The real beauty of a choke-input supply is that the choke stores energy during the whole rectifier cycle.
When the C1 voltage is highest, current increases into the choke from the rectifiers, and in turn feeds current into the output cap C2. Then, as voltage on C1 drops, due to falling amount of current available from the rectifiers, the current in the choke is held up by the action of the changing magnetic field around the choke (this is the energy-storage mechanism). So, although the current in the choke varies across the cycle, it is always positive (i.e. from in to out) due to the energy-storage action of the inductance. The storage action allows current to be taken from the rectifiers, even when the C1 voltage is very low.
At the output of the choke, C2 functions to stabilise the voltage. The current into C2 varies across the cycle - it is both positive and negative (into & out of the cap). But this current waveform is a gentle sinusoid, rather than heavy pulse.
The overall action of the choke-input supply is to take power from the rectifiers spread over the whole mains cycle. So we see a gentle sinusoid whose rms value is very close to the value of the dc load.
Compare with the cap-input filter:
In this case the rectifier current is dicated by the large cap after the rectifiers, C1.
It can only draw current when the recifier voltage is higher than the cap voltage, leading to heavy current pulses of short duration. These pulses have very high peak values, and hence high rms value. This inflicts more stress on the rectifiers, caps, and trafo. It also generates larger magnetic field pulses, which may couple into the signal wiring and components. It's usually best to keep cap-input filter supplies away from the signal chassis, if possible.
That would require a reversal of the choke current. This will not happen so long as the load current is positive or zero, and normal reverse-pulses (such as those reverse pulses given by a loudspeaker) are not enough to reverse the choke current.
That's not what I see and what prompted the question.
The inductor tries to maintain current flow, that's what inductors do, they resist change.
The capacitor C1 is holding charge. That charge depletes as the current flows through the inductor.
The inductor keeps flowing current and refuses to change much so that the charge on the capacitor drops all the way down to zero. Thatis what the plot shows the cap voltage changes from 45Vdc to 0Vdc.
What if the recharge circuit is not yet ready? The inductor will continue to suck/draw current from the source. The capacitor will start to become negatively charged.
What stops that happening?
The inductor won't just stop drawing current.
Is there some mechanism in the operation that actually prevents the capacitor becoming negatively charged?
The inductor tries to maintain current flow, that's what inductors do, they resist change.
The capacitor C1 is holding charge. That charge depletes as the current flows through the inductor.
The inductor keeps flowing current and refuses to change much so that the charge on the capacitor drops all the way down to zero. Thatis what the plot shows the cap voltage changes from 45Vdc to 0Vdc.
What if the recharge circuit is not yet ready? The inductor will continue to suck/draw current from the source. The capacitor will start to become negatively charged.
What stops that happening?
The inductor won't just stop drawing current.
Is there some mechanism in the operation that actually prevents the capacitor becoming negatively charged?
Hi Eli,
Can you expand on this..Ie a small diagram or example circuit..thank's
Regards
M. Gregg
Oh, you're referring to C1 [cap after the rectifier].
Yes, if the power is detached suddenly, the choke will try to maintain the current. But in pulling C1 negative, the bridge rectifier will become forward-biased, and clamp the voltage to ~1V (schottky bridge) 1.2V (PN bridge). Draw out the rectifier diodes to see the current path.
The choke's energy is dissipated in the rectifier diodes.
-1.2V briefly is safe for most electrolytics, but in any case C1 must be carefully selected - to withstand the rms ripple-current. In the case of B+ supplies, an FKP(foil) or MKP is a good bet. With polypropylene, negative voltages of any amplitude are no problem.
The majority would dissipate in the choke ESR.
Edit: plus the load on the output filter capacitor, as the choke current is effectively trying to keep the output DC voltage constant.
Last edited:
You can see an example in the circuit I gave early in the thread. L4 and C2 are the HF-VHF filter parts, and can be tailored to suit the noise expected.
For example, I used some of the parts named to get a transport product through EMC [conducted emissions] testing recently. In this case , a 6-hole ferrite bead [Wuerth 742 750 43] in both + & - lines, and a Murata 22nF 2kV [DEB series safety cap] in the C2 position dramatically cut emissions at 1.5-30MHz+ .... 30dB or more. if you need more insertion-loss at VHF, then SMD capacitors are needed. At lower frequencies, the main choke should suffice, or add a NiZn-cored wound choke of 1 to 10mH.
At low voltages [up to 100V], the Murata BNX series filter blocks give very high insertion loss across wide spectrum.
Hello,
It seems many problems that arise with using diodes can be tackled by using a choke input with a correct bleeder to assure the minimum current flow it needs to function properly.
I have been using choke inputs for a decade in most of my devices.
My tube line amp has a choke input with a Tango choke designed for choke input. I just did order a lundahl ll2743 which can be connected in what lundahl calls common mode connection. It has double the inductance of the Tango ( which is a very good choke) Hoping the lundahl will perform better because of common mode connection and the 41H which allows lower bleeder current.
I did read that the bleeder current should be 10% or lower from the total current drawn.
Will 41H also have bigger energy storage than the 20H Tango. Physically they are about thje same seize.
The lundahl is a so called anode choke so it will probably filter better because it is not built for 100/120 cycles/hertz but it specs are given for 30 hertz and it should work for a wider frequency rage just like an output transformer
The big advantage of choke input that transformer diodes and caps connected to it will have easier life.
Greetings, eduard
It seems many problems that arise with using diodes can be tackled by using a choke input with a correct bleeder to assure the minimum current flow it needs to function properly.
I have been using choke inputs for a decade in most of my devices.
My tube line amp has a choke input with a Tango choke designed for choke input. I just did order a lundahl ll2743 which can be connected in what lundahl calls common mode connection. It has double the inductance of the Tango ( which is a very good choke) Hoping the lundahl will perform better because of common mode connection and the 41H which allows lower bleeder current.
I did read that the bleeder current should be 10% or lower from the total current drawn.
Will 41H also have bigger energy storage than the 20H Tango. Physically they are about thje same seize.
The lundahl is a so called anode choke so it will probably filter better because it is not built for 100/120 cycles/hertz but it specs are given for 30 hertz and it should work for a wider frequency rage just like an output transformer
The big advantage of choke input that transformer diodes and caps connected to it will have easier life.
Greetings, eduard
That makes a bit of sense.
The first capacitor is not necessarily 0Vdc to 45Vdc, but could be as low as -1.2Vdc to 45Vdc.
The choke does continue drawing current from the source, pulling the capacitor into reverse voltage until the diodes provide a bypass route preventing the cap going any further into reverse voltage.
I have said repeatedly that in a CRC, or CLC, PSU that the first C must be adequately rated for ripple current, because it's duty in this location is severe.
Last edited:
22nf 2KV
Farnell FKP1U022206F00KSSD
Link:
FKP1U022206F00KSSD - WIMA - Film Capacitor, FKP1 Series, 0.022 µF, ± 10%, PP (Polypropylene), 2 kV | Farnell element14
The other was discontinued. (just for interest)
Regards
M. Gregg
Farnell FKP1U022206F00KSSD
Link:
FKP1U022206F00KSSD - WIMA - Film Capacitor, FKP1 Series, 0.022 µF, ± 10%, PP (Polypropylene), 2 kV | Farnell element14
The other was discontinued. (just for interest)
Regards
M. Gregg
Capacitor AC voltage rating at a good frequency is the rating of interest for a poly cap that is often used for the first C. An electrolytic won't work for a small uF first C
Ahhh yes, post 100 where that left-field simulation came in to the conversation which was meant for a low current bias supply but ended up simulating a 5A output with 1N4007 diodes and high PT secondary resistance. I agree that that hypothetical simulation would require an exotic electrolytic beyond most commercial offerings to cope with a 5A output.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.