• E3443.png


    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.​

What makes choke input chokes special?

Mostly they are designed to not buzz so much when used as an input choke.

They may be designed as swinging chokes where they will have more inductance with a light load and have less inductance with a heavy load.
The point of the swinging choke is to just satisfy the minimum inductance number that is required for light loading.
 
They are designed to handle large amounts of AC ripple. The inductance is high at turn on, and drops as the load increases. It's important to note, however, that a choke-input supply must present a certain minimum load at turn on. You'll usually see a hefty bleeder resistor following the first cap, to provide something like 10-20mA minimum current at turn on.
 
Assuming you are talking about a line frequency choke input rectifier, a choke input prevents the high current pulses that a capacitor input rectifier creates. In a cap input circuit, the rectifiers only conduct for a short pulse at the sine wave peak and therefore the cap voltage charges to that voltage. An inductor input rectifier tends towards the average voltage instead of the peak voltage and the current tends towards conduction over the whole AC cycle. For an accurate prediction, you need to run a simulation in SPICE, including the output loading. Note that a choke input rectifier requires a current path (diode) for the off-cycle where the choke is discharging.
This is similar to the much smaller chokes used in PWM power supplies and class-D amplifiers, where the switching charges and discharges the output choke in a low amplitude triangular current wave centered on the DC or audio frequency average.
 
Mostly they are designed to not buzz so much when used as an input choke.

They may be designed as swinging chokes where they will have more inductance with a light load and have less inductance with a heavy load.
The point of the swinging choke is to just satisfy the minimum inductance number that is required for light loading.
They are designed to handle large amounts of AC ripple. The inductance is high at turn on, and drops as the load increases. It's important to note, however, that a choke-input supply must present a certain minimum load at turn on. You'll usually see a hefty bleeder resistor following the first cap, to provide something like 10-20mA minimum current at turn on.

As a long-time professional servicer, I know how and why chokes work, fortunately.
But these two comments would stun the hell out of somebody else.
Both sound like they contradict each other.
 
As a long-time professional servicer, I know how and why chokes work, fortunately.
But these two comments would stun the hell out of somebody else.
Both sound like they contradict each other.
Huh? We both said the same thing. High inductance with a light load, inductance drops as load increases. In my comment I guess I could clarify that by "turn-on" I meant the circuit tubes are not yet conducting and current ramps up as they do.
 
Assuming you are talking about a line frequency choke input rectifier, a choke input prevents the high current pulses that a capacitor input rectifier creates. In a cap input circuit, the rectifiers only conduct for a short pulse at the sine wave peak and therefore the cap voltage charges to that voltage. An inductor input rectifier tends towards the average voltage instead of the peak voltage and the current tends towards conduction over the whole AC cycle. For an accurate prediction, you need to run a simulation in SPICE, including the output loading. Note that a choke input rectifier requires a current path (diode) for the off-cycle where the choke is discharging.
This is similar to the much smaller chokes used in PWM power supplies and class-D amplifiers, where the switching charges and discharges the output choke in a low amplitude triangular current wave centered on the DC or audio frequency average.
Thanks, that's a much better explanation than mine, and very helpful to me.
 
The operating conditions of the choke are quite different when comparing CLC versus LC filtering - a key difference being the much higher Vac across the choke for choke input (LC) application.

One way to appreciate that is to consider the BH curve of the core and what is happening. For a choke, a DC current moves the BH operating point away from 0,0 origin and out to an operating point that is closer to where the core is starting to saturate. Too high a DC current moves the operating point in to the saturating part of the BH curve. A CLC filter operates the choke with a relatively low Vac, meaning the operating point is out at the DC biased region and oscillates around the DC point in a minor loop that causes the choke to present a certain level of incremental inductance - which generally starts to fall in value as the DC current gets larger and the operating BH region moves in to the saturation region.

An LC filter forces the BH operating point to cycle over a much larger loop, which can extend in to the saturation region for part of the cycle, as well as extend back towards the origin for the other half of the cycle - and the inductance value effectively becomes an averaging effect over the total BH operating loop - which generally starts to fall in value as the DC current gets larger and the operating BH region partly extends in to the saturation region. The current waveform passing through the choke is easier to appreciate for a CLC filter as the inductance influence is more 'ideal', whereas for a CL filter the waveform becomes more distorted as the DC current biases the BH operating curve more in to the saturation region.

Some may be aware of how a typical power transformer has a magnetising current waveform that is noticeably 'distorted' (ie. not a nice sinewave), and how that waveform can become very peaky on one side if the power transformer primary is exposed to DC current (especially for toroidal transformers which can audibly buzz) - that is a good example of the transformer core BH curve being pushed in to the saturation region for part of the cycle which effectively lowers the instantaneous inductance at that time and that then allows winding current to peak up rapidly.

A swinging choke characteristic aims to increase the effective inductance as the choke current approaches zero during part of the cycle, to avoid a zero current operating condition which causes transient issues with choke voltage.

Simulating choke operation is prone to misleading results and waveforms when operation extends in to the saturation region, or when a core designed for swinging choke application is used, as most simulations use a simple linear inductance model (no matter where operation is on a BH curve).
 
Note that a choke designed for a choke-input filter will have an insulation rating to suit the large continuous Vac applied across the winding. A choke designed for CLC filtering does not need that extra insulation capability, as it normally has only a few 10's of volts working across it. Unless you have or are buying a choke that was purposefully made for an LC application, and with a known voltage rating, then caveat emptor.
 
Note that a choke input rectifier requires a current path (diode) for the off-cycle where the choke is discharging.
Really? Remember that in the tube heydays LC filtering was common especially with mercury filled rectifiers (80, 866 etc.). I've never seen an additional grounded diode there to »discharge« the choke.

Btw, hats off to all of you! Where from did you know that the thread opener was referring to PSU's? As I read the thread title, I was scratching my head about the necessity of chokes at an amplifier's signal input :unsure:...

Best regards!
 
What is special: the regulation. There are several diagrams of a choke that will regulate a power supply well within several volts. [I expect that this regulation will stop the moment you put a diode across it, backwards... I never tested that ]
The regulation works good, filters out AC line spikes too. It can be calculated of course but for now I don't know where (I designed mine in the age before internet, with printed paper and charts some might remember)
I made a choke LC with 1H to feed a 814 tetrode - worked good but indeed, the whole did hum, I did not mount the choke properly. I had some 470 uF / 500V after the choke. Quite a jolt at start up!
 
Rectifier valve datasheets often include a dc output voltage versus load current regulation chart for different forms of L-input and C-input filtering.

Choke input filters do not need a freewheeling diode - that function is provided by the other diode (in eg. a full-wave rectifier).

The concern with choke input filtering is that the diodes undergo a 'hard' commutation - not so noticeable for valve diodes, but much more of an issue for ss diodes.
 

Depanatoru

Member
2006-07-06 10:27 am
That "concern" is false , is doing a soft switching for the diodes ( bridge or not , silicon or tubes ) because is limiting the current spikes in the filter capacitor ...
And is working only in positive DC voltage , is not switching to negative voltage like in a SMPS , so there is not need for any diode even in the bridge rectifier :headshot:
 
In a SMPS there's also a rectifier, full wave or single pulse, that prevents the filter input choke from negative pulses.

Anyway, trobbins said:

Choke input filters do not need a freewheeling diode - that function is provided by the other diode (in eg. a full-wave rectifier).

The current from the rectifier also is cut off when the AC voltage crosses the zero line, 'cause at this point theoretically no diode is conducting. BUT: The choke tries everything to keep the current flowing. It raises the voltage to keep the same diode (!) conducting!

Best regards!