12 Volt DC Heaters

[Tage]

So have been looking at building a fairly sensitive pre-amp and wanted to use DC Voltage for the heaters. I looked into the regulated DC heater supply on Merlin's website and thought it could easily suit my needs as I won't be exceeding 1Amp of Current draw. But I am wondering if I can modify his circuit to give me 12.6VDC. I plan to swap the transformer for a 20V unit, change the regulator to a 7812. Would this method work and give me the +12.6V I wan't? I saw a similar but more complex arrangement used on some Ashdown amps, but they go +12V and -12V through 4 heater elements. I appreciate the feedback.

[oc disorder]

That should work ok Tage.

A couple of things to note from Merlins website (and I hope Merlin doesn't mind me quoting)

1." three-pin voltage regulators usually require an input voltage that is at least 2.5V above the output voltage in order to work."
So you need approx 15 volts D.C. going into the regulator
and
2. "A 5V regulator can have its output raised to ~6.3V by elevating its ground terminal by 1.3V; the voltage drop across a pair of silicon diodes is almost perfect for this."
If you fit a diode between the common of the regulator and ground the same way he has elevated the 5v reg this will increase the output by aprox .6 volts hopefully giving you the 12.6 you require.

I find sometimes under load the reg drops to just under 12 volts.
12 volts will run the filaments/heaters just fine anyway.

[Tage]

Yeah I figured I can elevate the ground with one diode and I should be good as it will have a drop of about 0.6V. Also would I be better off to go with a 16V transformer instead of a 20V? I'll make sure it has a VA rating of 20. I'm hoping to build the whole assembly on a board and be able to duplicate it for various builds.

[trobbins]

The regulator dissipates the the voltage drop from supply to 12.6V - that can get pretty significant - you will need to understand how to manage that dissipation - and the transformer selection is part of the process.

[Gregg]

The other way to get the exact voltage you want is to use an adjustable voltage regulator like LM317.
Using 16V PT is a better idea because you will have to dissipate less heat which means smaller heatsink for the regulator.

[loudthud]

Proper design of a regulated power supply is quite involved. You start with specifying what range of line voltages you want the output to maintain regulation. In the USA that might be 130V to 110V. Next, select a tranformer that might work. You'll need to know the primary and secondary resistances and the turns ratio or open circuit voltages. Plug these numbers into PSUD-II. Next, select a filter cap. You'll need to know the ESR. For a rectifier(s) select something that has atleast 2X the current rating for the load. Run a simulation but be sure to select the line frequency. Use 50Hz if your amp will be used internationally. Check the RMS secondary current to insure you are within the transformers' ratings.

Check the regulator chips' spec sheet to find the minimum input voltage you'll need. Is the lowest part of the ripple on the filter cap above that voltage? Now you have to adjust the transformer parameters for the low line voltage condition and check that you still have enough voltage to run the regulator chip. If you don't have enough voltage, you'll have to increase the transformer voltage, use a bigger filter cap, try Schottky rectifiers, or a low dropout regulator.

Lastly, run the simulation again at the high line voltage condition and gasp when you figure out how big the heatsink will need to be. You need to calulate the thermal impedance of the heatsink to keep the die temperature of the regulator within reason. Also check the capacitors' peak voltage and RMS ripple current rating.

[exclamationmark]

You should get the voltage drop across the regulator as close as you can to 2.5 volts (or the minimum difference between vin and vout) by picking a suitable transformer, if you want to run lots of current. Any extra voltage will be dissipated as heat. A TO-220 package can handle about 1 watt without a heatsink, probably close to 4-5 watts with a good heatsink. You can also get LDO regulators that have the junction voltage down near the range of 0.6 volts too.

For some reason using regulators in any of my circuits evokes a sense of terror whenever I smell them heating up, fearing that something has shorted out (not a totally irrational fear at that too!).

[Tage]

Alright I 'll do that as soon as I can, thanks for the advice Loudthud. I'd forgotten all about the duncan amps designer. Also what exactly constitutes a "low" ESR rating?

[loudthud]

The higher you go in capacitance and ripple current rating, the ESR tends to go lower. At 1000uF, ESR for a normal service cap line is around 0.2 ohms. Special high current lines or caps intended for switching power supplys would go lower. At 10000uF, most caps are in the .05 ohm range. A Digikey catalog is a good resource for browsing listings. For a 1 amp power supply, ESR is not too important but you need a number to plug into PSUD-II, 2 ohms is the default.

[Tage]

Yeah I saw there default of 2 ohms and figured it should be an alright place to start, now my only problem is the data sheet for my intended transformer, Hammond mfg #229C16, doesn't spec the resistances. I may just order it and then measure the pri and sec resistances.

[guiltless]

This is the thread I have been looking for!!!

I am looking at trying to do a similar setup. Though I am wanting to use 4 5691's. Their datasheet is showing that they pull .6A each (an I am sure more when they are just warming up).

So I was wondering if a something like a Hammond 166Q12 or 167Q12 (12.6v C.T. 6A Max) for a PT and an LM338T http://search.digikey.com/scripts/Dk...name=LM338T-ND work? And then a quad of 1N5408 3A diodes doing the rectifier duty.

After rectification would I be near the necessary 14v-15v needed for the LM338T? Or should I be looking at something like the 166Q14 instead for the 14v C.T?

[guiltless]

Nevermind. It looks like that regulator would need a huge heat sink to be able to put out the current that I may need.

[exclamationmark]

Typically the cold-resistance of a tube is 5-10 times lower than at operating temperature. ie, it draws 5-10 times more current for the first couple of seconds at startup. You may consider putting a NTC thermistor in series with the heaters you intend to run (directly after the regulator). I built an amp powered entirely by a SMPS and thermistors were a necessity, otherwise it would simply go into a lockup cycle upon startup due to the power supply thinking it was shorting out!

[jpfamps]

I've used 3-terminal regulators in various amps for DC filament supplies.

Some general observations:

1) A 12VDC supply makes more sense than a 6VDC supply due to the lower current requirements; this makes everything else so much easier.

2) I use Schottky diodes for the rectifier as these have lower forward voltage drop and cleaner turn off characteristics than standard rectifier diodes.

3) If using an adjustable 3-terminal regulator eg LM317 to supply all the current, then soft start can easily be implemented with the addition a resistor and transistor, ie cheaply! See the National Semiconductor LM317 application notes for details. Alternatively, if using a fixed 3-terminal regulator, or transistor around the regulator to supply most of the current, soft start can be implemented using a MOSFET, a couple of resistors and a cap.

[mooreamps]

then soft start can easily be implemented with the addition a resistor and transistor, ie cheaply!

I do the soft start with a resistor, and a cap. "even cheaper"... I have a print of that ; somewhere......

-g