AC mains to scope input

[Phil Allison]

** After just a little research - it is not hard to see why connecting a
scope directly to the AC mains is such a bad idea.

The vast majority of analogue and digital scopes have switched, passive
input attenuators using small resistors, trim capacitors and wafer switches
employing several banks. All these components have max voltage ratings that
do not include supply voltages like 240 volts AC.

Possibly the most vulnerable component to high voltages is the tiny trim
capacitor connected * directly across the input * to equalise capacitance
loading for all ranges - it is crucial that a scope's input present a
constant value of C ( usually 15 to 20 pF) in parallel with 1 Mohms on all
ranges so that the trimmer on a 10:1 probe can be adjusted once and left.

Wafer switches are also not rated for high voltages and lack sufficient
insulation and clearances to survive the spike voltages that regularly
appear on the AC supply. Insulation breakdown in the trimmer or switch
would be sudden and bit spectacular when full AC mains power is available.

Many scopes have a max input rating of 400 volts ( DC+AC) peak - which
means they are capable of accepting such peaks on a wave being viewed
without damage, despite the fact the actual peaks may not be shown. It also
means that when AC coupled, the max DC voltage should not exceed 400 volts
as that is the rating of the series input capacitor.

The worst case of ACCEPTABLE input overload is with 400 volts at the input
and the attenuator set to pass the voltage straight through to the FET
preamp. Makers typically fit a series resistor of 100kohms to 470kohms
between the switch and the gate of the first FET to limit current to a few
mA - then add one or more diodes to direct this current safely into the
+/- DC supply rails. These diodes must be low leakage types so not to create
an DC offset at the gate of the FET.

This high value series resistor needs a cap in parallel to pass high
frequencies that would otherwise be lost due stray capacitance to ground at
the gate of the FET - typically a 10nF cap is used and this too must be
rated for at least 400 volts.

In short - whenever scoping the AC supply make damn sure to use a 10:1
probe and be DAMN certain the switch is the right position.



.... Phil

 

[Phil Allison]

"Phil Allison"

In short - whenever scoping the AC supply make damn sure to use a 10:1
probe and be DAMN certain the switch is the right position.

** Just to be clear, the switch is the one on the side of the probe that has
3 positions: x1, X10 & GND.



.... Phil

 

[Phil Allison]

"Bob Myers"
fungus wrote:
"Phil Allison"

wrote:


Thanks for that...once again things are more
complex than I imagined.

That was good advice in any event; you should ALWAYS be in the habit of
spending a good deal more time setting up the measurement than actually
making it. "Grab a probe and stick it on" is a recipe for getting bad
measurements and/or frying lots of nice, expensive equipment.


** When trouble shooting electronics, probes are often flying thick and fast
and mistakes are bound to be made.

The arrival of affordable, auto-ranging DMMs in the 80s was a real boon -
particularly so as the ohms ranges were immune from damage up to 240VAC on
most. One gets tired of replacing the low ohms resistor in an analogue
multimeter every few weeks.

But one thing I have learnt to be very careful with is *current*
easurement - forgetting to put the red probe back in the correct hole for
volts, ohms etc is a recipe for disaster.

The Fluke 70 series 2 that I mainly use has no current ranges and this is a
good thing !!

Also, wherever possible, I find a resistor ( or even a length of wire) and
measure the voltage drop across it on a DMM that can resolve 0.1mV DC or AC.



..... Phil

 

[Phil Allison]

"George Herold"
"Phil Allison"

** After just a little research - it is not hard to see why connecting a
scope directly to the AC mains is such a bad idea.

The vast majority of analogue and digital scopes have switched, passive
input attenuators using small resistors, trim capacitors and wafer
switches
employing several banks. All these components have max voltage ratings
that
do not include supply voltages like 240 volts AC.

Possibly the most vulnerable component to high voltages is the tiny trim
capacitor connected * directly across the input * to equalise capacitance
loading for all ranges - it is crucial that a scope's input present a
constant value of C ( usually 15 to 20 pF) in parallel with 1 Mohms on all
ranges so that the trimmer on a 10:1 probe can be adjusted once and left.

Wafer switches are also not rated for high voltages and lack sufficient
insulation and clearances to survive the spike voltages that regularly
appear on the AC supply. Insulation breakdown in the trimmer or switch
would be sudden and bit spectacular when full AC mains power is available.

Many scopes have a max input rating of 400 volts ( DC+AC) peak - which
means they are capable of accepting such peaks on a wave being viewed
without damage, despite the fact the actual peaks may not be shown. It
also
means that when AC coupled, the max DC voltage should not exceed 400 volts
as that is the rating of the series input capacitor.

The worst case of ACCEPTABLE input overload is with 400 volts at the input
and the attenuator set to pass the voltage straight through to the FET
preamp. Makers typically fit a series resistor of 100kohms to 470kohms
between the switch and the gate of the first FET to limit current to a few
mA - then add one or more diodes to direct this current safely into the
+/- DC supply rails. These diodes must be low leakage types so not to
create
an DC offset at the gate of the FET.

Cool, do they use the CB junction of a transistor?


** Not low leakage enough and low breakdown voltage.

Something like this is the go:

http://www.centralsemi.com/PDFs/products/1n3595.pdf

Max 1nA at room temp and 125V.

Some scope makers use the gate junction of a JFET - cos that has extremely
low leakage.

(none of the x10 probes I use have a switch! I hate that thing.)

** I hate swapping probes...



.... Phil

 

[Phil Allison]

"fungus"

My reasoning was that a 'scope *has* to
have a very high impedance input. You don't
want it sucking current out of delicate little
circuits. Even at 10k ohms (quite low) you'll
only get 24mA from 240V A/C mains and
that's not enough to fry much.


** 240VAC across 10kohms is nearly 6 watts !!!

Enough to fry plenty.

Get real.


...... Phil

 

[Fred Abse]

(none of the x10 probes I use have a switch! I hate that thing.)

Me neither. How many times do you need a x1 probe?

After enduring the fragility of Tek probes for years, I settled on German
made PMK ones. 500MHz with readout switching, and a slew of nice
accessories included. Better than half the price of a comparable Tek probe.

 

[[email protected]]

My reasoning was that a 'scope *has* to
have a very high impedance input. You don't
want it sucking current out of delicate little
circuits. Even at 10k ohms (quite low) you'll
only get 24mA from 240V A/C mains and
that's not enough to fry much.

Not all scopes have a high impedance. Many have a 50 ohm input impedance. You
don't want to be hooking them to the 240V mains. ;-)

 

[Phil Allison]

"George Herold"
" Phil Allison "
OK, I've heard that the CB leakage of transistors is spec'ed at ~10
nA. But if you measure a few it's much better. I got something like
10 pA for a 2n3904 at 10 volts. 'course the reverse voltage will not
be enough for a 'scope input.

** See above, the diode is nowhere near the input.

The reverse voltage is only that between the gate of the JFET and each DC
supply rail - often only 6 to 12 volts.

The peak signal level at this point rarely exceeds 100mV.


.... Phil

 

[Fred Abse]

For low-level signals. Or for using with a DVM!

Not that often, for me, then. 10, or even 50 mV per division is mostly
enough. I can always pinch a 1:1 probe off a DVM for the odd occasion I
don't mind the capacitive loading.

Switched probes, never.

 

[[email protected]]

It'd at least blow out the 50 ohm termination.
(unless you'd used the x10 probe :^)
But personally I never liked 'scopes with switchable 50 ohm inputs.
It was just another setting I had to check.
What's wrong with a Tee and 50 ohm load...
or inline terminator?

Speed. The highest speed amplifiers (this goes up with time) have 50 ohm
inputs. Active probes are the norm in some applications.

 

[Fred Abse]

My Rigol came with some very nice switched probes. All the regular
(passive) Tek probes I have here are switchable 1x/10x.

I've found that Tek probes have a habit of breaking at the tip. The PMK
ones I settled on have replaceable tips, choice of rigid or spring-loaded.
I don't like the spring-loaded ones much.

 

[[email protected]]

I've found that Tek probes have a habit of breaking at the tip. The PMK
ones I settled on have replaceable tips, choice of rigid or spring-loaded.
I don't like the spring-loaded ones much.

You're not supposed to use them as darts. I've broken a lot of probes but I
don't think I've ever broken a tip.

 

[Fred Abse]

You're not supposed to use them as darts. I've broken a lot of probes but I
don't think I've ever broken a tip.

It's the teeny little molding that holds the tip in that cracks and falls
out, complete with tip.