measurements are taken with reference to GND.

Oh you bring back a bad memory of the day I was setting something up on a computer controlled test stand and had a scope ground clip short a power supply. The current ran thru an IEEE-488 bus that connected together about 100K of equipment that immediately burned open traces and made it all fail. I trucked a power supply, signal generator, spectrum analyzer and maybe some other stuff down to calibration with repair tags.

All this on a Saturday where I had dropped by to complete one last task, and as I was salaried I wasn't even getting paid!

Surprisingly enough the computer itself did not fail so I was able to borrow enough equipment to continue on testing. Come Monday morning my boss (a real bastich) came by to find out what happened. I told him of the failure, that the equipment was tagged and out for repair, I had a work-around and was just completing the tests so we could ship.

He nodded, and asked but one question: "This was Saturday?"

"Yes" I said. He then nodded and walked away.

WHEW!

 

"It depends" so you sometimes may (but usually not) connect a lab supply output to earth ground.

In the situation I'm thinking of, 3 wires from the pig tail entered the case, one went to the transformer, one went to the transformer through a fuse, and the common got bolted between the transformer frame and the case.

On the other side (secondary) of the transformer, all three wires were soldered to posts on the circuit board, with the center tap wire going to a post marked GND.

 

A picture is worth a thousand words

Is this isolated?

 

Yes, the picture makes it clearer. You may choose to call pin 2 of the transformer "circuit ground" in which case the output of the bridge will deliver positive and negative voltages with respect to this "circuit ground". However your voltages are still "floating".

You may choose to connect pin 2 to the chassis. (The green wire (GND) from the power cable must always be connected to the chassis).
Now your "circuit ground" is the same as chassis ground and earth ground. Your supply is no longer floating but referenced to earth ground.

In both cases, your power supply is isolated from AC mains.

(On another topic, serial transmission isolation such as isolated RS-485 networks is a different thing all together).

 

In both cases, your power supply is isolated from AC mains.

(On another topic, serial transmission isolation such as isolated RS-485 networks is a different thing all together).

And if I hook up an oscilloscope I would put the probe ground to circuit ground, or would that join circuit ground to chassis ground through the OScope??

As for Serial transmission... back the trolly up... I ain't there yet

 

As it stands there is no connection from the mains earth and a circuit reference ground (common) at terminal 2.

And this is a very low hazard case anyway with such a low secondary voltage.

Suppose you wanted to do some oscilloscope observations of the circuit or perhaps other circuitry connected to the rectifier. As part of the CRO setup for measurement, you may connect the scope ground to this secondary side ground. If the CRO ground is connected to the CRO frame, which may also be connected to mains earth, you then would have a situation in which the secondary side is connected to mains earth for the duration of the CRO being connected. This doesn't present a hazard to the observer as such, but it does show that one can inadvertently make connections that weren't consciously anticipated.

 

Thanks Allot, for now I can't see me going past 30V but next year is a different matter

 

(And if I hook up an oscilloscope ...). This is a very good question.

Consider the two cases:
(1) centre tap connected to chassis and
(2) centre tap not connected.
Assume your scope is already grounded through it's power cord.

Case 1: Your supply is referenced to earth ground and so is your scope. You may choose to use the tip of the probe only or you can also connect the ground clip to the chassis or circuit ground, no place else.

Case 2: Your supply is floating. When you connect the scope probe alone you will observe a large 60Hz signal (for North America). You need to connect the ground clip (or the ground post on the scope) to the circuit ground (centre tap). As I pointed out in my second example of using an isolation transformer, you may also connect the scope ground to any point in the circuit if you want to observe the voltage drop across any component. Don't do this in Case 1.

The voltage level (your example of 30V) is irrelevant for most of this discussion.

 

Another thing I should point out about your circuit diagram.
Make sure that the LIVE or HOT line is the fused and switched line.
(In North America, it is the BLACK wire from the power cord. I do not know the colour codes for other countries).
You do NOT put the fuse and switch on the NEUTRAL side (WHITE wire).

For wiring colour codes: check out
http://www.allaboutcircuits.com/vol_5/chpt_2/2.html

 

In general we always use an isolation transformer. PC switching supplies tend to be not isolated. They remove the transformer to reduce weight and cost.

Mr Chips,

It is wrong to say that PC power supplies are not isolated. It would be a very dangerous world out there if every desktop PC had its motherboard and I/O interfaces all directly connected to the mains circuits.

PC Power supplies eliminate the large iron core transformers that provided isolation of yesteryear by using high frequency switching techniques so that the primary mains side of the power supply can be isolated from the secondary output side through a transformer made with high permeability powered metal cores.

Michael Karas

 

I agree that the low voltage side is isolated from the mains through the high frequency transformer. But I still stand that the switching supply is not isolated from AC mains. That is one of the many reasons I do not repair switching supplies.

 

I agree that the low voltage side is isolated from the mains through the high frequency transformer.

But I still stand that the switching supply is not isolated from AC mains. That is one of the many reasons I do not repair switching supplies.

Think of all the wall warts (same as PC supplies) that are SMPSs. These are considered "isolated" by ISO and UL standards. I think that the concept of electrical isolation is from the user's end of the device, not what components that may be attached to the power line.

But I still stand that the switching supply is not isolated from AC mains. That is one of the many reasons I do not repair switching supplies.

I think if you are probing around the primary side of an open SMPS you are absolutely right. With that concept, TBayBoy's supply is not isolated because the switch and fuse are connected to the line.

But, if you're a user of a properly housed, transformer-coupled Linear or SMPS, it is "isolated".

Just a matter of prospective.

Ken