[AMRadio] Swinging chokes

Mon Feb 18 22:10:12 EST 2002

I've been following this thread from the sidelines.  There was a 
rip-roaring discussion on tuned chokes on the AMPS reflector 
(amps at contesting.com) a while back. Rich (Measures) was advocating the 
tuned-choke method but contradicting himself occasionally and everyone got 
all stirred-up. Happens all the time on that reflector when Rich gets 
involved.  Anyhow, there was some excellent input from others that had 
built or attempted tuned chokes and you should read some of those 
posts.  From memory, Tom Rauch reported that he was damn near killed when a 
tuned choke he built exploded - a design he copied from Bill Orr's 
handbook. A couple of guys from the U.K. (well respected, both RF design 
engineers) added further useful commentary on the difficulties and pitfalls 
of tuned chokes. One thing that sticks in my mind was the comment that you 
definitely DO NOT want to tune the thing to resonance at the ripple 
frequency (120 hz in our case). Apparently you want to tune it to a just 
slightly higher frequency. Obviously, a in parallel tuned circuit the 
voltage goes high and the current goes low at resonance - just want you 
want for filtering/blocking action. Typical of any parallel tuned "tank 
circuit".  However without properly designed components in the case of a HV 
power supply tuned system things can get out of hand. There are 
oscillations and increased circulating current and goofy things happening 
to the iron core because of DC + AC circulating current, and then the 
capacitor's current/voltage rating is exceeded and the chokes 
insulation-to-core rating is exceeded and the whole thing ends up as one 
big molten pool. So my question was, "if you can't tune it to 120 Hz but 
you're trying to filter 120 Hz ripple, what the hell good is it ?". The 
answer was (as I recall it) that like any coarse low pass filter, the 
attenuation *approaching* resonance increases; the attenuation curve does 
not have steep sides in a filter like this. So by having it resonant just 
above 120 Hz (I forget just how far above), the 120 Hz ripple is still 
significantly attenuated but resonance at 120 Hz is avoided. The result : 
the ability to still use a much smaller choke compared to an untuned 
system. Another question that came to mind during all of this was why would 
anybody bother with this if it was such a pain the butt. Someone pointed 
out that this method was developed and used by the big boys (Collins, 
Harris, Continental, and others) for use in large AM and FM transmitters 
where the load current was constant. In such a case the whole thing was 
much easier to implement and resulted in significant cost and space 
reductions (smaller chokes) because of the large scale of things in those 
applications.  Collins later used it in one of their amateur amps and Henry 
followed suit in their amateur amps. But if you study it, amateur SSB and 
CW amps are a poor application for a tuned choke. AM is a different story.

And that's as much as I remember.  Sorry for typing so much stuff.  There 
is also a small section on this subject in Single Sideband Principles and 
Circuits by Bruene, Schoenike and Sabin (I have a copy).

Phil (I know diddly about AM - that's why I sit on the side most of the time)

At 08:34 PM 2/18/2002 -0500, you wrote:
>Bob,
>
>I guess you noticed that Measures advocates that any choke other than a 
>tuned choke is bad news in a power supply because of the poor transient 
>response. Although he does hedge a little when he says that the problem is 
>not as great with a larger filter capacitor.
>
>It seems to me that with a tuned choke, once you draw some current and the 
>inductance drops a little and goes out of the resonant mode, you are back 
>to having a regular choke input supply. And at that point I would think 
>that the same transient problems would exist as if the choke was not tuned.
>Or maybe a regular choke (not a swinging choke) that is tuned will stay in 
>resonance longer as more current is drawn?
>
>There must be some critical value of filter capacitor that needs to be 
>used when a choke is involved to eliminate the transient problems.  I have 
>never seen anything written on that. I also have seen very little written 
>on the transient problems either. Power supplies are usually looked at in 
>a static mode rather than dynamically.
>
>73
>Gary  K4FMX
>
>
>Bob Bruhns wrote:
>
> > Hi Gary,
> >
> > I can't get to the qsl.net link tonight, but I checked out the 
> measures/rebuttal.
> >
> > Inductance will cause a transient voltage drop when a load is switched 
> in, if there is not enough capacitance.  But the time frame is not 
> seconds, it is 1/120 second.  And because inductance is always used in 
> this kind of filtering, this effect will exist whether inductive input or 
> capacitive input is used.  In fact, the reduced inductance of the 
> swinging choke under load might be expected to reduce this effect 
> compared to a non-swinging choke in a choke-input filter.  And the high 
> peak currents of a capacitor-input filter will cause a constant voltage 
> drop as you increase load current.
> >
> > The reason that low-current swinging-choke resonation helps is that it 
> reduces the current drawn by the 120 Hz ripple at low current, which 
> effectively increases the inductance of the choke at low current, and 
> this keeps the filter output closer to the 0.9 x RMS instead of riding up 
> to 1.4 x RMS.  At higher current, the inductance is reduced, so the 
> resonance is off-frequency and therefore no longer effective.  But at 
> that point, the load current is high enough so that the inductance of the 
> choke is high enough to keep the output at 0.9 x RMS.  So the resonating 
> cap helps at the low-current end of operation.
> >
> > At the high-current end, heavy saturation of the swinging choke core 
> causes the inductance to drop more than the current rises, so you start 
> to drop below critical inductance.  With proper component selection, the 
> resulting tendency to increase from 0.9 x RMS toward 1.4 x RMS will tend 
> to cancel the voltage drop caused by line, transformer and rectifier 
> resistances.
> >
> > The approximate critical inductance for a choke input filter is 
> L(Henries) = R(load)/1000.  With a swinging choke, you should figure 
> R(load) from the supply voltage and current at the low-current end of 
> operation.  The inductance value may be reduced somewhat if the tuning 
> cap is used.  You'll have to experiment.
> >
> >   Bacon, WA3WDR
> >
> > Gary Schafer wrote:
> > >
> > > Hi Bob,
> > >
> > > I understand the things that you have stated.  From what I have seen 
> there seems to be some difference in the characteristics between a 
> "tuned" choke and a non tuned swinging choke. I have attached a couple of 
> links that I found on the subject. One is written by Richard Measures. 
> Look down the page to the part of the article on POWER SUPPLIES, 
> filters.  Interesting, but I am not sure it really happens with the 
> swinging choke the way it is described
> > >
> > > The other discusses tuned chokes and also a swinging choke.
> > >
> > > http://www.qsl.net/i0jx/supply.html
> > >
> > > http://www.vcnet.com/measures/Rebuttal.html
> > >
> > > Go down to POWER SUPPLIES in the article.
> > >
> > > 73
> > > Gary  K4FMX
> > >
> > > Bob Bruhns wrote:
> > >
> > > > Hi Gary,
> > > >
> > > > A swinging choke is meant for varying loads.  It provides a degree 
> of system regulation with a varying load.  If we could not use it for a 
> class B amp meant for SSB, we could not use it for a class B modulator 
> either.  However, in an AM system, a choke input filter is easier to 
> design if the supply powers both the modulator and the RF amp, because 
> supply current will be much higher during the resting modulator condition.
> > > >
> > > > A swinging choke is designed to saturate gradually as more DC 
> current flows through it.  This is natural for ferromagnetic cores, only 
> requiring some optimization in the design of the core (generally under-sizing).
> > > >
> > > > A choke-input supply requires a minimum inductance versus 
> current.  If there is enough inductance, the choke averages the current 
> during conduction, and the DC output is about 0.9 times the RMS AC 
> applied between the rectifier diode and ground.  But if the minimum 
> amount of inductance (critical inductance) is not present, the choke is 
> not effective, and the filter operates more like a capacitor-input filter 
> (DC out equals 1.4 times the RMS AC applied between the rectifier diode 
> and ground; DC rides the crest value of the AC waveform).
> > > >
> > > > If the choke inductance falls as more current is drawn, it will 
> eventually reach the point where there is not quite enough inductance for 
> it to operate purely as an inductive-input filter.  What happens?  When 
> you are drawing more current, you get an increase in output voltage over 
> what you would have gotten with more inductance.  BUT - when you are 
> drawing more current, you get more voltage drop in the AC mains, from 
> power transformer transformer resistance, etc, and the available raw 
> voltage drops somewhat.  So, if you match the swinging choke to the 
> maximum current and the voltage drop from other parts of the system, 
> these two effects approximately cancel, and you can actually get better 
> regulation.  I consider this a natural.
> > > >
> > > > Below a minimum current, the choke will also have too little 
> inductance, and the output voltage will rise.  But you have amplifier 
> resting current and power supply bleeder current, and you can improve 
> the  operation of the swinging choke by adding a small parallel capacitor 
> across the swinging choke.  This capacitor is chosen to resonate with the 
> choke at two times the mains frequency when the minimum current is 
> drawn.  The Collins KW-1 does this, and it works...  BUT, remember that 
> the capacitor sees a large voltage swing, so use one that has enough 
> voltage rating to be a filter cap in the circuit!
> > > >
> > > > To prevent damage in the event of a choke or resonating capacitor 
> failure, make sure the rest of the system can withstand short term 1.6X 
> overvoltage.  Uncle Ed (WA3WPUN) had the resonating capacitor fail in his 
> KW-1, and the resulting overvoltage caused damage in the rest of the 
> transmitter.
> > > >
> > > >   Bacon, WA3WDR
> > > >
> > > > Gary Schafer wrote:
> > > > >
> > > > > Thanks for the response Bill. What I was trying to find out was 
> if the
> > > > > particular transformer was rated  220 or 120 volts. It has no 
> markings and
> > > > > only a single wound primary. I did find out how to check. By 
> applying voltage
> > > > > with a variac and monitoring the primary current for a rapid 
> increase which
> > > > > would indicate core saturation. If it made it to 220 volts 
> without saturation
> > > > > (about 250 ma current draw) then it was ok for 220 volt 
> operation. If the
> > > > > primary was intended for 120 volts it should saturate at around 
> 150 volts.
> > > > > Current would rapidly rise to several amps. I tried this with a 
> small filament
> > > > > transformer (rated for 120 volts) also and indeed found that it 
> would saturate
> > > > > at around 150 volts.
> > > > >
> > > > > Speaking of choke input, I wonder if a swinging choke only, will 
> cause
> > > > > problems on SSB. The supply that the transformer is going into 
> has a 20-4
> > > > > henry swinging choke. It was used that way before but I am not 
> sure it is a
> > > > > good idea. I seem to remember from the early days that a swinging 
> choke was
> > > > > not recommended as power supply dynamic regulation would suffer 
> badly ?
> > > > >
> > > > > 73
> > > > > Gary  K4FMX
> > > > >
> > > > > wwd at netheaven.com wrote:
> > > > >
> > > > > > Resend: Did this get out to Gary? Anyone? (testing, testing) 
> .   .   .
> > > > > >
> > > > > > Putting 240vac on the full primary will get you via a 
> solid-state bridge
> > > > > > (choke input) about 3400 vdc. Which is probably just what you 
> wanted for a
> > > > > > 'big' rig. I'd insulate the core from the chassis with sturdy 
> standoffs.
> > > > > > Also caution signs since a secondary short to core will leave 
> it hot to
> > > > > > chassis! Always  go in first with your hv probe! It will 
> probably last
> > > > > > forever. Good luck. Bill N2KQA
> > > > > >
> > > > > > > Gary Schafer <gschafer at mediaone.net> said:
> > > > > >
> > > > > > >I wonder if anyone can help identify a transformer for me.
> > > > > >
> > > > > > >It is a plate transformer with what looks like a hypersil (sp) 
> core. It
> > > > > > >has a single primary with 3 taps. Not sure if it is rated for 
> 120 or 240
> > > > > > >volts. With 120 volts I get 1900, 1770 or 1685 volts out. No 
> center tap
> > > > > > >on secondary.
> > > > > > >The only numbers on it are:  AA 3883  and S179027. Do those 
> numbers mean
> > > > > > >anything to anyone ?
> > > > > >
> > > > > > >It is probably mid 70s vintage.
> > > > > >
> > > > > > >Also does anyone know how I can tell if it is ok to run on 240 
> volts
> > > > > > >rather than 120 volts other than amount of smoke?
> > > > > >
> > > > > > >Thanks
> > > > > > >Gary  K4FMX
> > > > > >
> > > > > > --
> > > > > > -----------------------------------------------------------
> > > > > > wwd at netheaven.com
> > > > > > -----------------------------------------------------------
> > > > > >
> > > > > > _______________________________________________
> > > > > > AMRadio mailing list
> > > > > > AMRadio at mailman.qth.net
> > > > > > http://mailman.qth.net/mailman/listinfo/amradio
> > > > >
> > > > > _______________________________________________
> > > > > AMRadio mailing list
> > > > > AMRadio at mailman.qth.net
> > > > > http://mailman.qth.net/mailman/listinfo/amradio
> > > > _______________________________________________
> > > > AMRadio mailing list
> > > > AMRadio at mailman.qth.net
> > > > http://mailman.qth.net/mailman/listinfo/amradio
> > >
> > > _______________________________________________
> > > AMRadio mailing list
> > > AMRadio at mailman.qth.net
> > > http://mailman.qth.net/mailman/listinfo/amradio
> > _______________________________________________
> > AMRadio mailing list
> > AMRadio at mailman.qth.net
> > http://mailman.qth.net/mailman/listinfo/amradio
>
>_______________________________________________
>AMRadio mailing list
>AMRadio at mailman.qth.net
>http://mailman.qth.net/mailman/listinfo/amradio