Choke input filters

[Geezer]

The subject of choke input filters came up on another thread.
I could put them to use with several PT's I have from old gear, as they put out tremendous voltages (like 500-0-500).

I have several questions:

1) Where can I find drawings or schematics showing the configuration of choke input filters?

2) How would I go about choosing the correct choke to use?

About 1/2 way down this page ( http://www.angela.com/catalog/transformers/Hammond_Transformers.html ) there are lots of affordable units (150 series, $10-20)....which one would I use? Or would ANY of those be appropriate to use at all?

I've read the info on Aiken's site ( http://www.aikenamps.com/Chokes.html ) but it didn't help much with the above questions.

TIA,  Jeff

[Geezer]

In applying what Aiken says in the linked article, it looks to me like the units listed below would be good "general use" chokes for input filters, am I correct?
They have fairly high inductance, current handling and voltage ratings, and fairly low DCR. Is that what I should be looking for?

159P 10H, 125ma, 155 ohms, 500VDC. EACH $21.
159Q 7H, 150ma, 100 ohms, 500VDC. EACH $21.
159S 4H, 225ma, 65 ohms, 500VDC. EACH $21.
159T 2.5H, 300ma, 43 ohms, 500VDC. EACH $21.

I suppose I would need to calculate (add up) the current requirements of the circuit in order to choose the correct one (?)

[triode]

>> I  suppose I would need to calculate (add up) the current requirements of the circuit in order to choose the correct one (?)

When it gets down to it... yes. What you should really be doing is adding up currents and figuring out what kind of
percent regulation you want. Also of note, there is an upper limit to the size of the choke you can use, as if you are not pulling
enouhg continuous current, you can not create a big enough B field around the choke for it to do proper regulation, and it all
goes to hell.

OTOH, most people around here will just say "stick anything in there, it works", which seems to be a general consensus when
it comes to chokes as of late.

My $0.02, go through the trouble and check it out, it may do you good.

[Lucid_Alice]

If the tranny is putting out 500-0-500 a 500v choke may be too small voltage wise.

[Geezer]

If the tranny is putting out 500-0-500 a 500v choke may be too small voltage wise.

So do I need to be looking @ 600V units, or 1000V+ units?
Is it the total voltage, end to end (1000V) that I'm concerned with, or is it just 1/2 of the winding (500V) ?
What if the (unloaded) voltage is actually 480-0-480 ? It's been a while since I fooled with these PT's....I'd written them off as unusable.

Thanks,
Jeff

[triode]

Man, the voltage rating of a choke is the max continuous potential you can
have across the leads of the choke itself, NOT from
choke to ground.

So, if you use a choke input right after the rectifier, and you have a 500-0-500
tranny, you may see 707 right after the diodes, but you would see 450 right after
the choke, so the voltage on the choke itself is 250V. (This is on paper, it does not
literally come out to those numbers, but that is the idea).

A 400V choke would be fine... it will never see that kind of voltage unless you
short the B+ to ground right after it, and then you will pop a fuse before the insulation
in the choke breaks down. Hell, a 300V choke would make it, even in your 500-0-500
application.

[dwp]

OK stupid question here, what would be wrong with just adding a 5W resistor in front of the first filter caps, say series with the diodes, 1K, 5K, etc.

BTW, am reverting my 5E8 chassis build to tube rectifier. Partially scared of the choke inputs, mostly no room, breadpan chassis. First Filter cap wouldn't need to be 100UF, eh? am going 50UF.....

[HotBluePlates]

OTOH, most people around here will just say "stick anything in there, it works", which seems to be a general consensus when it comes to chokes as of late.

People do say that regarding a cap-input supply, but only because a choke in those circuits only has to support the current of the preamp tubes and output tube screens.

Also of note, there is an upper limit to the size of the choke you can use, as if you are not pulling  
enouhg continuous current, you can not create a big enough B field around the choke for it to do proper regulation, and it all goes to hell.

That's a crucial point. Rule of thumb, the voltage at the first filter cap of a choke-input supply is 0.9 times the RMS voltage you're rectifying. For 500vac, you're talking about getting 450v. If you don't draw enough current, the action of the choke doesn't take place, and the voltage reverts to about 1.4 times the RMS voltage. That would put you at 700v. Not good.... Sounds like a standby to allow the heaters to warm up would be a very imprtant thing to have, not for the tubes (although they could be damaged) but for your filter caps. I'd still feel better about rating that first filter cap for the worst-case voltage.

I went to a reliable source (Morgan Jone's book Valve Amplifiers) and looked up some choke-input supply info. He notes that in the "old days" the solution was to use a swinging choke. You have a very tough time finding those these days, so that's probably out of the question.

He gives a detailed formula for the minimum current draw needed, but then goes on to note an approximation for 50/60Hz power supplies. It's
Current_min. (in milliamps) = V_In (RMS) / L (in henries).
So for your 500vac transformer (which is RMS volts), you'd divide that 500v by the inductane rating of the choke you intend to use to see what the minimum current draw must be. As an example, for a 20H choke, 500v/20H = 25 milliamps of minimum current draw. The implication is you want a very high inductance rating for your choke.

He also notes that it would seem the choke only needs to be rated for the maximum d.c. load current, but that the choke really needs to be rated in excess of this. Why? The choke generates a magnetic flux in the core proportionate to the size of the current passing through it, and if that flux is too great, it saturates the core and the inductance of the choke falls to zero. Bam! You're back at 700v.

The current that the choke will support is the max d.c. load current and the instantaneous a.c. charging current (meaning the current used to charge the filter caps). So that's
I_DC + I_AC = I_{total peak current}.

He derives a formula for the a.c. peak current. For a 60Hz wall outlet, it's
I_{AC (positive peak)} = V_{In (RMS)} / [1386 * L)
Vin is the RMS transformer voltage again, and L is choke inductance in Henries.

Here's an example he gave in the book, but with a slightly different formula for 50Hz operation:

"A Class A power amplifier using a pair of push-pull 845 valves requires a raw HT of 1100v at 218mA, and a 10H 350mA choke is available, but is this adequate? The transformer supplying the choke has an output voltage of 1224V_RMS, and using the equation for a 50Hz mains:
I_AC = V_In / 1155 * L = 1224 / (1155 * 10) = 106mA
I_{total peak current} = I_DC + I_AC = 218mA + 106mA = 324mA

The total peak current is 324mA, so the 350mA is just barely sufficient ... "

That shows that the choke needs to be rated in excess of the peak d.c. load current, which is the current the amp draws from the power supply at max power, not at idle. Once again, a choke with more inductance tends to ease your required ratings.

But in all, you need to flesh out the design a bit more to properly rate the choke. Looking at this info, it's a very bad idea to just "slap something in" when you're thinking about a choke-input power supply.

Jones also notes that regarding the power transformer, it simply needs to be rated for the d.c. load current. He gives an explanation for why, but it's enough to know that's the case.

He does also note that the input choke creates voltage spikes seen by the power transformer. The solution is to have a snubber network, and he notes that the apparently best way of going about this is to fit 2 caps across the choke, with their "center-tap" connected to 0v. What this means is to use two 0.22uF 1kV caps connected in series, then take the outer leads and connect those to the choke to put this pair of caps in parallel to the choke, then connect the junction of the 2 caps to ground.

[HotBluePlates]

Last thing, chokes can buzz due to a vibration that the current through them can cause. The choke must not have a loose shroud, and must be bolted firmly to the chassis. If you do wind up with a vibration/buzzing problem, you can mount the choke on rubber grommets and/or use some foam material between the choke and chassis to dampen the noise.

[HotBluePlates]

OK stupid question here, what would be wrong with just adding a 5W resistor in front of the first filter caps, say series with the diodes, 1K, 5K, etc.

Main problem: spewing heat needlessly. The other possile problem is that it will create big supply sag (which in some cases may be a plus), and that could reduce the power output that your amp is capable of. What I mean is your output tubes (assuming a class AB amp) have big peak currents that might be 2 to 4 times the amount of idle current. Those have to come through that big resistor, and the supply voltage will drop even more than it does at idle. It's actually dynamically fighting against you getting full power output.

But it does work in a pinch.

BTW, am reverting my 5E8 chassis build to tube rectifier. Partially scared of the choke inputs, mostly no room, breadpan chassis. First Filter cap wouldn't need to be 100UF, eh? am going 50UF.....

Is there a question in there? Just kidding, but I don't know what you're starting from or going to so I can't comment on what might or might not work in your amp.

Oh yeah... barring any additions/corrections from PRR, this needs to be archived, since the topic only comes up once on a blue moon, but then we can't find the old post where the info was already written. It is a good way to use otherwise unusable power transformers, but it has to be designed right, with the choke properly selected.

[jjasilli]

da Geezer:

I see your problem.  First I would download & check out the Duncan amps power supply calculator:

http://www.duncanamps.com/psud2/index.html

If still in doubt about the voltage and other rating of your input choke, call Angela and hope for some advice.  You need to call (or fax) to place an order there anyway.

I have seen input choke schematics somewhere on the web, but without actual values.  It's easy enough to visualize:  The choke comes first in series with the B+ supply form the rectifier; then comes the first filter cap, etc.  Also, reference material like "Inside Tube Amps", by Dan Torres, seem to agree with other posters that the critical factor is that the choke can handle the ma requirements of all the tubes (plates, and screens too for the power tubes).

[dwp]

sorry for not qualifying my statement, HBP, no question is in there. I was involved in the previous thread referenced above, well it started as a "how bout two 5AR4's, or drop B+ for 4 EL84's", (5E8A/4 octal chassis) The more times I stuff a design in this breadpan chassis, the more I wonder why........Because it's here? I keep saying no more tweed repro's!

[Geezer]

Once I have my exact circuit picked out, I'll gather all the #'s & post them on the forum & get you guys to help pick out a suitable choke.... ;)

When it looks like all who wish to contribute have done so, I'll move this over to the archives.

Thanks for the great info!

Geez'r

[PRR]

> Also of note, there is an upper limit to the size of the choke you can use,

A lower limit of inductance which will work.

If the inductance is less than the critical value, it works like a cap-input filter, voltage rises.

The minimum inductance at zero current is infinite. This means the L-input filter tends to rise to 1.414*VAC until the tubes warm up. That means your filter caps have to be rated for the hgher voltage, but will cruise at about 2/3rd of that voltage.

There are ways to calculate, and rules of thumb, but if you have a PC then just use Duncan's PSC.

> the voltage rating of a choke is the max continuous potential you can have across the leads of the choke itself, NOT from choke to ground.

The "500V" rating is the winding-to-case rating. They are telling you that if you have 501V from winding to case, and it arcs-over, it isn't their fault. They will sell you a 1,000V or 2,000V rating, if you pay for it. (Well, you used to be able to buy various ratings; in this sad time, you take what you can find.)

It won't actually arc-over at 501V. With old part-specs and new materials, a "500V" choke may take 1,000V forever. Or it might take 1,000V only until you get month out into the Iraq desert, and then it arcs-out, proving there was a reason to rate it conservatively.

There is a trick. Put the choke in the ground leg, instead of the hot leg. A "500V" choke can stand over 1,000VAC this way.

[jjasilli]

How about this: instead of a big power resisitor -- a reverse polarity zener diode in the ground leg of the PT can drop about 100 volts.  (I don't know if you can put these puppies in series for higher voltage rating, and/or parallel for larger voltage drop).

This gets you going in the right direction.  It relieves some of the duty that your input choke must handle, and puts less strain on the filter caps at start-up.  It also avoids the limitations of power resistors in the power tube B+ supply, for which chokes are preferred in the first place.