Paralleled
Tube
Calculator

Last update
10 Aug 2024


Go Home


2003
Tube DIY


Current
Tube DIY


Evil
DIY Twin:
AmpSecond


Calculators



Paralleled
Tubes V2

 

Single Ended
Tube amp

Pre-load


Ratings
per Tube:

^- Rated Watts


^- Rated Volts


^- Ib0mA

Onset of Current Distortion



Enter Tube
Constants:


^- ohms
R_Plate

At Vgrid=0V


^- Ib0mA Used

Trade Less Power
for less Distortion



^- Ebmin Used

If not known
use 10% of
Rated Volts


Enter
Tube Load:

^-# Parallel tubes


^- Ohms Total
Primary Load


^- Desired plate
watts per tube



Calculate

^- Pout /Pbias

^- Total
    Plate Loss

  Eb1 Check

RDH

Best Load Ohms




HEADS UP !



Calculated

Results

ARE NOT

UPDATED

when the

Pre-Load

button

is pressed.


Re-click the

Calculate

button

to update

the Results.





> The First Row is using the User Inputed Bias Watts per Tube <
>> The Second Row is using 100% Rated Plate Watts. <<
>>> The Third Row is using Rated Plate Voltage,
but with No Power Limit! <<<
Load
Per
Tube
Best
Damp
ing
Factor
Plate
Loss
Per
Tube
%
Rated
Watts
Best
lb1 mA
per
Tube
Best
Vbias
(Eb1)
volts
Eb1
check
Vrms
on
plate
Pout
watts
per
tube
Pout
watts
total
Total
mA
Out
pk-pk

































.
. .
.  .
.    .
.      .
.        .





2A3 load line used to explain paralleling of tubes

   !!! WHAT !!!
These are different results than I get from my other tube design books!!!
Click here to find out WHY!
..
Usage
JavaScript must not be blocked on your browser for this to work.

First use the pull down menu above the "Pre-load" button to pick a tube
and then click on the "Pre-load" button to load the tube data and a suggested load,
or just enter all the tube data and tube loading yourself.

Second, click on the "Calculate" button.
Clicking on the "Calculate" button fills the table and plots the plot.

What do I do with all of this information?

When looking at the output power, remember that our ears hear volume logarithmically. For something to sound twice as loud requires ten times the power (10dB). A 20% difference in power is 0.8dB. That amount isn't a change in volume, but if that power difference is located in one frequency range, it can be a difference in "tone." Keep an eye on the DF (Damping Factor), for that impacts the "tone" or  sound of bass frequencies and impacts how the speaker performs at its crossover's impedance peaks and dips.

The calculator outputs three different bias points for a given load impedance.

The first is where you get to pick the steady state bias power for each tube.

The second is with all tubes running at 100% of rated power which can cause the Eb1 bias voltage to exceed the rating for tube.

The third is at 100% of rated plate voltage for the tube. This bias point can cause the steady state power loss to exceed what is rated for the tube.

On the bottom left, there is a suggested "RDH (Radiotron Designer's Handbook)" Best Load" resistance that will get you close to the maximum output power for the group of tubes being used. I prefer to use something 33-50% higher than this in order to deal with 8 ohm speakers being closer to 6 ohms in real life (i.e. use a 6800 to 7500 ohm transformer instead of the recommended 5000 ohm transformer.) The RDH best load resistance value will change if you change the number of tubes in parallel. To see the new number, you'll have to hit "CALCULATE" again. The "pre-loaded" values automatically display the RDH best load, but if the plate dissipation or number of tubes in parallel is changed, the total load resistance on the bank of tubes will not automatically update. If the recommended RDH value is negative, lower Ib0.  You may also need to adjust Ebmin if you don't like the results.

The DF (Damping Factor) is primary side damping factor that assumes that the output transformer is wound with super conducting (zero ohm) wire on the primary and secondary windings. We can't buy this part yet. As an approximation, when the primary resistance of the transformer is 1/4 the effective plate resistance of the tube, you are at the point of diminishing returns. The effective plate resistance is the Rplate for one tube / the total number of tubes in parallel / any reduction in plate resistance from primary side feedback. 

Reminders:

Eb1 is the steady state DC bias voltage from anode to cathode.

Ib1 is the DC steady state anode bias current.

Eb1 * Ib1 is your static DC power loss per tube. I recommend keeping this at least below 85% of the rated tube wattage, particularly with paralleled tubes. Also allow for plenty of air flow around the bank of paralleled tubes and insure each tube has its own Grid-Stop resistor.

Pout/Pbias is your plate efficiency. Single ended tubes are normally bias in class A. Anything over 30% (0.30) is "par for the course."

Example plot
Above is an example plot showing what the typical results will look like.

  • The left sloped portion of the purple "Tube 0V Grid" line shows the curve fit for plate's V and I. The curve fit for the "Tube 0V Grid" line must have high accuracy above the Ib1 current, i.e. the dc bias point, for the equations to work. Don't worry about fitting data points for the curve down by the Ib0 current, your curve fit software will help you figure out what Ebmin should be. The "Tube 0V Grid" curve is the plate voltage and current with Vgrid = 0V which defines the tube's plate resistance at this grid voltage.

  • After the turquoise colored "Load Line" curve hits the "Tube 0V Grid" line, Ib1 and Eb1 are then shown then on the purple "Tube 0V Grid" plot. 

  • Maximum output power occurs when the current swing from where the "Load Line" intersects the "Vgrid=0V line distortion point" to the "Ib1" bias point is equal to the current swing from "Ib1" bias point to the "Ib0" distortion point. This results in having the same output swing before distorting whether the current is going up or down from the DC bias point.

  • The "Load Line" should be below the brown "Tube's Power Rating" at all voltages to avoid stressing the tube.

  • The plots are calculated with finite voltage steps that change depending on the tube and the bias point. This can cause the plots to not perfectly match the data in the table.

Bad Bias Point for a tube
Above is an example of a "Bad Load Line."

General rules for Paralleling Tubes:
Do not "DC couple" both the Plate and Cathode of paralleled tubes at the same time. It makes it hard for the tubes to bias up correctly. What will happen is that one super hero tube will try to draw too much current and one slack-off tube will not draw enough current. The super hero will wear out quickly and the slacker will make distortion.

This means that if all of the plates read "zero ohms" to each other and all of the cathodes read "zero ohms" to each other, the circuit is not likely to bias up and to operate correctly unless the tubes are perfectly matched over temperature and aging. So with magically matched tubes, yes it will work, but magically matched tubes only exist in the modeling world.

With a normal SET, the recommended way to mitigate this is to tie all the plates together with a wire and give each cathode its own bias resistor (and bypass cap).

If using a Parafeed setup there is a different "option", each tube gets its own "high DCR" Plate Choke (or a CCS). Each plate then gets its own parafeed cap to the output transformer. That way the cathodes can be tied together and the output transformer can be connected to the cathodes.
.

Links to other sites with relevant tube design discussions

Link to Wauwatosa Tube Factory's good description of how to draw a load line
(Opens in new tab.)

Link to "MH Audio" calculators on the "Way Back Machine"
(Very slow to open, but opens in new tab.)



Version 2

The Parallel Tube Script was written by MilliVolt and was based on my General load line web page.
The site and script has minor edits by VoltSecond.
Send any questions or comments via emails to VoltSecond through the Tube Asylum.
Charts are drawn with MIT's Public Domain "chart.js" version 2.9.4 Link to MIT's Chartjs.org

Change Log and Known Issues

There are display issues between Android phones, Android tablets and Windows desktops with the placement and overlap of the data. Also, inside a given hardware platform, there are display differences between the different browsers. What looks great on a desktop, fails miserably on phones and tablets. What sort of looks "OK" on a phone, looked unprofessional on a desktop (and that is being polite about it.) Refreshing the browser cache will get you the latest compromise between the different formats

Landscape view tends to look better than portrait view on small screens.

Ver 1.0        Initial release
Ver 1.1-1.3  Changed pre-loaded "total load" from a static 2566 ohms to be the RDH recommended load for maximum output power. Changed last calculated row to be at "max rated voltage."
Ver 1.3-1.8 Changes to resolve issues between data display on small Android phones and medium sized Android tablets and on Windows Desktops.
Ver 1.9       A brute force attempt to fix issues with browsers using old copies of .css and .js scripts.
Ver 2          Added a plot of the load line..