Since you're not trying to squeeze every last nickel out of the cost like a manufacturer would... why not just design for 100% duty cycle at full output? Maximum current will be listed on tube data sheets - idle and max are usually listed for each operating point. In actual use, the current might average a few percent over idle (unless it's a guitar amp...), but it will run cool and supply regulation will be good.
Power Transformer specs
Typical PP amplifiers run at 3 to 7 watts for 80% of their useage.
The tubes can actually provide a 3X max rated power for a couple hundred ms. If you size your power transformer to a 5% or less, no load to full load voltage drop, you will meet all of the operating conditions for the amp and allow yourself to limit capacitance to just what is needed for ripple filtering.
If you move your transformer regulation down to 3 % no load to full load voltage drop you will have what is termed a "brute force" power supply. Just exactly what the original amp designers from the 30's and 40's were looking for. This will also allow the tubes to produce their full power peak for transient signals
It was not until transistors and consumer electronics became popular that the power transformers were forced to become smaller, hotter, cheaper and unable to provide all of the amplifiers power needs, without voltage regulation and capacitor banks for storage of energy, instead of just filtration of ripple.
You will get more dynamic sound from the 3% or better power transformers for a PP amp or a SE amp. It will also cause you to buy a custom unit as almost no one makes these brutes for off the shelf sales.
Bud
Typical PP amplifiers run at 3 to 7 watts for 80% of their useage.
The tubes can actually provide a 3X max rated power for a couple hundred ms. If you size your power transformer to a 5% or less, no load to full load voltage drop, you will meet all of the operating conditions for the amp and allow yourself to limit capacitance to just what is needed for ripple filtering.
If you move your transformer regulation down to 3 % no load to full load voltage drop you will have what is termed a "brute force" power supply. Just exactly what the original amp designers from the 30's and 40's were looking for. This will also allow the tubes to produce their full power peak for transient signals
It was not until transistors and consumer electronics became popular that the power transformers were forced to become smaller, hotter, cheaper and unable to provide all of the amplifiers power needs, without voltage regulation and capacitor banks for storage of energy, instead of just filtration of ripple.
You will get more dynamic sound from the 3% or better power transformers for a PP amp or a SE amp. It will also cause you to buy a custom unit as almost no one makes these brutes for off the shelf sales.
Bud
WHen designing amp and power transformer always design at full power output...Idle means nothing and things change non-linearly from the idle position... There are way too many unknowns if you design from the perspective of idle conditions...The unknowns are mostly in the power transformer, such as where the designer choose to use as his peak flux density, choice of core material, heat rise, losses...ect..ect...
Chris
Chris
Lets see an example. 2 tubes in push-pull. Yellow is upper tube, purple is lower tube. How much of a current ability would you specify for your power transformer HT just for one channel, shooting for best dynamics into load? Lets assume +10mA for input and drive stage too.
Attachments
For DIY I use the following calculation:
For a class B amp with no idle current the average current drawn from the power supply is Ipk/Pi, So for a class B stage that should be able to continously deliver full output power the power supply should be designed for this current value.
For a class AB stage it is not that simple as the idle current is not 0 but an approximation I have used is to add 1/2 the value of the idle current to the average current for a class B stage, I know that this is wrong but it is my way to compensate for the lower efficiency of a class AB stage and in my case of OTL amps with low idle current it is close enough for dimensioning of the power supply.
The correct method is to integrate the area under a curve describing the behavior of current in an AB amp where current starts at the idle current value. (instead of 0 as in class B amp) and goes to Ipk, it is not very difficult but I have never bothered to do it.
Regards Hans
For a class B amp with no idle current the average current drawn from the power supply is Ipk/Pi, So for a class B stage that should be able to continously deliver full output power the power supply should be designed for this current value.
For a class AB stage it is not that simple as the idle current is not 0 but an approximation I have used is to add 1/2 the value of the idle current to the average current for a class B stage, I know that this is wrong but it is my way to compensate for the lower efficiency of a class AB stage and in my case of OTL amps with low idle current it is close enough for dimensioning of the power supply.
The correct method is to integrate the area under a curve describing the behavior of current in an AB amp where current starts at the idle current value. (instead of 0 as in class B amp) and goes to Ipk, it is not very difficult but I have never bothered to do it.
Regards Hans
One thing to keep in mind is you don't need exact figures...as long as you are pretty close and you err on the conservative side, then you are good-to go....
Look at it this way....You know Class A is "ideal" 50% efficiency and Class B is "ideal" 78.6% ..... Typically when you bias the amp in Class AB, you want greatest efficiency is Class AB, therfore you don't bias "deep" into AB....you bias closer to B, what i am saying is that when biasing the amp you start cold up against Class B and you bring the current up just till you observe the notch vanish....So conservatively say 60% to 70% at best .....
You could start with the Class B figure and add in the losses of the biasing currents...
You know your power output, so it's just a matter of figuring input current based on efficiency ... For example you have a 50W amp....Class AB1... I will consider 65% efficiency also some losses..So figure about 77W input.... This is all RMS, so RMS current is figured next.....
The power transformer will deliver whatever current you demand...it's temperature rise that is crucial here for that given current and whether or not the power transformer burns up.... SO rate it for full power output and you will be good...ALso keep in mind dynamics is the supply ability to deliver X amount of current in a given time frame... Amps/uS ......
CHris
Look at it this way....You know Class A is "ideal" 50% efficiency and Class B is "ideal" 78.6% ..... Typically when you bias the amp in Class AB, you want greatest efficiency is Class AB, therfore you don't bias "deep" into AB....you bias closer to B, what i am saying is that when biasing the amp you start cold up against Class B and you bring the current up just till you observe the notch vanish....So conservatively say 60% to 70% at best .....
You could start with the Class B figure and add in the losses of the biasing currents...
You know your power output, so it's just a matter of figuring input current based on efficiency ... For example you have a 50W amp....Class AB1... I will consider 65% efficiency also some losses..So figure about 77W input.... This is all RMS, so RMS current is figured next.....
The power transformer will deliver whatever current you demand...it's temperature rise that is crucial here for that given current and whether or not the power transformer burns up.... SO rate it for full power output and you will be good...ALso keep in mind dynamics is the supply ability to deliver X amount of current in a given time frame... Amps/uS ......
CHris
Getting to the end of specifying a power transformer
Almost.... You still need to consider heating and a number of other criteria.
Ask for no more than a 50 deg C rise over ambient with emphasis paced upon 40 deg C. This will give you long term stability in DCR and even longer term life. This is especially important with voltages above 250 Vac rms as Corona begins to become a non linear threat to insulation as Vac rises. Vacume impregnated and baked out varnish is imperative here.
In addition, for low emitted stray field, to limit the hum available for chassis pick up and transfer and direct excitation of outputs, interstage's or input splitters ask for no more than 12 kilogauss of core flux at your country's mains voltage and frequency.
Also ask for split well bobbin construction with primary in one well and all secondaries in the adjacent well and a safety shield between secondary B+ windings and low voltage high current windings that are not for driving a tube rectifier. The 5 volt rectifier winding, if any, should be on the B+ side of the solid copper shield with single wire to ground.
This construction will give you 80 db of common mode rejection for noise from the power grid, 75 % reduction of all electrostatic noises from the power grid and 75% rejection of noises between heater and plate windings.
If you have trouble conversing with a local mfg. email me with your specs and I will provide design criteria, numerically, that they can use.
Bud
Almost.... You still need to consider heating and a number of other criteria.
Ask for no more than a 50 deg C rise over ambient with emphasis paced upon 40 deg C. This will give you long term stability in DCR and even longer term life. This is especially important with voltages above 250 Vac rms as Corona begins to become a non linear threat to insulation as Vac rises. Vacume impregnated and baked out varnish is imperative here.
In addition, for low emitted stray field, to limit the hum available for chassis pick up and transfer and direct excitation of outputs, interstage's or input splitters ask for no more than 12 kilogauss of core flux at your country's mains voltage and frequency.
Also ask for split well bobbin construction with primary in one well and all secondaries in the adjacent well and a safety shield between secondary B+ windings and low voltage high current windings that are not for driving a tube rectifier. The 5 volt rectifier winding, if any, should be on the B+ side of the solid copper shield with single wire to ground.
This construction will give you 80 db of common mode rejection for noise from the power grid, 75 % reduction of all electrostatic noises from the power grid and 75% rejection of noises between heater and plate windings.
If you have trouble conversing with a local mfg. email me with your specs and I will provide design criteria, numerically, that they can use.
Bud
a devils choice for sure
I only design E/I powers. Toroids will give you a better utilization of space and are rumored to have less flux emissions. This is not really true and the flux emitted, all from the windings, can be more invasive than that from an E/I transformer. You will also pay an approximate 40% surcharge for a toroid, over an E/I power from the same mfg. territory. Just labor dollars per unit, as the materials costs are comparable.
Having said that, Toroids are sexy looking beasts, especially when stood up on edge....
Bud
I only design E/I powers. Toroids will give you a better utilization of space and are rumored to have less flux emissions. This is not really true and the flux emitted, all from the windings, can be more invasive than that from an E/I transformer. You will also pay an approximate 40% surcharge for a toroid, over an E/I power from the same mfg. territory. Just labor dollars per unit, as the materials costs are comparable.
Having said that, Toroids are sexy looking beasts, especially when stood up on edge....
Bud
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.