Rectifier tube and bias

[dscottguitars]

Hi,  another issue with switching in a rectifier tube-5U4GB.  The current in the 6L6's goes down quite a bit.  The bias voltage is about the same, 1-2v less, but the current drops from 40mA to 2mA.  This does not make sense to me. 


solid state rectifier:  490v on plates, -49v on the grids, 40mA of current
tube rectifier: 467v on plates, -48v on the grids, 27mA of current


Any info on tube rectifiers would be great,


Thanks....

[TIMBO]

Hi dsc, Do you have this http://www.webervst.com/tubes1/calcbias.htm, cause the plate voltage has drop you need to adjust the bias accordingly. The bias will always stay the same as it is a separate circuit and NOT affected by the rectifier.

[dscottguitars]

I expect a little change, I've done it before, but this is too much current loss.  I shouldn't have to rebias by throwing the switch.  It's not an uncommon feature in amps.  The voltage drop is less than 10% but the current drop is 30%.

[stratele52]

Hi,  another issue with switching in a rectifier tube-5U4GB.  The current in the 6L6's goes down quite a bit.  The bias voltage is about the same, 1-2v less, but the current drops from 40mA to 2mA.  This does not make sense to me. 


solid state rectifier:  490v on plates, -49v on the grids, 40mA of current
tube rectifier: 467v on plates, -48v on the grids, 27mA of current


Any info on tube rectifiers would be great,


Thanks....

40 ma to 2 ma or 40 ma to 27 ma ?


Something wrong with your 5U4 or with 6L6s circuit. Do you try another tube ? GZ34/5AR4 is better than 5U4.

[dscottguitars]

Didn't see the typo, it's 27mA...


I'll try another tube.



What's better about a GZ34/5AR4?

[eleventeen]

What's better about a 5AR4/GZ34?


Lots less voltage dropped across the tube, lots. 5V4 is another good tube to use for low-rectifier drop, and sometimes cheaper than GZ34s.


http://www.300guitars.com/articles/rectifier-tube-voltage-drop-chart/


Ironically, a 5U4GB (if you believe the linked chart) drops 15% more volts than a 5U4 *GA* (Who'd a thunk it?)


I'm not sure I believe that. You seem to be losing a mere 23 volts across your 5U4 [GB, I assume]


I am also not understanding why the "mild" changes you have going on are reducing the 6L6 current so much. But, I suspect it is because you are switching on and off the filament of your 5U4. More specifically, because you are switching only one side of the 5U4 filament and leaving the connection across the (cold) 5U4 filament from the (now SS powered) B+, thru the filament, and thru the 5v winding. Yeah, maybe it sounds goofy.


I would try these:


1: Switch to solid state mode and pull the 5U4 out of its socket. See if you get any noticeable changes in voltages. Easy.


2: Temporarily, break "the connection that never goes away", the connection where the filament of the 5U4 goes to the top of the filter caps by jury-rigging an SPST switch in there. Measure. <<--This is just for temp use. Use a heavy-duty switch like you might use for switching line volts, not a cheesy little 1/4" bushing pickup-selector-sized thing. Just hang the switch up in the air and put some tape around it or safety it in some way.


3: Goofiest suggestion of all: Reverse the phase of the 5v connection to the 5U4.


OK, now you can let the laughing die down.
 

[HotBluePlates]

Screen voltage will matter more than changed plate voltage. If screen voltage drops a lot then plate current will also drop.

I expect a little change, I've done it before, but this is too much current loss.  I shouldn't have to rebias by throwing the switch.

Then don't rebias. If the amp is safe on the solid-state setting, then it will be safe at the lower-voltage/lower-current setting. The amp will play fine even if you have lowish plate current.

[dscottguitars]

Hot blue:  that works out to about 12.6 watts of dissipation.  That's not too low?  It sounds fine to my ears, but I'd like a more experienced opinion...

[stratele52]

Hot blue:  that works out to about 12.6 watts of dissipation.  That's not too low?  It sounds fine to my ears, but I'd like a more experienced opinion...

Try some ear test until 70% plate dissipation . Leave bias the colder you can if sound is not better for a longer tube life .

[HotBluePlates]

Hot blue:  that works out to about 12.6 watts of dissipation.  That's not too low?  It sounds fine to my ears, but I'd like a more experienced opinion...

If it sounds fine, then it is fine. Nothing's burning up, so it's safe.


I've got amps I haven't biased in years (even with tube swaps). I just turn them on and play. As long as I don't see redplating, and it doesn't sad bad, then everything is good to go.

[dscottguitars]

Good to know... thanks all.


I took out the tube and the voltage didn't change.  I didn't try those other things though.


The screen voltage is the same as the plate voltage.


It's better in the low power mode too, which is strange.


Here are the voltages and such:


High power:
solid state: 493v on plates, 41.5mA current
tube: 460v on plates (6.7% drop), 27mA current (35% drop)


Low power:
solid state: 319v on plates, 53.3mA current
tube: 303v on plates (5%drop), 44.2mA current (17% drop)


Clearly the high power mode is not providing as much current to the 6L6's.  It's definitely running on the cool side but I don't think I'm in cutoff range. 


Anybody know approximately how low of current will be too low?


Also, could the bias circuit resistors I have chosen to get the voltage be changed to improve the current drop?  (not even sure if it plays any part, just wondering because I use a different load resistor at the end on the low power mode)

[HotBluePlates]

High power:
solid state: 493v on plates, 41.5mA current
tube: 460v on plates (6.7% drop), 27mA current (35% drop)

Low power:
solid state: 319v on plates, 53.3mA current
tube: 303v on plates (5%drop), 44.2mA current (17% drop)

Clearly the high power mode is not providing as much current to the 6L6's. ...

Also, could the bias circuit resistors I have chosen to get the voltage be changed to improve the current drop?

It seems the issue is just that the roughly-same negative bias voltage cuts offr the tube more when the screen voltage drops. Meaning, when the bias stays basically-constant and the screen voltage drops a given percentage, the plate current drops a greater percentage.


That is the normal function of the control grid. Be aware that I asked about screen voltage because there are 2 voltages responsible for setting your output tube plate current: positive screen voltage and negative control grid voltage. The plate voltage is largely immaterial with respect to plate current unless it gets down to ~40-50vdc.


The cure for this is a bias voltage which gets smaller when the plate voltage gets smaller; that would cause plate current to stay a similar amount. There's 2 ways to get that: rig something within the fixed-bias supply which changes (the bias voltage) when you drop plate voltage, or switch to cathode bias. For any sensible cathode bias resistor value, plate dissipation will stay largely the same when screen voltage drops because less plate current through the resistor creates a smaller voltage drop, which equates to a smaller bias voltage, which increases plate current.


But if you're using high-voltage and low load impedance for big class AB output power, cathode bias would be incompatible; the bias voltage across the cathode resistor will tend to collapse when the output stage is pushed to anything other than a brief cut-off (meaning, when it is running class AB and making the full output power).