DC Heater Question

[markr14850]

OK, I see your point, but I thought we were talking about the CT of the winding, not either end (so notionally at a constant voltage, but accepting that nothing is perfect...).

Better to leave the AC side floating, then, and elevate one of the DC legs?

Yeah, elevation probably won't matter at that point, from a noise perspective. But it's still a good idea if we need to mind heater-cathode voltage limits - for example, if there's a cathode follower somewhere.

[jaysg]

With regards to some math upstream, if you need less drop, Schottky rectifiers deliver ~0.3V drops compared to 1N4001's ~0.6V. Also I think "12V heaters," are happy at 12.6V ac or dc.

The last amp I built from scratch didn't come with a fixed relationship between the PT and OT. The headphone test allowed me to find the optimum orientation and placement for the iron. This may be the OP's actual issue, not heater noise per se. hth

[shoggoth]

With regards to some math upstream, if you need less drop, Schottky rectifiers deliver ~0.3V drops compared to 1N4001's ~0.6V.

You have to check the datasheets for whatever diode you're using. The forward voltage drop varies with current. This is no big deal in most applications, but when you are cutting it close with voltage regulators it's suddenly important.

Also, even if you seem to be meeting the minimum voltage of the regulator, remember to account for the drops from voltage ripple - if those take the DC below the minimum for the regulator, that ripple will make it through.

[JazzGuitarGimp]

With regards to some math upstream, if you need less drop, Schottky rectifiers deliver ~0.3V drops compared to 1N4001's ~0.6V.

You have to check the datasheets for whatever diode you're using. The forward voltage drop varies with current. This is no big deal in most applications, but when you are cutting it close with voltage regulators it's suddenly important.

Also, even if you seem to be meeting the minimum voltage of the regulator, remember to account for the drops from voltage ripple - if those take the DC below the minimum for the regulator, that ripple will make it through.

Good advice. In addition to all of this, I also throw-in a 10% to 15% deduction for low AC line conditions.

[soma_hero]

You're not going to be able to use a regulator with that supply unfortunately. They need to drop 2 or 2.5 volts, check the data sheets.

You'll need at least a 8v transformer, full-wave rectified. You've also got to worry about heat dissipation, because all that "waste" voltage will get converted to heat by the regulator, so you'll have to drop voltage before it gets to the regulator, and heatsink the regulator. And that 8v is a bare minimum btw, it leaves little wiggle room. I'd personally got for a 12.6 w/ a bridge rectifier, drop a lot of voltage with more diodes and/or resistors to get around 9.5v, and then regulate that down to 6.22v using a 6v regulator and few diodes as Merlin does here: http://www.valvewizard.co.uk/heater.html

Also remember if you use a bridge, you halve the potential amps the transformer can output.

Done it before. Works fine. Bridge rectified 6.3V gives over 8V.

If you don't want to heavy load the rectifier/regulator, just regulate V1. Run the rest of the preamp on AC. V1 is the most critical. Mesa normally runs DC only to V1. They don't regulate either

LM1085 dropout for 300ma is less than 0.8V. Heater wants 6.3 +/- 20%. Even if his regulated voltage is around 5.5v it'll be fine

[rp]

http://www.valvewizard.co.uk/heater.html

I'm looking at the Valve Wizard link and see this diagram, aren't the AC and DC side here both grounded?

So far this has been the most informative post on DC heaters on the web. Not the straight forward undertaking I had assumed, I must think this through more for my 5C8 maybe try a humdinger or elevated first. Maybe best just live w/ the slight hum, part of the authentic old school charm.

[markr14850]

http://www.valvewizard.co.uk/heater.html

I'm looking at the Valve Wizard link and see this diagram, aren't the AC and DC side here both grounded?

Looks to me like only the DC side is grounded. The filter cap neg is tied to the bridge neg.

So far this has been the most informative post on DC heaters on the web. Not the straight forward undertaking I had assumed, I must think this through more for my 5C8 maybe try a humdinger or elevated first. Maybe best just live w/ the slight hum, part of the authentic old school charm.

Yeah, lots of good points here. But I'd still say that if you have space in your chassis, it's worth it to just pay the $17 for the power supply unit I linked to before, and never have to worry about it again.

[shoggoth]

LM1085 dropout for 300ma is less than 0.8V

You're correct, that will work then. 7805+diode/7806 would not with their higher dropout voltages, which were the ones I was thinking of.

[bal704]

I'm going to try the humdinger. I've got a 500 ohm pot on order. Seems like the simplest fix if it works.

I've also got some LM1085's on order. I may try it on V1 only. I'm not sure if I have room for the $17 regulator in my chassis, but that's plan C.

I'll update this once I know more.

[soma_hero]

I build high gain amps a lot, with AC heaters and hum is not an issue compared to hiss. It is possible to get even a high gain amp quiet with AC only. You need to be careful with how to route the heater wiring, how and where it's referenced to ground etc..

Post a gut shot of your amp and maybe we can see if there may be other issues.

Using the LM1085 is easy. Just follow the data sheet schematic. Big cap (22,000uF) on the input after the bridge, bypass the adjust pin with a 220uF, use a protection diode and then a 220-470uF on the output.

[dorrisant]

Mesa normally runs DC only to V1. They don't regulate either

Do you know the diodes they used? Had one (DC-3) on my bench and didn't check. D'oh!!!

Tony

[martin manning]

I'm sure 1N4007's would work fine there. Note that the upper right diode in that MESA schematic is reversed.

I've posted the Marshall Studio 15 DC heater supply before, but here it is again. I have personal experience with this set-up- it is quiet, and it gets a lot of play-time. Both of these unregulated supplies have some ripple.

[bal704]

I got one of the LM1085 chips I ordered today. Based on the circuit on page one of the data sheet, figure 4, I can use some resistors to get the 6.3V DC I need based on the equation given.

If I use a wall wart for DC input, do I need to hook the negative of the wall wart (and Negative of the LM1085 circuit) to the chassis for a ground reference? Or can I just hook the negative of the wall wart to the negative on the LM1085 circuit and have the output voltage float?

Basically, with the wall wart, nothing is connected to the power in the amp. Wall wart into the LM1085 circuit, and the output to the heaters.

[fusionbear]

Just out of curiosity, have you guys seen the way the Gibson Goldtone GA-30 does the filaments on the preamp? They run 25.6VDC with a interesting looking current limiting circuit. When I find the schematic, I will post...

[Cameron]

I got one of the LM1085 chips I ordered today. Based on the circuit on page one of the data sheet, figure 4, I can use some resistors to get the 6.3V DC I need based on the equation given.

If I use a wall wart for DC input, do I need to hook the negative of the wall wart (and Negative of the LM1085 circuit) to the chassis for a ground reference? Or can I just hook the negative of the wall wart to the negative on the LM1085 circuit and have the output voltage float?

Basically, with the wall wart, nothing is connected to the power in the amp. Wall wart into the LM1085 circuit, and the output to the heaters.

Most of the "hum" you are hearing is probably the tube. Floating the filament will most always kill the hum. If you are hearing a "buzz" that would be more of a reason to use DC or adjust the layout.