Benchtop Power Supply Question

[eccerr0r]

still think those 10mm "midget" super fast HRC fuses are way too expensive and they're rated only 600V.

[stj]

i'm using 32x6.3mm 600v FFF rated

[momaka]

Originally Posted by stj

it doesnt,
an iso transformer can kill you as easy as mains
they are to isolate the device from ground for your scope.

Yup.

It actually kind of goes both ways, though.

For example, if you touch something energized in a device connected to an isolation transformer and there is nothing providing a ground reference to this device, then you won't get electrocuted if you "ground" that energized part of the device with your body. So in that case, the isolation transformer can provide a little extra safety (after all, there is no "Live" and "Neutral" at the output of an isolation transformer - you just have two AC outputs not referenced to ground.) However, if you touch two energized parts in the device that have enough potential difference between them (and ability to provide sustained high enough current), then you definitely can get electrocuted.

Now imagine your bench is equipped with GFCI.
Without an isolation transformer, if you touch any "Live" energized part in a device, then the GFCI can detect unequal current between L and N (i.e. some current "sinking" into ground) and thus should trip. But if you connect an isolation transformer behind the GFCI and you provide a ground reference to your device... then any current "sinking" into ground won't be detected by the GFCI, because as far as the GFCI is concerned, there is equal current going between Live and Neutral into the isolation transformer. So in a case like this where you do have GFCI and an isolation transformer, the GFCI will become ineffective, because it will not be able to "see" any "stray" current going to ground through your body or anything else.

Thus, I think an isolation transformer can go both ways in terms of safety, depending on how it's used. But like you say, the main point of an isolation transformer is not to protect the user, but rather to provide a line output that is not referenced to ground so you can connect and o-scope or other grounded test bench gear to the primary of a power supply. Once the o-scope or other test equipment provides a ground reference point on the circuit, though, then any other parts in the circuit can become dangerous to touch if they have high enough potential and current capability above the ground-referenced part of the circuit.

[eccerr0r]

Exactly hence the comment I'd never know if I was touching an energized node whether it was for probing (touching the node and touching the grounded oscilloscope) or not. However if just that GFCI was there, it'd trip and the circuit would go dead.

Thus indeed they are for different purposes and the GFCI really doesn't help you do debug, only certain classes of mishaps... and all the cases the GFCI can protect, the isolation transformer also covers without cutting off power...

Putting a GFCI after an isolation transformer is kind of pointless, isn't it? It would never detect a case that it should shut off. The GFCI needs to be between the iso transformer and mains power -- and even then it shouldn't detect anything nor would it be deadly as no current would flow anyway. At least you can plug in other equipment to the GFCI and have those things protect you, instead of loading down the iso transformer for no good reason.

[stj]

interesting theory.
i'm not sure if it would trip after a transformer.
it really depends on the return-path.
if the return is through an earthed tool or an antistatic mat/strap thats some assholes swear by then it could trip.

btw, in europe they are usually 30mA trip,
the 7mA ones are too damned sensitive.

[momaka]

Originally Posted by eccerr0r

Putting a GFCI after an isolation transformer is kind of pointless, isn't it? It would never detect a case that it should shut off.

I was just thinking about that and forgot to mention it.

Actually, it may not be pointless to have it after an isolation transformer.

After all, most GFCI work by comparing current draws between Live and Neutral... or just two AC lines, in case neither is referenced to ground, which would be the case with an isolation transformer. So technically speaking, a GFCI should actually work after an isolation transformer.

If you don't ground-reference any part of the device, then you can touch any one energized part without getting a shock (but not TWO energized parts with high enough potential and current capability... though that extends to any electronic circuit with or without GFCI or ISO transformer.) And if you do ground-reference any part of the device (again, o-scope or other test equipment), the GFCI would not trip until you conduct current to ground through your body. In that case, an imbalance between the two AC lines will be seen by the GFCI, and so it should trip.

So in short, having a GFCI *after* an isolation transformer may not be a bad idea after all. (Or am I missing something here?)

BTW, back when I was in university and rented a small duplex in a 50's house as my "dorm", none of the outlets had ground. They were all 2-prong. Looking at electrical codes, the only way I could have a legal 3-prong ("grounded") outlet was by using a GFCI outlet with ground left floating. So that's what I did during my stay for the one outlet where I had to plug my computer equipment.

[redwire]

The DMM fuses one thing I like is they are sand-filled and when you short mains there can be quite an arc due to cable and load inductance. This will basically chrome plate the inside of a vanilla 5x20mm fuse if the arc happens in the fuse.
But these fuses are highway robbery, I've never seen such a cash cow. I want to buy a roll of fuse wire and stick it to the man.
Take a generic 5x20mm 213 fast 5A fuse, at 200% 10A 5 seconds and 50A around 10msec.
When I look at circuit breakers, they are designed to be a little faster than the wiring melting and burning up, that's all.

With an isolation transformer, and your scope grounded to SMPS DC bus (-) so you can probe the primary... a ground fault or leakage will NOT trip the (upstream) GFIC because it's current flowing to GND, Line and Neutral remain symmetrical as far as current I think.
Placing the GFIC on an isolation transformer secondary should work?

[jiroy]

I think it's time to put some reminders :

1 mA faint tingle
5 mA slight, not painful but disturbing shock
6-25 mA painful shock, loss of muscular control
9-30 mA freezing current, muscular contraction, but can also be pushed away from the circuit
50 - 150 mA extreme pain, severe muscular contraction, respiratory arrest, death is possible
1.000 - 4.300 mA heart fibrillation, muscular contraction and nerve damage, death is likely
10.000 mA severe burns, cardiac arrest, death is probable
Main causes of electrical injury or electrocution

[eccerr0r]

If you put a GFCI after an isolation transformer, earth ground no longer is part of the "hot"/"neutral" of the secondary (after an iso transformer, both poles are "hot" sort of ... or neither because neither pole has any relation to GND due to isolation) so no current will flow from the iso secondary to GND. And since no current flows, it will never trip.

If you have a iso transformer plugged into a GFCI, it likewise will not trip either, same problem.

Note that neither case is a loss of protection, just that it wouldn't trip, or at least I don't see how it could trip as no current would flow to ground as there's no connection and thus you wouldn't get electrocuted either - and the GFCI won't protect you anyway if you put two fingers into both holes of a wall outlet.

[stj]

there are other sources of ground though.
like your soldering station.

[Per Hansson]

Originally Posted by stj

if the return is through an earthed tool or an antistatic mat/strap thats some assholes swear by then it could trip.

Assholes?
You mean like 100% of all production facilities and commercial repair facilities in the world?
BTW you are wrong: a wrist strap and mat have a 1MEG resistor in series with ground.
Otherwise it would provide absolutely no ESD protection, quite the opposite!

[stj]

all good till there is a ground failure at the box and inductive coupling puts several hundred volts at 10's of mA on your earth wire.

been there - felt it, then later tracked it down and fixed it.
several times at different locations.

earth is there for emergencies - connecting yourself to it is unwise.
use a negative ion generator if you want a static-free area.

[redwire]

For ESD, you don't want a negative ion generator because they spew out -ve HV charges which basically charges up everything.
For workbench use there are AC ionizers to neutralize charge but I have never seen one used, they are silly expensive.
I have to have a grounded (dissipative) ESD mat because the winter humidity gets so low I can easily make 10kV walking around, or in getting out of a chair. The LED thingy I made shows stray charges on myself all the time.

My building uses metal piping for earth-ground and it would be very hard to open-circuit that and have power PE up at some potential, but even if it did happen - so what the entire bench mat, soldering iron, scope etc are all at a common potential so I can't see getting zapped.
Many times I took a long run of wire and measured ohms from plumbing pipes to power outlet PE ground to make sure things are OK. DannyX has that groundless building I think.

[stj]

the usual failure point is where the box is then strapped - literally, to the incoming pipework from the street.
in one case i found that the damned gas company had replaced the iron pipe with a poly one and just left the fucking earth strap hanging without telling anybody!

[momaka]

Going back to the discussion of fuses...
I was wondering if using more than 1 fuses in series can achieve a higher voltage protection.

Originally Posted by eccerr0r

If you put a GFCI after an isolation transformer, earth ground no longer is part of the "hot"/"neutral" of the secondary (after an iso transformer, both poles are "hot" sort of ... or neither because neither pole has any relation to GND due to isolation) so no current will flow from the iso secondary to GND. And since no current flows, it will never trip.

If you have a iso transformer plugged into a GFCI, it likewise will not trip either, same problem.

Note that neither case is a loss of protection, just that it wouldn't trip, or at least I don't see how it could trip as no current would flow to ground as there's no connection and thus you wouldn't get electrocuted either - and the GFCI won't protect you anyway if you put two fingers into both holes of a wall outlet.

Hmmm...
What you say is true up there, and I also thought of that in my last post.
Nonetheless, I was thinking about this in bed yesterday morning, and I realize I may indeed have had a brain fart with thinking that a GFCI plugged in after an isolation transformer would do anything. Indeed you're right - it won't. And again, that has to do with the fact that a GFCI measures difference in current flows between "Live" and "Neutral" lines (which like you said, neither of which can be said to be Live nor Neutral, since neither is referenced to ground.) After all, you can't have current flowing out of one line from the ISO traffo and not have this current come back in through the other - that would violate KCL!

The only way a GFCI will work after being connected to an ISO transformer is if one of the output lines of the ISO transformer is connected to ground (or referenced with a low-enough resistance) BEFORE/upstream of the GFCI. But then that completely defeats the purpose of having the ISO transformer in the first place.

So yeah... GFCI after an ISO transformer = no gain in safety.

Originally Posted by redwire

DannyX has that groundless building I think.

Many buildings in [former] Soviet Block countries were built like this - particularly the "high-rise" concrete condos. Neutral was tied to ground only at one point, (typically in the basement where the power comes in) for the entire building. From that point forward, I think those buildings only have Neutral and 3 AC phases going to the apartments (each apartment only gets 1 phase though, IIRC.) It was also common practice in such buildings for electricians to tie the ground pin on outlets to Neutral - something that would be a hard violation both here and there now in modern buildings. It's not well-checked, though, so some electricians still wire it this way even in new buildings.

[stj]

Originally Posted by momaka

Going back to the discussion of fuses...
I was wondering if using more than 1 fuses in series can achieve a higher voltage protection.

i would say no.

once one fuse blows if it arc's across then your hoping the second fuse blows.
it probably wont because of reduced current, but even if it does it may still arc too.

sillica filled fuses are best - and not expensive.
HRC fuses usually arent the fastest though.

[eccerr0r]

The explanation I heard is assume this: 240V supply, two 120V fuses shorted across it. One of the fuses blows. How man volts does the burned fuse need to withstand?

That said, that Samsung TV I had used a 5mmx20mm 250V fuse to protect the Vboost line of the APFC boost circuitry...

[stj]

20mm fuses are common on 240v,
but remember that ceramic ones often contain sillica or boron to stop it just plating the tube with metal oxide.

now glass fuses are a different story, i wont use glass for mains voltage.

[jiroy]

Originally Posted by sam_sam_sam

Interesting

Enjoy these gems ...

[jiroy]

I finally finished the overvoltage protector and its main features :


-Voltage comparison : LM339 is used in applications where a comparison between two voltage signals is required. In addition with four of those comparators on board the device can compare four pairs of voltage signals at a time which comes in handy in some applications.

-Timer : IC HA17555 timer is a highly stable integrated circuit. It can function as an accurate time delay generator and free running multivibrator as well. When it is used as an oscillator, its frequency and duty cycle can be accurately adjusted with the help of external components like resistors and a capacitor.What is the minimum voltage for 555?
+4.5 volts
Supply- voltage operating range for the 555 is +4.5 volts (minimum) to +16 volts (maximum), and it is specified for operation between +5 volts and + 15 volts. The device will operate essentially the same over this range of voltages without change in timing period.
I only wish if they make a 3v. version of 555...

- 2 Zeners in parallel to feed 4 indicator leds .

- Drop down voltage through mylar 1mf 400 volts and 4 diodes 1n4007 rectifying and 10 kohms resistor ,etc ..The main DC voltage is 12v .

- 3 potentiometers , 1 for time delay and two for high and low AC.