some hum in my new 5e3

[macdillard]

5F2A grounding

Here is what I did on my 5F2A and it does not hum...I used a bus ground, started at the PT to the Caps to the PI to the pots and down the line to the inputs..Ground lug at the input jack...I also elevated the heater circuit by 19.5 VDC from the cathode of the power tube...It does not hum.

[Paul P]

I did the same but grounded the 'buss' at the AC mains inlet, with similar
results. I think what's important is that things be ordered and there not
be any parallel paths. My heaters are elevated a bit over 40v.

Paul P

[tubeswell]

Here is what I did on my 5F2A and it does not hum...I used a bus ground, started at the PT to the Caps to the PI to the pots and down the line to the inputs..Ground lug at the input jack...I also elevated the heater circuit by 19.5 VDC from the cathode of the power tube...It does not hum.

Mine doesn't hum hardly at all either - but I am curious about the heater biasing. I have heard that it is 'good practice' to put a resistor in series with the heater CT , or put a voltage divider between the heater CT and teh cathode, when connecting it to the output tube cathode. Did you hook your Heater CT straight to the cathode, or did you put some padding in?

[tubeswell]

on the schematics that give detailed voltages, you can just calculate the current(since you can obtain V and R; V/R=I). For the PI probably about 1mA per triode. The (current) levels you see in an amp would be something like uA(microamp= one millionth of an amp) range for pickup and grids(across grid--not through), single mA (one thousandth) for triodes, tens of mA(static and higher when drawing signal for output tubes), around 100-something mA for a B+ winding(max.--depends what the load is--example, say 40mA x 4 output tubes plus couple mA for pres = around 163mA), Amps for heaters, spk. out(at max. signal=max.current). Basically you're supposed to try to do what that Henry Ott guy said (try to make youself concious of what is happening in the ground line)--and you can see an extension/application to tubes in the Kimura article.

Thanks Dai h

Since posting earlier, I did some sums to calculate the current of the various stage's cathodes for the 5G9 and I got about 1mA for most triodes (as you note). Also, I understand the screen current draw a 6V6 is around 4mA and the plate current is between 25-40mA ( I take it that I would double these for a pair?). I think I have worked out that the bias current draw is 0.5mA (-69V/138,000 Ohms). Therefore I presume (if I followed Stokes' logic) that I should ground the PI, Bias and trem grounds with the pre-amp (and maybe even the screen decoupling cap ground as well?) What is this Kimura artcle you refer to?

[dai h.]

I agree about the PI but maybe I could be missing something (or... he was referring to the PI filter cap ground??).

the Kimura article is this:

http://ampgarage.com/forum/viewtopic...220&highlight=

It's an article on tube amp grounding by a Japanese author which I liked a lot and thought was important enough to translate for English speakers. (The pics are missing at the amp garage page since the author switched to a different ISP. The original article--parts 1, 2, and 3 are here: )

http://www.op316.com/tubes/tips/b410.htm

http://www.op316.com/tubes/tips/b420.htm

http://www.op316.com/tubes/tips/b430.htm

and also, there are a bunch of related links. I've said this before (but I feel it bears repeating, so I will say it again)--if you *basically* know how the pickup, a tube, and a power supply works and you do some reading on the subject (grounding) and understand the basic theory an average person should be able to do things (for a simple amp such as champ to say Bassman/Marshall) such as correct mistakes or improve a grounding scheme or come up with one--mainly due to the fact that there is no heavy math involved (so not "rocket science" stuff but easily within your grasp with a bit of effort). I think you will come to understand that things like using a buss, star, plane, using the chassis for a portion of the ground line, etc. by themselves do not guarantee low hum (or a noisy amp) and you can succeed or fail by the way that you connect things together (meaning connecting points together the right way is more important than a particular "star grounding", "buss grounding", etc. "method").

Or that things that seem sort of different could be sort of the same. For example what Bruce says about the brass plate in Fenders being sort of a star--just a big gigantic one. If you "star" connections the idea is to avoid common impedance/resistance (reading into this means the lower the R between/among these starred points = low common R--the better because the less the common R the less a voltage beyond zero is able to develop across the R). If you made the point physically bigger and bigger, surface area becomes bigger therefore R is *lower* (and lower the bigger you go with the area) which means the goal of low common R is achieved.

(want to write a bit more but need to take a break for some anti-inflammatory stuff for my fingers... )

[dai h.]

some stuff I drew to help myself make sense of things:

this one was an exercise in making myself concious of what currents were flowing in the ground line:

http://www.flickr.com/photos/7878340@N05/1467701518/

this one is trying to sort out the logic behind some things like a plane and a buss:

http://www.flickr.com/photos/7878340@N05/2323154407/

the stuff on the right for example is how Soldano(and others) have the input jack ground and pot grounds wired. Why is it wired that way (I asked myself)? and 1. they are very low current returns (uA;microamp range) so it makes sense to group them together, and 2. it could be seen as sort of equivalent to a sideways "star" IF the connections between them have a low R (accomplished with using a thick wire) 3. trying to physically star all the pots to the input jack ground directly would not be desirable because you'd end up with this weird physical clump of jack and pots, so it makes sense to stretch the line out so the controls are where you want them to be (in sensible locations) 4. if you tried to "star" with individual wires from ea. pot back to the input jack, you'd end up with longer distances which is not desirable from the point of trying to keep the R low

anyway, I guess I am a bit fixated with this subject...

[macdillard]

Mine doesn't hum hardly at all either - but I am curious about the heater biasing. I have heard that it is 'good practice' to put a resistor in series with the heater CT , or put a voltage divider between the heater CT and teh cathode, when connecting it to the output tube cathode. Did you hook your Heater CT straight to the cathode, or did you put some padding in?

I was using a PT that did not have a CT for the heaters. I used 100 ohm resistors to create a CT and connected that to the cathode. I have also on other builds just connected the CT to the cathode. Go over to AX84.com and take a look at the P1Extreme (Phil's) build schematic and layout. That is what I did.

[tubeswell]

I agree about the PI but maybe I could be missing something (or... he was referring to the PI filter cap ground??).

He was referring to the PI cathode ground - I was asking about that, and about the Trem cathode grounds and the bias ground

http://www.ampwares.com/ffg/schem/tr..._5g9_schem.gif

In the layout pic, they all appear to be going to a 'brass plate'

http://www.ampwares.com/ffg/schem/tr...5g9_layout.gif

I was wondering whether the trem and bias ground would be better with the output tube cathode ground etc in a split-ground setup, simply because the bias and trem are connected to the output tube part of the amp. Does that logic make any sense, or shoudl I go by the relative amounts of current present at each ground.

Also what is the most ideal situation w.r.t. the screen grid decoupling cap ground? - should it be connected with the pre-amp decoupling cap ground (as advocated by Stokes), or the reservoir cap ground?

[tubeswell]

It could be seen as sort of equivalent to a sideways "star" IF the connections between them have a low R (accomplished with using a thick wire) 3. trying to physically star all the pots to the input jack ground directly would not be desirable because you'd end up with this weird physical clump of jack and pots, so it makes sense to stretch the line out so the controls are where you want them to be (in sensible locations) 4. if you tried to "star" with individual wires from ea. pot back to the input jack, you'd end up with longer distances which is not desirable from the point of trying to keep the R low

So I could use thick multi-stranded wire for the ground (which is a heap of wires in parallel, thereby lowering the R)?

[dai h.]

I'm saying IF the connections are low R (which a thick multi-stranded wire would be), then a "buss" can be sort of equivalent to a star connection. If you look at PCBs for example I think you see this a lot (wide trace that is sort of a buss or series-looking connections for grounds). The lower the R in general the better of course, BUT obvious disadvantages are more cost, more space, possibly more mechanical strain on lugs, turrets, etc. Also remember another way to get lower R is to shorten the distance between points. In some cases people do "extreme grounding", for example (if memory serves) kg in his BAGA amp had some gigantic piece of wire for the ground. Something apparently so thick he needs to use a blowtorch to solder to it. So it has an advantage (helps to get very low R) but as above some disadvantages (hard to solder, takes up space, mechanically could be a strain). Also, I found out recently that apparently in some very expensive recording consoles, the grounds use lots of copper. Also some of the better amp makers which use PCBs such as THD or CAE/Suhr--the designers might say something like using a heavier copper foil for the PCB is the "correct" way to do PCBs (what would that accomplish? it would help to lower R when using ground planes, right?).