"Ian Bell" wrote in message
...
David Looser wrote:

There shouldn't be "god awful current spikes" unless the conduction angle
is excessively short, which isn't good for the transformer, the
rectifiers or the capacitors. Since both the a.c. conductors are,
presumably, co-routed, the magnetic fields will very largely cancel
anyway.

Well, if you aim initially for 10% ripple then that will be the conduction
angle to a first approximation and the current spikes will be 10 times the
load current.


You were talking about a 200mA DC load, so how big are these spikes then? 3A
peak maybe?. hardly sounds like "god awful" to me. And you can use a twisted
pair for the ac so that the magnetic fields cancel out. I presume you are
not running the power cable close to the audio wiring? Are you using a
low-impedance balanced audio input to this amp of yours? if not I suggest
that doing that will make far more difference to your hum levels than
worrying about ac in the power cable.

David.

On Fri, 25 Dec 2009 09:03:41 -0500, Arny Krueger wrote:

Legacy crowbar - fire a thyratron.

If you can guarantee 99.9999% that it will fire whenever you want it to.

BTDTGTTS

--
John Stumbles

Men are from Mars, Women are from Venus and Pop Psychologists are from Uranus

On Wed, 30 Dec 2009 23:59:58 +0000, Ian Bell wrote:

John Stumbles wrote:
On Thu, 24 Dec 2009 14:43:30 +0000, Ian Bell wrote:

I am trying to select a relay for a delayed HT switch (which will also
discharge the HT when off). Most relays I can find have contacts rated
at 250VAC which translates into a peak of about 350V. However, data is
scarce on what dc voltage these relays can switch. So far I have found
only one that gives a dc current versus voltage curve and that stops at
210V dc (and 200mA) and I really want to be able to switch up to 350V at
up to 200mA. The rest just give a dc voltage at max current value.

The reason DC switching ratings are lower than AC ratings is that, when
breaking a circuit, AC limits the duration of arcing as the current
through the contacts drops to zero twice a cycle thus helping to
extinguish the arc. With DC the contacts have to open wide enough to
extinguish the arc on their own.


Indeed, I have found a very informative application note by Tyco that
explains wvery well what happens when contacts make and when they break.

You say:

The PSU is remote and the relay has an interlock to turn off the HT if
the PSU HT output lead is disconnected. I need to either disconnect the
HT or bleed it very quickly to avoid a possible shock hazard.

I assume we have the following setup in the PSU:

relay
HT+
to -------------o COM
LOAD /
o o----------- HT+ SUPPLY
NC | NO
|
-----
| R | - discharge resistor
-----
|
GROUND -----------+---------------- GROUND

Not quite, swap the LOAD and SUPPLY and it is correct - in other words
the HT supply goes to the common and is switched either to the load or
the discharge resistor because nearly all the charge storage is in the
HT supply not the load.

Could you not arrange it as above? I mean that all the above components
are in the PSU box and the node marked 'HT+ to LOAD' is the pole on the
output connector of the PSU to which the amplifier/load is connected. Thus
as soon as the relay falls back to its rest position the output lead is
connected via the discharge resistor to ground (discharging any residual
capacitance in the lead electronics) and the internal HT smoothing
capacitors within the PSU can be allowed to discharge under their own
steam, or with a small, high-value bleed resistor.

Doesn't that do what you want?

--
John Stumbles

Procrastinate now!

"John Stumbles" wrote in
message
On Fri, 25 Dec 2009 09:03:41 -0500, Arny Krueger wrote:

Legacy crowbar - fire a thyratron.

If you can guarantee 99.9999% that it will fire whenever
you want it to.

BTDTGTTS

If its made with tubes, we can't even guarantee that the entire piece of
equipment will work 99.9999% of the time.

IOW the criteria of 99.9999% is unrealistically high.

John Stumbles wrote:
On Wed, 30 Dec 2009 23:59:58 +0000, Ian Bell wrote:

John Stumbles wrote:
On Thu, 24 Dec 2009 14:43:30 +0000, Ian Bell wrote:

I am trying to select a relay for a delayed HT switch (which will also
discharge the HT when off). Most relays I can find have contacts rated
at 250VAC which translates into a peak of about 350V. However, data is
scarce on what dc voltage these relays can switch. So far I have found
only one that gives a dc current versus voltage curve and that stops at
210V dc (and 200mA) and I really want to be able to switch up to 350V at
up to 200mA. The rest just give a dc voltage at max current value.
The reason DC switching ratings are lower than AC ratings is that, when
breaking a circuit, AC limits the duration of arcing as the current
through the contacts drops to zero twice a cycle thus helping to
extinguish the arc. With DC the contacts have to open wide enough to
extinguish the arc on their own.

Indeed, I have found a very informative application note by Tyco that
explains wvery well what happens when contacts make and when they break.

You say:

The PSU is remote and the relay has an interlock to turn off the HT if
the PSU HT output lead is disconnected. I need to either disconnect the
HT or bleed it very quickly to avoid a possible shock hazard.
I assume we have the following setup in the PSU:

relay
HT+
to -------------o COM
LOAD /
o o----------- HT+ SUPPLY
NC | NO
|
-----
| R | - discharge resistor
-----
|
GROUND -----------+---------------- GROUND
Not quite, swap the LOAD and SUPPLY and it is correct - in other words
the HT supply goes to the common and is switched either to the load or
the discharge resistor because nearly all the charge storage is in the
HT supply not the load.

Could you not arrange it as above? I mean that all the above components
are in the PSU box and the node marked 'HT+ to LOAD' is the pole on the
output connector of the PSU to which the amplifier/load is connected.

All those components will be in the power supply and the connection to
the electronics made as you state.

Thus
as soon as the relay falls back to its rest position the output lead is
connected via the discharge resistor to ground (discharging any residual
capacitance in the lead electronics) and the internal HT smoothing
capacitors within the PSU can be allowed to discharge under their own
steam, or with a small, high-value bleed resistor.

Doesn't that do what you want?


Pretty much but I was trying to avoid having a bleed resistor
permanently across the supply. The PSU has a lot of stored charge and a
bleed resistor that takes only a nominal 1% of the load current would
take about 5 minutes to discharge it. That in itself is not a problem as
the PSU is in an enclosure but when I am testing it, it is a pain to
have to wait that long.

Cheers

Ian

On Wed, 06 Jan 2010 16:54:59 +0000, Ian Bell wrote:

Pretty much but I was trying to avoid having a bleed resistor
permanently across the supply. The PSU has a lot of stored charge and a
bleed resistor that takes only a nominal 1% of the load current would
take about 5 minutes to discharge it. That in itself is not a problem as
the PSU is in an enclosure but when I am testing it, it is a pain to
have to wait that long.

You seem to be changing the design criteria. As I understood it the design
brief (if you like) was that the connections on the outside of the box had
to assume safe potentials in the time it took a British Standard Idiot to
unplug the connector from the power amp and plug his finger in there
instead. Now you seem to be saying you want the potentials within the
PSU box to drop to safe levels in some arbitrarily short time so you can
stick your finger on conductors inside the PSU box without harm. That's a
different problem. I think what I was proposing addresses the original
problem, but if you want to make it safe for yourself to prod around in
the PSU arbitrarily I suspect that's going to be a non-trivial problem to
solve.

--
John Stumbles

Time flies like an arrow
Fruit flies like a banana
Tits like coconuts

In article , John Stumbles
wrote:
On Wed, 06 Jan 2010 16:54:59 +0000, Ian Bell wrote:

Pretty much but I was trying to avoid having a bleed resistor
permanently across the supply. The PSU has a lot of stored charge and
a bleed resistor that takes only a nominal 1% of the load current
would take about 5 minutes to discharge it. That in itself is not a
problem as the PSU is in an enclosure but when I am testing it, it is
a pain to have to wait that long.

That's a different problem. I think what I was proposing
addresses the original problem, but if you want to make it safe for
yourself to prod around in the PSU arbitrarily I suspect that's going to
be a non-trivial problem to solve.

The simplest solution is to have an additional, low value, bleed resistor
soldered across the caps during the period of development. Then clip them
off when finished, check the behaviour is still OK, and stop.

But as you say, your situation as someone developing and testing a circuit
is quite different to that of the end user. So the primary requirement is
to take due care and use whatever extra safety methods seem relevant during
development.

Slainte,

Jim

--
Please use the address on the audiomisc page if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

"Ian Bell" wrote in message
...

Pretty much but I was trying to avoid having a bleed resistor permanently
across the supply. The PSU has a lot of stored charge and a bleed resistor
that takes only a nominal 1% of the load current would take about 5
minutes to discharge it. That in itself is not a problem as the PSU is in
an enclosure but when I am testing it, it is a pain to have to wait that
long.


If you switch off the supply to the HT rectifier before the heaters go off
then the load current itself will discharge the smoothing capacitors.

David.

John Stumbles wrote:
On Wed, 06 Jan 2010 16:54:59 +0000, Ian Bell wrote:

Pretty much but I was trying to avoid having a bleed resistor
permanently across the supply. The PSU has a lot of stored charge and a
bleed resistor that takes only a nominal 1% of the load current would
take about 5 minutes to discharge it. That in itself is not a problem as
the PSU is in an enclosure but when I am testing it, it is a pain to
have to wait that long.

You seem to be changing the design criteria. As I understood it the design
brief (if you like) was that the connections on the outside of the box had
to assume safe potentials in the time it took a British Standard Idiot to
unplug the connector from the power amp and plug his finger in there
instead. Now you seem to be saying you want the potentials within the
PSU box to drop to safe levels in some arbitrarily short time so you can
stick your finger on conductors inside the PSU box without harm. That's a
different problem. I think what I was proposing addresses the original
problem, but if you want to make it safe for yourself to prod around in
the PSU arbitrarily I suspect that's going to be a non-trivial problem to
solve.


I have not changed the brief. Originally it was simply to delay the
switch on of HT until the heaters had warmed up; that and nothing more.
It then occurred to me that when the relay dropped out a contact could
be used to discharge the PSU caps more rapidly than a bleed resistor. It
also occurred to me that I could include the relay coil in a loop in the
dc power cable so if it were disconnected the HT would be disconnected too.

The delay of the HT until heaters warm up is the requirement. The other
'features' came out as possibilities in using a relay to meet the
requirement.

Hope that clarifies things.

Cheers

ian

David Looser wrote:
"Ian Bell" wrote in message
...
Pretty much but I was trying to avoid having a bleed resistor permanently
across the supply. The PSU has a lot of stored charge and a bleed resistor
that takes only a nominal 1% of the load current would take about 5
minutes to discharge it. That in itself is not a problem as the PSU is in
an enclosure but when I am testing it, it is a pain to have to wait that
long.


If you switch off the supply to the HT rectifier before the heaters go off
then the load current itself will discharge the smoothing capacitors.



Normally that would be the case but this power supply has a lot more
capacitive smoothing than the average and there is still significant
stored charge once the heaters have cooled.

Cheers

Ian

David.