Power supply inductors (what specs?)

[BrianGT]

I was down at Victoria Magnetics picking up a couple of transformers, and talked to the owner, John about possibly making power supply inductors. He has all of the wire from making the transformers, and said that he could make inductors if I was to give him a design. He even suggested that they make 12ga wire in a square form, and would help eliminate the air gaps in the windings. Does anyone have any clue about the best design for power supply inductors? What is a good tool to figure out how much inductance to put in a power supply? How many watts?

All that I have seen about inductors in Super BQ's soz:
http://www.geocities.com/super_bq/soz3.html

If we can come up with a design, this could be a good source for power supply inductors.

--
Brian

 

[eLarson]

The only rule of thumb I've seen for specifying the size of an inductor in a power supply came from the 1970 edition of the Radio Amateur's Handbook That formula was given to ensure that the inductance was sufficient to cause the filter to act as a proper choke-input supply. They didn't say how they derived it and I've been too lazy to figure it where it might have come from.

Nonetheless, here it is again:

L (given in henries) = (desired output voltage) / (current draw in milliamps)

So for a 32 volt ouput and a 2 amp current draw you would need an inductance of ...

32 / 2000 = 16 mH

It would have to be able to cope with a constant 2A draw. Current squared times its internal resistance can get you a first pass at the power handling needs.

If you aren't looking for a choke-input supply, then you can pick pretty much any value you like. I try to ensure that the combination of the inductance, capacitance and internal resistance of the choke (+ any actual resistors) is it at least critically damped.

Erik

 

[Nelson Pass]

16 milliHenries might end up with more resistance than
you want. I find that in practice, anything at all helps,
but 2 mH is a pretty effective value.

 

[stadams]

Inductor use?

Are you designing the inductor for use as the primary filtering component or as part of a larger filter network (i.e. PI filter)? Mr. Pass used a PI filter in Zen-V2 I believe after complaints of a high noise floor with the original Zen amplifier. He prescribed a value of 2-mH which maybe excessively high if calculating inductance based on ripple voltage after the first filter capacitor.

My only suggestions would be a value that would work in most projects that we see on the board--Zen's, Aleph, etc.--and that they have as low of a DCR as possible.

Later,

 

[roddyama]

Hi Brian,

You want a choke that can handle the maximum current/rail (peak) and the continuous current/rail. The wattage it will have to handle is determined by the DC resistance and the continuous current (I*I*R). For an Aleph2 you should, IMO, use at least 12 gauge windings, but the trick is to get enough inductance with as low a DC resistance as possible because its wasted watts. The voltage drop across the choke will fluctuate causing the rail voltage to fluctuate with variations in current draw.

How much is enough? I think there’s a lot of leeway in the amount of inductance need to suppress the ripple noise. I believe Mr. Pass suggests a 2.2 iron core. You can check some of these threads.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=4974&highlight=choke+psu+aleph
http://www.diyaudio.com/forums/show...age=15&highlight=choke psu aleph&pagenumber=3

Stick with the PI PSU. The Choke Input PSU is a whole different animal.

There’s also a link to the Duncan PSU Designer II that can be helpful to quickly see proposed variations.

Hope this helps,
Rodd Yamas***a

PS Mr. Pass chimed in before I posted, but I had to run so I didn't go back through my post.

 

[BrianGT]

I am looking for making inductors for a PI filter for the Aleph 2. It would be in CLC formation. Will 2mH be enough? I am sure that by using heavy gauge wire, the resistance can be kept low. Should it be an air core inductor, or a iron core inductor? I was thinking that a pair of aircore inductors would be fine.

--
Brian

 

[Bertie]

Another question on PSU chokes, is it good, bad or indifferent to use common mode chokes, one half in each supply rail?

I'm talking about a C-L-L configuration in each rail.

 

[roddyama]

Brian,

An air core of about 2mH should work fine. I tried to find the thread where someone used 1mH and it work ok in lowering the ripple and besides, 2mH is what Mr. Pass suggests. I’m going to use 1mH Alpha Air Cores with a DC resistance of 0.12Ohms for my Aleph5. I may go 2mH yet, but I haven’t decided if I want to spend the $140 for more Alphas. Here’s where I got the Alphas.

http://www.partsconnexion.com/catalog/inductors.html

You could roll your own. It’s pretty tedious work unless you have a machine and they’ll never be as neat as you would like them to be. I made the coils for my first x-over for my horns. They worked fine, but it was 16 gauge. 12 gauge will be a lot tougher to work with. I know you’ve been DIYin’ everything you can so you’ll probably make your own no matter what I say, so I’ll shutup about it.

Bertie,

Take a look at the link I posted earlier in this thread. The bottom line on Common Mode Chokes is that they won’t work for PSU’s.

Rodd Yamas***a

 

[Variac]

This has been discussed a lot. You could search and read for an hour!(worth doing!) Even for 2 mH you need at least 12 ga wire. 10 is better. Nelson likes the steel core Erse from Zylatron. They are appealing cause of the low resistance and can handle a lot of current.

 

[Bertie]

I've just realised what a numbskull question asking about common mode chokes was, sorry.

Getting back to the real world and engaging brain am I right in thinking that air core chokes are best, followed by iron then ferrite?

Also anyone know where I can get some 2mH 4 amp chokes for use in the Aleph 5 PSU in the UK?

 

[Salas]

coils and pi's

Hello,

In my post of zen revisited pcb I describe a 3mh, 2 Ohm aircore.
It worked best. Sometimes Rdc is a tool to shape ripple and damp harmonics. Plus you can use it to control the V+ value if your trans and bridges stray upwards. Especially in a Zen too much V+, and you get it on the sinks to no avail.
Dont be afraid of Rdc in class A. Tube stuff incorporates it spledindly for ages. Just play and shape your final sound checking out 1,2,3 mh cheap aircoils. Some combination will come up right with your individual PSU and sound will flow. Cut and try.

Thanx

Nick Salamouras

 

[Coulomb]

I use Accusonic coils. 3.2 mH Iron core 18 Gauge wound. He has 16 Gauge wound as well, I am not sure about 14 or 12 Gauge.

They were $4.63 CAD Each.

Copntact Info.
Accusonic
Vince
416 663 3135
134 Martin Ross Rd.
North York, On
Canada

 

[stadams]

DCR

Nick,

I agree with your statements however, I would rather have an extremely low DCR if possible. This allows the designer of the supply to optimize the power supply response by choosing the necessary R (type or device) for proper damping as alluded to by Erik.

I try to ensure that the combination of the inductance, capacitance and internal resistance of the choke (+ any actual resistors) is it at least critically damped.

If this is of no concern, the benefit is that you can generally find much cheaper inductors as the DCR rises.

Later,

 

[audionut]

Incorporating a 2mh air cored choke in a CLC type of filter, would the choke filter out the high frequency ringing from the bridge rectifier? or do you use snubbers on the rectifiers?

 

[Clun]

Toroid as filter choke?

I see that many talks about air core chokes but how about toroid types for speaker crossover can they be used as psu chokes?

/Clun

 

[BrianGT]

You could roll your own. It’s pretty tedious work unless you have a machine and they’ll never be as neat as you would like them to be. I made the coils for my first x-over for my horns. They worked fine, but it was 16 gauge. 12 gauge will be a lot tougher to work with.

Rodd Yamas***a

Rodd,

I wasn't going to do it myself, but the guy in charge of Victoria Magnetics told me he could make me inductors if I told him how I wanted them made and what specs. He has all of the transformer making machines, and would to make nice tightly would inductors if I told him the specs.

--
Brian

 

[Bertie]

I've wound myself some chokes on a toroid core using 1.7mm diameter wire, 18 turns, giving me 2.4mH @100Hz and a DC resistance of 0.064 ohms.

I'm going with these in combination with 47,000 uF before the choke and 22,000uF after per rail in my Aleph 5 monoblocks.

It's got 2 chances, and if it doesn't work I guess I'll be busy a little longer!

 

[Salas]

Hello again,

I do agree that using a low Rdc choke and then judge the proper
resistance to add for damping is a better approach if you are going to be thorough, and free from this embeded element.
In a Zen though, to my ears, the 1-3 Ohm Rdc of cheap normally wound aircore coil for speakers -avoid metal cores- is just what the doctor has ordered.
Avoid snubs, -they sound zingy- ,yes the coil do filter the nasties towards your amp.
It will not choke em out of your powerline though.
So for the less adventurous and ill bench equipped of us, trying out 2 or 3 cheap coils is a good alternative.

Cheers

Nikos

 

[Petter]

Toroid

Typically you need an air gap in a coil. A toroid does not per se have an air-gap. To fix this you basically have 3 choices:

1. Choose a low H material (typically not practical). One such material is air or plastic. Not recommended.
2. Choose a material with distributed air-gap such as www.mag-inc.com Kool-mu
3. Cut the core and add spacers, then glue back together. I have a few of these.

Petter

 

[Salas]

Hello Petter,

You have a strong point there. I wanna add:
Only an air gap sounds non muddy with cores on DC. An air gap in SE tube amps output transformers is essential for example. There they loose coupling and go big, but its no concern to this thread.
Fellas,
In a Zen if you think you must check using a very low Rdc core follow Petter. Dont miss the damping check with some resistors after applying the CLC. Maybe you can tune the sound better.
Ah, another thing. Always use the smaller cap before the L.

Thanx

Nikos