Schematic and parts list for DC heater supply?

[Structo]

So you have the cathodes of the diodes connected together right?
The striped side.
Then at that point to the + end of the first cap, right?

Then you take a wire from the + side of the cap where the diodes connect for your + side of the heater wiring.
The negative side of the heater wiring would be chassis ground from that point.

Are you measuring with your red probe where the two diodes connect to the + side of the cap?

Then the negative - probe to chassis ground?

[markr14850]

Is your secondary center tapped and the tap grounded?

[dehughes]

So you have the cathodes of the diodes connected together right?
The striped side.
Then at that point to the + end of the first cap, right?

Then you take a wire from the + side of the cap where the diodes connect for your + side of the heater wiring.
The negative side of the heater wiring would be chassis ground from that point.

Are you measuring with your red probe where the two diodes connect to the + side of the cap?

Then the negative - probe to chassis ground?

Yes on all counts. I've not wired the + side up to the tube heaters yet, as I wanted to check the DC voltage before hand, in case I needed to drop some.

I'm just stumped as to why I'm reading less than 1vDC on the + end of things....

[markr14850]

I can't tell by looking at the picture, but it sounds like you don't have a complete circuit.

Check what you have against the schematics here:
http://en.wikipedia.org/wiki/Rectifier

[paulster]

Sounds like you don't have your centre-tap connected to ground.

Got clearer photos showing the wiring of the diodes, caps and centre-tap (presumably outside the chassis since it doesn't come through the same grommet)?

[Structo]

What is the secondary voltage supposed to be?

I haven't re-read all the previous posts but I kind of thought you were going to go with a 12v transformer and regulate it down to 6.3 vdc or so.

[Sonny ReVerb]

[dehughes]

Sounds like you don't have your centre-tap connected to ground.

Got clearer photos showing the wiring of the diodes, caps and centre-tap (presumably outside the chassis since it doesn't come through the same grommet)?

I thought I wasn't supposed to ground the CT...I thought the grounding was to happen on DC side of this? Unless I've misunderstood the previous postings....my understanding was that it was a bad idea to ground the CT on the filament transformer, and that I should ground the DC side instead. Here is what you said:

On your 6.3Vac supply to your power tubes, ground it as normal (centre-tap).

On your new 6.3Vdc supply, which is completely independent, cap off the centre-tap of its transformer and ground the DC side only.

You'll end up with the power tubes on ground-referenced AC and the preamp tubes on ground-referenced DC.

That'll be the quietest option.

Lemme know what I'm not understanding...

Okay...maybe I'm getting it:

Two types of rectification with diodes...1) full wave with a CT, or 2) full wave bridge. One uses full wave bridge if you don't have a CT....otherwise, one just connects the CT and then runs diodes to pass the separate halves of the wave, then a big cap to store up the DC and minimize the ripple. Am I close?

In my case, then, being as I have a CT on the filament transformer secondary, I'd ground that to the same point as my smoothing caps. Otherwise, I need to create a bridge rectifier for the sake of a ground reference (or is it negative DC?).

Sorry, it's my first day....

[paulster]

I thought I wasn't supposed to ground the CT...I thought the grounding was to happen on DC side of this? Unless I've misunderstood the previous postings....my understanding was that it was a bad idea to ground the CT on the filament transformer, and that I should ground the DC side instead.

Correct, but this was when using a transformer with 6.3V across the entire secondary winding (3.15-0-3.15V effectively) and a bridge rectifier. In this case you wanted to treat the transformer as a 0-6.3V transformer, which is why you needed to forget the centre-tap (as it would sit at an effective potential of 3.15Vdc and would therefore short out one half of the secondary each half-cycle if it was tied to ground). For reference, in a bridge rectifier circuit the bridge is alternating which end of the secondary winding is connected to ground on every half cycle.

Okay...maybe I'm getting it:

Two types of rectification with diodes...1) full wave with a CT, or 2) full wave bridge. One uses full wave bridge if you don't have a CT....otherwise, one just connects the CT and then runs diodes to pass the separate halves of the wave, then a big cap to store up the DC and minimize the ripple. Am I close?

You're spot on. With full wave rectification you're only using one half of the secondary at a time, and the common point is the centre-tap, which you use as your 0Vdc reference point.

In my case, then, being as I have a CT on the filament transformer secondary, I'd ground that to the same point as my smoothing caps.

Bingo. The centre-tap is 0Vdc for your application.

Otherwise, I need to create a bridge rectifier for the sake of a ground reference (or is it negative DC?).

True, but you'd then end up with 12.6Vdc (give or take), which you'd only use if you went for the regulated PSU board option to give lots of headroom. It'll be a little tough on the heaters otherwise!

As a side note, the reason many people (incorrectly, in my opinion) refer to bridge rectifiers as voltage doublers is because they use the entire length of the secondary winding at the same time, whereas the older style centre-tap to ground and then two diodes (or a rectifier tube) uses only half the secondary between the centre tap and one end of the winding at any one time and therefore produces half of the possible voltage. If this is your (old-fashioned) reference point then the bridge rectifier sounds like it's doing some fancy voltage doubling trick. It isn't.

[dehughes]

Paulster...really...I know that Newcastle Brown is probably your "common beer" over there, but if you were over here I'd take you out and buy you as many pints as you'd like. Thanks SO MUCH man...greatly appreciated. You need some honorary Amp Garage doctorate or something...

I'll go wire it up "correctly" and see what I get...

[dehughes]

So, now that I wired things up correctly (thanks Paul) I'm getting 9.56vDC after rectification....waaaaay to much for my poor little tubes.

How best should I reduce that?

[paulster]

Do you have an 8ohm dummy load at all? That's broadly the right resistor value to simulate the kind of load your tubes will put on it.

Once it's loaded down a bit you can see how much you genuinely need to reduce it by, whereas at the moment you've got no load and a cap charged up to the peak of the AC waveform, with the AC being higher than usual as well because of a no load situation.

[dehughes]

Do you have an 8ohm dummy load at all? That's broadly the right resistor value to simulate the kind of load your tubes will put on it.

Once it's loaded down a bit you can see how much you genuinely need to reduce it by, whereas at the moment you've got no load and a cap charged up to the peak of the AC waveform, with the AC being higher than usual as well because of a no load situation.

I have an Airbrake that I built to function at about 8 ohms...but I don't think I have any 8 ohm resistors lying about....

Or, I suppose I could put in the tubes for a second and see...

Still though, what are my options for reducing the voltage once I figure out where they land?

[paulster]

It depends how much it needs to come down under load. Either a regular diode or two to drop up to about 1.5V, or a series resistor or power zener diode if it needs more than that.

[dehughes]

It depends how much it needs to come down under load. Either a regular diode or two to drop up to about 1.5V, or a series resistor or power zener diode if it needs more than that.

Right on. I think I might have a 10 ohm resistor lying about. So, I'd connect that from the + to ground in order to simulate a load? Or just connect it to the + and not ground the other end, and then measure?