# Power supply sag when calculating load lines

Hi All!

I've been looking around Google, etc, trying to find any general rule of thumb with regards to power supply sag when calculating load lines. The measured B+ when the tubes are cold always drops when the tubes begin conducting, barring a regulated HT supply. Maybe I'm over-complicating things, but it would seem that each power supply change (resistor value, cap value, etc), will yield a different plate voltage, giving different bias level needs, etc. Is there a general rule of thumb for this? Thanks for any and all insight!

#### rayma

The B+ is the operating DC voltage with the normal loading, not with an open circuit load.
You have to determine that, and supply it in the calculation.

Last edited:

Thanks for the reply! That's basically what I'm doing now, was hoping there was a simpler way or general rule of thumb on assumed percentage of sag for general load line calculations.

#### rayma

Not really. The sag under load is highly dependent upon the nonlinear behavior of the unregulated power supply.
Don't forget AC line variation, capacitor, resistor, and transformer tolerances, etc. About the only rule of thumb
that would help is that transformers typically will sag a certain amount from open circuit to full rated load.
See the data sheet, often around 10%.

#### 6A3sUMMER

Power transformer primary DCR and secondary DCR have sag; solid state rectifiers have very little sag; vacuum tube rectifiers have some sag; choke DCR has sag; small capacitance caps have sag; resistors from cap to cap for isolation and filtering have sag.

Some vacuum tubes require a minimum resistance from the secondary (but the primary DCR x step up ratio, and secondary DCR sometimes meet that minimum recommended value. The rectifier's minimum required series resistance goes up:
With more secondary voltage,
With more first capacitor capacitance,

Generalization:
Choke input filters have less sag.
Tube rectifiers with choke input filters, usually do not need series resistors; that makes the voltage sag less under load.
Cap input filters have more sag.

#### wg_ski

The problem is choosing a load line, given a power transformer you have, or choosing what UNloaded voltage to go buy/make, given a load. To get this right, you basically need to build the power supply and load test it. Not with one DC/AC load, but with several to generate a curve of exactly what it drops TO given how many mA of B+ (and AC of heaters). Then you can choose what load line works best with that power transformer, or what to rewind it (or re-spec a similar one) TO if you already have an output trafo. An accurate simulation model would work, but to be accurate you need DCR and leakage inductance values. Even one with zero DCR will sag. I prefer to measure the donor core before rewinding, knowing i’ll get the same % drop at a given VA if the core utilization is similar. I’ve also resorted to using multiple secondary taps, and using that as trim. Especially if I’m increasing the turns per volt to quiet one down.

#### Hearinspace

Paid Member
I don't know if this will help you or not but you did say "any and all" insight and your post made me remember one particular thing in my thinking that I had to change when I started doing this , which was my native approach of doing everything as a sequential process.

It took me a while to see that the entire circuit diagram with all its parts values represents the already arrived at steady state the circuit is being designed for. What you are calling sagged is not sagged. It's running where it's supposed to be.

Figure out your desired load lines and then design the rest of the circuit as if it is already supporting them

Needing time to settle, there may well be over-voltage issues on start-up but in a simple tube circuit they are often overcome in simple ways like parts voltage ratings , or uncomplicated sub-circuits. Start with drawing what you want to see after it's all warmed up. Then look at how it gets there. If you try to deal with circuit problems before you have a circuit it's a recipe for paralysis.

As for the power transformer regulation, PSUD is your friend . . . . . . most of the time.

. . . . or maybe you're doing a guitar amp? Sag then has a different role . . . . . . .

1 user

#### 6A3sUMMER

Generalization for un-regulated B+ supplies:

Class A single ended, and Class A push pull amplifiers. . .
With no signal, with small signal, then with large signal, B+ does Not Sag.

Class AB push pull . . .
With no signal, with small signal, B+ does Not sag; But then with large signal, B+ Does Sag.

The word Sag most often refers to B+ voltage during the amplifier operation, where the B+ is reduced during large signals.
The Sag definitely changes the linearity and long term maximum power out of the amplifier.

The word Sag is less often used to note the loaded and unload B+ during the design of the amplifier to get the quiescent B+ Voltage you want.
Just use the term loaded versus unloaded. Much more letters to type than sag, but is more correct.

Over-Voltage during warm up is another term, it applies more to safety and reliability of the parts.

Just Saying

Last edited:

#### pblix

Sag is OK intermittently. If your B+ sags "a lot", you need a bigger amp with more headroom. In general though, its a good practice to design your power supplies as stiff as you can. Avoid the use of dropping resistors before the output transformer in your pi filter and avoid tube rectifiers. If you need less voltage, get a different transformer or design an amp around tubes that better fits your transformer. Low DCR chokes are OK in SE amps as the difference in current draw (between idle and driven hard) is less, and the common mode rejection from your output transformer is lost, so you need the extra filtering the choke provides. SE amps and class A PP are not immune, they too draw more current when pushed hard, but in class AB it's a feature causing their much improved efficiency. As a result, the stiff PSU is most important in pentode push-pull class AB than any other common typology I'm aware of. Of course, none of this is true in guitar amps, as sag is a feature that adds to your toan.

#### 6A3sUMMER

pblix,

It should be noted that if you Regulate the Screen voltage of a Pentode or Beam Power tube, with B+ voltage changing, the tube plate current changes very little . . .

A typical example: A pentode or beam power tube that has a 20k plate impedance rp, and a 10V change in B+ only changes the plate current by 0.5 mA!
0.5mA out of 50mA (Only 1% change)

A Triode Wired Pentode or Beam Power tubes is far different. Typical plate rp = 1.5k; 10v/1.5k = 6.6 mA out of 50mA (13.3% change).

Ultra Linear is different yet. You need to regulate the B+, because a change in screen voltage at the UL tap, Will change the plate current by a lot.
Or, just live with the change in current.

Guitar Amps, yes - - - Instruments & Amps. Now you know why this is the Tubes / Valves thread.

No design is immune to every factor you can dream up . . . one factor will always get you.
Just like the COVID "Vaccine"; It is not a vaccine by the Old Definition. Look it up!

Current differences versus signal amplitude in a true Class A single ended, or a true Class A push pull are Very small.
They are easily made up for by a large B+ filter capacitor for the output stage.
Of course there are exceptions . . . the constant droning of a 32 foot organ pipe, for minutes on end.

Happy analyzing, and Happy designing!

Two different terms:
Sag during soft to loud music.
Just my take on the common literature

Need lots less B+, do not use a dropping resistor, instead use a Choke input filter.
By the way . . . generalization: choke input filters have less Sag.

Last edited:

I appreciate all the replies! The system is an old Lowery organ amp using 6SL7's with a 6V6 PP output. The original design used a 5U4, I'm using solid-state. The iron, sockets, and tubes are all the original Lowery items with the exception of adding one octal socket to accommodate a Marshally-style preamp design.

#### rayma

The B+ is going to be higher than the original design spec, due to the higher AC line (120V instead of 115V),
and also from the silicon rectifier instead of the 5U4 (which drops a lot).

#### 6A3sUMMER

Is this organ amplifier going to be changed into a Guitar Amplifier?

The solutions, goals, and results of a Hi Fi / Stereo amplifier, should be different from a Guitar amplifier.
That is why there are 2 Thread Areas:
Tubes / Valves
Instruments & Amps
. . . Just saying

The B+ is going to be higher than the original design spec, due to the higher AC line (120V instead of 115V),
and also from the silicon rectifier instead of the 5U4 (which drops a lot).
All true and taken into account. The short answer to my original question about generalizations is "no", with the following general exception:
Not really. The sag under load is highly dependent upon the nonlinear behavior of the unregulated power supply.
Don't forget AC line variation, capacitor, resistor, and transformer tolerances, etc. About the only rule of thumb
that would help is that transformers typically will sag a certain amount from open circuit to full rated load.
See the data sheet, often around 10%.
This is what I am generally seeing under full load, approximately 10% drop, so I will base my calculations on that.

To 6A3sUMMER, yes, it has been converted to use as a guitar amp. My impression was that all tube discussions fell under the tube/valve category, but admins can feel free to relocate the topic as they see fit.

#### 6A3sUMMER

Generalization:

Some tube guitar amplifier owners know which threads areas have most of the answers.
I like reading about guitar tube amplifiers in the Tubes / Valves threads areas, but I appreciate it when a poster says up-front: this is a guitar amplifier, or is being changed into a guitar amplifier.
That is because the considerations of a HI Fi Stereo amplifier is one set of criteria, and considerations of a guitar amplifier is a different set of criteria.

Another subject:
My power mains varies between 117VAC and 123VAC. I am blessed to have it that good.

Last edited: