[Prune]

I have DC heaters on four EL34 tubes and I want to regulate them. Now, the
transformer has two windings, so I could have put them in parallel then
rectify, with all heaters in parallel. However, the diode drops don't give
me enough margin. Instead, I put the windings in series since when I
rectify I only get half the overall diode drop. Then I have to put two
heaters in series, and parallel with the other two in series, so a 2x2
arrangement.

Now my problem is how to do regulation on this.

I posted my question on a forum and someone replied with the following:
""E" series valve heaters should always be run in parallel; similarly "U"
and "P" series should always be run in series."

Can some one explain this? I've found no reference, but the issue about
potentially mismatched heaters bothers me. I don't see in a case of
mismatch why one way should be better than the other, and what would make a
difference between the series that was claimed.

Assuming he was correct, however, one way I found to split is to regulate
two parallel-pairs separately, as in this simulation:
http://img167.imageshack.us/img167/9999/hrrw2.png

Beyond this, the above is a voltage regulator. But I've read that current
regulators should be used with heaters. I'm not sure if this is just to
prevent inrush current on cold heaters, or has other benefits. I'm also
not sure how it affects that series/parallel thing that I was told about
"E" series tubes.

So guys please clear this up. What I don't want is a discussion about why
I chose to do DC heaters; this is not a power amp and I'm aiming for ultra-
low noise.

[Eeyore]

Prune wrote:

I have DC heaters on four EL34 tubes and I want to regulate them. Now, the
transformer has two windings, so I could have put them in parallel then
rectify, with all heaters in parallel. However, the diode drops don't give
me enough margin. Instead, I put the windings in series since when I
rectify I only get half the overall diode drop. Then I have to put two
heaters in series, and parallel with the other two in series, so a 2x2
arrangement.

Now my problem is how to do regulation on this.

What are the voltages (and currents) of the relevant windings on your
transformer and I'll talk you through it.

Graham

[Eeyore]

Prune wrote:

the issue about potentially mismatched heaters bothers me.

What do you mean by "mismatched heaters" ?

Graham

[Eeyore]

Prune wrote:

But I've read that current regulators should be used with heaters. I'm not
sure if this is just to prevent inrush current on cold heaters, or has other
benefits.

I'm not sure how much the heater resistance changes fron cold to hot but I doubt
it's phenomenal. In any event you'd only want a 'current limited' supply.
Heaters should be run in normal use at the specified *voltage*.

Graham

[Prune]

Eeyore wrote in
:



Prune wrote:

I have DC heaters on four EL34 tubes and I want to regulate them.
Now, the transformer has two windings, so I could have put them in
parallel then rectify, with all heaters in parallel. However, the
diode drops don't give me enough margin. Instead, I put the windings
in series since when I rectify I only get half the overall diode
drop. Then I have to put two heaters in series, and parallel with
the other two in series, so a 2x2 arrangement.

Now my problem is how to do regulation on this.

What are the voltages (and currents) of the relevant windings on your
transformer and I'll talk you through it.

The two windings are the same. I get about 15 V DC after low-dropout
bridge rectifier into CRC. About half volt drop under half load (two
heaters in series). What would leave at least 13 V for all four, assuming
mains doesn't sag too much. EL34 are 6.3 V heaters, but I'd assume they'll
be fine at a bit lower, so the margin I've left of about a volt should be
fine with low Rds_on MOSFETs as the regulator pass device.
The voltages are by memory so I'll doublecheck them tomorrow. By the way,
the other windings of the transformer were unloaded. I'm not sure about
this, but can I assume that loading the other windings won't cause voltage
to sag on the heater ones?

[Prune]

Eeyore wrote in
:



Prune wrote:

the issue about potentially mismatched heaters bothers me.

What do you mean by "mismatched heaters" ?


I think there was as much as 8% difference in cold heater resistance among
the four tubes, even though I bought them as a factory matched quad
(they're new Svetlanas). Thus, if they were run in parallel, one would
have more current through it. If they were run in series, one would drop
more voltage.

[Prune]

Eeyore wrote in
:

Prune wrote:

But I've read that current regulators should be used with heaters.
I'm not sure if this is just to prevent inrush current on cold
heaters, or has other benefits.

I'm not sure how much the heater resistance changes fron cold to hot
but I doubt it's phenomenal. In any event you'd only want a 'current
limited' supply. Heaters should be run in normal use at the specified
*voltage*.

If I parallel a capacitor by the regulator's voltage reference, it would
cause a slow rampup of voltage, so large inrush current would be avoided.

[Eeyore]

Prune wrote:

Eeyore wrote
Prune wrote:

the issue about potentially mismatched heaters bothers me.

What do you mean by "mismatched heaters" ?

I think there was as much as 8% difference in cold heater resistance among
the four tubes

That sounds within typical manufacturing tolerances. No cause for alarm there.


even though I bought them as a factory matched quad (they're new Svetlanas).
Thus, if they were run in parallel, one would have more current through it.

???


If they were run in series, one would drop more voltage.

You don't run them in series. That's asking for trouble.

Graham

[Eeyore]

Prune wrote:

Eeyore wrote
Prune wrote:

But I've read that current regulators should be used with heaters.
I'm not sure if this is just to prevent inrush current on cold
heaters, or has other benefits.

I'm not sure how much the heater resistance changes fron cold to hot
but I doubt it's phenomenal. In any event you'd only want a 'current
limited' supply. Heaters should be run in normal use at the specified
*voltage*.

If I parallel a capacitor by the regulator's voltage reference, it would
cause a slow rampup of voltage, so large inrush current would be avoided.

Sure, that's a good idea anyway. Beware of 'power-off' problems caused by
capacitance there btw.

Graham

[Patrick Turner]

Prune wrote:

I have DC heaters on four EL34 tubes and I want to regulate them. Now, the
transformer has two windings, so I could have put them in parallel then
rectify, with all heaters in parallel. However, the diode drops don't give
me enough margin. Instead, I put the windings in series since when I
rectify I only get half the overall diode drop. Then I have to put two
heaters in series, and parallel with the other two in series, so a 2x2
arrangement.

Now my problem is how to do regulation on this.

Just why you want to use dc on EL34 output tube heaters is a mystery.

I have NEVER found a hum problem due to leaky heater to cathode
conditions.

And I would say I am a fairly well experienced tech.

However, since you insist in asking for comment,
let us suppose you have two 6.3Vrms windings which give 12.6Vrms when in
series.

The peak voltage is 17.7V.
When rectified via a bridge, there is a drop across two diodes = 1.4V
approx,
so the no load voltage dc at a cap would be 16.36Vdc.
But under load, winding losses at say 3.2 amps dc output might result in
another 1.5V,
so you get 14.8Vdc and this is only 2.2V above the 12.6Vdc that you
want,
and to regulate, you need the unregulated raw supply at least 3V above
the 12.6V,
preferably 5V. With only 2.2V, forget trying to regulate.
It would be better to simply have a CLC filter, with C1 = 10,000 uF,
so ripple voltage at 3.2Adc = 0.7Vrms, and can be 20mH,
and the next C would be 10,000 uF.
The ripple voltage after the LC filter = 18mVrms, and mainly free of
rectifier
switching harmonics.
The DC resistance of the choke needs to be less than 0.5 ohms.
So a careful design using a gapped core or E&I lams should be used.

You should then place some series R of about 0.47 ohms in series from
the transformer winding to C1
to reduce the peak charge currents.

Hopefully you will find you get your 12.6Vdc at 3.2Adc, suitable for a
quad of EL34.



I posted my question on a forum and someone replied with the following:
""E" series valve heaters should always be run in parallel; similarly "U"
and "P" series should always be run in series."

Can some one explain this?

I can't explain the E, U or P distinctions.

But in some amps I have used a pair of 12.6Vrms windings to give
a 25.2Vrms winding with a CT to supply the heating power to a dozen
6550 used in the amps at
http://www.turneraudio.com.au/300monobloc.html

By having two phases of 12.6Vrms, the heater current is a heck
of a lot smaller than if I had just 6.3V for all of the tubes.

I've found no reference, but the issue about
potentially mismatched heaters bothers me. I don't see in a case of
mismatch why one way should be better than the other, and what would make a
difference between the series that was claimed.

Heaters can be seriesed or paralleled and mismatches between brands is
negligible
I have found, and in some amps I have seen 4 different brands of EL34
and
the sound and measured performance was all fine.
In some amps I have seen 25Vrms used for 4 tubes in series.
This means that if a heater goes open in one, the rest are all turned
off
and the amp doesn't suffer further damage.
But if one heater shorts, the voltage to the rest is increased, and one
will fail open
and then cause silence

Assuming he was correct, however, one way I found to split is to regulate
two parallel-pairs separately, as in this simulation:
http://img167.imageshack.us/img167/9999/hrrw2.png

Simulation? simulation of just what?

Why would anyone go to so much trouble with dc heater currents in output
tubes?


Beyond this, the above is a voltage regulator. But I've read that current
regulators should be used with heaters.

Current regs are not any better than voltage regs.

Tube heaters are very easily designed to give the wanted current when
hot
when a set voltage is applied; they never were designed around current.
Engineers agreed a voltage controlled world was the way to go and it all
starts with
mains AC volatges, which when transformed can provide all manner of
wanted FIXED voltages,
and it is simply not easy to make a world where voltage can be any darn
thing and current
remain fixed.

However, one can indeed avoid inrush current due to cold heaters
by having the dc supply a fixed current, but you need to have a
fixed voltage well above the working voltage of the heaters when the
correct current
has brought the heaters up to the wanted 900C.



I'm not sure if this is just to
prevent inrush current on cold heaters, or has other benefits. I'm also
not sure how it affects that series/parallel thing that I was told about
"E" series tubes.

So guys please clear this up. What I don't want is a discussion about why
I chose to do DC heaters; this is not a power amp and I'm aiming for ultra-
low noise.

Whatever it is you are building may not need regulated DC heaters.
Well filtered DC will do. CLCLC will give remarkably low ripple
voltages, and the regulation
due to the naturalo transformer regulation should be enough.

You are trying to force something not meant for creating DC to make DC,
and there simply isn't enough headroom for regulation, and
even if there were, the heat lost in a drop of from say +16V to +12.6V
= 3.4V x 3.2A = 10.9 watts.

Patrick Turner.

[Ian Iveson]

Prune wrote

I have DC heaters on four EL34 tubes and I want to regulate them.
Now, the
transformer has two windings, so I could have put them in parallel
then
rectify, with all heaters in parallel. However, the diode drops
don't give
me enough margin. Instead, I put the windings in series since when
I
rectify I only get half the overall diode drop. Then I have to put
two
heaters in series, and parallel with the other two in series, so a
2x2
arrangement.

Now my problem is how to do regulation on this.

I posted my question on a forum and someone replied with the
following:
""E" series valve heaters should always be run in parallel;
similarly "U"
and "P" series should always be run in series."

Can some one explain this? I've found no reference, but the issue
about
potentially mismatched heaters bothers me. I don't see in a case of
mismatch why one way should be better than the other, and what would
make a
difference between the series that was claimed.

Assuming he was correct, however, one way I found to split is to
regulate
two parallel-pairs separately, as in this simulation:
http://img167.imageshack.us/img167/9999/hrrw2.png

Beyond this, the above is a voltage regulator. But I've read that
current
regulators should be used with heaters. I'm not sure if this is
just to
prevent inrush current on cold heaters, or has other benefits. I'm
also
not sure how it affects that series/parallel thing that I was told
about
"E" series tubes.

So guys please clear this up. What I don't want is a discussion
about why
I chose to do DC heaters; this is not a power amp and I'm aiming for
ultra-
low noise.

I think it is bad practice to combine series and parallel in one
chain, and would prefer either regulated voltage for parallel heaters,
or regulated current for series connection.

I wouldn't know how the power sharing problem might work out in
practice coz I haven't tried it. However, Morgan Jones warns against
your proposed arrangement, and he has done quite a lot of measuring.

Consider what happens if one heater fails to an open circuit, in the
case of current regulation, or to a short, in the case of voltage
regulation.

I had assumed that the E P and U designations only indicate intended
usage, for the sake of convenience.

cheers, Ian

[Eeyore]

Ian Iveson wrote:

I wouldn't know how the power sharing problem might work out in
practice coz I haven't tried it.

And also because you're a clueless idiot who consistently offers bad advice.

Graham

[west[_3_]]

"Eeyore" wrote in message
...


Ian Iveson wrote:

I wouldn't know how the power sharing problem might work out in
practice coz I haven't tried it.

And also because you're a clueless idiot who consistently offers bad
advice.

Graham

ditto, way to go Graham.

west

[Prune]

Eeyore wrote in
:

If I parallel a capacitor by the regulator's voltage reference, it
would cause a slow rampup of voltage, so large inrush current would
be avoided.

Sure, that's a good idea anyway. Beware of 'power-off' problems
caused by capacitance there btw.

Can you elaborate on the power-off problems? I'm not sure what form they
would take.

[Prune]

Patrick Turner wrote in
:

It would be better to simply have a CLC filter, with C1 = 10,000 uF,
so ripple voltage at 3.2Adc = 0.7Vrms, and can be 20mH,
and the next C would be 10,000 uF.
The ripple voltage after the LC filter = 18mVrms, and mainly free of
rectifier
switching harmonics.
The DC resistance of the choke needs to be less than 0.5 ohms.
So a careful design using a gapped core or E&I lams should be used.

Rectifier harmonics is not an issue since I put a snubber on the secondary,
and the switching is clean on the scope. I understand of course that
passive filter is enough sine the load is constant, so regulator isn't
needed. But if I use CLC, my problem is figuring out the inductor. I have
various EI cores lying around, but I'm not sure how big one I need here.

For example, I have one with 1.5 cm^2 center-bar crossection. Measures 1
ohm and 50 mH. So if I rewind with a larger gauge wire, is that enough, or
do I need larger core? I actually have two of these so I could do CLCLC or
alternatively have an L on the + and - side.

And with CLC, I guess I need a time delay relay to limit inrush current.

BTW, the tube is cathode driven and the cathode swings between about -380
and -290 V. Do I just leave the heater supply floating, or how do I
reference it to the cathode?

Thanks.