DC Heaters on V1 - Does the physical location of the Rectifier matter?

[SnickSound]

NB: I'm not asking whether I should use DC heaters or not, I will be doing it. Gotta try everything once.

So yeah, I'll be feeding DC to the filaments on V1 only. I got the bridge rectifier, the massive cap and a 5W resistor to drop the voltage back down to where I need it. I tested it on breadboard and it works just fine.

Now that I'm building the amp for real, I'm just wondering if it matters whether I locate the rectifier and cap close to the tube or as far as possible from it? I've read about rectifiers potentially introducing noise, but unsure if they mean that it can "radiate" EM interference to nearby components, or just the switching noise it introduces to the circuit.

[jjasilli]

Everything to do with power, like the bridge rectifier, should be physically located on the PT side of the chassis, and not near the signal input side.  If you must locate the Bridge & heater rectification circuit away from the PT side, then run all its grounds back to a grounding point on the PT end of the chassis. 

[SnickSound]

Everything to do with power, like the bridge rectifier, should be physically located on the PT side of the chassis, and not near the signal input side.  If you must locate the Bridge & heater rectification circuit away from the PT side, then run all its grounds back to a grounding point on the PT end of the chassis.

Thanks! Kind of what I figured, although I wondered if having filtering close to the tube could be beneficial.

As for grounding, I reference my filaments to the power tube cathodes (it's cathode biased) so I definitely cannot connect anything to chassis ground anyway.

[jjasilli]

I wondered if having filtering close to the tube could be beneficial.
I don't think it matters where DC heater filter caps are located -- the topic has never come up. For B+ filtering it only matters that the filter cap is close enough, say within 1' of the tube.  Fender keeps the filter caps near the PT; Marshal locates them them near ea. tube they supply.  Buckdancer's choice so long as they're close enough to the tube.  Unlike B+, the heater current is not part of the signal chain, except for unwanted induction.  So I think  it's best not to ground any part of the heater circuit, including any DC rectification circuitry, near the signal path -- to avoid that induction.


As for grounding, I reference my filaments to the power tube cathodes. 
OK.  There's a number of valid grounding schemes.  But I would apply that to the Bridge and heater rectification circuitry.

[SnickSound]

As for grounding, I reference my filaments to the power tube cathodes. 
OK.  There's a number of valid grounding schemes.  But I would apply that to the Bridge and heater rectification circuitry.

I should have specified, I do this to get DC elevation (about 25-27VDC above ground) on all my filaments, usually makes the amp quieter (and reduces the heater-cathode voltage offset on the cathode follower stage)

Don't know if it matters that V1 will end up elevated AND running DC.

[jjasilli]

I'm not aware of DC filament voltage ever being elevated from ground.  I can't get my head around it, but maybe it's problematical.  Note that some amp guru's state that DC heater supplies can be noisier than AC, if:  i) not sufficiently filtered; and ii) the DC leads are particularly subject to picking up ambient noise, which would be cancelled in an AC supply with twisted pairs.

[jjasilli]

Further important research:  "A common beginner's error is to try to add a ground reference on both sides of the rectifier. This will short out the rectifier! Only one ground reference is required (centre tap, humdinger, elevation, whatever). It can be on either side of the rectifier; the other side will received its reference through the rectifier. (Actually, you can spot this mistake in the Mesa Dual Rectifier and Carvin Legacy amps which foolishly use artifical centre taps on both sides of the rectifier, so maybe it's not such a beginner's error!) If you do add a rectifier it is a good idea to add a fuse too, as rectifiers have a nasty habit of failing short. You can avoid the need for a very high-value fuse by putting it after the AC heaters but before the rectifier."  http://www.valvewizard.co.uk/heater.html

[Willabe]

I'm not aware of DC filament voltage ever being elevated from ground. 

TUT (KOC) books recommend elevating the heaters whether AC or DC.      

And Kevin likes to go 70 to 80dcv for elevation.

[jjasilli]

Yes, see my Reply 6 referencing Merlin.

[SnickSound]

Further important research:  "A common beginner's error is to try to add a ground reference on both sides of the rectifier. This will short out the rectifier! Only one ground reference is required (centre tap, humdinger, elevation, whatever). It can be on either side of the rectifier; the other side will received its reference through the rectifier. (Actually, you can spot this mistake in the Mesa Dual Rectifier and Carvin Legacy amps which foolishly use artifical centre taps on both sides of the rectifier, so maybe it's not such a beginner's error!) If you do add a rectifier it is a good idea to add a fuse too, as rectifiers have a nasty habit of failing short. You can avoid the need for a very high-value fuse by putting it after the AC heaters but before the rectifier."  http://www.valvewizard.co.uk/heater.html

Oh yeah, I read that article through quite a few times. DC filament supply is indeed fused, got a massive 4700uF cap (plus a .1uF film caps because why not), a 1 ohm resistor to get the voltage just right, and the leads are twisted (just easier to run them this way instead of two distinct wires, have to run the return wires since I can't use the ground buss or chassis)

As I map this out in my head, I can't really see a reason why DC + elevation would be a problem. You just never hear about it because most people will only do one.

Anyway, if it ever turns out to be a problem, I can run it off my relay supply which is ground referenced. Before regulation it has enough voltage to run the filaments in series too.

[sluckey]

Anyway, if it ever turns out to be a problem, I can run it off my relay supply which is ground referenced. Before regulation it has enough voltage to run the filaments in series too.

Bingo! Assuming the AC source for the relay power supply is NOT sharing your filament winding, this sounds like a better option.

[trobbins]

If you rectify a PT winding (eg. to get a heater dc voltage supply) then the rectifier and filter cap are best located as close to the PT as practical, as it is the wiring from the PT to diodes to filter cap that conduct the high peak currents.  From the filter cap out to the valve heater, the voltage and current on those wires has very little AC and hence don't radiate or capacitor couple hum over to preamp circuitry.

But that's the ideal situation where the heater winding is not connected to anything else, and one side of the DC heater supply is also connected to a ground, or elevated DC supply (so as not to float the whole heater circuit, and maintain the heater voltage at a constant DC voltage level with respect to amp circuit ground).

It can start to get uglier if the heater winding is used for other purposes, such as for AC heater powering other tubes.  That's because the AC heater is then floating around with a higher AC voltage wrt to 0V than if it was just simply referenced to ground through a humdinger or winding CT.  Most valve circuits other than the input stage valve can live with that higher heater hum signal, as the signal level is higher.

[SnickSound]

If you rectify a PT winding (eg. to get a heater dc voltage supply) then the rectifier and filter cap are best located as close to the PT as practical, as it is the wiring from the PT to diodes to filter cap that conduct the high peak currents.  From the filter cap out to the valve heater, the voltage and current on those wires has very little AC and hence don't radiate or capacitor couple hum over to preamp circuitry.

But that's the ideal situation where the heater winding is not connected to anything else, and one side of the DC heater supply is also connected to a ground, or elevated DC supply (so as not to float the whole heater circuit, and maintain the heater voltage at a constant DC voltage level with respect to amp circuit ground).

It can start to get uglier if the heater winding is used for other purposes, such as for AC heater powering other tubes.  That's because the AC heater is then floating around with a higher AC voltage wrt to 0V than if it was just simply referenced to ground through a humdinger or winding CT.  Most valve circuits other than the input stage valve can live with that higher heater hum signal, as the signal level is higher.

Thanks for all the info

I've finished wiring 90% of the amp, what's left to do is that first stage of the overdrive channel (the one that uses the DC'd filaments). The clean channel is quiet as a mouse despite the extra gain stage used for the FX Loop recovery.

If my elevated DC heaters cause issues, I have two options: continue the elevated AC heater lines to V1, or feed it DC from the relay supply. The relay supply is a Delon voltage doubler fed by the 5VAC tap (I'm not using a tube rectifier so it was free to use). Before being regulated down for the relays, the DC coming off the filter cap (dual 2200uF) is just a bit over 12V... so I could feed the filaments in series instead of parallel without doing anything else. Or regulate it down to 6V if I so desired, but that's a bit overkill for filaments.

[SnickSound]

Well I'll be damned... despite the rectifier and PT being about 15" away from the input tube, I did get a buzzing sound. The kind that is clearly a mains related noise, but not hum per se (I'm guessing diode switching noise). I tried increasing filtering to 10000uF, adding a film cap, etc.

Eventually I admitted defeated and switched to AC filaments... problem fixed. Buzzing sound gone. No hum. Despite the SLOesque level of gain.

I built the filament DC supply exactly as described by Merlin. Mesa does the same thing with the Dual Rec... maybe that mistake of double referencing to ground actually pays off?

Anyhoo... there's basically no incentive to use DC filaments (there's no hum or buzz), so I won't bother. One less fuse to worry about.

All that's left, if I'm being picky, it a bit more hiss than I'm comfortable with. Will try a beefier plate resistor.

[Willabe]

I like to use this.

And I take the dcv from the very last filter cap, where the dcv will be the cleanest, most filtered. I apply it to the 1st preamp tube and put the humdinger pot at the very end of the chassis by that 1st preamp tube.

A member here tested the dc stand off voltage. He found that on some tubes not much difference in noise until he took the dcv up to 70 and 80 volts. 

[SnickSound]

I like to use this.

And I take the dcv from the very last filter cap, where the dcv will be the cleanest, most filtered. I apply it to the 1st preamp tube and put the humdinger pot at the very end of the chassis by that 1st preamp tube.

A member here tested the dc stand off voltage. He found that on some tubes not much difference in noise until he took the dcv up to 70 and 80 volts.

Yeah I did have the heaters elevated, which is why I couldn't reference the DC side to ground. But I was using the power tube's cathode resistor for free elevation, but that only gave me 25VDC above ground.
So I went ahead and added a proper DC elevation circuit (220k + 68k, 22uF bypass) and now I'm about 70V above ground and it's significantly quieter.

That couple with a beefier 2W plate resistor on V1a, as well as a beefier but lower value grid stopper on the input, and it's starting to be very well behaved for such a high gain build. It helps that my input jack is about 1" from the tube socket (input jack on the back, it's the future!)

[jjasilli]