Calculate Power Supply Dropping Resistors

[mwelch55]

Is there a way to calculate the size of the resistors between nodes?  I built a high-gain 25 watt amp using 5881 output tubes with 340 volts on the plate.  I put the same preamp in a 50 watt amp using EL34 output tubes with 490 volts on the plate.  In order to get the B+ voltage low enough for the preamp, I'll need to use different (larger) resistor values in my 50 watt power supply.  On my previous build (25 watt), I guessed at it and changed the resistors out until I had the right voltage in my preamp.  I eventually got there, but I would like to be able to calculate the resistor values.

How do you calculate the value of the power supply resistors to get the proper B+ voltages?

Mike

[Ed_Chambley]

Loading makes the calculation much harder than using a string of resistors. simply remove the extra resistors.

[sluckey]

On my previous build (25 watt), I guessed at it and changed the resistors out until I had the right voltage in my preamp.

That's how I do it. Fastest way if you have a good stock of resistors.

[silverfox]

There are lots of videos, this one is an air corp training video. Seems to do a good job; And they had to as training new recruits back then, you might have a full range of personnel, all walks of life to educate so the material had to be a good common denominator.

silverfox.

[jjasilli]

Sluckey & Ed have lots of experience, so their guessing could be pretty close.  If you know your amp is similar to a published schematic, you can check that.  If you're facing utter mystery, Ohm's Law is your friend.  V = R x I.  The Voltage Drop across a resistor = Resistance in Ohms x Current in Amps. 


Check the tube spec charts to find the current draw for the tubes in your amp.  For power tubes both the Plates and Screens draw current.  For preamp tubes you can simply assume at first that they each draw 1mA = .001 Amps.  Add up all the current draw downstream from your B+ resistor.  Determine what voltage you want on the downstream side of the resistor.  If you have 350VDC on the upstream side of the resistor, and want 300VDC on the downstream side, the Voltage Drop = 50.


So, V_drop = R x I; 50 = R x I; R = 50/I


If there are 3 preamp tubes downstream from the resistor, then I = .003; 50/.003 = about 17K Ohms. 


If you wind up with, say, 280VDC instead of your target of 300:  You now have enough info to calculate the exact current draw given the actual Voltage Drop of 70.  Then recalculate Ohm's Law for the exact R value.  No guesswork is required (except for the standard, tentaive guesstimate of preamp current draw).

[PRR]

> How do you calculate the value of the power supply resistors to get the proper B+ voltages?

You go around in circles.

First observe a build that works but is not happy. Ensure you really have appropriate values throughout. Now observe the V and I in each stage. You can reduce each tube+Rs to a simple resistor. You can assume this equivalent resistance hardly changes (maybe 10%) even if the B+ node is changed from 200V to 400V.

Now you have a series string of drop-resistors and a bunch of stage-resistors hanging off. There's various network theorems to get an "exact" analysis. However if they are all low, the cave-man hack is to change an early 10K drop-resistor to 2K. That will bring-up all later voltages. Enough? Too much? Maybe take a second guess. Or it may bring-up stage 3 good but stage 1 is low. Pick a drop-R between 3 and 1 and reduce its value.

Having estimated all voltages and currents, steady power rating is easily computed. Always double this; with fat caps going very over-Watts is wise due to start-up surge.

Go through your new drop-R values and proposed decoupling caps and check each R-C works out well below the audio band. You need to keep wall-AC buzz out of all stages, especially early stages; you need to keep late-stage power wobble from sneaking into early stages to howl-around. There's detailed calculations. At present e-cap prices, just over-sizing all caps is cheaper than hard thinking.

[shooter]

the cave-man hack

my hairy knuckle way is decide about what I expect my power tube(s) needs, 400v, 100mA.  R=e/I = 4k, so I hang a 5K 50W off the PA 1st tap, quick measure, move to last tap, and do a quick measure.  Even then, every build I've done, I've still re-tweaked 90% of the time

[PRR]

> I hang a 5K 50W off the ... last tap

The usual 12AX7 with 1.5K and 100K stage "looks like" 330K.

A half-dozen such stages looks like 55K. Usually only 1 or 2 such stages are at the very end of the line.

I suggest that 5K is a way-heavy load for the end of most B+ strings.

[shooter]

is a way-heavy load

I use it to "set/check" the pa, the far end test is just to make sure if I screwed up, i'll find it before wiring more stuff. 

[92Volts]

If you have the preamp working in the low-voltage build you can measure voltage drops in that example and adjust resistors from the original value to drop more/less voltage as needed.

Also, you only need to worry about the first resistor that feeds the entire preamp... since you aren't interested in changing voltage drop between each stage of the preamp.

Say powertube B+ is 340v and 1k resistor drops that to 320v to the preamp. 20v is dropped across the resistor.

If B+ rises to 500v you need to burn off 180v to feed the preamp 320v. Make the resistor 9x larger for 9x voltage drop. 9k ohms (9.1k ohms is a close standard value).

You're done!



If you wanted, you could use Ohm's law to figure out the current (20ma in this example) then use that to calculate the resistance needed to drop the same amount of voltage. But you don't really need to.

Though you'll want to be able to figure the current to find power through the resistor, 3.8W in this example. Most preamps draw (way) less than 20ma so you'll be working with higher resistance values, but hopefully less power than this example.

[mwelch55]

It sounds like my method is not far from the way some others do it.  I will do some testing using some of the great ideas discussed here and see what method works best for me.

Thank you for the replies and the information.

Mike