Further Grid Leak Bias Questions

[rp]

Back to messing around with the 5C8 - this time the grid leak channel (my build has both grid leak and cathode bias and with a 6SL7 replacing the 6SC7). Yes, elevated @60V, 6SL7 better that 6SC7, everything is good there. I noticed when turned up with nothing plugged in the hum is loud, even with one triode used and the other unused. It's exactly like having an open ungrounded grid type hum. I was wondering if the grounding switch on the jacks worked the same with grid leak bias given that there's the input cap before the switch. I found this post concerning exactly this: http://music-electronics-forum.com/t30471/ -> #10

A typical input scheme on old Supros is a 47K to 100K resistor right off the jack hot, into a cap that varied from .01 to .05. [As a side note, there really is no need to go larger than .02 here, and doing so allows for hotter signals to "swamp" the input grid just like an oversized coupling cap, leading to what sounds like blocking distortion.] The grid leak, usually 6.8M, follows the cap, and then there was usually a 2.2K soldered from that junction right to the tube pin. To allow the shorting jack to work effectively, I've found it best to run a short wire from the ground 'switch' on the jack to the junction of the cap, grid leak and 2.2K resistor. this quiets it down a lot when nothing is plugged in because it allows the shorting jack to effectively short to ground, or within 2.2K of ground, as the jack is intended. EFK 08-20-2012, 11:25 PM #10

I want to solve this last little bug-a-boo in this amp and want to try this but it doesn't seen right. If I ground the junction he mentions I'd be grounding the grid leak bias, the tube would be in runaway when nothing is plugged into either side. In fact any way I try to ground the grid on a grid leak amp I'm be losing the bias, no? What am I missing, will the way it's described above actually work? What is the point of a value 2.2k as a stopper? Just to have the lowest value to ground and still act as a stopper? IOW why not 10K or 33K, or 1K for that matter?

If not anyone got a proper work-around to shorting the grid leak bias using just the jacks when unplugged? I guess I can not worry about it, Fender never did, but it'd be nice to get this amp buttoned up 100%.

Also what is the difference to varying the grid leak resistors say from 4.7 to 5.6, 6.8 10M? Or it doesn't matter here as long as a charge is built up.

[martin manning]

Here's the effect of varying grid leak resistor value on bias for a 12AX7: http://ampgarage.com/forum/viewtopic.ph ... 331#329331 , which addresses your last question. Varying the grid leak from 1M to 10M changed the grid bias by only ~0.2V. I could run that experiment again with a 6SL7 to get a definitive answer, but it'll be a couple of days before I can get to it.

Re grounding the grid, if you draw a 100k load line on the plot linked above, from (300V, 0mA) to (0V, 3mA) you can see that there is no way the plate dissipation (red line) will be exceeded at Vg = 0, so grounding the grid shouldn't harm the tube. 6SL7 is much like a 12AX7, and in your case the current will be more limited since you have a 250k plate load shared by two triodes. You could measure the voltage drop across the 250k to get current. I'd measure it as is, then with with one grid shorted to ground, then the other, then both.

Be aware that there might get a bit of added noise from the bit of wire added from the grid to the input jack's shorting contact, which becomes an antenna when the contact is opened.

[rp]

I could run that experiment again with a 6SL7 to get a definitive answer, but it'll be a couple of days before I can get to it.

Buongiorno Sig. Martino, sorry I can't offer you a capucciono. Sure if you wish run the math. I kind of intuited that it couldn't matter much. If I got it right the bias is from electrons hitting the grid and building up charge. I can't really see how the grid resistor could have a very precise scalable effect here.

you can see that there is no way the plate dissipation (red line) will be exceeded at Vg = 0, so grounding the grid shouldn't harm the tube.

I trust you but it just seems weird that you can leave a tube, even a preamp tube unbiased. I will check the current as I want see the difference.

Be aware that there might get a bit of added noise from the bit of wire added from the grid to the input jack's shorting contact, which becomes an antenna when the contact is opened.

Ahead of you there, I hope. It should be just an inch or two of straight wire.

About the 2.2k the quoted poster mentioned. Why is that needed, couldn't I just put the switch right to the pin?[/quote]

[martin manning]

...About the 2.2k the quoted poster mentioned. Why is that needed, couldn't I just put the switch right to the pin?

It won't affect audible frequencies, but it could prevent (ultrasonic) oscillation.

[trobbins]

I think varying the grid leak resistance, along with the input coupling cap, will vary the attack and decay times of the built up bias voltage as signal is passed. The grid current characteristics of a tube can vary markedly with samples, especially older tubes, so it would be luck if copied circuit values meant you were hearing what the original designer was hearing.

Which I think is why many give up on grid-leak bias - as it not very repeatable. You pretty much have to measure anode voltage to know where you stand with no signal, and then have some kind of benchmark anode DC level with a benchmark input signal level, so you can recreate a particular setup.

Good layout of the tube base and input jack, to keep them close, but far away and shielded from other noise may help (although that pretty much stands for any amp).

[rp]

Well the M-E forum dude was right, cool thanks, it works as promised, finally hum free. BUT it ain’t that easy, never is. Current measured across 220K shared plate resistor below (Fluke meter) seems no big deal, but the voltages are way messed. ~65V Vp is actually the Fender layout value, the 6SL7s seem to run a touch lower than the 6SC7. I was surprised it worked well with such low voltage, even more surprised with 39V, it's like tube car audio here Reason I checked the voltages was because using a dummy plug in the unused jack while playing (i.e. both ungrounded) I could hear the volume become noticeably stronger and fuller.

Both Open - 57.5V / 4.0mA
Both Grounded - 24.2V / 5.0m
Open & Closed and visa vera - 39.2V / 4.5mA, 38.6V / 4.5mA

Bummer, looks like the jack grid to ground switching can’t be done w/o splitting the triodes up with separate 100K resistors which would also require running an second cap up to the volume pot. Nuts!

Good layout of the tube base and input jack, to keep them close, but far away and shielded from other noise may help (although that pretty much stands for any amp)

Yes, for sure. Not sure why but the grid leak set up is more microphonic than the cathode biased side, even when I flip the tubes. Also, here's a pic, very little difference in wire length btw V1a and b but one side has more noise.

I learned more with this build than all the other combined. My final lessons is - don't strap the dang anodes together, even if copying a vintage circuit, you'll regret it down the road!

[trobbins]

Just some comments - perhaps of some assistance.

I use a shielded, insulated, switched mono phono socket - similar to this:
https://www.jaycar.com.au/6-5mm-stereo- ... t/p/PS0184

The extra shielding gives me some added confidence.

I can't really make out the ground path in your photo - the 6SL7 cathodes don't appear to be directly connected to your single point local ground for the grid leaks - I guess that is because you take the cathodes back to the filter cap neg leg, which is on that thick bus ground.

It may be worth trying a large'ish cap between grid circuit and input socket switched ground (instead of a link) as a way of AC grounding the grid with no input plug - that shouldn't change the DC grid bias, or affect performance. Even the larger 0.1uF as used by Fender is equivalent to 26k ohm at 60Hz.

There seems to be in and out heater cables - although of no significant issue, its good to wire the input stage heater by itself (not part of a daisy chain).

It looks like the input valve socket is rubber 'sprung' - but the leads from the socket are 'hard' with no flex (eg. the pin 1 grid stopper).

Ciao, Tim

[martin manning]

There seems to be in and out heater cables - although of no significant issue, its good to wire the input stage heater by itself (not part of a daisy chain).

Tim, sounds like you are suggesting running a dedicated pair of leads from the PT filament winding to the input tube's heater. If so, how is that better than just making the input tube the last one in the chain?

[trobbins]

Sorry for ambiguous description - yes it should have been clearer (eg. end of daisy chain).

[rp]

I use a shielded, insulated, switched mono phono socket - similar to this:

No idea those existed, thanks.

I can't really make out the ground path in your photo - the 6SL7 cathodes don't appear to be directly connected to your single point local ground for the grid leaks

I grounded the grid leak resitors to the lug nearest the inputs, it’s were the bus terminates and grounds to the chassis. It’s within 2 inches of the cathode grounds, that should do I figured.

It may be worth trying a large'ish cap between grid circuit and input socket switched ground.

That sounds possible, the switching pops loud too, it should fix that.

One more question, is this how I end the tyranny of the strapped anodes? Just making sure it’s as simple as this:

[martin manning]

Re separating the anode load, as far as the idle point goes you'd need to use two 440k resistors, one for each triode, to get the same current. I think I would experiment a bit. Maybe start with something close to 440k, but try 220k and see how that works.

The input jack Tim linked will be very familiar to the Dumble crowd, as it looks just like the Kobikon part used to switch the FET stage in.

[rp]

Makes sense Martin now that I think it through, though it's damn weird using 440K as plate resistors, hard to wrap my head around this stuff, especially when in I wander off the paint by numbers path. Glad I waited on a reply before doing all the work and trashing two good 1W Allen Bradleys! Now, do I have some nice 440K ABs? Maybe best to just jury-rig it all with some cheap carbon films and see first. Good thing I'm bored - but idle hands do the devil's work. I should learn to leave well enough alone.

[martin manning]

I think it will be cleaner and louder after separating the triodes, and you may not like it, so yes experiment using expendable parts. I haven't forgotten about doing the grid leak bias test for 6SL7... I'll probably have some time tomorrow.

[rp]

I think it will be cleaner and louder after separating the triodes

Sounds good actually, and now I'm anticipating your 6SL7 results. I like my preamps clean and output stage dirty. Though with this amp I'd just like it to sound like a 1953 5C8, but it's veered off. I just wanted it to sound like this:

https://www.youtube.com/watch?v=AL-JKM4Hn-4

[martin manning]

The amp in the YT is def on the clean side; sounds good too. What are the supply and plate voltages on the 6SL7 now?