3 Sep
2008
3 Sep
'08
9:56 a.m.
William Donzelli wrote:
Can't get much cheaper than effectively free for a well regulated, high current DC
power supply using a modified PC power supply that isn't otherwise being used. Plus,
the idea of using tubes with 5V filaments instead of 6.3/12.6V reduces the competition
with typical audio tube customers and makes the use of a high-current 5VDC supply
possible. For the homebrew 1.5A adjustable DC supplies, I use LM317HV based units plugged
into a homebrew chassis where I've mounted a range of three different inexpensive,
surplus, multi-secondary transformers with output voltages ranging from 70VCT to 12.6VCT.
The chassis has binding posts for each secondary and I just plug the banana plugs for the
various inexpensive LM317HV adjustable DC supplies into whatever AC outputs most closely
match the DC voltages I need at the time in order to reduce the power dissipation
requirements of the LM317HVs.
I've got
several homemade variable DC supplies,
but they are all limited to 1.5A output
which would be
exceeded with just three 5BK7A filaments @ 600mA each.
An old filament transformer would solve all your problems
in this area. Not expensive. 
3 Sep
3 Sep
10:25 a.m.
On 3 Sep 2008 at 9:56, William Blair wrote:
Can't get much cheaper than effectively free for a
well regulated,
high current DC power supply using a modified PC power supply that
isn't otherwise being used. Plus, the idea of using tubes with 5V
filaments instead of 6.3/12.6V reduces the competition with typical
audio tube customers and makes the use of a high-current 5VDC supply
possible.
For digital circuits (I assume that's what we're still discussing),
why is a DC heater supply necessary?
If you do decide to use a regulated DC supply with overcurrent
protection, keep in mind that cold heater inrush current can be
substantially greater than the nominal heater current rating.
Cheers,
Chuck
11:56 a.m.
Can't get much cheaper than effectively free for a
well regulated, high current DC power supply using a modified PC power supply that
isn't otherwise being used. Plus, the idea of using tubes with 5V filaments instead
of 6.3/12.6V reduces the competition with typical audio tube customers and makes the use
of a high-current 5VDC supply possible.
For a few bucks, you could get a 5 Volt transformer with a decent
current rating of 10 Amps, and not have to worry about the number of
tubes you need to feed for a while.
Filament transformers come in many voltages other than 6.3.
--
Will
1:45 p.m.
William Donzelli wrote:
Ya, it would be more authentic but the filament supply would then weigh 20 lbs
instead of 2.
Can't get
much cheaper than effectively free for a well regulated, high current DC power supply
using a modified PC power supply that isn't otherwise being used. Plus, the idea of
using tubes with 5V filaments instead of 6.3/12.6V reduces the competition with typical
audio tube customers and makes the use of a high-current 5VDC supply possible.
For a few bucks, you could get a 5 Volt transformer with a decent
current rating of 10 Amps, and not have to worry about the number of
tubes you need to feed for a while.
Filament transformers come in many voltages other than 6.3.
1:22 p.m.
On 3 Sep 2008 at 12:45, Brent Hilpert wrote:
Ya, it would be more authentic but the filament supply
would then weigh 20 lbs
instead of 2.
Sure, but a nice beefy transformer will give you many decades of
service without problems. High-ampere 5V transformers shouldn't be
uncommon, as many 5v rectifier filaments (such as the 5U4) were
pretty hungry (the 5U4 is rated at 3 amps).
Cheers,
Chuck
3:58 p.m.
Chuck Guzis wrote:
On 3 Sep 2008 at 12:45, Brent Hilpert wrote:
William Donzelli wrote:
Ya, it would be more authentic but the filament
supply would then weigh 20 lbs
instead of 2.
Sure, but a nice beefy transformer will give you many decades of
service without problems. High-ampere 5V transformers shouldn't be
uncommon, as many 5v rectifier filaments (such as the 5U4) were
pretty hungry (the 5U4 is rated at 3 amps). Or use a smaller transformer. Don't make me break
out the Radio Masters.
Oh, yes, for example, we do have some variety of such at the radio museum.
Transmitters with parallelled 5V fil. rectifiers can be a source of larger ones.
Joking aside, and I don't want to speak for William, but I think at this point
his project is in the experimental stage .. whatever is at hand will do.
On the other hand, for the ABC ASM reconstruction I was working on I was
targetting using only period-appropriate components. I'm still debating whether
to cheat and use a silicon diode for the C- bias rectifier (it could perhaps
have been a selenium) or go to the bother of throwing in another rectifier
tube, which will also require the hassle of an additional isolated filament supply.
3:48 p.m.
On 3 Sep 2008 at 14:58, Brent Hilpert wrote:
On the other hand, for the ABC ASM reconstruction I
was working on I
was targetting using only period-appropriate components. I'm still
debating whether to cheat and use a silicon diode for the C- bias
rectifier (it could perhaps have been a selenium) or go to the bother
of throwing in another rectifier tube, which will also require the
hassle of an additional isolated filament supply.
In the old days, the most common bias supply in many applications was
a simple dry battery (e.g. the 4.5v "C" battery). If you want to use
an AC-powered bias supply, there's no particular reason that you need
to use a separate filament transformer--use a rectifier with an
indirectly-heated cathode (and sufficient heater-to-cathode breakdown
voltage). Since bias supplies don't generally supply lots of
current, even a small-signal diode, such as a 6AL5 should work fine
(heater-cathode max voltage is 330).
You could, of course, use selenium or copper-oxide rectifiers--Cougar
Electronics (supplies the elevator industry) still makes the selenium
jobs (you can't simply drop in a silicon diode in a lot of old
equipment without redesigning it). I don't know if any manufacturers
of copper-oxide rectifiers still exist.
Cheers,
Chuck
11:45 p.m.
Chuck Guzis wrote:
On 3 Sep 2008 at 14:58, Brent Hilpert wrote:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Well, that was the hitch. I was looking at using a 6X5 (ind. heated)
but it was insufficient due to the fact that there is a polarization to
the H-C V limits (weird, but it's there, something to do with the oxide
formation I'm guessing).
On the other hand, for the ABC ASM reconstruction
I was working on I
was targetting using only period-appropriate components. I'm still
debating whether to cheat and use a silicon diode for the C- bias
rectifier (it could perhaps have been a selenium) or go to the bother
of throwing in another rectifier tube, which will also require the
hassle of an additional isolated filament supply.
In the old days, the most common bias supply in many applications was
a simple dry battery (e.g. the 4.5v "C" battery). If you want to use
an AC-powered bias supply, there's no particular reason that you need
to use a separate filament transformer--use a rectifier with an
indirectly-heated cathode
(and sufficient heater-to-cathode breakdown voltage). Since bias supplies don't generally supply lots
of
current, even a small-signal diode, such as a 6AL5 should work fine
(heater-cathode max voltage is 330).
I don't think the 6AL5 was around in 1939/40, but in the same spirit
I was considering a 6H6. The problem there (even though the H-C V permits
it to be used as a low current C- rect) was that the C- (and B+) for
the logic need to be regulated and by the time you account for the extra
current for a shunt regulator (gas tube) and some more for a resistor
divider, it was on the margins or exceeded what the little guys
could handle.
It was one of those annoying design problems where every optimisation
you try leads to some minor hitch and you have to revert to what seems
like an overblown solution.
You could, of course, use selenium or copper-oxide
rectifiers--Cougar
Electronics (supplies the elevator industry) still makes the selenium
jobs (you can't simply drop in a silicon diode in a lot of old
equipment without redesigning it). I don't know if any manufacturers
of copper-oxide rectifiers still exist.
Didn't know anybody was making them still. I could try scavenging a
selenium rect at the radio museum, but the old ones I run across are
usually in questionable condition.
11:19 p.m.
On 3 Sep 2008 at 22:45, Brent Hilpert wrote:
Well, that was the hitch. I was looking at using a 6X5
(ind. heated)
but it was insufficient due to the fact that there is a polarization to
the H-C V limits (weird, but it's there, something to do with the oxide
formation I'm guessing).
Well, there are always the gas rectifiers (0Z4) with no need for a
heater supply, but generates a lot of RF "hash". Or the line-voltage
heater (e.g. 117Z6) variety of rectifier, allowing you to dispense
with a filament transformer altogether.
Or you could just drop a couple of strips of aluminum into a solution
of bicarbonate of soda...
Didn't know anybody was making them still. I could
try scavenging a
selenium rect at the radio museum, but the old ones I run across are
usually in questionable condition.
ISTR that selenium rectifiers degrade when they're not being used and
need to be reformed by operating under a load for a time. Sort of
like wet electrolytics. Tony or Will might remember more on the
subject.
Me, I could never stand the smell of one going south.
Cheers,
Chuck
4 Sep
4 Sep
11:44 a.m.
For a few
bucks, you could get a 5 Volt transformer with a decent
current rating of 10 Amps, and not have to worry about the number of
tubes you need to feed for a while.
Filament transformers come in many voltages other than 6.3.
Ya, it would be more authentic but the filament supply would then weigh 20 lbs
instead of 2. 
2:16 p.m.
On Thursday 04 September 2008, Tony Duell wrote:
And why is this a problem? It's hardly going to be a portable device,
is it?
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
They're about the only company I know of that takes a 1900lb rack of
machines, and then puts another 200lbs of steel plates on the bottom to
add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a bad
back thanks to my IBM gear.
Pat
--
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
For a few bucks, you could get a 5 Volt transformer
with a decent
current rating of 10 Amps, and not have to worry about the number
of tubes you need to feed for a while.
Filament transformers come in many voltages other than 6.3.
Ya, it would be more authentic but the filament supply would then
weigh 20 lbs instead of 2. 
5 Sep
5 Sep
4:19 a.m.
On Thu, 2008-09-04 at 17:16 -0400, Patrick Finnegan wrote:
On Thursday 04 September 2008, Tony Duell wrote:
And why is this a problem? It's hardly going to be a portable device,
is it?
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
They're about the only company I know of that takes a 1900lb rack of
machines, and then puts another 200lbs of steel plates on the bottom to
add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a bad
back thanks to my IBM gear.
Of course, having 100kg of ballast in the bottom does give you the
rather handy feature of making the rack not tip over and kill you when
you pull a machine at the top...
Gordon
For a few bucks, you could get a 5 Volt transformer
with a decent
current rating of 10 Amps, and not have to worry about the number
of tubes you need to feed for a while.
Filament transformers come in many voltages other than 6.3.
Ya, it would be more authentic but the filament supply would then
weigh 20 lbs instead of 2. 
5 a.m.
On Fri, Sep 05, 2008 at 12:19:33PM +0100, Gordon J. C. Pearce wrote:
On Thu, 2008-09-04 at 17:16 -0400, Patrick Finnegan
wrote:
I've seen plenty of DEC cabinets with a couple hundred pounds of
counterweight in the bottom _plus_ anti-tip feet. Fortunately, I've
never had the pleasure of one tipping over on me, so I can understand
why they do it.
The closest I've personally experienced was trying to extract an 11/750
out of a minivan solo. I managed not to crush me or crunch the CPU, but
it was tough.
I _really_ don't want an SA482 cab falling on top of me.
-ethan
--
Ethan Dicks, A-333-S Current South Pole Weather at 5-Sep-2008 at 11:50 Z
South Pole Station
PSC 468 Box 400 Temp -34.1 F (-36.7 C) Windchill -66.4 F (-54.7 C)
APO AP 96598 Wind 17.0 kts Grid 7 Barometer 672.1 mb (10928 ft)
Ethan.Dicks at usap.gov http://penguincentral.com/penguincentral.html
They're about the only company I know of that
takes a 1900lb rack of
machines, and then puts another 200lbs of steel plates on the bottom to
add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a bad
back thanks to my IBM gear.
Of course, having 100kg of ballast in the bottom does give you the
rather handy feature of making the rack not tip over and kill you when
you pull a machine at the top... 
6:54 a.m.
New subject: stability / was vac tubes / was Re: Schematics of Atanasoff-Berry
>>>> "Ethan" == Ethan Dicks
<ethan.dicks at usap.gov> writes:
> Of course, having 100kg of ballast in the bottom
does give you the
> rather handy feature of making the rack not tip over and kill you
> when you pull a machine at the top...
Ethan> I've seen plenty of DEC cabinets with a couple hundred pounds
Ethan> of counterweight in the bottom _plus_ anti-tip feet.
Ethan> Fortunately, I've never had the pleasure of one tipping over
Ethan> on me, so I can understand why they do it.
There's a legend (possibly urban) that says you could make a whole row
of 3330 (?) drives topple over by sending them then "unload drive"
command all at the same time. That would spin down the drive and
extend the slide-out cabinet (so you could open the top lid). Drives
were bolted together for stability, but unloading the whole low would
defeat that -- supposedly.
Never got a chance to try it... :-)
paul
7:34 a.m.
New subject: stability / was vac tubes / was Re: Schematics of Atanasoff-Berry
There's a legend (possibly urban) that says you
could make a whole row
of 3330 (?) drives topple over by sending them then "unload drive"
command all at the same time. That would spin down the drive and
extend the slide-out cabinet (so you could open the top lid). Drives
were bolted together for stability, but unloading the whole low would
defeat that -- supposedly.
No, you can not do that.
--
Will
7:45 a.m.
New subject: stability / was vac tubes / was Re: Schematics of Atanasoff-Berry
There's a
legend (possibly urban) that says you could make a whole row
of 3330 (?) drives topple over by sending them then "unload drive"
command all at the same time. That would spin down the drive and
extend the slide-out cabinet (so you could open the top lid). Drives
were bolted together for stability, but unloading the whole low would
defeat that -- supposedly.
No, you can not do that.
6:59 a.m.
On Fri, Sep 05, 2008 at 12:19:33PM +0100, Gordon J. C. Pearce wrote:
On Thu, 2008-09-04 at 17:16 -0400, Patrick Finnegan
wrote:
Exactly. That's why HP also added a nice load of heavy steel plates to
the bottom of our HP racks in a previous job. The rack contained some
HP N- and L-class PA-RISC machines. Damn heavy iron - we needed 3 people to
actually lift the damn things into their rack locations. You really don't
want the whole rack to tip over when you pull one of them out for
hardware maintainance/replacement.
Regards,
Alex.
--
"Opportunity is missed by most people because it is dressed in overalls and
looks like work." -- Thomas A. Edison
On Thursday 04 September 2008, Tony Duell wrote:
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
They're about the only company I know of that takes a 1900lb rack of
machines, and then puts another 200lbs of steel plates on the bottom to
add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a bad
back thanks to my IBM gear.
Of course, having 100kg of ballast in the bottom does give you the
rather handy feature of making the rack not tip over and kill you when
you pull a machine at the top... > For
a few bucks, you could get a 5 Volt transformer with a decent
> current rating of 10 Amps, and not have to worry about the number
> of tubes you need to feed for a while.
>
> Filament transformers come in many voltages other than 6.3.
Ya, it would be more authentic but the filament supply would then
weigh 20 lbs instead of 2.
And why is this a problem? It's hardly going to be a portable device,
is it?
9:55 a.m.
On Friday 05 September 2008, Alexander Schreiber wrote:
On Fri, Sep 05, 2008 at 12:19:33PM +0100, Gordon J. C.
Pearce wrote:
If you think that's heavy, you've never lifted IBM hardware. For a
(relatively modern) 8U machine, an IBM SP high node is amazingly
heavy/dense.
Pat
--
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
On Thu, 2008-09-04 at 17:16 -0400, Patrick
Finnegan wrote:
Exactly. That's why HP also added a nice load of heavy steel plates
to the bottom of our HP racks in a previous job. The rack contained
some HP N- and L-class PA-RISC machines. Damn heavy iron - we needed
3 people to actually lift the damn things into their rack locations.
You really don't want the whole rack to tip over when you pull one of
them out for hardware maintainance/replacement. On Thursday 04 September 2008, Tony Duell wrote:
Of course, having 100kg of ballast in the bottom does give you the
rather handy feature of making the rack not tip over and kill you
when you pull a machine at the top... > > For a few bucks, you could get a 5 Volt
transformer with a
> > decent current rating of 10 Amps, and not have to worry
> > about the number of tubes you need to feed for a while.
> >
> > Filament transformers come in many voltages other than 6.3.
>
> Ya, it would be more authentic but the filament supply would
> then weigh 20 lbs instead of 2.
And why is this a problem? It's hardly going to be a portable
device, is it?
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
They're about the only company I know of that takes a 1900lb rack
of machines, and then puts another 200lbs of steel plates on the
bottom to add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a
bad back thanks to my IBM gear.
9:52 a.m.
On Friday 05 September 2008, Gordon J. C. Pearce wrote:
On Thu, 2008-09-04 at 17:16 -0400, Patrick Finnegan
wrote:
Between the fixed, 75kg or so power supply in the bottom and the other
three 200kg nodes in the rack, it's really not necessary, unless you
are stupid and are removing the nodes from the rack, from the bottom up
instead of top down.. And, in that case, I don't think 100kg of steel
plates in the bottom is going to help you compared to the 200kg node
that's nearly 2m up in the air.
Pat
--
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
On Thursday 04 September 2008, Tony Duell wrote:
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
They're about the only company I know of that takes a 1900lb rack
of machines, and then puts another 200lbs of steel plates on the
bottom to add weight. (IBM SP frames)
Now that I think about it, I'm surprised that I don't yet have a
bad back thanks to my IBM gear.
Of course, having 100kg of ballast in the bottom does give you the
rather handy feature of making the rack not tip over and kill you
when you pull a machine at the top... > For
a few bucks, you could get a 5 Volt transformer with a
> decent current rating of 10 Amps, and not have to worry about
> the number of tubes you need to feed for a while.
>
> Filament transformers come in many voltages other than 6.3.
Ya, it would be more authentic but the filament supply would
then weigh 20 lbs instead of 2.
And why is this a problem? It's hardly going to be a portable
device, is it?
1:13 p.m.
Ya, it would be more authentic but the filament supply
would then
weigh 20 lbs instead of 2.
And why is this a problem? It's hardly going to be a portable device,
is it?
5:21 p.m.
On Friday 05 September 2008, Tony Duell wrote:
Tony, your reasoning sounds like what IBM uses to decide how much
something should weigh.
You've still not explained why this is a problem.
I really don't understand this modern love with making things as
small and light as possile. I'd rather have things that are heavy
enough to stay put (in the case of a rack unit, not to topple, or
even feel wobbly when units are extended ot the maintenance position)
and that are large enough for me to be able to work on them.
Bigger and heavier means less items in your collection to fit in the
space, and a smaller number before the floor falls out from underneath
of it. I'd hope you could appreciate that. :)
Another problem I seem to have is that the heavier an item is (and
harder it is to move myself), the less I seem to be able to find people
to help move it.
I in no way said that I wanted everything made of lightweight (and
easily breakable) plastic, it's just annoying when things are severely
over-enginered and thus increase the weight, or use heavier components
when lighter ones would work in a perfectly acceptable manner.
Pat
--
Purdue University Research Computing --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
Ya, it would
be more authentic but the filament supply would
then weigh 20 lbs instead of 2.
And why is this a problem? It's hardly going to be a portable
device, is it?
6 Sep
6 Sep
1:34 p.m.
> I really don't understand this modern love
with making things as
> small and light as possile. I'd rather have things that are heavy
> enough to stay put (in the case of a rack unit, not to topple, or
> even feel wobbly when units are extended ot the maintenance position)
> and that are large enough for me to be able to work on them.
[For the record, I consider the OP's comment that he wants to use a PC
PSU becuase he has one or that it's convenient, or.. to be entirely
reasonable.]
Bigger and heavier means less items in your collection to fit in the
space, and a smaller number before the floor falls out from underneath
of it. I'd hope you could appreciate that. :)
So by that arguemnt we should all collect handhelds, not desktop and
rack-mounted machines :-)
Another problem I seem to have is that the heavier an
item is (and
harder it is to move myself), the less I seem to be able to find people
to help move it.
Fortunately I don't mind having to dismantle something to move it. Many
of my larger machines came into the house in small-ish pieces. I
remember, for example, dismantling a DEC RA60 on the back of a friend's
pick-up truck (fortunately the RA60 is dismantled from the top down, so
this wasn't hard), taking the bits inside and then putting them back
together.
Gettign back to the OP's heater supply, I don't think a mains-input
heater transformer to supply 5 normal-sized receiving valves is going to
make the device impossile to lift. If you used ECC83s (12AX7s), I think
you could use a 20VA transformer with no trouhle at all, and that's
hardly large or heavy.
I in no way said that I wanted everything made of
lightweight (and
easily breakable) plastic, it's just annoying when things are severely
over-enginered and thus increase the weight, or use heavier components
when lighter ones would work in a perfectly acceptable manner.
Well, there are limits, but I must admit I tend to admire solid
overengineered devices.
-tony
5951
days inactive
5954
days old
23 comments
11 participants
participants (11)
-
als＠thangorodrim.de -
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
ethan.dicks＠usap.gov -
geneb＠deltasoft.com -
gordonjcp＠gjcp.net -
hilpert＠cs.ubc.ca -
pat＠computer-refuge.org -
Paul_Koning＠Dell.com -
wbblair3＠yahoo.com -
wdonzelli＠gmail.com