AC Isolation Transformer, Autotransformer and Dim Bulb. What Order and Why?

[t1d]

I have these devices. In what order should I arrange them and why?

I am newer to the AC game. Thank you for your help!

[james_s]

What are you trying to do? Normally if I'm using a dim bulb I would put it after the isolation transformer, but it doesn't actually make much difference. By autotransformer do you mean a variable autotransformer? If that's the case then that goes first, plugged directly into the wall in most cases.

[Ian.M]

By "autotransformer", I assume you mean variable autotransformer, aka: Variac.   

The lineup should be (in order from mains supply to D.U.T^*):

• Isolation transformer (if used at all)
• Variac
• Dim Bulb

Reasoning:
Using an isolation transformer disables both upstream and downstream GFCI (RCD) protection^# and the notional safety of 'isolation' making contact with a single downstream circuit conductor 'safe' is removed as soon as you connect any test equipment that introduces a ground, e.g. your scope probe's ground clip.  If you need to probe waveforms around the primary side of a PSU, for safety, you should be using a >500V Cat II rated isolated differential probe, or a handheld scopemeter with comparable ratings.

The isolation transformer is most usually only rated for your nominal mains voltage +10%, so should *NOT* be used downstream of a Variac capable of boosting the supply voltage, as excess voltage will cause saturation resulting in grossly excessive current.   If your Variac doesn't boost or your isolation transformer is rated for the max. Variac output voltage then you can put the Variac first as James suggests, which can be beneficial if the isolation transformer isn't really 'beefy' enough (VA) for the load.

You shouldn't pass the magnetizing current of the isolating transformer and/or Variac through the dim bulb, as it will pre-load it increasing the supply impedance significantly, possibly resulting in incorrect operation of the D.U.T.  Also its frequently beneficial to insert the dim bulb downstream of the D.U.T's  rectifier and reservoir capacitor(s) as there is often enough energy stored in the capacitor(s) to blow the chopper transistor(s).  If you insert the dim bulb between the capacitor(s) and the feed to the switched mode transformer, it will often act fast enough to save the transistor(s) if there is a fault in the SMPS drive circuit.  N.B. If using a dim bulb after the reservoir capacitor(s), you need to disconnect all output loads as the dim bulb drastically limits the available power.  Also, if the PSU has a voltage doubling rectifier, (not uncommon in 115V or 100V countries) you'll need a 240V dim bulb, as the primary side DC bus voltage can be up to 360V.

* D.U.T. = device under test

# See https://www.eevblog.com/forum/beginners/(another)-isolation-question/msg3128246/#msg3128246

[t1d]

Thanks, James and Ian, for your efforts to help me. Much appreciated!

@ Ian
- The "why" information that you provided is exactly what I needed.
- Yes, the popular notion is that the Iso Tran will protect the oscilloscope. But, your explanation of the ground path shows why that is errant thinking. You likely saved my bacon, right there. Thank you!
- Inserting the dim bulb directly into the DUT's supply path is a neat trick that I may have never thought of = Kudos!
- My system was already arranged in the order you suggest, just based on my assumptions. But, I know that assumptions will get you into trouble. I guess that is why the question has always nagged at me to learn the details... I will be printing out your post and hanging it in my lab, next to the Iso Tran...

[Ian.M]

The isolation transformer *DOES* protect the oscilloscope (used with an ordinary ground-referenced probe), however the  instant you connect the probe ground, you remove all safety for the technician, as all parts of the primary side of the SMPSU apart from the now grounded negative rail become 'Live', with no GFCI (RCD) protection.

You *can* work on such a circuit, but the risk of electrocution is significant, (hundreds of volts DC is in no way 'safe'), and IMHO is unacceptably high for anyone who isn't trained and qualified to work on live high voltage DC circuits.  If you must without the formal training, the only reasonably safe way to do so is to hook up test equipment with the mains supply off, and stand back before applying power, make your measurements, then isolate (disconnect power to) the D.U.T. again before touching the D.U.T. and any probes attached to it.

[t1d]

The isolation transformer *DOES* protect the oscilloscope (used with an ordinary ground-referenced probe), however the  instant you connect the probe ground, you remove all safety for the technician, as all parts of the primary side of the SMPSU apart from the now grounded negative rail become 'Live', with no GFCI (RCD) protection.

You *can* work on such a circuit, but the risk of electrocution is significant, (hundreds of volts DC is in no way 'safe'), and IMHO is unacceptably high for anyone who isn't trained and qualified to work on live high voltage DC circuits.  If you must without the formal training, the only reasonably safe way to do so is to hook up test equipment with the mains supply off, and stand back before applying power, make your measurements, then isolate the D.U.T. again before touching the D.U.T. and any probes attached to it.

I will have to think on this, to sort it out. Here is what I do understand...
- Depending on how and where a scope ground plug is connected, the iso tran may not provide the protection the technician is expecting.
- The mid-section of any form of switch mode system (inverter, converter, etc.) can have extremely high voltage levels lurking to bite you, hard.

[bdunham7]

- Depending on how and where a scope ground plug is connected, the iso tran may not provide the protection the technician is expecting.

Yes, if you use any test instrument that is ground referenced, you lose your isolation.  Not that it is wise to count on it in the first place.

- The mid-section of any form of switch mode system (inverter, converter, etc.) can have extremely high voltage levels lurking to bite you, hard.

Yes and I referred to these hazards in another post as unmitigable.  Neither an RCD or an isolation transformer will protect you, although you can gain some protection by not have any part of the DUT or your instruments grounded.

[bdunham7]

The isolation transformer is most usually only rated for your nominal mains voltage +10%, so should *NOT* be used downstream of a Variac capable of boosting the supply voltage, as excess voltage will cause saturation resulting in grossly excessive current.   If your Variac doesn't boost or your isolation transformer is rated for the max. Variac output voltage then you can put the Variac first as James suggests, which can be beneficial if the isolation transformer isn't really 'beefy' enough (VA) for the load.

If the isolation transformer could be saturated by the variac, then that would be a fair point and I suppose this is worth checking.  However, I haven't seen that to be the case in real life (and being in 60Hz land may be part of the reason) and the problem with putting the isolation transformer first is indeed excessive sag under even moderate loads.

You shouldn't pass the magnetizing current of the isolating transformer and/or Variac through the dim bulb, as it will pre-load it increasing the supply impedance significantly, possibly resulting in incorrect operation of the D.U.T. 

There are cases where you want this behavior as a deliberate design feature.  If the bulbs are a tiny bit warm, your maximum surge current will be lower.  But generally I agree-bulbs last, if you use bulbs. 

[james_s]

I will have to think on this, to sort it out. Here is what I do understand...
- Depending on how and where a scope ground plug is connected, the iso tran may not provide the protection the technician is expecting.
- The mid-section of any form of switch mode system (inverter, converter, etc.) can have extremely high voltage levels lurking to bite you, hard.

The reason you need an isolation transformer to protect the scope is because you are working on a circuit that is not isolated and "ground" in the circuit is not the same as earth ground, it could be floating at hundreds of volts above true ground which is what the ground of your scope probe is at so when you connect the two it goes bang.

When you have an isolated device it is relatively safe to you because you are tied to ground while the circuit is floating, no current can flow through you to the earth because the isolation prevents a complete circuit. When you connect the scope ground to something in the circuit suddenly that node is tied to earth ground and other nodes that have a significantly different potential relative to the point you have grounded can give you a shock. Since you are at ground potential and so is the scope, the ground lead on the scope can be thought of as being an extension of your finger that is always now resting on one node of the circuit. If another of your fingers touches a different part of that circuit it is like poking two fingers into the isolated device and touching different points, you get a shock.

[BrokenYugo]

You *can* work on such a circuit, but the risk of electrocution is significant, (hundreds of volts DC is in no way 'safe'), and IMHO is unacceptably high for anyone who isn't trained and qualified to work on live high voltage DC circuits.  If you must without the formal training, the only reasonably safe way to do so is to hook up test equipment with the mains supply off, and stand back before applying power, make your measurements, then isolate the D.U.T. again before touching the D.U.T. and any probes attached to it.

I'd go so far to say that this is how anything with hazardous voltage present should be measured when possible, regardless of skill level. Familiarity breeds complacency and there's little room for that when one wrong move can kill.

[t1d]

This is all such useful and extremely important information, as it directly relates to personal safety. I am very grateful for all of the effort you are making to reply and to help me. Thank you.

Though I am still lacking in SMPS learning, I recently realized that SM designs carry very high voltage. That is part of the reason that I began to build up an isolated AC supply.

Then, it occurred to me that SM designs are in most all modern electronic devices... including innocent looking small wall warts. I was so moved/freightened that I wrote to my DIY buddies and cautioned them, passionately, to get the needed safety equipment... Meaning, the components of an AC power supply, etc.

I even built an e-cap stinger, for one of them that is less electronics oriented, because he often repairs air conditioners. Air conditioner e-caps are very mean animals.

I also upgraded my personal DIY e-cap stinger... I simply added high ohms and wattage resistors across the lead inputs of a cheap multimeter. Now, even though the voltage readout values are wrong, I can at least see when the voltage reads zero. I added caution labels, so the meter does not get used for other purposes. And, I know not to use it for super high voltages, because of the cheap leads, glass fuses and lack of impact bracing. But, it is somewhat better than nothing and I have a proper Brymen 869s, for the more dangerous situations.

As I write this, it comes to mind that I could make an adapter for my Brymen. Just a double banana male to female connector with the needed resistors connected across the posts... Plug the adapter into the meter and the leads into the adapter. I would need to encapsulate the resistors, for explosion protection, but the other robust protections already built into the meter would then be in play. Hmm... I just high-jacked my own thread. lol

[tggzzz]

Dim bulbs and SMPSs are not always a good combination.

SMPSs are constant power output devices, supplying whatever current the load requires at the specified voltage. If you reduce the input voltage, the SMPS will therefore increase the input current. Components such as transistors may object to that.

[t1d]

Dim bulbs and SMPSs are not always a good combination.

SMPSs are constant power output devices, supplying whatever current the load requires at the specified voltage. If you reduce the input voltage, the SMPS will therefore increase the input current. Components such as transistors may object to that.

Excellent point.

[jeroen79]

A series light bulb will help limit the current.
By doing so it will also drop the input voltage which may cause a smps to try to draw more current.
If this gets into some runaway situation the bulb will still limit the current while the voltage drops near 0.

A variac can be more dangerous in this aspect.
If you slowly bring up the voltage you would start with a high current.
This could damage both the dut as the variac.

I think it is useful to draw out any setup you may want to use.
Then you can see how things are connected (or not) when you perform some action.

Also get to know your equipment.
Which connections are floating? Which are tied to earth? Which are tied to another potential?

[tggzzz]

A series light bulb will help limit the current.
By doing so it will also drop the input voltage which may cause a smps to try to draw more current.
If this gets into some runaway situation the bulb will still limit the current while the voltage drops near 0.

That would depend on the balance between components' behaviour, which isn't easy to assess.

A variac can be more dangerous in this aspect.

Agreed.

Neither are so bad for old-style linear PSUs, but cargo-cult science leads people to use the techniques on SMPSs.

[bdunham7]

Dim bulbs and SMPSs are not always a good combination.

SMPSs are constant power output devices, supplying whatever current the load requires at the specified voltage. If you reduce the input voltage, the SMPS will therefore increase the input current. Components such as transistors may object to that.

I'm not a proponent or user of dim bulbs (although I've been accused of being a dim bulb) but I think the key to their use is to select the correct bulb or bulbs for the requirements of the DUT.  A dim bulb is a current foldback type of protection not a soft-start.  I don't see why it wouldn't work to prevent catastrophic damage to even an SMPS.  If you have the right bulb, the SMPS should see full voltage immediately unless it has a short or overload of some sort, in which case it will see very low voltage.  It's not going to ever see half-voltage if you do it right.

[tggzzz]

Dim bulbs and SMPSs are not always a good combination.

SMPSs are constant power output devices, supplying whatever current the load requires at the specified voltage. If you reduce the input voltage, the SMPS will therefore increase the input current. Components such as transistors may object to that.

I'm not a proponent or user of dim bulbs (although I've been accused of being a dim bulb) but I think the key to their use is to select the correct bulb or bulbs for the requirements of the DUT.  A dim bulb is a current foldback type of protection not a soft-start.  I don't see why it wouldn't work to prevent catastrophic damage to even an SMPS.  If you have the right bulb, the SMPS should see full voltage immediately unless it has a short or overload of some sort, in which case it will see very low voltage.  It's not going to ever see half-voltage if you do it right.

I'm more concerned that a dim bulb or variac could cause catastrophic failure in an SMPS.

You would have to really know what you are doing to choose the right (incandescant) bulb for the DUT. As you imply, it certainly isn't the case that one wattage fits all.

[james_s]

You really do have to be careful with a SMPS, as a general rule I do not use a variac when working on one. I will occasionally use a dim bulb but it depends on the specific power supply and the suspected fault. In one case I was working on the power supply for a Tek TDS400 series and I tried using a dim bulb and it did actually cause the chopper transistor to fail. For analog stuff they are helpful though and can easily prevent for example burning out the power transformer in an antique radio due to a short on the secondary side.