DC Heaters and artificial center tap

[JacksonRR]

I'm getting some conflicting advice and was hoping one or a few of you gentlemen could clear this up. One person told me to put pin4 to ground in addition to the two 100ohm resistors that are creating the artificial center tap. I showed that PCB design to someone else and he told me it was effed and it should be a closed system that doesn't mix my 12VDC current with ground and the two 100ohm resistors should be the only path out, possible two 47ohm and a balancing pot. Now they've both given me stellar advice in the past and know more than I do so it's confusing. Thanks in advance for any help. I do appreciate it.

[pdf64]

There are a couple of 'similar threads' towards the bottom of the page which look they may have provided a detailed answer.
If not, and I've understood your proposed arrangement correctly, there would almost certainly be smoke with both an ac ground reference and a dc ground reference.
Whatever answer you get, it's best to prototype unknowns before committing them to PCB.

[defaced]

Agreed. Prototype it.

[jmaf]

When you have a DC heater, and no other reference to the chassis ground, you can choose to ground either side of DC directly. Well filtered or regulated DC won't inject any hum.

(Edit: just like if you elevate the heaters, you must remove any center tap or connection from the AC side if you choose to ground + or - DC)

[JacksonRR]

Thanks JMAF

[defaced]

Nice edit BTW.

IMO, post the schem/layout - or at least a snippet of it. I'm the guy who said the layout was effed because the layout does not match the schematic.

And no, it's not a simple question because there is no "right" answer. There are many ways to skin the DC heater cat, including how JMAF outlined which is a perfectly acceptable solution to the probelm. The most desirable way will be dependent on the circuit and layout.

[jmaf]

Nice edit BTW.

IMO, post the schem/layout - or at least a snippet of it. I'm the guy who said the layout was effed because the layout does not match the schematic.

And no, it's not a simple question because there is no "right" answer. There are many ways to skin the DC heater cat, including how JMAF outlined which is a perfectly acceptable solution to the probelm. The most desirable way will be dependent on the circuit and layout.

You're right. I had many problems myself with regulated DC supplies, and there are several combinations that work.

Last time, I grounded the regulator's body(output) so the chassis was +12VDC with respect to the heater's 0V. I hate thinking backwards....so to add a capacitor at the end of the heater line(experimenting), I had to have positive to chassis, etc, each case is a different case....

[JacksonRR]

Nice edit BTW.

Well, I read myself afterwards and could see my frustration coming through from behind whatever point I was trying to make. Not a good thing to display when you are looking for help or answers.

[JacksonRR]

[defaced]

Yep, I know that frustration. We've all been there.

I'll take a look at this new layout tonight and see if I can offer any useful comments.

[Gregg]

When you have a separate winding for your DC heaters you just ground the negative side and that's it. No resistors and/or artificial ground point is necessary.

[Steve A.]

As far as I know the 100 ohm virtual center tap thing would only be applicable to an AC filament supply with two leads, which I believe are out of phase with each other. [Perhaps someone can confirm or condemn my hunch that the two leads of the PT secondary power would be 60hz (or 50hz) sine waves that are out of phase with each other- not that it is relevant to the issue at hand but that is what popped into my head. ;-) ]

When I've used the trick from the Mesa Boogie IIIC+ to obtain a 6vdc filament supply to energize V1 along with one or two relays, if I used a footswitch to activate a relay the leads had to be isolated from chassis ground. I attached my ac_to_dc.gif drawing based on the MkIII+ dc filament/relay supply which references both + and - to chassis. I ground. The drawing includes the 100 ohm resistors to ground on the ac filament supply so if that is already wired in elsewhere you can omit those two resistors (otherwise you will have 50 ohms from each leg to ground). You would need to disconnect a CT for the filament supply. (I've used that circuit in many amps when I need 6vdc for a relay or want to use a DC filament supply for one or more preamp tubes on DC.)

Steve Ahola

[Gregg]

From my experience you don't even need all those resistors. I used couple of times Mark IV (ver. A) style regulated DC filament supply but at some point I decided to remove the resistors and guess what. The amp was still dead quiet without them. Then in another amp I tried that again - grounding only the negative side of the rectifier and still no hum whatsoever.

[stevecg2]

AC filament hums with Center tap grounded

AC filament hums with Center tap grounded. If I float the heater it is dead quiet. Ground the center tap and it hums. This is a new build. I have floating DC to preamp tubes and floating AC to output tubes. I am puzzled. Tried to ground to other places other than the star ground, same results. Functionaly all seems good, amp sounds quite good. Fender Deluxe type preamp and Marshall 2204 preamp, relay swiched to 6V6 PP output.
Question: Is it ok to float the DC and AC heaters?

[jmaf]

It's probably grounded both at the center tap and via 100 ohm resistors at the same time....creating a ground loop. The only reason to reference the heaters to ground is to quiet the amp, so if actually floating them quiets the amp, it's ok to do. But if it were me, I'd look for the root cause, because what you describe is not normal behavior.