questions RE power supply dropping resistors

[12AX7]

In particular the preamp and PI nodes. Is there any reason to use resistors of the same value or is it considered proper to use various values? to get voltages where u want them for each node? As an example, i have a 4.7k for the PI because i want it clean, a 18k at V2 because i want the gain stages to be at low enough voltage to sag, and another 4.7k to V1 because it;s already dropped a lot at V2. I don't know how "proper" that is but it sounds good like that but i'm wondering is it would be better to drop them all equally like 3 10k's or such.
Also, why is it so few amps use parallel nodes? Seems like the best way to precisely adjust voltages, especially is for example you wanted V1 to have higher voltage than V2.

[HotBluePlates]

... Is there any reason to use resistors of ... value or is it ... to get voltages where u want them for each node? ...

I edited your question to basically be the answer.

What is Ohm's Law?  Volts = Current x Resistance, or rearranged it is Resistance = Volts / Current.
Here, "Volts" is the volts dropped across the power supply resistor.

So you start with 400v and want 300v, and know the preamp will pull 6mA of current through the dropping resistor:  (400v - 300v) / 0.006A = 100v / 0.006A = ~17kΩ.  Round that up the to standard value of 18kΩ, and you have your required dropping resistor.


Other use is ripple-filter.

The power supply node after the dropping resistor gets cleaner DC in proportion to R x C, so the resistor value can't be too small or it doesn't give us good filtering.  "Good" is subjective, and filtering matters most around the output tube screen (and maybe phase inverter).  There's usually beaucoup filtering by the time the supply reaches V1.

[12AX7]

I'm not really asking about how to determine what voltages i'll get with given resistor values, but whether there is a reason to make the values all the same vs doing it as i did. In other words, i could make all three 10k and when added together the total would be about the same as two 4.7k's and one 18k. But that of course would not allow the PI to be at a higher voltage than if i left it at 4.7k and made the nodes later on higher values. The question is whether there is some reason to make them all the same and add up to about the same total droppage as i have now. Thats the way all or most amps seem to do it so i'm simply questioning my uninformed self as to whether there is a reason it is done that way that might be beneficial compared to the way i did it.
To be clear as to why i did it as i did, i wanted much lower voltage at V1 and 2 than i was getting with the 4.7k for the PI node so i made the next resistor huge to insure low voltage at V2 and the next one 4.7k again to insure V1 voltage doesn't go too low. I'm just wondering if my method is flawer because i don't see that done in other amps.

[thetragichero]

Is there any reason to use resistors of the same value?

same reason leo fender used 600v caps for tone caps in guitars: it's cheaper to stock one value of component than many values

[HotBluePlates]

I'm not really asking about how to determine what voltages i'll get with given resistor values, but whether there is a reason to make the values all the same ...

My answer was not "how to determine what voltages."

It was "people select a resistor value to give them the voltage they want; here's an example where I start with a voltage & find a resistor value that gives the voltage I want."

... The question is whether there is some reason to make them all the same ...

No, there is no reason to make them all the same value.

Unless "all the same value" results in the desired voltage for each node.

[12AX7]

No, there is no reason to make them all the same value.

Unless "all the same value" results in the desired voltage for each node.

Thanks, so i guess it's a case of most manufacturers not going to lengths to find the perfect voltages or it just doesn't matter a lot as long as  the overall voltages are close enough. I tried doing 10k across the board after i posted this and it was pretty close but i prefer it the way i had it a bit more, but main reason was to insure the loudest cleanest power section since it's not a amp with much headroom.

[echuta13]

I'm just reinforcing what HBP has already said.  Your power resistor values should be a value that gives you the voltages you want for the different stages. 

If you're starting with an already classic amp design as a foundation you could use their power supply as a reference perhaps, but only as a beginning, not an end. 
Keep in mind that most designs were based on economy and we have the luxury of not having that limit.   

[PRR]

You also want enough ripple and bass filtering along the way. Input stages are more sensitive to buzz. Bigger capacitors cost more, but bigger resistors do not. Multiple R-C stages are very effective. Big currents in the final stage sneak-back into input stages and the system is ripe for oscillation. Because the R-C chain is low-pass this will be bass or sub-bass, motorboat.

[Williamblake]

My guess being, once you figured out paralell nodes you go serial nodes. I am stuck with paralell nodes but i do SE only. So maybe it is the sag?

[HotBluePlates]

... so i guess it's a case of most manufacturers not going to lengths to find the perfect voltages ...

"Perfect" for what?  "Perfect" in what way?  Because I promise that most (all?) manufacturers decide they've reached an "optimum" with the choices they put into production.

Young Amp Aficionados (myself included, at one time) lack a sense of proportion, and "how much is good enough."

     1 Cycle refining Gold with nitric acid, aqua regia and SMB will get 99% purity.
     Getting 99.999% pure Gold may take 5 cycles through the refining process.
     
     The result is mathematically 1% more valuable, but requires 500% of the cost/time to achieve.

     The spot price on 1oz Gold (net metal content) is the same whether it is 99% pure 24kt, or 99.999% pure 24kt, or 91.7% pure 22kt, or 90% pure "crown gold."  So the extra degree of perfection doesn't yield a return proportional to the effort/cost-to-produce.


  - An Economics class would teach this as the Law of Diminishing Returns.

  - In Audio, we might point out that hearing & logarithms & decibels means "10x More" is perceived as "Double."

  - In Electronics, we might note that most parts actually used have 2 Significant Digits plus a multiplier.  That should be a strong hint about "how perfect" we actually need to get with things.