Cheap B+ boost

[Engineer[_2_]]

Hi, Vacuumlanders,
Not sure if I invented this but I've not seen it before... but it's
quite obvious when you see it!

I had a physically large vintage 300-0-300 VAC Hammond power
transformer, along with some other decent used "iron", that I wanted
to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
all out of the spares or scrap box, "20" is the targeted rms watts
out.) However, the B+ at about 320 VDC was too low for cathode biased
6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
rectification as I like to see the glowing tubes... why else build
these things?! I needed a cheap B+ boost rather than a new PT!

I boosted the B+ by adding a scrap 30 VAC 1/2 amp mains transformer
plus a s/s FW bridge rectifier in the 300-0-300 power transformer
centre tap to ground. There's no need to filter the 30 VDC as all it
does is add positive-going 1/2 cycles in phase to each 1/2 cycle going
to the tube rectifier plates. The small transformer has to be sized
to take the full ripple current of the main power supply. In my case
the steady DC B+ curent was around 100 mA with a peak current of just
over 300 mA - actually lower than I expected for a 50 uF capacitative
input CLC filter, presumably due to the total impedance ahead of the
5U4.

Another way is to use a 30-0-30 VAC boost transformer out of a
scrapped s/s stereo receiver. In this case you have two choices:
1. Use the full 60 VAC and a FW rectifier to get a 60 VDC boost as
above (leave CT open), or
2. Use two diodes from the 30-0-30, CT to ground, to get a 30 VDC
boost to the B+ and two more diodes to get -30 VDC (filter this
heavily for 45 VDC!) to get a negative fixed bias supply source. Of
course, this is the same as using the FW rectifier on the 60 volt
total, grounding the CT, and using the "plus" from the rectifier for B
+ boost and the "minus" from the FW rectifier for fixed bias, as
above.

Use a s/s rectifier salvaged from a dead PC power supply. I have not
yet explored the fixed bias idea for my "Scratch 20". If I end up
using the 30-0-30 transformer I'll likely just take the +60 VDC boost
and keep cathode bias (around 24 volts depending on continuous audio
power level in class AB1.)

There seems to be no downside to this scheme, except chassis space,
and it is much cheaper than a higher voltage PT.
Cheers,
Roger

[John L Stewart]

;929170']Hi, Vacuumlanders,
Not sure if I invented this but I've not seen it before... but it's
quite obvious when you see it!

I had a physically large vintage 300-0-300 VAC Hammond power
transformer, along with some other decent used "iron", that I wanted
to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
all out of the spares or scrap box, "20" is the targeted rms watts
out.) However, the B+ at about 320 VDC was too low for cathode biased
6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
rectification as I like to see the glowing tubes... why else build
these things?! I needed a cheap B+ boost rather than a new PT!

I boosted the B+ by adding a scrap 30 VAC 1/2 amp mains transformer
plus a s/s FW bridge rectifier in the 300-0-300 power transformer
centre tap to ground. There's no need to filter the 30 VDC as all it
does is add positive-going 1/2 cycles in phase to each 1/2 cycle going
to the tube rectifier plates. The small transformer has to be sized
to take the full ripple current of the main power supply. In my case
the steady DC B+ curent was around 100 mA with a peak current of just
over 300 mA - actually lower than I expected for a 50 uF capacitative
input CLC filter, presumably due to the total impedance ahead of the
5U4.

Another way is to use a 30-0-30 VAC boost transformer out of a
scrapped s/s stereo receiver. In this case you have two choices:
1. Use the full 60 VAC and a FW rectifier to get a 60 VDC boost as
above (leave CT open), or
2. Use two diodes from the 30-0-30, CT to ground, to get a 30 VDC
boost to the B+ and two more diodes to get -30 VDC (filter this
heavily for 45 VDC!) to get a negative fixed bias supply source. Of
course, this is the same as using the FW rectifier on the 60 volt
total, grounding the CT, and using the "plus" from the rectifier for B
+ boost and the "minus" from the FW rectifier for fixed bias, as
above.

Use a s/s rectifier salvaged from a dead PC power supply. I have not
yet explored the fixed bias idea for my "Scratch 20". If I end up
using the 30-0-30 transformer I'll likely just take the +60 VDC boost
and keep cathode bias (around 24 volts depending on continuous audio
power level in class AB1.)

There seems to be no downside to this scheme, except chassis space,
and it is much cheaper than a higher voltage PT.
Cheers,
Roger

Hi Roger- What you have is a 'stacked' HV PS, not new but as you have seen quite useful.

Have fun, John

[Phil Allison[_3_]]

"Engineer"
>
> I had a physically large vintage 300-0-300 VAC Hammond power
> transformer, along with some other decent used "iron", that I wanted
> to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
> all out of the spares or scrap box, "20" is the targeted rms watts
> out.) However, the B+ at about 320 VDC was too low for cathode biased
> 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
> rectification as I like to see the glowing tubes... why else build
> these things?! I needed a cheap B+ boost rather than a new PT!


** No 5AR4s in your junk box ??

Would get you another 40-50 volts DC.


..... Phil

[John L Stewart]

Hi Roger- If you put 10R in the -ve return lead from the 1st filter cap to the rectifier system you can observe the cap charging currents on you scope. If the peaks are equal or close all is OK.

Not equal indicates a 60 Hz component in the rectified output along with the expected 120 Hz & harmonics. The 60 Hz component is more difficult for the following PS filter to remove. It is the result of unequal resistances of the PS Transformer HV secondary windings. Unequal forward drop in the rectifier diodes also contribute to this problem (or might help).

Simply reversing the rectifier heater supply leads results in about a 15 db reduction in the 60 Hz component at no cost.

The attachment shows another refinement, easy to make.

Cheers, John

[Big Bad Bob]

On 04/05/11 22:08, Phil Allison so wittily quipped:
> "Engineer"
>>
>> I had a physically large vintage 300-0-300 VAC Hammond power
>> transformer, along with some other decent used "iron", that I wanted
>> to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
>> all out of the spares or scrap box, "20" is the targeted rms watts
>> out.) However, the B+ at about 320 VDC was too low for cathode biased
>> 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
>> rectification as I like to see the glowing tubes... why else build
>> these things?! I needed a cheap B+ boost rather than a new PT!
>
>
> ** No 5AR4s in your junk box ??
>
> Would get you another 40-50 volts DC.

Quite possibly, yes. But silicon rectifiers (1Kv rating should be
enough) would give you even more. Using tube rectifiers in power
supplies nowadays is an interesting 'retro' idea, if you really want to
do that, but it won't affect sound quality (and you get a slightly lower
B+). As for me, I'll use silicon rectifiers for the obvious reasons.
As a side-benefit, the 5V winding can then be rectified through a
'charge pump' voltage multiplier (6 stages probably enough) to make a
grid bias supply for the output tubes, maybe with an adjustment for the
desired class AB quiescent current, or a zener shunt regulator to keep
it consistent. Then you won't need as high of a B+ for the same power
since the cathode bias will be gone. Been there, done that, at any rate.

[Phil Allison[_3_]]

"Big Bad Boob"
Phil Allison so wittily quipped:
>> "Engineer"
>>>
>>> I had a physically large vintage 300-0-300 VAC Hammond power
>>> transformer, along with some other decent used "iron", that I wanted
>>> to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
>>> all out of the spares or scrap box, "20" is the targeted rms watts
>>> out.) However, the B+ at about 320 VDC was too low for cathode biased
>>> 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
>>> rectification as I like to see the glowing tubes... why else build
>>> these things?! I needed a cheap B+ boost rather than a new PT!
>>
>>
>> ** No 5AR4s in your junk box ??
>>
>> Would get you another 40-50 volts DC.
>
> Quite possibly, yes. But silicon rectifiers (1Kv rating should be enough)
> would give you even more.


** Try reading the last two sentences in the OP's post.



.... Phil

[Big Bad Bob]

On 04/06/11 16:31, Phil Allison so wittily quipped:
> ** Try reading the last two sentences in the OP's post.

try not being such an ASSHOLE

[Phil Allison[_3_]]

"Big Bad Boob"


** Drop dead.

[Big Bad Bob]

On 04/06/11 16:56, Phil Allison so wittily quipped:
> "Big Bad Boob"
>
>
> ** Drop dead.

you first. Or you're welcome to come meet me and try to invert that
situation, if you _DARE_

HA, ha ha ha hah

{what a troll}

[Big Bad Bob]

On 04/06/11 06:53, John L Stewart so wittily quipped:
> Hi Roger- If you put 10R in the -ve return lead from the 1st filter cap
> to the rectifier system you can observe the cap charging currents on you
> scope. If the peaks are equal or close all is OK.
>
> Not equal indicates a 60 Hz component in the rectified output along with
> the expected 120 Hz& harmonics. The 60 Hz component is more difficult
> for the following PS filter to remove. It is the result of unequal
> resistances of the PS Transformer HV secondary windings. Unequal forward
> drop in the rectifier diodes also contribute to this problem (or might
> help).
>
> Simply reversing the rectifier heater supply leads results in about a 15
> db reduction in the 60 Hz component at no cost.
>
> The attachment shows another refinement, easy to make.

assuming the 5V winding doesn't have a CT (the good ones do) but yeah,
the resistor matrix you showed there would help null the 60hz component
from the 5V supply in the event that there's no CT. Also you might have
to reverse the 5V leads AGAIN in the sample circuit for lowest hum
since it's somewhat 'one-sided' in the adjustment. For that I'd go with
all 3 being 20 ohms so you have a 'center' that's really in the center.

[Engineer[_2_]]

On Apr 6, 1:08*am, "Phil Allison" > wrote:
> "Engineer"
>
>
>
> > I had a physically large vintage 300-0-300 VAC Hammond power
> > transformer, along with some other decent used "iron", that I wanted
> > to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
> > all out of the spares or scrap box, "20" is the targeted rms watts
> > out.) *However, the B+ at about 320 VDC was too low for cathode biased
> > 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
> > rectification as I like to see the glowing tubes... why else build
> > these things?! *I needed a cheap B+ boost rather than a new PT!
>
> ** No 5AR4s in your junk box ??
>
> * *Would get you another 40-50 volts DC.
>
> .... *Phil

Phil, yes, a few... but for now I want to use a 5U4. Then I can keep
the over-valued 5AR4's for my pension plan!
Cheers,
Roger

[Phil Allison[_3_]]

"Engineer"
"Phil Allison"
> "Engineer"
>
> > I had a physically large vintage 300-0-300 VAC Hammond power
> > transformer, along with some other decent used "iron", that I wanted
> > to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
> > all out of the spares or scrap box, "20" is the targeted rms watts
> > out.) However, the B+ at about 320 VDC was too low for cathode biased
> > 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
> > rectification as I like to see the glowing tubes... why else build
> > these things?! I needed a cheap B+ boost rather than a new PT!
>
> ** No 5AR4s in your junk box ??
>
> Would get you another 40-50 volts DC.
>

Phil, yes, a few... but for now I want to use a 5U4. Then I can keep
the over-valued 5AR4's for my pension plan!


** Nothing stops you using a 5U4 and fitting diodes.

The tube does nothing but still looks cool.


..... Phil

[Engineer[_2_]]

On Apr 6, 9:08*pm, Big Bad Bob <BigBadBob-at-mrp3-
> wrote:
> On 04/06/11 06:53, John L Stewart so wittily quipped:
>
> > Hi Roger- If you put 10R in the -ve return lead from the 1st filter cap
> > to the rectifier system you can observe the cap charging currents on you
> > scope. If the peaks are equal or close all is OK.
>
> > Not equal indicates a 60 Hz component in the rectified output along with
> > the expected 120 Hz& *harmonics. The 60 Hz component is more difficult
> > for the following PS filter to remove. It is the result of unequal
> > resistances of the PS Transformer HV secondary windings. Unequal forward
> > drop in the rectifier diodes also contribute to this problem (or might
> > help).
>
> > Simply reversing the rectifier heater supply leads results in about a 15
> > db reduction in the 60 Hz component at no cost.
>
> > The attachment shows another refinement, easy to make.
>
> assuming the 5V winding doesn't have a CT (the good ones do) but yeah,
> the resistor matrix you showed there would help null the 60hz component
> from the 5V supply in the event that there's no CT. *Also you might have
> to reverse the 5V leads AGAIN in the sample circuit for lowest hum
> since it's somewhat 'one-sided' in the adjustment. *For that I'd go with
> all 3 being 20 ohms so you have a 'center' that's really in the center.

There is a 5 VAC CT for the 5U4, so this is not an issue.
Cheers,
Roger

[Big Bad Bob]

On 04/08/11 16:13, Phil Allison so wittily quipped:
> "Engineer"
> "Phil Allison"
>> "Engineer"
>>
>>> I had a physically large vintage 300-0-300 VAC Hammond power
>>> transformer, along with some other decent used "iron", that I wanted
>>> to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
>>> all out of the spares or scrap box, "20" is the targeted rms watts
>>> out.) However, the B+ at about 320 VDC was too low for cathode biased
>>> 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
>>> rectification as I like to see the glowing tubes... why else build
>>> these things?! I needed a cheap B+ boost rather than a new PT!
>>
>> ** No 5AR4s in your junk box ??
>>
>> Would get you another 40-50 volts DC.
>>
>
> Phil, yes, a few... but for now I want to use a 5U4. Then I can keep
> the over-valued 5AR4's for my pension plan!
>
>
> ** Nothing stops you using a 5U4 and fitting diodes.
>
> The tube does nothing but still looks cool.
>

a glowey 'do nothing' tube. heh.

[Big Bad Bob]

On 04/08/11 17:48, Engineer so wittily quipped:
> On Apr 6, 9:08 pm, Big Bad Bob<BigBadBob-at-mrp3-
> > wrote:
>> On 04/06/11 06:53, John L Stewart so wittily quipped:
>>
>>> Hi Roger- If you put 10R in the -ve return lead from the 1st filter cap
>>> to the rectifier system you can observe the cap charging currents on you
>>> scope. If the peaks are equal or close all is OK.
>>
>>> Not equal indicates a 60 Hz component in the rectified output along with
>>> the expected 120 Hz& harmonics. The 60 Hz component is more difficult
>>> for the following PS filter to remove. It is the result of unequal
>>> resistances of the PS Transformer HV secondary windings. Unequal forward
>>> drop in the rectifier diodes also contribute to this problem (or might
>>> help).
>>
>>> Simply reversing the rectifier heater supply leads results in about a 15
>>> db reduction in the 60 Hz component at no cost.
>>
>>> The attachment shows another refinement, easy to make.
>>
>> assuming the 5V winding doesn't have a CT (the good ones do) but yeah,
>> the resistor matrix you showed there would help null the 60hz component
>> from the 5V supply in the event that there's no CT. Also you might have
>> to reverse the 5V leads AGAIN in the sample circuit for lowest hum
>> since it's somewhat 'one-sided' in the adjustment. For that I'd go with
>> all 3 being 20 ohms so you have a 'center' that's really in the center.
>
> There is a 5 VAC CT for the 5U4, so this is not an issue.

excellent - you're all set then. So to compensate, can you use a fixed
bias supply (extra transformer maybe?) for the final stage? That might
do better than trying to boost B+.

If you want to be a purist, a tube diode should work ok on the bias
supply, especially since it's low current, and would give you an extra
set of 'glowey things'.

/me trying to think up a rectifier + regulator circuit for negative bias
voltage using 6AV6 or 6AT6 tube. that'd be fun.

[Phil Allison[_3_]]

"Big Bad Boob"

> /me trying to think up a rectifier + regulator circuit for negative bias
> voltage using 6AV6 or 6AT6 tube. that'd be fun.


** Isn't it better practice NOT to regulate the negative bias voltage fed
to OP tubes ?

Rather, just let it move up and down with the HT as the incoming AC supply
varies over time.

Keeps the plate bias current closer to constant ....




..... Phil

[Big Bad Bob]

On 04/11/11 00:49, Phil Allison so wittily quipped:
> "Big Bad Boob"
>
>> /me trying to think up a rectifier + regulator circuit for negative bias
>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>
>
> ** Isn't it better practice NOT to regulate the negative bias voltage fed
> to OP tubes ?
>
> Rather, just let it move up and down with the HT as the incoming AC supply
> varies over time.
>
> Keeps the plate bias current closer to constant ....

uh, not sure if that's the case so much. Changes in grid volts will
have a larger effect than plate volts.

FYI - a nice dual diode tube might be 6AL5 - double checking my idea,
this tube has > 300V cathode:heater voltage tolerance, and <10V drop
across the diode at 60ma - so for a bias supply (or any other low
current full wave rectification need) it would work VERY well. One site
was advertising them at ~$5 so not a high priced tube.

[unfortunately a regulator using 6AV6 would look a bit convoluted -
better off hiding a zener diode under the chassis]

[Phil Allison[_3_]]

"Big Bad Boob"
Phil Allison
>
>>> /me trying to think up a rectifier + regulator circuit for negative bias
>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>
>>
>> ** Isn't it better practice NOT to regulate the negative bias voltage
>> fed
>> to OP tubes ?
>>
>> Rather, just let it move up and down with the HT as the incoming AC
>> supply
>> varies over time.
>>
>> Keeps the plate bias current closer to constant ....
>
> uh, not sure if that's the case so much.


** You are not sure if your arse is on fire.

> Changes in grid volts will have a larger effect than plate volts.


** So you have no experience with real tube amplifiers.

Thought so.


..... Phil

[John L Stewart]

On Mon, 11 Apr 2011 01:15:24 -0700, Big Bad Bob
> wrote:

>On 04/11/11 00:49, Phil Allison so wittily quipped:
>> "Big Bad Boob"
>>
>>> /me trying to think up a rectifier + regulator circuit for negative bias
>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>
>>
>> ** Isn't it better practice NOT to regulate the negative bias voltage fed
>> to OP tubes ?
>>
>> Rather, just let it move up and down with the HT as the incoming AC supply
>> varies over time.
>>
>> Keeps the plate bias current closer to constant ....
>
>uh, not sure if that's the case so much. Changes in grid volts will
>have a larger effect than plate volts.

The OP said he's using the 6L6 so it isn't "plate volts" that matter
so much as the screen and, as for the grid being 'more sensitive', you
might have noticed that, for high gain tubes, grid bias is also a hell
of a lot smaller than the screen supply.

Best is to regulate both with floating both coming in second.

The worst thing you can do is regulate one but not the other.

>FYI - a nice dual diode tube might be 6AL5 - double checking my idea,
>this tube has > 300V cathode:heater voltage tolerance, and <10V drop
>across the diode at 60ma - so for a bias supply (or any other low
>current full wave rectification need) it would work VERY well. One site
>was advertising them at ~$5 so not a high priced tube.
>
>[unfortunately a regulator using 6AV6 would look a bit convoluted -
>better off hiding a zener diode under the chassis]

-----------------------------------------------------------------

A few years ago I put together an experimental 2X6V6 Amp wherein the G1 fixed bias was partially regulated by tapping off the negative supply to an 0D3. The tap off point was determined by applying to the equation the triode mu of the 6V6 (G2 vs G1). That way the bias tracked the HT well & no problems. The PP OP runs Class AB2.

Hope all that makes sense, John

[Big Bad Bob]

On 04/11/11 03:35, Phil Allison so wittily quipped:
> "Big Bad Boob"
> Phil Allison
>>
>>>> /me trying to think up a rectifier + regulator circuit for negative bias
>>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>>
>>>
>>> ** Isn't it better practice NOT to regulate the negative bias voltage
>>> fed
>>> to OP tubes ?
>>>
>>> Rather, just let it move up and down with the HT as the incoming AC
>>> supply
>>> varies over time.
>>>
>>> Keeps the plate bias current closer to constant ....
>>
>> uh, not sure if that's the case so much.
>
>
> ** You are not sure if your arse is on fire.
>
>> Changes in grid volts will have a larger effect than plate volts.
>
>
> ** So you have no experience with real tube amplifiers.
>
> Thought so.

you are an idiot. it is not possible to hold a reasonable conversation
with you. If you treat people in real life the way you treat people
(like me) here, you would frequently be the recipient of many
potentially disfiguring wounds to various parts of your body. Saying
you are an asshole, or a ****wit would only encourage your disruptive
'trolling' behavior. But I like pet trolls, so keep up the good work.

Asshole. ****wit.

[Big Bad Bob]

On 04/11/11 04:22, flipper so wittily quipped:
> On Mon, 11 Apr 2011 01:15:24 -0700, Big Bad Bob
> > wrote:
>
>> On 04/11/11 00:49, Phil Allison so wittily quipped:
>>> "Big Bad Boob"

That 'Phil Allison' guy really is an idiot. I suppose his testicles
feel bigger every time he insults me.

>>>> /me trying to think up a rectifier + regulator circuit for negative bias
>>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>>
>>>
>>> ** Isn't it better practice NOT to regulate the negative bias voltage fed
>>> to OP tubes ?
>>>
>>> Rather, just let it move up and down with the HT as the incoming AC supply
>>> varies over time.
>>>
>>> Keeps the plate bias current closer to constant ....
>>
>> uh, not sure if that's the case so much. Changes in grid volts will
>> have a larger effect than plate volts.
>
> The OP said he's using the 6L6 so it isn't "plate volts" that matter
> so much as the screen and, as for the grid being 'more sensitive', you
> might have noticed that, for high gain tubes, grid bias is also a hell
> of a lot smaller than the screen supply.

screen has an effect, yes. but it's not _THAT_ much, even at 'worst
case' levels. A 10% change in volts gives you about a 10%-20% change in
quiescent current, more or less, depending on where you are [and
assuming AB bias]. It's actually LESS of an effect for the change in G1
volts (after me checking it out), but what the G1 volt change does
that's a lot WORSE is shift where you are on the operating curve, which
is likely to increase distortion and perhaps give you other problems.
So as long as your design can handle a bit higher average plate
dissipation, you're ok with a 10% shift UP in B+, and a shift down would
simply force you into AB1 instead of AB2 [or similar]. But if you're
that concerned about it, you can always use (or add) self-biasing.

Here's the data sheet I was looking at. It's been a while since I
messed with this sort of thing, so it was a nice refresher. Typical
design with screen taps would put plate and g2 at equal potential 'at
rest', and I looked at the 400V numbers which would be 'worst case'.

http://html.alldatasheet.com/html-pdf/229316/GEC/6L6-GC/348/6/6L6-GC.html
(see page 6, 8)

> Best is to regulate both with floating both coming in second.

I like THIS idea best. However, Patrick has mentioned that Australian
line voltages often vary more than 10% of the normal '220' rating, so
regulating B+ might be a bit expensive without switching power supplies.
Regulating the bias, however, wouldn't be (shunt zener).

Then again you've just presented another reason to use switching power
supplies in tube amps, something I've been kicking around as big-iron
power transformers are more expensive than output transformers.

> The worst thing you can do is regulate one but not the other.

not really. The worst you can do is assume that line volts is always
what the label says it should be. 2nd worst would be to regulate the
bias but NOT allow adjustments. But that's my $0.10 worth. Your
mileage may vary. Normal circumstances would be +/- 5%. Patrick's
+15%/-10% situation is a bit out of the ordinary, still worth
considering when you design something for world distribution.

[GRe]

"Big Bad Bob" > wrote in message
m...
> On 04/11/11 04:22, flipper so wittily quipped:
>> On Mon, 11 Apr 2011 01:15:24 -0700, Big Bad Bob
>> > wrote:
>>
>>> On 04/11/11 00:49, Phil Allison so wittily quipped:
>>>> "Big Bad Boob"
>
> That 'Phil Allison' guy really is an idiot. I suppose his testicles feel
> bigger every time he insults me.
>
>>>>> /me trying to think up a rectifier + regulator circuit for negative
>>>>> bias
>>>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>>>
>>>>
>>>> ** Isn't it better practice NOT to regulate the negative bias voltage
>>>> fed
>>>> to OP tubes ?
>>>>
>>>> Rather, just let it move up and down with the HT as the incoming AC
>>>> supply
>>>> varies over time.
>>>>
>>>> Keeps the plate bias current closer to constant ....
>>>
>>> uh, not sure if that's the case so much. Changes in grid volts will
>>> have a larger effect than plate volts.
>>
>> The OP said he's using the 6L6 so it isn't "plate volts" that matter
>> so much as the screen and, as for the grid being 'more sensitive', you
>> might have noticed that, for high gain tubes, grid bias is also a hell
>> of a lot smaller than the screen supply.
>
> screen has an effect, yes. but it's not _THAT_ much, even at 'worst case'
> levels. A 10% change in volts gives you about a 10%-20% change in
> quiescent current, more or less, depending on where you are [and assuming
> AB bias]. It's actually LESS of an effect for the change in G1 volts
> (after me checking it out), but what the G1 volt change does that's a lot
> WORSE is shift where you are on the operating curve, which is likely to
> increase distortion and perhaps give you other problems. So as long as
> your design can handle a bit higher average plate dissipation, you're ok
> with a 10% shift UP in B+, and a shift down would simply force you into
> AB1 instead of AB2 [or similar]. But if you're that concerned about it,
> you can always use (or add) self-biasing.
>
> Here's the data sheet I was looking at. It's been a while since I messed
> with this sort of thing, so it was a nice refresher. Typical design with
> screen taps would put plate and g2 at equal potential 'at rest', and I
> looked at the 400V numbers which would be 'worst case'.
>
> http://html.alldatasheet.com/html-pdf/229316/GEC/6L6-GC/348/6/6L6-GC.html
> (see page 6, 8)


Here's an analysis of the subject by means of page 8
from the 6L6 data sheet, like the one you posted:

Va= c= +400V, Vg2= 300V ±50V, Vg1= c= -7.5V, than:
Ia= 224mA for Vg1=c while Vg2= 350V
Ia= 173mA for Vg1=c while Vg2= 300V
Ia= 123mA for Vg1=c while Vg2= 250V
Or, in words:
for a voltage variation of Vg2= 300V ±16.7%, the
variation in anode current Ia= 173mA ±29% while
the grid voltage Vg1 is kept constant at -7.5V.
(see the red lines in the 6L6 -Vg/Ia graph)

Va= c= +400V, Vg2= 300V ±50V, Ia= c= 173mA, than:
Vg1= -12.8V for Ia=c while Vg2= 350V
Vg1= -7.5V for Ia=c while Vg2= 300V
Vg1= -2.1V for Ia=c while Vg2= 250V
Or, in words:
for a voltage variation of Vg2= 300V ±16.7%, the
grid must be varied from -7.5V ±71% to keep the
anode current Ia constant at 173mA
(see the blue lines in the 6L6 -Vg/Ia graph)

From the above it is evident that % screen voltage
variation for the 6L6 has a much greater influence
on the anode current than % grid voltage variation.

In other words, while a mains voltage fluctuation
of ±16.7% impacts on anode current with ±29%, one
will need to vary Vg1 with ±71% to counter act.
Above does not account for Ia change due Va change,
but if accounted for, it'll only worsen things.

Sorry, but I think you'll need another refresher.
For the 6L6 graph see ABSE: 6L6 -Vg=f(Ia) graph.
If ± does not display correctly, read +/-

Gio

PS, the quality of this data sheet is a bit better:
http://frank.pocnet.net/sheets/093/6/6L6GC.pdf

[Big Bad Bob]

On 04/12/11 00:59, flipper so wittily quipped:
> On Mon, 11 Apr 2011 11:34:46 -0700, Big Bad Bob
> > wrote:
>> It's actually LESS of an effect for the change in G1
>> volts (after me checking it out), but what the G1 volt change does
>> that's a lot WORSE is shift where you are on the operating curve, which
>> is likely to increase distortion and perhaps give you other problems.
>
> Piffle. Changing screen voltage *already* changes your 'operating
> point'.

'operating point' with respect to curve SHAPE, actually. you want the
flatter parts, not the rounder ones, for your quiescent point. Just
above the knee is pretty good or AB1, a bit higher for AB2. Raising the
bias to compensate for grid 2 voltage increases wouldn't reduce the
current as much as it would shift you 'closer to the knee' and that
would create more distortion, especially at low signal levels.

Higher plate power dissipation, interestingly enough, would probably NOT
cook the amp, but may cause you to replace power output tubes more often
[like when people 'tweek' an amplifier to get a few more watts out of
it], as long as the output transformer doesn't fail. Some manufacturers
create their own versions of the standard power output tubes for the
purpose of 'tweeking' an amp to get more power like that, so in addition
to tweeking the amp, you are buying slightly better output tubes that
can handle the higher power dissipation. As for me I'll just buy a
better amp, but there are people out there that re-work amps like that
(probably 'cause it's cheaper to re-work than buy a new amp, or maybe
because some musicians are too stoned to know what they're doing).

[Big Bad Bob]

On 04/12/11 08:02, GRe so wittily quipped:
> "Big Bad > wrote in message
> m...
> Here's an analysis of the subject by means of page 8
> from the 6L6 data sheet, like the one you posted:

[snip details]

> In other words, while a mains voltage fluctuation
> of ±16.7% impacts on anode current with ±29%, one
> will need to vary Vg1 with ±71% to counter act.
> Above does not account for Ia change due Va change,
> but if accounted for, it'll only worsen things.

there's a general relationship there that also depends on what the G1
bias is. Increased G1 bias (to -40V let's say) significantly changes
that. So if you're operating class 'A' then yeah, what you said. For
AB1 or AB2, not so much.

> Sorry, but I think you'll need another refresher.

don't think so. I read the curves differently than you do. Being
buried in the details doesn't help. Getting the general operating range
of the device, and finding the ideal spot for your design, does.

As you can see there are many points at which you can operate, and the
curves help you find that, and tell you what will happen at the middle
and extremes of operating conditions (so you plot all of those, but not
necessarily the in-betweens, for your design). Typical operation at
G1=0V would involve very low plate volts since you'd be at the max
current point, and with screen taps, it wouldn't change at all (for all
practical purposes). At the min current level (let's say G1=-80V)
you're at close to 2 x B+, but very low current, and of course near
cutoff (and not likely to be much different). At the quiescent point,
which is probably 20-50ma, you'd see a slight change, may an extra 20ma
or 30ma in quiescent current [per device] for an increase of 40V in B+
(assuming Vp = Vg2). This would slightly alter the behavior of the
circuit, give you a 'tolerable' increase in average plate dissipation
(16W per tube, worst case), and may reduce tube life. A similar
increase in bias volts (increase of 10%, 40 to 44V) would put you closer
to the 'knee' in the curve, VERY possibly altering the level of
crossover distortion.

that was my point.

pre-emptive answers to potential questions...

Q. have you ever designed an amplifier that worked like this?
A. yes (I built it for someone else who gave me some of the parts and a
chassis and asked me to build an amp around it)

Q. did it operate well?
A. yes. It used 4x6L6 for about 70WRMS, with fixed bias.

Q. did you guess or use the curves from the GE tube book to pick the
operating point?
A. being the good engineer that I am, I used the book.

[Big Bad Bob]

On 04/12/11 14:30, flipper so wittily quipped:
> On Tue, 12 Apr 2011 08:43:56 -0700, Big Bad Bob
> > wrote:
>
>> On 04/12/11 00:59, flipper so wittily quipped:
>>> On Mon, 11 Apr 2011 11:34:46 -0700, Big Bad Bob
>>> > wrote:
>>>> It's actually LESS of an effect for the change in G1
>>>> volts (after me checking it out), but what the G1 volt change does
>>>> that's a lot WORSE is shift where you are on the operating curve, which
>>>> is likely to increase distortion and perhaps give you other problems.
>>>
>>> Piffle. Changing screen voltage *already* changes your 'operating
>>> point'.
>>
>> 'operating point' with respect to curve SHAPE, actually. you want the
>> flatter parts, not the rounder ones, for your quiescent point. Just
>> above the knee is pretty good or AB1, a bit higher for AB2. Raising the
>> bias to compensate for grid 2 voltage increases wouldn't reduce the
>> current as much as it would shift you 'closer to the knee' and that
>> would create more distortion, especially at low signal levels.
>
> Nonsense. You aren't biased 'at the knee' or anywhere even remotely
> near it.

uh, whatever. do it YOUR way then, have fun.

> As I said, from the distortion aspect 'more current' is 'better' in
> both AB and AB2, because you then have more 'A' in the mix, but if you
> could 'afford' the current you'd BIAS it there to begin with and
> having to allow for high line voltage just makes things worse than you
> could otherwise bias at.

well, there are a lot more things to consider than simply adding more
current. Ideally only one tube conducts at a time, with a small AB1/AB2
'crossover' to compensate for component variability. Unfortunately the
bottom part of the Vg1/Ip curve goes asymptotic instead of simply
cutting off. So you pick a bias point that sets up your quiescent
current properly and gives you proper circuit gain, while minimizing the
effects of the nonlinear regions. Negative impacts of shifting this
point appear to outweigh those of changing B+.

<snip general side commentary regarding people who are fault-finders.
It's petty and irritating. done.>

well anyway, the problem _IS_ solved by regulating all of your voltages,
right? I'm still not sure why you're so vehemently disagreeing with me,
but whatever. Then again, great minds have always encountered violent
opposition, so there ya go. Einstein said that.

[Fred Dag]

;929175']"Engineer"
>
> I had a physically large vintage 300-0-300 VAC Hammond power
> transformer, along with some other decent used "iron", that I wanted
> to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
> all out of the spares or scrap box, "20" is the targeted rms watts
> out.) However, the B+ at about 320 VDC was too low for cathode biased
> 6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
> rectification as I like to see the glowing tubes... why else build
> these things?! I needed a cheap B+ boost rather than a new PT!


** No 5AR4s in your junk box ??

Would get you another 40-50 volts DC.


..... Phil
And cook your toast a lot faster darling.

[John L Stewart]

pre-emptive answers to potential questions...

Q. have you ever designed an amplifier that worked like this?
A. yes (I built it for someone else who gave me some of the parts and a
chassis and asked me to build an amp around it)

Q. did it operate well?
A. yes. It used 4x6L6 for about 70WRMS, with fixed bias.

Q. did you guess or use the curves from the GE tube book to pick the
operating point?
A. being the good engineer that I am, I used the book.[/QUOTE]

--------------------------------------------------------

Got any performance data?

[John L Stewart]

pre-emptive answers to potential questions...

Q. have you ever designed an amplifier that worked like this?
A. yes (I built it for someone else who gave me some of the parts and a
chassis and asked me to build an amp around it)

Q. did it operate well?
A. yes. It used 4x6L6 for about 70WRMS, with fixed bias.

Q. did you guess or use the curves from the GE tube book to pick the
operating point?
A. being the good engineer that I am, I used the book.[/QUOTE]

-----------------------------------------------------------------------

What kind of engineer are you? Degree in Fahrenheit or from a valid degree granting university?

Just curious, some are in their minds only! Illegal to pretend in some jurisdictions.

Cheers to all, John L Stewart PEng (legal)

[Fred Dag]

;929440']"Big Bad Boob"
Phil Allison
>
>>> /me trying to think up a rectifier + regulator circuit for negative bias
>>> voltage using 6AV6 or 6AT6 tube. that'd be fun.
>>
>>
>> ** Isn't it better practice NOT to regulate the negative bias voltage
>> fed
>> to OP tubes ?
>>
>> Rather, just let it move up and down with the HT as the incoming AC
>> supply
>> varies over time.
>>
>> Keeps the plate bias current closer to constant ....
>
> uh, not sure if that's the case so much.


** You are not sure if your arse is on fire.

> Changes in grid volts will have a larger effect than plate volts.


** So you have no experience with real tube amplifiers.

Thought so.


..... Phil
So, you have extensive experience in toaster repair? Thought so !

[Numbat01]

;929170']Hi, Vacuumlanders,
Not sure if I invented this but I've not seen it before... but it's
quite obvious when you see it!

I had a physically large vintage 300-0-300 VAC Hammond power
transformer, along with some other decent used "iron", that I wanted
to use in my home-brew "Scratch 20" mono amplifier ("Scratch" means
all out of the spares or scrap box, "20" is the targeted rms watts
out.) However, the B+ at about 320 VDC was too low for cathode biased
6L6 output tubes with a 5U4 rectifier. I did not want to use s/s B+
rectification as I like to see the glowing tubes... why else build
these things?! I needed a cheap B+ boost rather than a new PT!

I boosted the B+ by adding a scrap 30 VAC 1/2 amp mains transformer
plus a s/s FW bridge rectifier in the 300-0-300 power transformer
centre tap to ground. There's .....actually lower than I expected for a 50 uF capacitative input CLC filter, presumably due to the total impedance ahead of the 5U4.

Another way is to use a 30-0-30 VAC boost transformer out of a
scrapped s/s stereo receiver. In this case you have two choices:
1. Use the full 60 VAC and a FW rectifier to get a 60 VDC boost as
above (leave CT open), or
2. Use two diodes from the 30-0-30, CT to ground, to get a 30 VDC
boost to the B+ and two more diodes to get -30 VDC (filter this
heavily for 45 VDC!).....

Use a s/s rectifier salvaged from a dead PC power supply. I have not
yet explored the fixed bias idea for my "Scratch 20"....

There seems to be no downside to this scheme, except chassis space,
and it is much cheaper than a higher voltage PT.
Cheers,
Roger

Um, hate to say this, I know its almost Heresy, but have you considered putting a 1N5408 diode, one across each plate to cathode of the 5U4. This effectively places a semiconductor diode across each section of the thermionic rectifier. Hence you get reduced forward conduction voltage drop but still retain the warm rosy glow of the valve. Indeed it can be useless low emission unit but in this case it will make no difference as its the 1N5408 thats doing all the heavy lifting.

[Big Bad Bob]

On 04/15/11 17:02, John L Stewart so wittily quipped:
> What kind of engineer are you?

a good one.