I set bias on my bread-boarded Bassman (fitted with 6V6s) at 20mA (measured across 1 ohm on the 6V6 cathode-to-ground leg), no input ("quiescent"). But with a 1 kHz signal at 150mV, max volume gives me 80mA (80mV across the 1ohm). With Vplate of 380, this is 30 watts. Isn't this much more than rated for the 6V6? Is tube data an average? The 150 mV is p-p from my scope. Tubes arent red-plating.... I want to simulate normal guitar, which when playing will probably :"average" much less than 150 mV? Thanks!
I'm wondering about the bias point. A 6v6 dissipates 14W max. At 380 plate volts the tube is biased at about 7.7 watts. This sets the bias point at only 55% of max. Isn't the general bias point between 60 and 75%? Is my math correct? 14X.55=7.7 watts. 7.7watts is only 55% of 14 watts. I seems to be biased cold. It seems the bias point would be between around 8.4W(60%) and 10.5W(75%).Or 22mA and 27.6mA. This is a newer area of learning for me. Please tell me if I'm wrong but I did follow the procedure outlined in Jeffrey Falla's book. He even published the tables for the 6V6 for different voltages and current values.
B-tube: I have Falla's book (How to hot-rod). 14 Watts is from the tube data sheet. I guess 20mA at 380V is around 55% (its not on the chart on pg 62), but my concern is that current goes way up when the volume is turned up. I think the "bias point" (percentage you pick) is just whatever "cushion" you feel comfortable with, how driven you want the power stage to be/sound, and how many spare 6V6s you have around....
70% is only a rule of thumb a lot of guys like, but there is no requirement for that. If the amp sounds good at 55%, then run it that way, simple as that.
Idle is set at idle, period. Current readings while amplifying signal are not a part of bias setting.
Think of bias like setting the idle speed on your car engine. Maybe it should be 500-700 rpm or something, but as soon as you step on the throttle, the rpms go up. When you amp is working, it conducts more current.
Idle is set at idle, period. Current readings while amplifying signal are not a part of bias setting.
Think of bias like setting the idle speed on your car engine. Maybe it should be 500-700 rpm or something, but as soon as you step on the throttle, the rpms go up. When you amp is working, it conducts more current.
There are two different types of plate dissipation to consider...
At idle, which is DC plate current ...
Then when you operate the tube, this is dynamic AC plate dissipation.... which is the average power based on integrating the waveform over the degrees of ON time for that valve... This is non-linear since the path the tube travels is elliptical and the gm varies...
You can approximate it based on a linear load line of the plate load...
Keep in mind that measuring current at the cathodes was from the old days of power triodes... Those 6V6 tubes will dump out 4 to 5 mA of screen current into the cathode at idle and when you crank the amp into square wave region you will get roughly 24mA per screen summing into the cathode...so you will get error in measuring plate current reading at cathode....
Also make sure to account for the dip in B+ voltage when signal is cranking through the amp...this can be measured..
What happens at IDLE stays st IDLE ..... I have seen amps run cooler, due to a big plate load, when signal is running through them and then at idle they would overheat...
At idle, which is DC plate current ...
Then when you operate the tube, this is dynamic AC plate dissipation.... which is the average power based on integrating the waveform over the degrees of ON time for that valve... This is non-linear since the path the tube travels is elliptical and the gm varies...
You can approximate it based on a linear load line of the plate load...
Keep in mind that measuring current at the cathodes was from the old days of power triodes... Those 6V6 tubes will dump out 4 to 5 mA of screen current into the cathode at idle and when you crank the amp into square wave region you will get roughly 24mA per screen summing into the cathode...so you will get error in measuring plate current reading at cathode....
Also make sure to account for the dip in B+ voltage when signal is cranking through the amp...this can be measured..
What happens at IDLE stays st IDLE ..... I have seen amps run cooler, due to a big plate load, when signal is running through them and then at idle they would overheat...
Last edited:
Yes, I do understand the difference between quiescent and dynamic measurements. I guess I didn't state clearly about whether or not the colder bias point contributed to the OP's question of the possible over-wattage of the tubes. I was just questioning the bias point as a starting off point of the overall operation of the power amp for this discussion. Yes, we can bias the tubes as hot or cold as we want within reason.
If you look at GE's tube data sheet, it gives 92mA for max. plate current and 13.5mA for screen current for a total of 105.5mA cathode current for TWO class AB1 biased tubes or 52.75mA per tube. This is for 285V for screen and plate. With a higher plate voltage, wouldn't max overall current have to be less by ohm's law to stay at or under 14W? The OP said he got 80mA per tube for 160mA total. This does seem high.I understand plate voltage drops during higher power, but would it drop almost 100V? Is there something I am missing in interpreting the data sheets? Also, wouldn't we need to know the screen voltage he is using to arrive at an overall answer? The screen voltage has more of an effect on performance than plate voltage.
I am wondering if he measured BOTH tubes' full load current. It is possible one tube it stronger than the other and the one he measured is higher than the other so the total is less than 80+80=160?
I'm not questioning your knowledge, just trying to learn and get to the next level. Thanks for all your info.
If you look at GE's tube data sheet, it gives 92mA for max. plate current and 13.5mA for screen current for a total of 105.5mA cathode current for TWO class AB1 biased tubes or 52.75mA per tube. This is for 285V for screen and plate. With a higher plate voltage, wouldn't max overall current have to be less by ohm's law to stay at or under 14W? The OP said he got 80mA per tube for 160mA total. This does seem high.I understand plate voltage drops during higher power, but would it drop almost 100V? Is there something I am missing in interpreting the data sheets? Also, wouldn't we need to know the screen voltage he is using to arrive at an overall answer? The screen voltage has more of an effect on performance than plate voltage.
I am wondering if he measured BOTH tubes' full load current. It is possible one tube it stronger than the other and the one he measured is higher than the other so the total is less than 80+80=160?
I'm not questioning your knowledge, just trying to learn and get to the next level. Thanks for all your info.
Does this scenario make sense? I went down to my workshop and fired up my Hot Rod Deville. It measured 475V on the plate and I biased it at 65mA for both tubes-quiescent state. I then dimed the gain and master had my buddy play a full E chord- all 6 strings. The plate voltage dropped to about 420. This is a factor of .8842. Now if we apply this factor to the OP's amp his full load voltage would be 336(380X.8842) His full load amperage was 80mA. Cerrem said there could be 24mA per screen. 80-24=56mA left for the plate. 336vX.056=18.8 watts plate watts. Too much. These measurements were made on a very good digital meter, not the cheap $14 Radio Shack variety. The cathode is at ground, the plate is(hypothetically)336v, so isn't 336 volts dropped across the tube? Ohm's law? W=ExI
It was also brought up that the tube only conducts for 1/2 of the cycle. Isn't the 14watt rating already taking this into consideration? The rating for class A is around 5-6 watts, much lower because the tube conducts for the full cycle. The 14 watt rating for class AB1 has been arrived at knowing that the tube will be off for 1/2 cycle. So when we use the 14 watt rating, this has already been taken into account. It seems if we state that it can handle more power because it rests for 1/2 cycle, we are twice taking into account that it rests. It has already been done for us with that 14 watt rating.
If I'm way off, please let me know where. I'm a retired electrician and some of these concepts are new, but ohm's law seems to be at work here. I'm just trying to learn more because the tube amp bug has bitten rather hard. Thanks in advance for your help and understanding.
It was also brought up that the tube only conducts for 1/2 of the cycle. Isn't the 14watt rating already taking this into consideration? The rating for class A is around 5-6 watts, much lower because the tube conducts for the full cycle. The 14 watt rating for class AB1 has been arrived at knowing that the tube will be off for 1/2 cycle. So when we use the 14 watt rating, this has already been taken into account. It seems if we state that it can handle more power because it rests for 1/2 cycle, we are twice taking into account that it rests. It has already been done for us with that 14 watt rating.
If I'm way off, please let me know where. I'm a retired electrician and some of these concepts are new, but ohm's law seems to be at work here. I'm just trying to learn more because the tube amp bug has bitten rather hard. Thanks in advance for your help and understanding.
I think it my be helpful to properly measure what happens in the amp....
Drive in a sine wave at various amplitudes and measure the B+ voltage as well as screen voltage drop AFTER the screen resistor...
Your power supply will have 4 major points of operation to measure...
1. STAND-BY VOLTAGE..... no load voltage with standby switch...
2. IDLE VOLTAGE: this is typically where it is biased..not always optimum..
3. FULL CLEAN POWER OUTPUT: This is where the sine wave is at the hairy edge of clipping but still sine wave..
4. FULLY CRANKED AMP: This is Square Wave output, this is where the screens will draw some appreciable currents...depending on time constants of the Phase Inverter you may go off of 50% Duty Cycle..
Drive in a sine wave at various amplitudes and measure the B+ voltage as well as screen voltage drop AFTER the screen resistor...
Your power supply will have 4 major points of operation to measure...
1. STAND-BY VOLTAGE..... no load voltage with standby switch...
2. IDLE VOLTAGE: this is typically where it is biased..not always optimum..
3. FULL CLEAN POWER OUTPUT: This is where the sine wave is at the hairy edge of clipping but still sine wave..
4. FULLY CRANKED AMP: This is Square Wave output, this is where the screens will draw some appreciable currents...depending on time constants of the Phase Inverter you may go off of 50% Duty Cycle..
I totally agree,cerrem. That is what I was getting at when I posted we need to know screen voltage. We need more information to judge if the tubes are being over dissipated. That is why I did the measurements on my amp. Also, because of the similarities between 6L6 and 6V6. If to OP had EL84s, it could be different. There are some differences between true pentodes like the EL84 and beam power tetrodes like the 6L6 and 6V6. I don't know if it is significant enough but I tried to keep it as close as I could.
Your #4 example would that also lead to bias excursion of the 6V6s causing a shift to a more negative bias voltage? This drives the power tubes closer to class B, correct? I just finished reading about that in Richard Kuehnel's excellent book on guitar power amps. I hope I understood that section.
Your #4 example would that also lead to bias excursion of the 6V6s causing a shift to a more negative bias voltage? This drives the power tubes closer to class B, correct? I just finished reading about that in Richard Kuehnel's excellent book on guitar power amps. I hope I understood that section.
There are too many unknowns to make that call..
Each amp is a case by case situation.... almost all guitar amps have un-regulated voltage supplies, with non linear current loads,ie tubes.. This makes for lots of non linear situations.... Many times the bias supply takes a dip along with the rest of the voltages and maintains biasing in Class AB...
Here is a classic 100W Marshall example.... 480V Plate and screen at idle...
400V Screen and 430V Plate voltages at full clean power output at roughly 100W...in this case 97W ...this also depends on the condition of the filter caps as well as the gm of the output tubes...
Power Factor is major problem these days with the mains voltage and will mess with your DC voltages in a MAJOR way....
So, LATUBEGEEK,I think we need more info. A schematic would also help. Too many variables now to determine what is happening. Get these voltage measurements. Also if you can, get measurements of your bias voltages at mild power, full clean power and fully cranked power. See if it holds or goes more negative. Maybe max out your input at 100mV if you are concerned about over wattage of power tubes or a short test at the 150mA as before.
I noticed you had a question mark in your first post about "normal playing voltage input". In Merlin Blencowe's excellent book on preamps,he has 50mV RMS for single coil strumming and 100mV RMS for humbucker strumming avg. I highly recommend this book for those who have the basic concepts down but want to get to the next level of understanding. Easy to understand and minimal math. Best for those in the mid-level of understanding circuits
Here is some food for thought when a guitar amp is pinned into Square Wave....
The 6V6 output tubes spend roughly 98% of their ON-Time at roughly 50V .....
And it's technically no longer behaving as a Beam Pentode.... but rather a TRIODE with a plate resistance of around 300 Ohms...
The flat top of the square wave is in the TRIODE region of the curves and neglecting rise and fall times, is spending all it's time there....just follow the plate load ..ie load line right slam into the curves..
Now the transformer parameters that worked when the tube was behaving as pentode no longer apply...
The 6V6 output tubes spend roughly 98% of their ON-Time at roughly 50V .....
And it's technically no longer behaving as a Beam Pentode.... but rather a TRIODE with a plate resistance of around 300 Ohms...
The flat top of the square wave is in the TRIODE region of the curves and neglecting rise and fall times, is spending all it's time there....just follow the plate load ..ie load line right slam into the curves..
Now the transformer parameters that worked when the tube was behaving as pentode no longer apply...
Last edited:
Thanks very much for that bit of info. It really does change everything if the signal is a total square wave. Doesn't this only happen with very high gain amps? It seems to me that a great deal of amps cannot provide that kind of drive, like most vintage Fenders and earlier Marshalls and amps of that type. Many oscillograms I have seen of distorted signals is only a flattening of the tops of the input waveform with a rounding of the transition, not the near vertical high frequency transient of a square wave. But yes, definitely something to be considered. I guess we will have to see what LATUBEGEEK comes up with.
On the contrary...
The early Marshalls and Fenders start square waving when you strum hard with volume knobs around 4...
Thanks for setting me straight on that one. I went back to the oscillograms I referred to and they were set to just after the onset of clipping, not full dimed settings. This is the next area of learning for me, power amps; lots of new info to learn! Then I wonder why in the little experiment I ran on my Hot Rod Deville did I only get a drop of plate voltage from 475 to 420 with both pre and power amp settings full up. Tone controls were at noon, no reverb.Surely this amp has more gain that the older fenders. If the power amp spends most of its time at 50 volts when square waved, why did I get a reading of 420v?
Last edited:
OK I'm back.....got sidetracked into some guitar pickup work. Just read all your awesome posts! I'm glad to have initiated such informative dialog! Some questions I can answer: Was only measuring one tube (have only 2 meter), was measuring plate voltage (which dropped less than 10v from.quiescent to max....have diodes in the rectifier socket but expected more). Input is a Sig gen at 100mv p-p on scope, at 1KHz, so I get solid measurements compared to a guitar (did banging that E chord sag your B+?). Anyway, I will run more test, measuring screen voltage & current, on BOTH tubes, AND look at bias voltage changes if any from 0-dime. Thanks!
It dropped from 475v to around 420v; the meter jumped from 420-430v. See post #9. It is still my feeling that if each of your tubes draw 80mA, for a total of 160mA, especially since your plate voltage dropped less than 10v, they are being over dissipated, but getting screen voltages is a must. We need to establish how much screen current is flowing. Then we will know how much the plate current is by subtracting screen current from cathode current. Also, see if your output is a square wave at full power-all volumes dimed. See cerrem's suggested measurements.
Last edited:
Your input signal is good from scope....
If you want to really look at what happening.... use a scope... One CH is high voltage probe at plate and the other a current sense probe also on the plate wire....
Some scopes allow X-Y operation, therefore you can view the load-line ...which can be elliptical into reactive loads..or fairly flat into resistive load...
If you want to really look at what happening.... use a scope... One CH is high voltage probe at plate and the other a current sense probe also on the plate wire....
Some scopes allow X-Y operation, therefore you can view the load-line ...which can be elliptical into reactive loads..or fairly flat into resistive load...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.