Grid bias w/tremolo?

[yorgle]

Greetings amp folk.  I'm in the process of building a tremolo circuit for my junk-box, octal SE franken-amp built mostly from AA5 radio parts.  It uses a 12sj7 and 12sq7 as cascaded pre-amps feeding into a 35l6 output with a volume and tone control.  The tubes (along with the heaters) are powered by a 117v isolation transformer and 35z5 rectifier.  The amp works well and sounds fantastic with my Epi Wildkat, but I'm longing for some warble so I've started tinkering with tremolo ideas.  I've got 2 extra octal sockets on the chassis and an extra 12sq7 which should suffice for a LFO.  My question is whether a vibrochamp-style trem circuit will work with my grid-biased pre-amp setup.  I think I should be able to inject the trem signal to the grid of the output tube, but I'd prefer to have the signal warbled at the preamp stage.  If I inject the trem at the grid of the 2nd stage preamp, will it matter to the trem that the cathode connects straight to ground? 

[yorgle]

Ok... I've done a LOT of reading through these most excellent boards and based on what I think I've learned, I'll take a stab at an answer to my own question.  I predict that the wobbulation from a 12sq7 (even if I manage to get it to oscillate at such low b+) are not going to do much to vary the current flowing through a preamp cathode that doesn't have at least some resistance to ground.  Am I right?

[shooter]

not going to do much to vary the current flowing through a preamp cathode that doesn't have at least some resistance to ground

If the cathode is at ground the grid controls current, so that needs to be your wiggle point.  How that's done, find schematics that do it and steal 

work your way through this thread, might gleam something that helps

http://el34world.com/Forum/index.php?topic=25086.0

[yorgle]

Thanks for the tips, Shooter.  That makes sense and confirms that some of this theory stuff is starting to sink in.  Now I just need to get an 12sq7 to oscillate with only 130VDC available from from B+.  I've tried biasing the tube with a red LED and a zillion different plate resistors from 100R to 470K but so far no success-- with the 470K plate resistor, the LED just barely glows with no hint of blinking.    I've read that a 12ax7 can oscillate with as little as 100V (not sure that's B+ or measured at the plate), but if that's true, then I suppose I could swap one of the octal sockets for a 7-pin (although, I'd like to keep this thing all-octal for irrational, fetish reasons, of course). 

[sluckey]

I'm confused. In another thread you said you built an optocoupler and the sound was "smooth, rich wobblulations".

[shooter]

swap one of the octal sockets for a 7-pin

I'd try  9pins for an AX7 no extra soldering on the 2 pins that don't get holes 

[yorgle]

I'm confused. In another thread you said you built an optocoupler and the sound was "smooth, rich wobblulations".

That was a different amp- made from an old Magnavox phono amp.

[yorgle]

swap one of the octal sockets for a 7-pin

I'd try  9pins for an AX7 no extra soldering on the 2 pins that don't get holes 

Yeah, that certainly seems like a better idea. 

[yorgle]

Ok, I give up.  Can anyone recall a widow-maker amp that had tremolo?  Maybe I can find a schematic and see how they got these tubes to oscillate without step-up transformers?  Or was there no such beast?

[sluckey]

Look at the cheap stuff in the schematic library...

https://el34world.com/charts/Schematics/files/Kay/Kay500.pdf

https://el34world.com/charts/Schematics/files/Kent/Kent5999.pdf

Probably more but this should get you pointed toward a solution.

[jjasilli]

If I inject the trem at the grid of the 2nd stage preamp, will it matter to the trem that the cathode connects straight to ground?
This question is not clearly thought-out.  You need to post a schematic.

Preamp trem usually inputs to a cathode.  If there is no cathode R to lift the trem voltage from ground, then the trem voltage is defeated.  If the cathode R is bypassed by a cap, then the VibroChamp circuit must be used, with a split cathode resistance, or the oscillating trem voltage (AC) will bleed to ground through the bypass cap.

Alternatively, you could wobble the grid of the power tube.  The difference in trem output voltage is significant: a typical preamp cathode has less than 2V.  A power tube may have 10 or more volts on G1 to be wobbled.  Too much V swing on a power tube grid may deprive it of adequate bias, causing it to redplate.

[sluckey]

I think the main issue is getting the oscillator to work with low B+. Where to inject the trem signal is kinda a moot point until the oscillator is working.

[jjasilli]

Oh!  Then maybe use a small 9V or whatever PT to power a SS trem circuit, like for a trem pedal, inside the amp.  Or a small tranny to step up the voltage to power a reverb tube: 117V doubled = 234VAC rectified > 331VDC B+ supply voltage to a tube oscillator.


Don't know how you're powering filaments???

[kagliostro]

Here you can see some schematic of SE amp with tremolo

http://el34world.com/Forum/index.php?topic=20727.msg218834#msg218834

Franco

[jjasilli]

Actually the Kay 500 amp looks promising.

[yorgle]

Thanks for all of the help, folks.  Looking at the examples you guys suggested I see these amps use 12av6 tubes for oscillation.  So I went back and compared the datasheet for my 12sq7 to the 12av6 and if I read correctly, it looks like I would need at least 180 volts for my 12sq7 to get past the magic gain of 29, whereas, the 12av6 can hit 35 with only 90 volts.  I've got lots of stuff in my junk drawer, including a few 12ax7's, so I started building a socket adapter from a dead octal tube and a 9 pin socket.  Once that's done, I can try plugging in a 12ax7 and see if can get that to wobbulate.  If it does, then I'll at least know my LFO circuit is correct and I'll have to decide between my commitment to keeping this thing all-octal or swapping in a bigger power transformer.

[yorgle]

Here's a schematic of my monstrosity thus far. Please disregard the lack of plate resistor on the trem tube and power switch/fuse- I see I forgot to draw those 

[sluckey]

That schematic has several deal breaker flaws in it. The amp will not work properly if at all if it is wired IAW that schematic. I'll start with the trem oscillator. I've never seen a trem oscillator with an intensity control in the grid. It's actually in the feedback loop for the oscillator. That poor tube is trying to build a signal out of nothing but you are limiting the feedback signal with that 220K and pot. Self defeating. Dump that circuit and just copy the circuit in your other amp that has a working oscillator. Or copy the Kay circuit. Or any other working oscillator.

Now for the amplifier chain... You'r input stage is being operated as a grid leak bias amplifier. This would benefit from a coupling cap between the input and the grid to prevent the negative bias on the grid from leaking off through the low resistance guitar circuit. But wait. You show a triode but the 12SJ7 is a pentode! You must have additional circuitry for the two extra grids.

The second triode has no grid return resistor. The tube cannot operate properly without one.

The volume control cannot function properly as a volume control. As shown, it's just a variable grid stopper. That third leg needs to connect to ground.

[yorgle]

I appreciate all the input, folks.  I feel bad to have wasted time by posting that schematic.  After I compared it more closely to my amp I see that (as sluckey generously pointed out) it's got a few significant errors- eg, the volume pot actually is grounded, I do have a coupling cap at the input to the pentode, and the 12sj7 is wired as a pentode, so the amp does actually work.  I will take the advice though and set this project next to my other amp (which has a working LFO) so I can visually compare the two trem circuits.  I will also add the grid return resistor.   

[yorgle]

Got it working!  After rewiring the LFO to match sluckey’s tremor-later (which is what I used in my other amp) it still didn’t work.  So I swapped in a 275volt power transformer from another amp which more than doubled the b+. Immediately, the 12sq7 started to oscillate!  Plate voltage is now 165vdc.  With the isolation transformer, the highest I could get was about 140, so obviously, the threshold for the 12sq7 using a red LED for cathode bias is somewhere in that range. 

[PRR]

> the highest I could get was about 140, so obviously, the threshold for the 12sq7 using a red LED for cathode bias is somewhere in that range.

Aside from the drawing/build "errors" (no plate load, unwanted grid pot), an old-fashion bias 12SQ7/12AX7 sure will oscillate with less than 100V.

I think the issue is the LED. Say you have 140V of supply, and maybe 70V at plate. The amplification factor of this/these tubes is 100. 70V/100 makes about 0.7V of bias needed to turn-off, something less to conduct good. However the LED is nearly fixed at 1.6V. So the tube is hard-OFF.

The cathode bias MUST be related to the supply and tube voltage. For B+ of 250V to 350V, the 1.6V LED bias does work good. But as you proved, by 140V it is really marginal. (Sluckey doesn't fool much with low-Volt amps, he worked with kiloVolt bottles for a job.)

At 140V or so a single Silicon diode (1N4007, 1N4148) may do well (but lacks "glow"). 1K with a fat electrolytic cap also works (until the cap rots). In the $9.95 amps they just grounded the cathode and let the tube run wide-open (aint likely to die at 140V and good plate resistor). That's not the cleanest wave-form, but hey it wobbles.

[yorgle]

Thanks for the input, PRR.  I see what you mean about the LED being the limiting factor, but rather than mess any further trying to make low voltage wobbling, I'm going to keep the LED trem and look for a step up transformer that will give me 275- 285 b+ for the oscillator.  I spent some time last night tweaking things and discovered that connecting the ground side of the photo-resistor to the side of the depth pot that has the 100K resistor to ground (instead of directly to ground) nicely softens or rounds off the wobs (singular of wobbles?) for a very pleasant effect. 

[yorgle]

Here is the revised schematic which, God willing, should reflect what I've actually constructed thus far.  What do you guys think?  Comments, criticisms, and advice are always welcome.  Right now it's running off one side of a 275-0-275 PT borrowed from another amp so I had to temporarily run the heater string directly from the mains.  The heaters don't seem to mind the extra few volts from the wall, but depending on what I find for a step-up transformer, if I have to stick with the mains for the heaters I may add a limiting resistor to cool things a bit.  Obviously, I need to add a fuse as well.  The lack of an intensity pot for the tremolo was intentional- I don't have a 50K pot handy and I'm happy with the level of wobble I have so I may just leave it fixed as is.  Also, the resistor off the grid of the second preamp tube-- do I need one there?  It sounds and works great without it.

[yorgle]

This weekend I added a foot switch for the tremolo which allowed me to compare the amp’s overall sound both with and without the effect.   I was surprised by just how much volume and mid frequency the tremolo eats.  Inserting a 20k resistor into the trem signal line seems to help a bit, so maybe I need to rethink my omission of an intensity pot, which I presume would allow me to dial in a happy balance between wobbles and tone. 

[yorgle]

I broke down and ordered a 50K intensity pot (with on/off switch) and changed the trem injection point from the grid of the 1st preamp tube to a lug on the volume pot.  In the process, I was able to shorten some of the signal wiring and the changes seem to have greatly reduced the tremolo "tone suck."

Now that the trem is working, I want to figure out more productive use for the remaining "extra" tube, which at the moment is simply serving as a glorified resistor for the heater string.  A single tube reverb seems a logical progression for this amp project, but so far, I've not found any octal one-tube reverb circuits. I'm thinking along the lines of a 12sl7 dual triode- which seems like it could possibly fill the place of the 12ax7 typically seen in such circuits.