• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Choke sizing in PSU vs physical space

Jump to Latest
Windcrest77

Windcrest77

Member
Joined 2019
#1
I've read here and there "get the most henrys you can fit in your chassis" when it comes to chokes. Lets say I want an LCRCLC filter, and I want the supply to be relatively stiff to around 300mA. I find a 16H 300mA choke that fits mechanically for L1, and great, I can also fit another 16H 300mA choke for L2. Why on earth would I buy a 5H or 9H 300mA choke when these 16H guys fit just fine? The Duncan PSU model shows significantly less ripple with 16H chokes. Am I missing something when it comes to compromising choke henrys against chassis restrictions? The larger chokes also don't need as large capacitors. I asked AI to compute an ideal reactance capacitor size for an LC filter at 120hZ with 16H choke, it came back with a nice 110uF, nice that is DC link territory size wise film caps. With a 5H choke it computed a capacitor size much larger which puts me into using electrolytics, which I am holding on the non-compromising side. Everything is a compromise at some point even against the mechanicals of the chassis. Should I buy these 16H chokes, or go back to the drawing board?
 
#2
A choke size is determined by:
-Core type
-Core and window size
-Wire gauge, Rdc of choke
-Rated inductance
-Maximum Idc before saturation (both go together)
-Maximum ripple at the required frequency before saturation (both go together)

You can have have both 5H and 16H @300mA on the same core and the same filling of the window. The 5H choke might have been wound with a thicker wire, so lower Rdc. If Rdc turned out to be the same as the 16H, meaning identical turn count, then there would be flux density headroom, which could be used for ripple voltage accross it (suitable for choke input duty). If it wouldn't be used with a higher ripple voltage, then it could withstand more Idc, 960mA.
 
#6
It is choke input. Surely a 16H 300mA choke would hold up in first position? I can even fit their 12H 400mA version if that is better. Or their 10H 500mA version. All of which have the same physical size. So 16H 300mA, 12H 400mA, or 10H 500mA all fit. Bot the power transformer is rated 350mA the expectation is to make a 300mA PSU, but if oversizing the first choke current rating and reducing the H to 12 or 10 is better I'd do that. Is critical inductance a minimum or a figure to meet exactly?
 
#7
AI answer for the capacitor size with the 12H 400mA choke.

question:

"I have an LCRCLC choke input power supply filter. The AC is at 120 hertz. Compute an ideal capacitor reactance size if the chokes were 12 Henry."

I think I'll go with the 12H 400mA choke as L1 as choke input oversizing, the capacitor size is 150uF not unreasonable to parallel a couple of DC link MKP which is great. Then maybe the 16H as L2. Both chokes are identical in size.

In an LCLC choke input filter, the capacitor reactance (X_C) plays a key role in shaping voltage ripple and current smoothing. Reactance is calculated using:


[X_C = \frac{1}{2 \pi f C}]


where:


  • (f = 120) Hz (the AC frequency)
  • (C) is the capacitor value we need to compute

For an optimal design, the capacitor should ideally have a reactance that complements the choke inductance (L = 12) H. Inductive reactance is given by:


[X_L = 2 \pi f L]


Plugging in the values:


[X_L = 2 \pi \times 120 \times 12 = 9043.2 , \Omega]


To balance the circuit effectively, capacitive reactance can be designed to be in a similar range. Let’s rearrange the capacitor reactance equation:


[C = \frac{1}{2 \pi f X_C}]


Setting (X_C \approx X_L):


[C = \frac{1}{2 \pi \times 120 \times 9043.2}]


[C \approx 146.8 , \mu F]


So, an ideal capacitor size would be around 147 microfarads to match the inductive reactance effectively. Would you like to discuss how this affects ripple filtering or power factor?
Click to expand...
 
#9
Windcrest77 said:
"I have an LCRCLC choke input power supply filter. The AC is at 120 hertz. Compute an ideal capacitor reactance size if the chokes were 12 Henry."
Click to expand...
Figuring in reactance is an interesting concept for an anode supply. What are your design goals? Maximum energy efficiency? At what ripple at the output? Have you checked if maybe a C input filter might get better efficiency?
 
#10
schiirrn said:
Use the "stepped load" feature in PSUD2 and check your supply for ringing. Most relevant part here is the last LC. Calculate values or DCR, L, and C to get as close as possible to a Q of 0.7.
Click to expand...

I'm not sure how to interpret this graph. Transitioning from 50mA to 150mA, too much ringing here? Using a ms timeframe the AC voltage ripple seems to be under .0002 or .2 mV. If this ringing is too bad, what can I adjust to lessen it? The choke sizes I have to choose from at 300mA or more from this mfg are 5H, 9H, 10H, 12H, 16H. Should I just experiment on this stepped model plugging in these choke values?

1749412021106.png
 
Last edited:
#15
Ah, I see the idea here, to be able to analyze a small change in load, I'm learning a lot. OK here is a 1% increase and a .5% increase. I can see now the span of the ringing, voltage-wise. In class A this wouldn't matter much, but I can see now how reactance in the PSU could really participate in the sound of an amp varying the current draw. I'm learning so maybe I'm wrong. You mentioned the second choke ringing being significant, I see why now. I will play around with maybe the lower H chokes available to me here along the line of controlling the ringing. Yeah the 16H choke seems to ring .1 v thereabouts, the 5H and the 9H .01 v thereabouts.


1749415080790.png


1749415210720.png
 
#16
It will also matter for class A. This is only anecdotal but it makes an objective difference when the ringing is gone and the step is a clean step: increase in dynamics for one.
Use the smaller choke for L2, add series resistance (in the parasitics window, increase the 66R, that can later be added as a resistor) and increase capacitance and see what happens. Be warned that with 12H you will need a lot more C for sharp corner then you might expect or like...
Change the step in current draw to 10%.
Add more delay at startup and have the step later, so the start up ringing and the load change are clearly separate.
 
Last edited:
You must log in or register to reply here.
Top Bottom