Good Wednesday everyone,
I'm looking to play with some 6083s in P-P. This is what I found online:
AF POWER AMPLIFIER - CLASS B
PLATE VOLTAGE 600
GRID 3 VOLTAGE 0
GRID 2 VOLTAGE 250
GRID 1 VOLTAGE -33
PEAK TO PEAK SIGNAL 66
ZERO SIGNAL PLATE CURRENT 2x29 ma
MAX SIGNAL PLATE CURRENT 2X102 ma
ZERO SIGNAL GRID 2 CURRENT 2x11 ma
MAX SIGNAL GRID 2 CURRENT 2x28 ma
LOAD PLATE TO PLATE 6320 Ω
MAX POWER OUTPUT 82 W
How can this be Class B if Grid 1 is not driven positive?
Thanks
Ray
I'm looking to play with some 6083s in P-P. This is what I found online:
AF POWER AMPLIFIER - CLASS B
PLATE VOLTAGE 600
GRID 3 VOLTAGE 0
GRID 2 VOLTAGE 250
GRID 1 VOLTAGE -33
PEAK TO PEAK SIGNAL 66
ZERO SIGNAL PLATE CURRENT 2x29 ma
MAX SIGNAL PLATE CURRENT 2X102 ma
ZERO SIGNAL GRID 2 CURRENT 2x11 ma
MAX SIGNAL GRID 2 CURRENT 2x28 ma
LOAD PLATE TO PLATE 6320 Ω
MAX POWER OUTPUT 82 W
How can this be Class B if Grid 1 is not driven positive?
Thanks
Ray
It is not a matter of if the grid is driven positive. Supposedly with class B each tube or transistor handles exactly half of the signal.
You cannot easily have class B without some bias to avoid crossover distortion. Class B might be closer to 36 volts negative grid with maybe 5ma each tube.
In the provided numbers, each tube plate is dissipating 17 watts at zero signal. Also 2.5 watts on the screens.
I would consider this class AB1. Also the numbers show the control grid not driven positive.
You would call it class AB2 if you drove the control grid positive.
You cannot easily have class B without some bias to avoid crossover distortion. Class B might be closer to 36 volts negative grid with maybe 5ma each tube.
In the provided numbers, each tube plate is dissipating 17 watts at zero signal. Also 2.5 watts on the screens.
I would consider this class AB1. Also the numbers show the control grid not driven positive.
You would call it class AB2 if you drove the control grid positive.
The strict definition of "class B" is that each output device conducts for EXACTLY 180 degrees, or half the audio cycle. This applies to tubes, transistors, mosfets, and other amplifying things. The grid bias has nothing to do with amplifier class, though many transmitter tubes require positive grid operation to work in class B. Pure class B results in lots of crossover distortion, so it's rarely used except in some LoFi modulators that run negative feedback or make lots of power so that they spend little time in the crossover region. The term has been incorrectly applied to many modulators like this one which operated clearly well into the class AB1 region.
A true class B modulator will have zero low idle current. A realistic amplifier will have a very low idle current. This one has 29 mA per tube which puts it well into class AB.
Class A and Class AB are further divided into no grid current flows (A1 or AB1) or grid current flows for part or all of the cycle (A2 or AB2). This one runs right up to the AB1 limit but does not enter AB2. Drive it a bit harder and it is AB2 if the driver can handle the grid current.
A true class B modulator will have zero low idle current. A realistic amplifier will have a very low idle current. This one has 29 mA per tube which puts it well into class AB.
Class A and Class AB are further divided into no grid current flows (A1 or AB1) or grid current flows for part or all of the cycle (A2 or AB2). This one runs right up to the AB1 limit but does not enter AB2. Drive it a bit harder and it is AB2 if the driver can handle the grid current.
With the data sheet saying 45 watts for the plate dissipation, would you run the idle current up to perhaps 35-40 ma/ tube?
Why? The specified idle current is determined by whatever is necessary to attain class B. Increase it and you're out of B into AB.
This is a simple question with a complicated answer.
First off, nobody would want to listen to a pure class B amplifier unless some means of reducing the crossover distortion is applied. A modulator in an AM transmitter has no volume control. It is adjusted such that the peaks just touch clipping and asymmetric amplitude compression is applied to keep the average audio power well above the level of uncompressed material. An AM transmitter run at 100% modulation sees the transmitted RF power drop to near zero on negative modulation peaks. This kills reception, so the compressor is adjusted to keep the RF power up, and thus keep the modulator out of the crossover region for more than a millisecond or two. A HiFi amplifier is not run this way unless it is being used as a PA on the dance floor.
Crossover distortion is created by nonlinearities in the output devices when they are operated at low current in the region where one device passes control of the output current to the other device, the crossover region. Every tube is different with respect to the type and amount of distortion it generates in the low current region. Tubes that were never intended for, and therefore not optimized for operation at low currents tend to be worse than a typical audio output tube. There have been many crossover reduction schemes invented in the past both for tube and SS applications, but the usual two still remain the most popular. They are, crank up the idle current, and apply negative feedback. If you are building that dance floor PA, then pile on the GNFB, it may actually help the over compressed music you are going to stuff through it. For a home HiFi amp many things need to be considered.
How is the amp going to be used? When I paired a 125 WPC version of Pete's Engineers Amp with 99 dB horn speakers I simply cranked up the idle current such that the 24 watt sweep tubes ran at 18 to 20 watts. The amp rarely saw excursions above 1 WPC, and crossover was clearly audible at idle dissipations below about 15 watts per tube. GNFB was not used or needed. When I took that same amp outdoors and did use it as the PA for a rock band where it carried everything except for the bass guitar and drums, I ran the idle current down to about 12 watts per tube. Here a little bit of crossover distortion would not be noticed except by those close to the speakers, and their ears were already into compression.
Why is this necessary?
A pure class A amplifier runs an idle current that is typically 75 to 90% of the tube's maximum dissipation. At idle (no signal) ALL of this power is turned into heat (dissipated) since none is turned into music. The plate efficiency is ZERO, so a class A amp sees maximum dissipation at idle. Since the AVERAGE current through the tube doesn't change much from idle to full power, the class A amp will see minimum dissipation at maximum power output.
A pure class B amp has zero idle current. it dissipates no plate power at idle. In most cases the dissipation rises as the output level is increased, and reaches maximum at a point just below clipping. Once clipping occurs the tubes begin to act more like switches than linear devices and dissipation drops off.
A class AB amp behaves somewhere in between these extremes. Maximum tube dissipation occurs somewhere between idle and full unclipped power depending on how hot the idle current is set. In my case those 24 watt tubes idling at 20 watts reached maximum dissipation somewhere below half power which was still well above the speaker blowing level, so I had no issues. If your amp will see similar use, low average power in normal use, feel free to crank up the current. Measure the low level distortion, at 1KHz and other points and adjust the idle as needed to reach the lowest THD. There is little or nothing to be gained by going beyond this level on most amps. Some amps may benefit from a bit more current to extend the region where the amp stays in class A.
If you are going to beat the amp hard where it will see considerable time in the region where the average power output is 40 to 80% of maximum undistorted power then measuring and plotting dissipation VS power output should be done to avoid melting tubes in normal use. Measuring and plotting screen grid dissipation should also be done since it is often the real power limiter.
I have included a spreadsheet I made where I extracted over 30 watts from a pair of 50C5 radio tubes. This was done in AB2 in a conventional grid driven design. The 50C5 tube was derived from the octal 6W6. It is rated for 5 to 6 watts of plate dissipation depending on which data sheet you read. The 50C5 is intended for use in a table radio at a period in time where a radio may be on for several hours a day, every day. It runs class A and its typical operating conditions have it running at 100% of the plate dissipation rating. Its plate is the same size as the plate in some 10 to 12 watt tubes like the 6AQ5, so I see no problem running it at 10 watts. the screen grid is however rated for 1.2 watts and must be respected.
Two different pairs were tested at 340 volts of B+ (way over spec) and 5 watts of idle dissipation. You will notice that as the drive is increased the plate efficiency increases continuously. This is typical since the amp is converting more of the power supply's energy into audio. You will see that the plate dissipation peaks somewhere between 15 and 20 watts of power output, but the screen grid dissipation increases with power output, and "hockey stick's" at the point where the THD reaches 3%. This amp must be built in a manner that the screen dissipation is not exceeded, or the tubes WILL self destruct. I let it run overnight at 20 watts and found it making 20.02 watts the next morning. I make similar plots on anything that will be run hard to avoid meltdowns later. As seen here, plate voltage specs can often be exceeded provided that everything else in the design can handle the voltage (2 to 4 X the B+ voltage). The plate dissipation can be exceeded on peaks, but the average dissipation should be kept in spec. This as stated, depends on the use case of the amp. Screen grids are fragile things. running them hard will result in outgassing (impurities being released into the tube). This can cause an eventual tube runaway condition, especially with some new production tubes.
First off, nobody would want to listen to a pure class B amplifier unless some means of reducing the crossover distortion is applied. A modulator in an AM transmitter has no volume control. It is adjusted such that the peaks just touch clipping and asymmetric amplitude compression is applied to keep the average audio power well above the level of uncompressed material. An AM transmitter run at 100% modulation sees the transmitted RF power drop to near zero on negative modulation peaks. This kills reception, so the compressor is adjusted to keep the RF power up, and thus keep the modulator out of the crossover region for more than a millisecond or two. A HiFi amplifier is not run this way unless it is being used as a PA on the dance floor.
Crossover distortion is created by nonlinearities in the output devices when they are operated at low current in the region where one device passes control of the output current to the other device, the crossover region. Every tube is different with respect to the type and amount of distortion it generates in the low current region. Tubes that were never intended for, and therefore not optimized for operation at low currents tend to be worse than a typical audio output tube. There have been many crossover reduction schemes invented in the past both for tube and SS applications, but the usual two still remain the most popular. They are, crank up the idle current, and apply negative feedback. If you are building that dance floor PA, then pile on the GNFB, it may actually help the over compressed music you are going to stuff through it. For a home HiFi amp many things need to be considered.
How is the amp going to be used? When I paired a 125 WPC version of Pete's Engineers Amp with 99 dB horn speakers I simply cranked up the idle current such that the 24 watt sweep tubes ran at 18 to 20 watts. The amp rarely saw excursions above 1 WPC, and crossover was clearly audible at idle dissipations below about 15 watts per tube. GNFB was not used or needed. When I took that same amp outdoors and did use it as the PA for a rock band where it carried everything except for the bass guitar and drums, I ran the idle current down to about 12 watts per tube. Here a little bit of crossover distortion would not be noticed except by those close to the speakers, and their ears were already into compression.
Why is this necessary?
A pure class A amplifier runs an idle current that is typically 75 to 90% of the tube's maximum dissipation. At idle (no signal) ALL of this power is turned into heat (dissipated) since none is turned into music. The plate efficiency is ZERO, so a class A amp sees maximum dissipation at idle. Since the AVERAGE current through the tube doesn't change much from idle to full power, the class A amp will see minimum dissipation at maximum power output.
A pure class B amp has zero idle current. it dissipates no plate power at idle. In most cases the dissipation rises as the output level is increased, and reaches maximum at a point just below clipping. Once clipping occurs the tubes begin to act more like switches than linear devices and dissipation drops off.
A class AB amp behaves somewhere in between these extremes. Maximum tube dissipation occurs somewhere between idle and full unclipped power depending on how hot the idle current is set. In my case those 24 watt tubes idling at 20 watts reached maximum dissipation somewhere below half power which was still well above the speaker blowing level, so I had no issues. If your amp will see similar use, low average power in normal use, feel free to crank up the current. Measure the low level distortion, at 1KHz and other points and adjust the idle as needed to reach the lowest THD. There is little or nothing to be gained by going beyond this level on most amps. Some amps may benefit from a bit more current to extend the region where the amp stays in class A.
If you are going to beat the amp hard where it will see considerable time in the region where the average power output is 40 to 80% of maximum undistorted power then measuring and plotting dissipation VS power output should be done to avoid melting tubes in normal use. Measuring and plotting screen grid dissipation should also be done since it is often the real power limiter.
I have included a spreadsheet I made where I extracted over 30 watts from a pair of 50C5 radio tubes. This was done in AB2 in a conventional grid driven design. The 50C5 tube was derived from the octal 6W6. It is rated for 5 to 6 watts of plate dissipation depending on which data sheet you read. The 50C5 is intended for use in a table radio at a period in time where a radio may be on for several hours a day, every day. It runs class A and its typical operating conditions have it running at 100% of the plate dissipation rating. Its plate is the same size as the plate in some 10 to 12 watt tubes like the 6AQ5, so I see no problem running it at 10 watts. the screen grid is however rated for 1.2 watts and must be respected.
Two different pairs were tested at 340 volts of B+ (way over spec) and 5 watts of idle dissipation. You will notice that as the drive is increased the plate efficiency increases continuously. This is typical since the amp is converting more of the power supply's energy into audio. You will see that the plate dissipation peaks somewhere between 15 and 20 watts of power output, but the screen grid dissipation increases with power output, and "hockey stick's" at the point where the THD reaches 3%. This amp must be built in a manner that the screen dissipation is not exceeded, or the tubes WILL self destruct. I let it run overnight at 20 watts and found it making 20.02 watts the next morning. I make similar plots on anything that will be run hard to avoid meltdowns later. As seen here, plate voltage specs can often be exceeded provided that everything else in the design can handle the voltage (2 to 4 X the B+ voltage). The plate dissipation can be exceeded on peaks, but the average dissipation should be kept in spec. This as stated, depends on the use case of the amp. Screen grids are fragile things. running them hard will result in outgassing (impurities being released into the tube). This can cause an eventual tube runaway condition, especially with some new production tubes.
I think the original numbers you posted are good if you want a fairly simple 80 watt class AB1 amplifier. You just need a good 6000 or so ohm output transformer and appropriate power supplies.
Maybe look at the improved dynaco stereo 70 drivers for this tube.
Maybe look at the improved dynaco stereo 70 drivers for this tube.
As I stated before you could probably crank up the idle current if you are running efficient speakers and don't play the amp at full volume all the time. If you are running Magnepans or other low efficiency speakers and need the AVERAGE power output in the 20 watt range, then you have less room to turn up the idle current. this decision doesn't need to be made until after the amp is built and tested. The decision that needs to be made before building is the size (VA capacity) of the power transformer. Again, this depends on the AVERAGE power output which depends on the use case of the amp.
Some will say that a class AB tube amp with 100 watts of total output will need 200 VA for the plate supply + heater and driver power in its power transformer. This is true if it was playing sine waves at full power all the time. Realistically one needs to consider the AVERAGE power output, which depends on the use case, the plate efficiency (assume 40 to 50% if not known) plus the usual heater and driver power, then add some margin to reduce heat and allow for cranking up the bias which reduces efficiency.
If you still have your Edcor 3300 ohm OPT's these tubes may work well at 3300 ohms with about 450 volts of plate voltage for 80+ watts. I get 75 watts from 6550's in triode on 450 volts using AB2 from the Edcor 3300 ohm OPT's. These tubes will NOT run in triode mode.
I was going to order some Edcor 6600, 100 watt transformers today as the spec sheet is looking for 6320. I am able to produce the 600 v for the plate at 500 ma. The power supply will be a dual supply, one for each channel, so power should not be lacking. Are you thinking I should target 475 v for the plate and use 3300? Either way I need to order transformers as my 3300s walked. What I have in my possession are two 4400s that are 100 w. And this 62 year old usually does not play his music at a “normal” level. ….
I want to add that I have the ability, assuming proper heatsinking, to adjust the B+ as I am using a Maida regulator.
I forgot to add that the screen supply is regulated. I'll be using an International Power IHB250-0.1.
If you use a separate supply for the screen grid, you MUST make sure that there is no way that the screen grid can get power without the plate having voltage applied. I got to see a pair of nice KT-88's go flash bang when the banana plug came out of my plate supply on the amp seen in the YouTube video.
I just remembered that you had a pair of Edcor 3300's. I have a pair and I have used them with several different tube types to get 50 to 150 watts. If you have to but OPT's then go with the 6600's and a known set of good operating conditions, the ones you posted from the data sheet. If I was building this, I would try the 4400's on somewhere around 500 to 525 volts, but I have variable supplies and can turn knobs to find the happy spot for most tube / OPT combinations.
Ya, what he just said. Everyone should look at what happens to screen grid current when plate voltage goes to zero. Starts ramping up once the plate voltage drops below the screen voltage.
Electrons from the cathode no longer want to jump on the plate once the voltage drops. Just as well hop on the screen.
I’ll try with the 4400s as my max B+ will be 500-510v. The IHB250 is adjustable from 215-265v @ 120ma.