[EE] Voltage accumulation

[Manu Abraham]

Hi,

I did not forget to put the sq brackets this time, Bob . :-)

Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.

Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.

The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.

Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.

Discharged the capacitor with a huge spark. Things pretty much settled down.

At least thoughts burned up a few hours, where that 80Vdc was coming from.
After that, sat back and laughed at myself.

Cheers,
Manu

[Matthew Miller]

You're seeing an effect called "bounce back". This commonly occurs
with high voltage caps.
Putting a 1M resistor across the cap will help with the safety here.
Lots of info in the answers to the stackexchange question:
https://electronics.stackexchange.com/questions/282972/capacitor-gaining-voltage-over-time

Take care.

Post by Manu Abraham
Hi,
I did not forget to put the sq brackets this time, Bob . :-)
Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.
Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.
The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.
Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.
Discharged the capacitor with a huge spark. Things pretty much settled down.
At least thoughts burned up a few hours, where that 80Vdc was coming from.
After that, sat back and laughed at myself.
Cheers,
Manu
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

Matt,

As far as I know, the effect you mention cannot produce a higher voltage
than the capacitor was charged to recently.

Sean

Post by Matthew Miller
You're seeing an effect called "bounce back". This commonly occurs
with high voltage caps.
Putting a 1M resistor across the cap will help with the safety here.
https://electronics.stackexchange.com/questions/282972/capacitor-gaining-voltage-over-time
Take care.

Post by Manu Abraham
Hi,
I did not forget to put the sq brackets this time, Bob . :-)
Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.
Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.
The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.
Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.
Discharged the capacitor with a huge spark. Things pretty much settled

down.

Post by Manu Abraham
At least thoughts burned up a few hours, where that 80Vdc was coming

from.

Post by Manu Abraham
After that, sat back and laughed at myself.
Cheers,
Manu
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Manu Abraham]

Hi Matt and Sean,

I have seen issues in the past where output capacitors in a boost
DC-DC converter gaining voltage, till it breaks down. Read somewhere,
to put a small dummy load resistor, which keeps it stable at the
required voltage. Can't recollect which one though. In the
explanation, IIRC the capacitor is getting charged due to harmonics
and or spikes (I dont know, if my memory serves correctly).

That said, this is not a switching power supply, nor are there any
unwanted spikes and or harmonics. With all that, the only thought was
that, till what voltage would it climb ?
The point to be noted is that, it reached 80V within a day or two
probably. If I leave a meter connected across, I can see the voltage
climbing very slowly.

Cheers,

Manu

Post by Sean Breheny
Matt,
As far as I know, the effect you mention cannot produce a higher voltage
than the capacitor was charged to recently.
Sean

Post by Matthew Miller
You're seeing an effect called "bounce back". This commonly occurs
with high voltage caps.
Putting a 1M resistor across the cap will help with the safety here.
https://electronics.stackexchange.com/questions/282972/capacitor-gaining-voltage-over-time
Take care.

Post by Manu Abraham
Hi,
I did not forget to put the sq brackets this time, Bob . :-)
Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.
Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.
The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.
Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.
Discharged the capacitor with a huge spark. Things pretty much settled

down.

Post by Manu Abraham
At least thoughts burned up a few hours, where that 80Vdc was coming

from.

Post by Manu Abraham
After that, sat back and laughed at myself.
Cheers,
Manu
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jason White]

I would strongly suspect that the transformer (if connected to mains) is
the source of the 80Vdc.

No mains connection is entirely free of transients and noise.

There also may be some nonideal characteristics in the transformer such as
capacitive coupling between the primary and the secondary.

[Manu Abraham]

Hi Jason,

Suddenly, I thought what you said could possibly be right. But then,
we are not talking high frequencies for that capacitive coupling,
unless there is a breakdown of the insulation. If that would be the
case, a mains line tester (with a neon lamp) has always shown that
kind of a situation. But that does not seem to be the scenario.

Now that said, wanted to have a look at spikes, if any, with a scope.
Wonder, wonder .. There seems to be some potential difference between
the capacitor GND and the GND of the scope. Nothing else connected in
between.
So there, you could be right, there could be a parasitic capacitive
coupling between primary and secondary. But that said, the neon tester
test didn't really show up either.

A bit weird though.

Cheers,
Manu


On Mon, Oct 22, 2018 at 7:18 PM Jason White

Post by Jason White
I would strongly suspect that the transformer (if connected to mains) is
the source of the 80Vdc.
No mains connection is entirely free of transients and noise.
There also may be some nonideal characteristics in the transformer such as
capacitive coupling between the primary and the secondary.
--
Jason White
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[David Van Horn]

Dielectric soakage.
https://www.electronicdesign.com/analog/whats-all-soakage-stuff-anyhow


-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Matthew Miller
Sent: Monday, October 22, 2018 5:17 AM
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: Re: [EE] Voltage accumulation

You're seeing an effect called "bounce back". This commonly occurs with high voltage caps.
Putting a 1M resistor across the cap will help with the safety here.
Lots of info in the answers to the stackexchange question:
https://electronics.stackexchange.com/questions/282972/capacitor-gaining-voltage-over-time

Take care.

Post by Manu Abraham
Hi,
I did not forget to put the sq brackets this time, Bob . :-)
Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.
Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.
The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.
Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.
Discharged the capacitor with a huge spark. Things pretty much settled down.
At least thoughts burned up a few hours, where that 80Vdc was coming from.
After that, sat back and laughed at myself.
Cheers,
Manu
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist

[AB Pearce - UKRI STFC]

The answer to this is obvious.

1. The transformer has no shield between the primary and secondary windings.

2. One side of the output is grounded to earth.

This produces a situation where there will be capacitive coupling between the transformer windings which while of high impedance at mains frequency is still low enough to allow a small current to flow. This is then rectified by the diodes on the transformer secondary and charges the capacitor.

The way around this is to have an earthed shield between the primary and secondary windings. A simple form of this is a single layer winding with only one end brought out for connection to ground. A properly done shield is a layer of copper tape with the two ends insulated from each other so it doesn't form a shorted turn, and then a wire is soldered onto it to bring the grounding connection out. The shield always has to be grounded to be effective, if it is not then the original scenario described below will exist.


-----Original Message-----
From: piclist-***@mit.edu <piclist-***@mit.edu> On Behalf Of Manu Abraham
Sent: 21 October 2018 11:03
To: Microcontroller discussion list - Public. <***@mit.edu>
Subject: [EE] Voltage accumulation

Hi,

I did not forget to put the sq brackets this time, Bob . :-)

Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.

Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.

The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.

Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.

Discharged the capacitor with a huge spark. Things pretty much settled down.

At least thoughts burned up a few hours, where that 80Vdc was coming from.
After that, sat back and laughed at myself.

Cheers,
Manu

[John Gardner]

..It's obvious...

Now that you point it out, it is - Thanks, Alan...

[Manu Abraham]

I've been wondering where the GND path could come in.

But well, it could exist. The only path: The capacitor (-) GND goes to
a PWM driver board, the GND of which goes to a DSP board, the GND of
which goes through the USB to the PC.

I just wonder whether the PC GND (0V) = Earth. If that's true, yes the
capacitor (-) would be at Earth potential. Wonder what the PC SMPS
would be doing ..

Thanks for the explanation.

Cheers,
Manu


On Mon, Oct 22, 2018 at 8:38 PM AB Pearce - UKRI STFC

Post by AB Pearce - UKRI STFC
The answer to this is obvious.
1. The transformer has no shield between the primary and secondary windings.
2. One side of the output is grounded to earth.
This produces a situation where there will be capacitive coupling between the transformer windings which while of high impedance at mains frequency is still low enough to allow a small current to flow. This is then rectified by the diodes on the transformer secondary and charges the capacitor.
The way around this is to have an earthed shield between the primary and secondary windings. A simple form of this is a single layer winding with only one end brought out for connection to ground. A properly done shield is a layer of copper tape with the two ends insulated from each other so it doesn't form a shorted turn, and then a wire is soldered onto it to bring the grounding connection out. The shield always has to be grounded to be effective, if it is not then the original scenario described below will exist.
-----Original Message-----
Sent: 21 October 2018 11:03
Subject: [EE] Voltage accumulation
Hi,
I did not forget to put the sq brackets this time, Bob . :-)
Not a query, but tripped on this one, wondered for a while, later on
alone realization dawned.
Thought about sharing this one on the list.
Have a 230V/12-0-12 CT transformer, of which 0 &12 connected to a
Bridge rectifier, the output of which is smoothed out by a 680uF/200V
capacitor, providing about 16VDC (12 x 1.4142). The power supply
worked well and good enough.
The transformer was left connected to the mains for a while, in absent
minded mode of operation. A few days later, just before wanting to
test a small board measured the voltage on the power supply, saw it
was about 80V.
Sat in thoughts for about an hour what was wrong, thought maybe the
transformer windings had sprung a leak. Thought about it for a moment,
then thought probably the capacitor was getting charged due to no
load.
Discharged the capacitor with a huge spark. Things pretty much settled down.
At least thoughts burned up a few hours, where that 80Vdc was coming from.
After that, sat back and laughed at myself.
Cheers,
Manu
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[RussellMc]

On Tue, 23 Oct 2018 at 04:07, AB Pearce - UKRI STFC <

Post by AB Pearce - UKRI STFC
The answer to this is obvious.
1. The transformer has no shield between the primary and secondary windings.
2. One side of the output is grounded to earth.
This produces a situation where there will be capacitive coupling between
the transformer windings which while of high impedance at mains frequency
is still low enough to allow a small current to flow. This is then
rectified by the diodes on the transformer secondary and charges the
capacitor.
I was going to suggest that as a likely reason, but with less confidence

:-).

______________________

Another nice "trap" is to have two "Y" capacitors on the input side from
each AC line to chassis ground, and have output ground also connected to
chassis ground, BUT no external ground connected to chassis ground. The
chassis is now at (about) VAC/2 and various other couplings may then allow
various magic to work.

With typical Y cap values you can typically feel the 1/2 AC chassis voltage
as a not-nice sharp but not too too painful bite on the backs of your
fingers. Touching the chassis with a firm finger contact MAY not cauise any
shock sensation.

Connecting the equipment to another item via a signal lead that is chassis
grounded at both ends (eg PC to printer with only one of the two items
having a mains ground) can result in destruction of the port at one and/or
other end and possibly more complete destruction of either item. Long long
ago I had a new Centronics-cable-connected printer destroyed on first being
plugged into a PC's printer-port because only one of the two had an earth
wire in its cord :-(. (That's one way to learn such lessons).



Russell

[Martin McCormick]

Post by RussellMc
Another nice "trap" is to have two "Y" capacitors on the input side from
each AC line to chassis ground, and have output ground also connected to
chassis ground, BUT no external ground connected to chassis ground. The
chassis is now at (about) VAC/2 and various other couplings may then allow
various magic to work.
With typical Y cap values you can typically feel the 1/2 AC chassis voltage
as a not-nice sharp but not too too painful bite on the backs of your
fingers. Touching the chassis with a firm finger contact MAY not cauise any
shock sensation.
Connecting the equipment to another item via a signal lead that is chassis
grounded at both ends (eg PC to printer with only one of the two items
having a mains ground) can result in destruction of the port at one and/or
other end and possibly more complete destruction of either item. Long long
ago I had a new Centronics-cable-connected printer destroyed on first being
plugged into a PC's printer-port because only one of the two had an earth
wire in its cord :-(. (That's one way to learn such lessons).

My last full-time job lasted from 1990 until retirement
in 2015. I worked as a systems engineer in the Network
Operations group for Oklahoma State University.

Times were certainly different in 1990 and OSU had a new
Ericsson phone and data system plus a campus-wide serial data
network. One could rent a telephone which had what was called a
TAUT or Telephone Access Unit Telephone which had an Ericsson
voice telephone unit and a module on the rear with a RS-232
25-pin connector that was female, just like a dial-up modem. It
even answered to Hayse commands for dialing, etc sent from a
connected computer.

I got a call, one day, from our help desk in which
someone told me that on some of those TAUT's customers were
reporting that the light signalling data transmission and
reception were flickering even when there were no data being
passed, not even any connection being made.

I asked the lady who was telling me all of this whether
or not people were getting shocked or had any damage to the TAUT
or their computer as this was concerning to those of us who know
why this might be happening.

As far as I know, nothing ever burned up or nobody got
shocked that way but I think we just had a big ground/Earth loop
between the telephone and whatever Earth was where the subscriber
was.

These telephones were not ISDN but an Ericsson
proprietary 128-KB/S data format in which 64 KB was the voice
channel and another 64 KB could be a second voice or data
channel. You could send and receive data at 9600 baud or, if you
didn't need voice, the TAUT would work at 19,200 baud. Back
then, we thought that was fast.

I never found out how much ground loop voltage was
involved but had that been an analog phone line, there probably
would have been quite a hum.

Martin McCormick WB5AGZ