Any issues with using a grid choke on a direct coupled amp?

[Jim R.]

Subject says it all -- does the absence of the coupling capacicitor/resistor have any effect on a grid choke in a direct coupled amp?

Thanks,

Jim

[Paul Joppa]

Well, normally a grid choke is used at the output tube grid, so it would ground the driver plate at DC. ...  :^)

You could use one at the input, of course. Because of the driver gain it is much more sensitive to hum pickup, so you'd need one with at least one mumetal shield unless you're awfully careful - and lucky.

[Grainger49]

Grid chokes have confused me.  I know they take the place of the load resistor at a tube's grid.  But what are the specifications that are best for a grid choke?

[Jim R.]

Paul,

Exactly what I was concerned about -- loading the grid of the 2a3 in the paramount.  Guess I could either rc couple it or save the grid chokes for the stereomour :-).

-- Jim

[Paul Joppa]

Jim - ?? I'm not following you here. The Paramount 2A3 is direct coupled; the 2A3 has no grid-to-ground resistor, and no interstage capacitor either. There is no need to load the grid of the 2A3, it is nailed in place by the driver plate voltage. The Stereomour does have RC coupling, and converting to LC coupling may well have benefits (as long as the grid choke doesn't pick up excessive magnetic-field hum).

Grainger -  the best grid choke has a high impedance at audio frequencies, and a low resistance at DC. The audio impedance should be at least 250K (the typical grid resistor) to keep distortion of the driver down, while the low DC resistance keeps the output tube's grid bias stable. By using an LC filter (coupling cap to grid choke) you can in theory get faster recovery from grid-current overload, compared to an RC coupling. This is IMHO why they sound good; I don't of course have enough data to "prove" that theory!

[Jim R.]

Hi Paul,

Ok, I'm essentially revealing a fairly large hole in my understanding of grid chokes, but my assumption was that they were used in place of te grid tie resistor, not the grid stopper, and I also had no idea that there was no grid tie on the power tube of the direct coupled 2a3 as I can't read the schematic.  Of course that's my issue, not yours.  I'm also sorely lacking in the concepts of direct coupling, but that too is a matter of dinding some documentation in a format I can use.  I tried all the normal sources for all the online tube manuals, the navy manual, etc. but they are all scanned images of the actual pages of these books, which don't do me any good, and sites like tubecad also rely heavily on graphics (but of course.)

Anyway, in an amp like the paramour or or 300b paramount, how is the grid choke connected -- at least I can clear that up, and work on the why later when I can get a better handle on the circuit as a whole.

As for cap coupllig the 2a3 paramount, I suppose I was assuming that I could simply replace the direct connection with an appropriately spec'ed rc coupling circuit (or is it RL, again my original assumption is that the grid tie was where the grid choke went) and then install the grid choke.    Failing that, I'd just use the grid chokes I have on order in my stereomour build.

Hope that clears up my assumptions and what I was trying to say.

-- Jim

[Grainger49]

Paul,

That makes perfect sense to me.  Thanks for the explanation.

[Paul Joppa]

Hi Jim-

Thanks for pointing that out, I hadn't really thought about where one might find a useful non-pictorial description of direct coupling. When I did, I couldn't come up with anything. So, here's an attempt at least:

(To keep the following simple, I am ignoring the grid stopper, or perhaps you can say I'm regarding it as part of the grid - anyhow when I refer to the grid connection I really mean the other end of the grid stopper.)

In an RC coupled 2-stage amplifier, the plate of the driver goes through a capacitor to the grid of the output tube, which is also connected to ground by the grid resistor (which is I think what you have called the "grid tie"). The grid resistor puts the output tube grid at ground potential for DC, since there is no DC current through the grid or the coupling capacitor. The plate current of the output tube also flows through its cathode resistor, positioning the cathode at a positive voltage (60v in the Stereomour) so the grid is 60v lower than the cathode.

Using a grid choke, it replaces the grid resistor (grid to ground). There is still a coupling capacitor to prevent the driver plate voltage from driving any DC current through the power tube's grid or the choke. Its low DC resistance keeps the output tube grid at ground for DC, while its high AC impedance prevents it from loading down the driver stage. The same cathode resistor is used to obtain a negative grid bias relative to the cathode.

In direct coupling, the plate of the driver is connected to the grid of the output tube directly, without any capacitor. There are no other connections to the output tube grid. If the driver plate is at +150v DC, then so is the output tube grid. In order to get a negative grid bias, the output tube's cathode resistor is much larger; to get 60v of bias the output tube cathode must be at +210v. For example, in Paramount the cathode resistor for RC coupling with 300Bs is 1000 ohms; for direct coupling with 2A3s it is 4000 ohms. (We do this with two 2000 ohm resistors, in parallel for 300B and in series for 2A3). The power supply voltage must be quite a bit higher when using direct coupling.

I hope that's at least a starting point.

[Jim R.]

Paul,

Again, wow, your descriptions and interwoven theory are right on the money and extremely helpful.  A couple more small, but significant light bulbs went on as a result of this description, and yes, I've been calling the grid resistor the grid-tie resistor as that is the terminology I learned, though it seems plenty simple enough to just drop the "tie" and be done with it :-).

This is great!  Thanks so much!

-- Jim

[saildoctor]

Slightly off topic but - I've got v1.0 Paramounts with the soft start upgrade.  I'm assuming it's the same as the current production.  I've set it up as direct coupled but left the 300k resistor on the driver board.  In the circuit this goes from the top of the driver plate/power tube grid to ground right?  While it doesn't seem to hurt anything being there (no longer any cap to make it an RC coupling) - is there any reason why it couldn't just be removed?

[Paul Joppa]

Yes, it is absolutely necessary in the direct coupled configuration. Not so much with cap coupling, but I'd keep it there anyhow.

The resistor keeps the output tube grid at ground before the driver heats up - the soft start can have some tiny leakage current, so this gives that somewhere safe to go. The main point is to prevent "pops" from the power tube; we determined this to be necessary by listening as we swapped various tubes. These "pops", in the worst case, can produce internal arcing in some power tubes, which can cause serious damage. The only two tubes that have had this problem in our experience happen to be among the best and most expensive tubes available, so I consider this component critically important. I do not claim to have tried all available tubes, so there may be others!

[saildoctor]

Good to know - thanks!  Makes sense.  Someone who shall remain nameless did manage to get my Philco 45's to arc once - flipped one of the power switches off and on with a cd playing. Seemed to still work fine though so I still use them.  I do have a new pair of Tung Sol's I'm going to swap in once I work in the pair of Orca speakers I'm getting.  Jim - you've gotten me thinking about my next upgrade down the line!  Replacing the driver grid resistor with a grid choke that is.

[Jim R.]

Kerry,

I was only planning on using the grid choke on the output tube -- just a bit afraid that any noise pickup issue with a grid choke on the driver tube will be exaggerated by the much higher gain of the driver tube.  Of course I haven't tried it, so I could be way off here.

I'm going to put the high nickel grid chokes on order in the SR-45s when I rebuild them and then take the existing gcni ones that are in the SR-45 currently and see if I can find a good home for them in the stereomour.

Those NOS american 45s are tough -- except for the baloon types, so I'm not surprised they took a lickin' and kept on tickin'.  When my dad was in college they used to make vacuum sensors out of them.

-- Jim