Hi

Have anyone tried a grid choke on the inputtube, instead of the 100k pot or resistor? I have a pair of Lundahl 1670 laying around, and would like to try it. But if someone have tried it and found it not satisfying, I'll skip it (http://www.lundahl.se/pdfs/datash/1670.pdf)


Cheers
Bjorn

Can you describe how a gridchoke works, and how you installed it into your Zen?

Howdy Yoda,

A grid choke solves some of the problems inherit to grid leak resistors. A grid leak resistor helps to set the input impedance of the stage, and gives the grid a path to ground. All tubes have a small DC voltage created on the grid by the flow of the electron emission from cathode to plate passing through it. The grid leak resistor gives this DC voltage a path to "leak" to ground. The problem is that it also bleeds off a wee bit of the AC music signal, and more at some frequencies than others.  

On tube data sheets, you'll often see a "Maximum Input Resistance" spec that lets you know the maximum resistance value that can be in series between the grid and ground. The grid leak and grid stopper resistor values together make up the input resistance.

For power tubes, there'll often be two specs, one for grid (fixed or adjustable) bias, and one for cathode bias. This spec is not to be exceeded, as the path to ground becomes too resistive. The tube can't leak off the grid voltage fast enough, and if it's a grid biased power tube, can go into bias runaway. The grid leak resistor value can be larger with a cathode biased power tube than a grid biased one, since any increased conduction by bias runaway here will be countered by an increase in bias by the cathode resistor, therefore decreasing the conduction.

A grid choke offers a low impedance path to ground for the DC voltage when compared to a sufficient value grid leak resistor. Inductors pass DC, but block AC on a logarithmic curve according to frequency. The grid choke value is high enough to be an infinite impedance all the way down to the lowest AC signal frequencies. It easily passes the unwanted DC voltage to ground, but blocks 100% of the AC signal, as Ludo noted. So, the input signal loss is zero with a grid choke, and any hash, noise, or harmonics present in ground return currents in the ground system cannot get into the tube stage.

Eddie

Thanks Eddie. Nice explanation.

nubz69, you can use the Lundahl 1670 listed in the first post.

bjorn

UPSCALE from the Hammond part Eddie gave is the MagneQuest BCP16.  It comes in 2 "flavors", Permalloy (Ni) and M6 (Silicon steel), for the lamination stack.  The BCP-16NI is $150/pair, while the BCP-16M6 is $99/pair.  With Ni lams the inductance is a WHOPPING 4800 Henries, while the M6 lams yield "only" 1000 Henries of inductance.

IMO, the place to use BCP16s in a SE84 is as replacements for the SV83 grid leak resistors.  All of the benefits Eddie outlined obtain and the net gain of the 1st stage increases.

Also will these chokes fit in the zen select chassie?  Is there a real reason to use the more expensive chokes?  The hammonds are realatively cheap.

OK guys,
this one looks fun and easy, so I'm going to do it. Couple of questions from a layman.

1). In the input stage, is it worth the extra cost for the Lundahl or Magnequest. I'm willing to spend if there'a a sonic improvement over the Hammond 156c recommended by Eddie. But if the Hammond does its thing sonically as well as the others, why waste the money. From what I've read, take out the 100K pot and replace with the grid choke. I would think the 1k grid stopper needs to stay as is. Can I still use the front input, or does the input signal need to develop across the grid choke. Also, can I put the pot back in place before the grid choke to have volume control. If I can, do I place the pot directly before the grid choke with the wiper feeding the positive leg of the choke, with the grid stopper between the choke and the grid, OR, do I place the grid stopper in series with the wiper before the positive leg of the choke. Also, does the value of the pot stay at 100k, or does it need to be changed because of the choke.

2). Eli's post got me thinking that two may be better than one here, and that I should replace the 330k grid leak resistor on the SV83 with a grid choke. If the performance will improve with both stages grid choke loaded, I will do it. Since this is not a 100k but a 330k, can the same chokes be utilised, or do thay need to be different.

Size doesn't matter, as a new chassis will be constructed. The only issue is performance. Any comments or thoughts will of course be greatly appreciated.

Thanks,                   Bill

Hi

I have just built a clone, so it's different from the original design.

It looks like this:





Inside:


The chokemod:




It doesn't look pretty, but it works

Cheers
Bjorn