another trem problem

[floyd]

Looking at this cathode bias trem circuit , can anyone see why the intensity is not activated until the pot is turned to about 8 ? After 8 , the trem is fine... it's almost like an on/off switch.

[sluckey]

Use a linear pot like the original PR did. You could probably just use the original 250K-L pot.

[sluckey]

Any luck?

[tubeswell]

I've said it before and I'll say it again. Bias-vary trem doesn't work as well on cathode-biased output stages, because the trem-wiggle is having to 'fight' the auto-bias characteristic of the cathode-biased output stage. (Cathode biasing 'auto-corrects' any changes in bias because if the tube current increases, the voltage drop across the cathode resistor also increases, which in-turn increases the grid-to-cathode voltage and lowers the tube current). So you need to hit the output tubes with even bigger changes in current from the LFO to get the trem effect.

[floyd]

Any luck?

          I'll probably have to order some linear taper pots.

[floyd]

I've said it before and I'll say it again. Bias-vary trem doesn't work as well on cathode-biased output stages, because the trem-wiggle is having to 'fight' the auto-bias characteristic of the cathode-biased output stage. (Cathode biasing 'auto-corrects' any changes in bias because if the tube current increases, the voltage drop across the cathode resistor also increases, which in-turn increases the grid-to-cathode voltage and lowers the tube current). So you need to hit the output tubes with even bigger changes in current from the LFO to get the trem effect.

   I'm not having a problem with enough intensity.. it's huge after 8 on the pot., ( I'm using an LED instead of cap/resistor , and a diode on the intensity pot ).

[tubeswell]

Yeah, but the problem is you have to get to the point where the current pulses from the LFO are strong enough to disrupt the auto-bias characteristic of the cathode-biased output tubes (which is about '8' on your depth pot). Here's an experiment - try setting the depth pot at some lower rotation like '5' or '6' and then switch off the main power (not the standby) of the amp, and see, as the amp's power dies off (but before the heaters cool right down), the trem temporarily kick-in as the output tube current decreases. (Caveat - I don't know how the diode on the depth pot is going to mess with this)


The only way I can see of changing the point of effectiveness in the depth pot rotation is to 'pre-load' the depth control with a tail resistor between the depth pot's ground lug and the ground return. Either that, or get a lower output impedance from the LFO stage with a buffer stage, like a cathode follower (or a source follower using a high-voltage mosfet like an IRF820) so that you have more current to drive the bias wiggle with.

[Willabe]

Yeah, but the problem is you have to get to the point where the current pulses from the LFO are strong enough to disrupt the auto-bias characteristic of the cathode-biased output tubes (which is about '8' on your depth pot).

Yep.


            Brad       

[sluckey]

I agree that the cathode biased 6V6 will be harder to wiggle due to the nature of the cathode bias operation. But, a 6V6 does not need extra current from the LFO to overcome the bias. It needs more voltage from the LFO. There is no grid current involved with a properly operating 6V6.

The cathode follower that you see on the output of many trem LFOs is not about providing more drive current to the output tube grid. It's about providing a buffer between the LFO high z output and the output tube lower z input. I know that the output tube grid z is fairly high, but it is low enough to cause the LFO to drop out (quit oscillating) at lower or higher frequencies. That LFO is much happier working into a 10M load than a 250K load. The hi z input and low z output of the cathode follower provides very good isolation between the LFO and the grid load.

There are several amps out there that don't use a buffer between the LFO and the cathode biased output stage. And the trem effect is more than adequate. I recently built one and it works very well. It works even better with the LED. The LFO output level v. freq. is more constant and that allows operating between 2 and 10Hz with no drop outs.

I'm fairly confident that a linear pot is all Floyd needs to get what he wants.

[tubeswell]

Yes its the changes in grid voltage that change the bias, and therefore its ultimately grid voltage that is important. However what I was openly pondering was not so much the voltage attenuation where a relatively high Z out is followed a relatively low Z in,  but being able to source more current to drive the voltage swings. So while I was ultimately thinking about voltage swings, it was from the point of view of current pulses. :-) 

[HotBluePlates]

... high Z out is followed a relatively low Z in,  but being able to source more current to drive the voltage swings. ...

A grid that is not driven to or beyond an instantaneous 0v (i.e., not drawing grid current) could be considered a near infinite input impedance. So the only thing that lowers it is the grid reference resistor (which are generally limited to about 220k or so because that is what tube manufacturers specify). However, that does not apply in the case of the tremolo or bias circuits, because they are not impedances to ground, but like wires to an open circuit (the grid).

I'm thinking the issue may be a result of bias-shift caused by the rectification effect of the output stage. This is where the d.c. voltage (bias) changes when a signal is applied, and because it's a d.c. change with an applied signal it is described as rectification.

Regardless, step back from your theory for a second. If it was true that the cathode bias is opposing the trem signal, how is the guitar signal causing an output to occur?

Now, maybe we could say the cathode bypass cap (if used) is not big enough to be an effective bypass down to a couple-Hz where the trem signal is, and that the local negative feedback of the cathode resistor is coming into play. If that is the cause of the apparent reduced trem strength (assuming it's happening), couldn't you look at other bypass caps or other methods to bias the tube that have no feedback down below 1Hz?

By "other methods" I'm thinking either resorting to solid-state to define the bias voltage, or the old-school d.c. heating of some preamp tubes by placing their heaters in series between the output tube cathode and ground. That second method is tricky, because you have to need something like 25.2v bias while the tubes idle at 150mA (for 2x 12A_7 tubes with heaters wired in series and using pins 4, 5 only).

[PRR]

> If it was true that the cathode bias is opposing the trem signal, how is the guitar signal causing an output to occur?

Signal is push-pull. Bias and trem are common-mode.

Also, at most the drop in Rk is perhaps half the grid swing. So just hit it twice as hard.

Cap value must matter. Someone otta try zero, 10u, 50u, and 2,000u.

> wires to an open circuit (the grid).

And the coupling caps. If they are not over-size, the loss is small, but not negligible.

[HotBluePlates]

> If it was true that the cathode bias is opposing the trem signal, how is the guitar signal causing an output to occur?

Signal is push-pull. Bias and trem are common-mode.

Doh!! That's a good point which I overlooked!

[sluckey]

I know that the output tube grid z is fairly high, but it is low enough to cause the LFO to drop out (quit oscillating) at lower or higher frequencies. That LFO is much happier working into a 10M load than a 250K load. The hi z input and low z output of the cathode follower provides very good isolation between the LFO and the grid load.

It's not really correct to say that the 'tube grid' impedance is low enough to affect the LFO. The output tube grid is hi z. As HBP says, it's the grid return resistors that set the load on the LFO. The Intensity pot becomes the real load on the LFO. Very often you will see 220K for grid returns on a 6V6. Then put a 250K Intensity pot under that and you're up to 470K, depending on the setting of the pot. Some amps, i.e., old Ampegs used a 1M intensity pot. So, by using a larger INT pot, you can lighten the load on the LFO. But at the same time you're increasing the grid return for the output tubes. You can only go so far, so the value of the INT pot becomes a compromise between satisfying the grid return parameters of the output tube and the load on the LFO. And that's where the cathode follower comes to the rescue. The CF allows the use of a reasonable grid return resistance for the output tubes while also providing a very hi z (light) load to the LFO. Now everything is happy. At least that's what I think the CF is all about. 

[floyd]

A 250K linear pot is all that's standing in our way.. then I'll report the result

[floyd]

A 1meg linear pot DID solve the problem mentioned in my first post ! Now.. I wonder how I can solve the slow trem start-up when footswitching in/out... ground the trem in a different place in the circuit ? BTW... in this circuit , even with a red LED , there is almost zero depth.. but WITH a diode on the intensity pot , the depth is huge.

[floyd]

I solved the slow footswitch trem start-up by switching the LED ground in/out.

[sluckey]

Was the LED flashing before you reversed the leads? Is it flashing now?

The LED anode lead is typically longer than the cathode lead. And the LED anode must connect to the tube cathode to work properly.

[Tom_Hull]

hi
what do the guys think about this??
a 470 k resistor might help .in place of the red wire line .

[sluckey]

I wonder how I can solve the slow trem start-up when footswitching in/out...

One way is with that 1M resistor that you changed on the PR schematic. I'm talking about the red 1M in the pic above. You have it connected to ground now. If you put it back where it was originally (cathode of LFO), you'll have a faster startup. Doing this will not take away any "Vox mojo".

Fender had that resistor grounded in earlier models but soon switched to connecting to the cathode specifically to speed up startup. Then, with the AB763 circuit, they made it even faster and bulletproof as far as startup is concerned. They modified the LFO grid circuit slightly, connecting to the bias supply for very fast, reliable startup. Look at the Deluxe Reverb AB763 to see the simple mod. This mod would be very easy on a PR.

[sluckey]

what do the guys think about this??
a 470 k resistor might help .in place of the red wire line .

Doing that would decrease the trem signal some, but since he's using a 1M INT pot and a LED, he should still have plenty of trem signal. But, with a 470K in series with the INT pot, another possibility is available for instant on trem. Simply move the footswitch to the top of the INT pot. This will kill the trem signal to the INT pot, but it allows the LFO to run continuously. Never wait on the LFO to start up again because it's already up and running. This is the approach I took with my Warbler and Revibe.

[floyd]

Was the LED flashing before you reversed the leads? Is it flashing now?

The LED anode lead is typically longer than the cathode lead. And the LED anode must connect to the tube cathode to work properly.

The LED WAS flashing when footswtched in , but even before , when foot switched out, the LED was on .. just not flashing. I like the new footswitch arrangement better now because the LED only comes on when I activate the trem with the footswitch.

[sluckey]

I solved the slow footswitch trem start-up by switching the LED ground in/out.

Man was I in another world when I first read that!  I get it now. The footswitch is switching the LED, Breaking the tube cathode circuit. I thought you swapped the leads of the LED. Sorry for my confusion.