B+ ground reference

[Kermit]

Up late on a Friday night, trying to better understand tube amps... Not trying to tweak, just to understand.

On the PSU diagram for the Crack, it shows the rectifier bridge grounded to the PSU ground.  Poking with my multimeter, I can see that the neg lead from my rectifier IS grounded to the chassis.

Where is that ground connection?   I think it is through the center pin of the terminal strip attached to the 12au7 socket. 

If that's the case, is the power supply grounded "through" the headphone jack?  Does that matter?  I'm fuzzy on ground loops - would it be good or bad to ground terminal 20 straight to chassis earth?

(You guys named it Crack for a reason!)

[diynewbie]

I'll take a stab at it.  The bridge rectifier and CRCRC ground path is T20 - T14 - T12 - Headphone Jack - Pot - T3 - chassis.  Depending on instructions, the transformer grounds from tab 4 to T14 path or T22 - transformer bell - T16 - chassis.  How'd I do?

[Kermit]

That's my understanding on the rectifier / CRCRC ground path as well.   I'm not following on the second part of your explanation - (T4 is B+, do you mean T3?).  Do you mean that the transformer body is earthed through T16 -> chassis ground loop (right above it), and the T3->T21 path is grounded via the chassis at T3?

The ground on the input side of the transformer is the N wire (which the crack doesn't connect to E, but house wiring does).  The only grounds (that I can find) on the output side are the center tap of the filament heater (for the low voltage windings) and via the T21 -> T3 path (high voltage windings).

[Paul Birkeland]

T3 is the connection of the chassis to the signal ground.

Yes, the headphone jack is part of the ground.  The signal levels at the headphone jack are relatively high to the residual noise left on the power supply at that point, so I would argue that there isn't anything to be improved by moving this around.  (Measurements would seem to agree)

It would be really bad to ground terminal 20 straight to the chassis.  You only want that connection in one place.  You do not want the chassis to carry signal current.

[Kermit]

Since the B+ output windings are grounded via T3, the chassis is part of it's path to ground.  Doesn't that mean that the chassis is already carrying some current?  Is it just that noise is low enough in this amp that there's no need to complicate things with a star or bus type grounding system?

[Deke609]

I believe the Crack grounding scheme is already "star grounded" - with T3 being the star.

If you're inclined to experiment with it, you might try wiring the -ve of the last PSU cap directly to T3 (bypassing the headphone jack, but leaving the connection from the jack to T3 in place) - so that each of the HV wiring and the output signal wiring has its own separate connection to T3.   I have no idea whether it would make an audible difference (for the reasons that PB stated) - but it might make an interesting experiment! 

cheers, Derek

[Paul Birkeland]

Doesn't that mean that the chassis is already carrying some current?  Is it just that noise is low enough in this amp that there's no need to complicate things with a star or bus type grounding system?

The ground system is a modified star. (The power supply is arranged as a bus and then tied to the star to reduce noise)

Current has to flow in a loop, the signal ground touches the chassis in one spot, therefore no current can flow.  If you believe that the chassis is carrying current, it's possible to isolate T3 from the chassis and float the ground, and the amp will still work fine (though I don't recommend doing this).

[Kermit]

Thanks!  Makes sense to me now.