Schematic and parts list for DC heater supply?

[paulster]

Don't bother with diodes - it's hassle to build a board for them when you can just as well get a bridge rectifier module like the one I pictured earlier in the thread.

Get as big a cap as you can that you can fit on the board you use; I'd recommend 10,000uF to minimise the ripple.

Oh, and to reiterate, don't connect the DC ground to the chassis unless you lift the AC centre-tap first. But you may well find that grounded DC is quieter. It is in all my builds, which all use DC heaters.

[collinsamps]

Or you can probably just lift the ground reference on the ac filaments and probably end up with it just as quiet.

One example.

http://people.cornell.edu/pages/mt24/Am ... aters.html

[dehughes]

Or you can probably just lift the ground reference on the ac filaments and probably end up with it just as quiet.

One example.

http://people.cornell.edu/pages/mt24/Am ... aters.html

Huh. Interesting. So I'd just build this on to the CT of the heater wiring, and call it good for the whole string of tubes? Never thought of that. I don't have a 10uf cap....would a 25uf/25v or 50v cap work?

Also, would I need to connect to the B+ line? If I moved the CT to the top of the power tube resistor and cap as recommended by others, then I'd not be connected to the B+ line.

[dehughes]

Don't bother with diodes - it's hassle to build a board for them when you can just as well get a bridge rectifier module like the one I pictured earlier in the thread.

Get as big a cap as you can that you can fit on the board you use; I'd recommend 10,000uF to minimise the ripple.

Oh, and to reiterate, don't connect the DC ground to the chassis unless you lift the AC centre-tap first. But you may well find that grounded DC is quieter. It is in all my builds, which all use DC heaters.

Okay. So you're saying bridge rectifier for the AC for the entire tube string, yes? To make sure I'm understanding it....the 6.3v heater wires would go to the bridge rectifier, which would be grounded (thus lift the 6.3v CT), and then connected to a 10000uf cap (25v or so?)? Is that correct?

As for the hassle, I'm fine putting in another terminal strip...that's not a big deal to me. I just want to do whatever makes this amp the quietest.

[collinsamps]

Or you can probably just lift the ground reference on the ac filaments and probably end up with it just as quiet.

One example.

http://people.cornell.edu/pages/mt24/Am ... aters.html

Huh. Interesting. So I'd just build this on to the CT of the heater wiring, and call it good for the whole string of tubes? Never thought of that. I don't have a 10uf cap....would a 25uf/25v or 50v cap work?

Also, would I need to connect to the B+ line? If I moved the CT to the top of the power tube resistor and cap as recommended by others, then I'd not be connected to the B+ line.

True, both are ways of getting to the same place. Your top of the cathode resiistor method requires less parts & is less costly of course.

[Structo]

That's what I was interested in too.
I've heard you can connect the heater supply center tap to the cathode side of the cathode resistor on a cathode biased amp.

I don't understand how that works?
Does it some how it elevate the voltage 30-40 volts but remains at around 6.3v because the difference in potential is still the same only elevated?

[dehughes]

Or you can probably just lift the ground reference on the ac filaments and probably end up with it just as quiet.

One example.

http://people.cornell.edu/pages/mt24/Am ... aters.html

Huh. Interesting. So I'd just build this on to the CT of the heater wiring, and call it good for the whole string of tubes? Never thought of that. I don't have a 10uf cap....would a 25uf/25v or 50v cap work?

Also, would I need to connect to the B+ line? If I moved the CT to the top of the power tube resistor and cap as recommended by others, then I'd not be connected to the B+ line.

True, both are ways of getting to the same place. Your top of the cathode resiistor method requires less parts & is less costly of course.

Right on. I'll try that first. So, just disconnect the heater CT from ground and run a wire over to the power tube cathode resistor....I'll try that tomorrow or Saturday....

[paulster]

Don't bother with diodes - it's hassle to build a board for them when you can just as well get a bridge rectifier module like the one I pictured earlier in the thread.

Get as big a cap as you can that you can fit on the board you use; I'd recommend 10,000uF to minimise the ripple.

Oh, and to reiterate, don't connect the DC ground to the chassis unless you lift the AC centre-tap first. But you may well find that grounded DC is quieter. It is in all my builds, which all use DC heaters.

Okay. So you're saying bridge rectifier for the AC for the entire tube string, yes? To make sure I'm understanding it....the 6.3v heater wires would go to the bridge rectifier, which would be grounded (thus lift the 6.3v CT), and then connected to a 10000uf cap (25v or so?)? Is that correct?

That's right, but you only want to run the preamp tubes or your first 12AX7 and your 6SN7. Don't try to run your power tube heaters this way or you'll be putting a lot of strain on your heater winding and you'll have loads of DC ripple anyway.

This sort of DC supply isn't particularly clean and you end up with a saw-tooth wave on top of the DC voltage as the cap is being continually topped up and then discharged slightly by the load. The extent of this ripple is defined by the current drawn (i.e. how many tubes you string off it) and the capacitance.

What I would suggest, if you haven't tried this, would be to run all of your tubes on AC and then try the lantern battery trick on just the 6SN7 and possibly the first 12AX7 if you have any residual hum getting in there as well.

If you're happy that your noise is only getting in at the front of the amp then I'd run only those tubes on DC. That will give you the cleanest DC you can get from this method.

In terms of the grounding you have a number of options. One is to leave the AC side grounded, which means your DC side is flipping between +/- 6.3Vdc. Option 2 is to lift the AC ground and ground the DC side. This means that the AC side is flipping between (effectively) +/- 6.3Vac. This is the method I use in all my amps and yields no noise whatsoever, but I run all my preamps off DC and only the power tubes off AC so you'll have to see how this pans out for you. This is why I was suggesting running the first 12AX7 in the same way as the 6SN7, because any noise induced there will be amplified right the way through the amp, so if you have to make a compromise then make it beyond this tube.

The third option, which you could explore before even trying DC heaters, is to elevate the heater ground reference as has already been suggested by others. You can do this with AC or DC heaters, but I think that you're likely to still have an issue with the octal tube simply because of the heater design.

Paul

[dehughes]

Thanks Paul. Much appreciated. My intent was to run only the 6SN7 and the first 12AX7 on DC, so I like your idea there.

I suppose I'd like to emulate the Komet Constellation in that regard, as that's what they did. Does anyone know how they did that? I know on the recto board they have the bridge recto for the amp as a whole, then two other diodes (connected to where?) which then lead to the 10000uf cap. I'm not sure what type of diodes they are, and what they connect to on the other side, though. I do have pictures of the circuit but can't tell where they connect in the circuit....

[paulster]

I suppose I'd like to emulate the Komet Constellation in that regard, as that's what they did. Does anyone know how they did that? I know on the recto board they have the bridge recto for the amp as a whole, then two other diodes (connected to where?) which then lead to the 10000uf cap. I'm not sure what type of diodes they are, and what they connect to on the other side, though. I do have pictures of the circuit but can't tell where they connect in the circuit....

I think the Constellation may have a 12V centre-tapped secondary on the PT as well. Looking at the photos they run the heavy 6.3V heater wires to the jewel light, and then the white/black wires run from there to the back of the chassis and then to the power tubes. From any of the pictures I can't see anything going to the two rectifier diodes for the DC heater supply.

That makes me think there is also a 12V supply, which would only require two diodes, and would give 6.3Vdc or thereabouts when rectified. Since this board is directly over the power transformer it would also explain why you can't see any wires heading to it.

Using the 6.3Vac supply and a pre-packaged bridge rectifier will basically get you to the same point. They have the advantage that they can pick the optimal ground for the DC and AC side independently, but that shouldn't cause you a problem anyway.

The only thing you might want to do if you make a little eyelet board for this is to leave a pair of eyelets in the positive supply rail that you can either link together or stick a diode in to drop a bit of voltage if your DC ends up around the 7V mark. A regular 1N4007 will drop 0.7V.

There's an example of how I've done a similar thing below. On the left is my DC regulator board (that Dan had mentioned in the other thread) that I run my preamps off, so that's the tightly regulated 6.3Vdc going off to the left. The right of that board has 6.3Vac coming in, and 6.3Vac (white wires) going off to my footswitch board. This has the bridge rectifier module that I was suggesting (in 3/32" eyelets) and then its filter cap. That gives me a DC supply for the footswitch that is decoupled from both the AC for the power tubes and the DC for the preamp tubes.

You could get away with a tiny board for all that you need to do, especially if you use a can cap like the Constellation and don't mount it on the board. In that case you could probably just use a piece of tag strip.

[Structo]

Or if your PT is too soft to handle the extra demand, you could always pick up a cheap filament xfrmr and make the supply out of tha.
Run the power tubes from the PT and the preamp from the new xfmr.

[dehughes]

Right on. This is where I think I'm going to land. I like the DIY part of building a DC supply (no offense to your suggestion, Phil) as opposed to a wall wart (though that'd be easier), and the parts aren't too costly, AND I'd get to learn a lot in doing it...which is really half of why I'm doing this.

So, two questions:

1) Paul, what are the actual part numbers of the items I need? You mentioned the bridge rectifier earlier...but I see you have diodes (what kind?) on your board. I'm going to compile a list of what I need to purchase. I have lots of turret strips of all lengths lying about, so I'm set there.

2) Structo, how would I know if the tranny is too weak to do what I'm intending? If it is, what type of filament transformer would I need to get (part number?). I don't even know where to start with filament trannies...


THANKS!

[markr14850]

If you go for single diodes rather than a one piece, I'd recommend these Schottky diodes: Mouser - 625-1N5818-E3. The low loss on these might let you run a regulator.

For the filter cap, I'd highly recommend these: 647-UHE1C103MHD. They are designed for high ripple situations.

If you want an LDO 6v regulator that (with a heatsink) will be good up to 1.5A, you could try this: Mouser - 513-NJM#7806FA. ($0.54 each!)

[paulster]

Paul, what are the actual part numbers of the items I need? You mentioned the bridge rectifier earlier...but I see you have diodes (what kind?) on your board. I'm going to compile a list of what I need to purchase. I have lots of turret strips of all lengths lying about, so I'm set there.

Here's some Mouser part numbers to get you going:
Bridge rectifier: 512-GBU8B
Capacitor (10,000uF 16V axial): 594-2222-021-15103
1N4007 diode (to drop 1V if needed): 833-1N4007-TP
3/32"x.156" eyelets (for bridge rectifier): 534-35
1/8"x.156" eyelets (for everything else): 534-45

This is assuming that you'll have the 10,000uF cap mounted on the board, and that your FR4 board is 1/8" thick for the eyelets I've specified. Otherwise adjust to suit.

The diodes on my board are for flyback voltage suppression from the off-board relays so you don't need to worry about these. The narrow block thing between where the white wires land on the board and the cap is the rectifier.

I've included a 1N4007 diode in the list which you might need if your rectified DC voltage is too high. It will go between the positive terminal of the cap and the + supply to the tube heaters and will drop about 0.7V. If you don't need it then use a wire link between the two eyelets instead.

Paul

[Structo]

The only way I know is if you have the PT specs, you then add up all the filament heater currents and if it approaches the max of your PT then it is going to probably drop some voltage due to the current demand.

But yours really should be able handle it.
I just don't know that much about what kind of load the DC supply will place on your PT.

BTW, a lot of guys used the filament transformer that Radio Shack carried for years.
There was some talk about them not selling them anymore.

They just want to sell cell phones and batteries anymore it seems......