Schematic and parts list for DC heater supply?

[moonrock]

I’ve tried this before with a Dumble style amp. I wasn’t ever able to get a quieter sounding DC heater method until I regulated the supply. The rectifier with a lot of capacitance still had ripple that injected noise into the system. Adding a regulator makes a dead quite supply and amp. However, the cost is that the regulator needs several additional volts on the supply side to properly regulate with no ripple. So that requires a different AC supply. The standard 6v AC won’t cut it. I bought a cheap 12v transformer from Radio Shack and it has worked great. I needed to add a power resistor in series to drop the voltage a little before it hit he regulator to keep the temperature of the regulator down. Good luck!
Bruce

[dehughes]

Right on. I'm getting it, slowly... Thanks so much.

1) The size of the cap will determine how much ripple (residual AC?) is present on the DC. So, for me, a high-value cap (10000uf?) would be preferable for running two tubes, as it would have less ripple than a smaller cap (1000-4700uf, etc...). This being because the more tubes run on the DC circuit, the more current needed, the more the cap would have to charge/discharge, thus putting more AC on top of the DC.

2) A regulator would smooth things out even more, though I'd need to watch the input voltage to make sure I don't work over the regulator.

3) A separate supply from a dedicated winding (or filament transformer) is preferable for my situation over tapping into the 6.3v AC winding. So, I should be looking into a hefty (2A?) 12v transformer.

Sound right?

[paulster]

1) The size of the cap will determine how much ripple (residual AC?) is present on the DC. So, for me, a high-value cap (10000uf?) would be preferable for running two tubes, as it would have less ripple than a smaller cap (1000-4700uf, etc...). This being because the more tubes run on the DC circuit, the more current needed, the more the cap would have to charge/discharge, thus putting more AC on top of the DC.

Exactly. It's not exact but you can broadly say that you'll have 10% ripple for 1A drawn at 10,000uF. You can halve the ripple by going to 20,000uF, or you'd about double it by going down to 4,700uF. You'll need at least 10,000uF for your application I'd say because of the susceptibility of the 6SN7.

2) A regulator would smooth things out even more, though I'd need to watch the input voltage to make sure I don't work over the regulator.

You need to ensure that you don't go over the regulator's maximum input voltage and, equally importantly, you need to ensure that your minimum DC voltage under load doesn't drop below the DC output you want plus the dropout voltage of the particular regulator. The dropout voltage is the voltage that will be lost due to inefficiencies in the regulator.

3) A separate supply from a dedicated winding (or filament transformer) is preferable for my situation over tapping into the 6.3v AC winding. So, I should be looking into a hefty (2A?) 12v transformer.

Only if you're going to use a regulator. It's easier with a separate winding as you can get plenty of headroom and don't have to use specialised regulators and/or diodes.

I use the 6.3V windings and rectify with very low forward voltage Schottky diodes (0.45V drop), use large amounts of filtering (30,000uF) and then use a very low dropout regulator (0.55V) so I can comfortably get regulated 6.3Vdc off 6.3Vac, but it would be easier to use regular diodes or a bridge rectifier block, less filtering and a less specialised regulator. I wanted to be able to retro-fit mine into any amp without having to add an additional tranny so I figured the effort was worthwhile.

For yours I'd still try the brute force regulation on a tag strip or something before you go overboard looking for your solution.

[dehughes]

Nice. Thanks Paul. I think I'm getting it.

So, when you say "brute force" regulation you mean...?

Also, if you'd be so kind, what parts did you use to create your DC off of the 6.3v AC? In reading your post on Page 1, you suggested "a low-dropout regulator and Schottky diodes to get 6.3Vdc with some headroom off 6.3Vac". If you'd be so kind, let me know what parts you're referring to (Mouser/Newark/etc...) so I'm sure to order the right ones.

What I'm thinking of doing is ordering a ton of diodes (I'm going to build other amps with SS rectifiers eventually....), a couple high uf caps, and a regulator and bridge rectifier chip or two so I can explore all the options. Just want to be sure I have the right parts, since I think I'm seeing several distinct options:

1) Bridge recto made of 4 diodes off of the 6.3v AC into a high uf cap.
2) 12v Filament tranny with a pair of diodes into a high uf cap.
3) Bridge recto block with regulator and Shottky diodes off the 6.3v AC, and a couple of high uf caps.

[markr14850]

TubeCad has some discussion on heater supplies here:
http://www.tubecad.com/2008/09/blog0149.htm
Note: It's mixed in with a bunch of discussion about high voltage regulation. Be sure to keep the two areas differentiated.

He describes his H-PS-1 low voltage regulator kit - which is very nice, and I'd even recommend it as a solution for you - if you have a place to fit it. It comes with all the parts you need. $24.

He has deeper discussion of high and low voltage DC supplies in the manual for his Janus regulator kit here:
http://www.tubecad.com/Product_PDFs/Janus%20Rev%20A.pdf
Great reading!

--Mark

[dehughes]

Killer Mark! Great reading... I just may go with that H-PS-1 unit, as I really want to get this amp finished ASAP....and that's a good price...

Would you recommend the 5v or 6.3v? Probably the 6.3v model.

[paulster]

So, when you say "brute force" regulation you mean...?

Enough capacitance that the ripple is insignificant in your application. Not the prettiest solution but very simple in terms of part count and can work quite well when you have a fixed load.

Also, if you'd be so kind, what parts did you use to create your DC off of the 6.3v AC? In reading your post on Page 1, you suggested "a low-dropout regulator and Schottky diodes to get 6.3Vdc with some headroom off 6.3Vac". If you'd be so kind, let me know what parts you're referring to (Mouser/Newark/etc...) so I'm sure to order the right ones.

I use Sipex SPX29302AU5-L regulators and MS503 diodes. I'd advise against trying to build a regulator using a variable voltage regulator unless you do it on a PCB though, because they need the reference resistors very close to the IC or the resistance in the wiring can become an issue and instability can sometimes occur.

Which brings me to:

He describes his H-PS-1 low voltage regulator kit - which is very nice, and I'd even recommend it as a solution for you - if you have a place to fit it. It comes with all the parts you need. $24.

That's a great price and looks like exactly what you need for a full-blown solution to this. It's essentially the same thing as I use but with a different regulator.

[dehughes]

Sweeeeeeeet. Thanks Paul! Greatly appreciated...

What type of transformer (input voltage? output voltage/current?) do you recommend I run that regulator board with?

[paulster]

That regulator has a 1.5V dropout voltage at 3A, and according to the data sheet about 1.1V at 1A.

That means you need between 7.4 and 7.8Vdc on the input to the regulator.

He doesn't state what diodes he uses but if we take a notional 1V forward drop then you'll want another 2Vdc available, so 9.8V. You also need to ensure that the ripple in the filtering doesn't drop below this, so you want about 12Vdc to be comfortable, since the capacitance isn't specified either.

12Vdc peak can be obtained from a 8.5Vac supply, so I'd look for a 9V or 12V transformer.

The higher the input voltage, the more headroom you'll have for power brownouts but the more heat that the regulator will have to dissipate. The regulator will be dissipating the current drawn by the preamp tubes multiplied by the difference between the input and output dc voltages in watts.

Its maximum input voltage is 30Vdc but you'll be nowhere near that.

If you're going to go to all this trouble then run all your preamp tubes off it. You might as well go for the quietest power supply you can after this work.

[dehughes]

That regulator has a 1.5V dropout voltage at 3A, and according to the data sheet about 1.1V at 1A.

That means you need between 7.4 and 7.8Vdc on the input to the regulator.

He doesn't state what diodes he uses but if we take a notional 1V forward drop then you'll want another 2Vdc available, so 9.8V. You also need to ensure that the ripple in the filtering doesn't drop below this, so you want about 12Vdc to be comfortable, since the capacitance isn't specified either.

12Vdc peak can be obtained from a 8.5Vac supply, so I'd look for a 9V or 12V transformer.

The higher the input voltage, the more headroom you'll have for power brownouts but the more heat that the regulator will have to dissipate. The regulator will be dissipating the current drawn by the preamp tubes multiplied by the difference between the input and output dc voltages in watts.

Its maximum input voltage is 30Vdc but you'll be nowhere near that.

If you're going to go to all this trouble then run all your preamp tubes off it. You might as well go for the quietest power supply you can after this work.

Right on. Thanks again Paul!

One channel has a 12AX7, the other has a 6SN7, and they both run into a 12AX7 PI. That's it for the preamp. Should I run all three on DC? I hadn't considered running the PI as well...just the first two...

[paulster]

If you're going to put the DC regulator card and an extra transformer in then I'd run the PI off DC as well, unless your wiring routing makes it impractical.

If you're just going to try strip-boarding a rectifier then I'd keep the current down to minimise the ripple and just do the 6SN7 and first 12AX7.

[dehughes]

If you're going to put the DC regulator card and an extra transformer in then I'd run the PI off DC as well, unless your wiring routing makes it impractical.

If you're just going to try strip-boarding a rectifier then I'd keep the current down to minimise the ripple and just do the 6SN7 and first 12AX7.

Wonderful. Thanks.

As for tranformers, would something like this work?

http://www.mouser.com/Search/ProductDet ... GgAA%3d%3d

Or this one?

http://www.mouser.com/Search/ProductDet ... Z%2fPn4%3d

Bigger? Smaller? More/less voltage? More/less current? Lemme know...

As well, I don't mind spending the $$$ to get the right part, for if I'm going to all this trouble then I want to do it well...

[paulster]

The first one should do what you want. You'll be drawing 1.1A running all the preamp tubes but Hammonds are conservatively rated and if it bogs down slightly then you'll still have loads of voltage headroom.

The regulator will basically be dropping 9V or so at 1.1A, so there'll be 10W of dissipation through its heatsink.

That's going to get pretty hot, which would be my concern. I'd maybe go for 546-187C10 as a transformer since that's 10V rather than 12.6V, which will get the dissipation down to a more reasonable 6W and still give you some headroom.

You can see why manufacturers stick to AC. It's not because AC is better but because DC is a PITA to get right!

[dehughes]

The first one should do what you want. You'll be drawing 1.1A running all the preamp tubes but Hammonds are conservatively rated and if it bogs down slightly then you'll still have loads of voltage headroom.

The regulator will basically be dropping 9V or so at 1.1A, so there'll be 10W of dissipation through its heatsink.

That's going to get pretty hot, which would be my concern. I'd maybe go for 546-187C10 as a transformer since that's 10V rather than 12.6V, which will get the dissipation down to a more reasonable 6W and still give you some headroom.

You can see why manufacturers stick to AC. It's not because AC is better but because DC is a PITA to get right!

Wonderful! Thanks Paul. I'll look for transformers which are more to your specs, and post back with ones that look promising.

I don't mind the extra work/expense so long as the end result is to my liking. Most thing in life worth doing or having require lots of effort/expense.

[dehughes]

Cool. So in looking at different transformers, the 187C10 seems to be a good fit for what I'm doing. I think I can mount it on the outside-side of the chassis, and then mount the circuit board on the inside side opposite the tranny, and both will be located close to the power supply and be out of the way of any preamp wiring, etc..

I'll put in the order tomorrow or Thursday, so if anyone has other tranny recommendations, lemme know, as I only want to buy one...and the right one. I'm still waiting to hear back from that Glassworks place, as to what type of tranny they'd recommend to run their H-PS-1....

THANKS!