The RCA spec sheet for the 5U4 has those calculations too. Same / similar to the above post by Depanatoru
The RCA spec sheet for the "real" 5U4 says it'll handle 40uf.
Not "swinging"? What physical construct makes the difference between a choke and one that swings? It dont mean a thing if it aint got that swing applies to both music and tube powersupplies - what a world.
You can learn things by trying to help. Maybe it's like the difference between a single ended OPT and a push pull.
Nope. The data sheet says no such thing. There's an example inthe Tung-Sol data sheet that uses a 4µf input capacitor buty that's just an example. the typical 5U4G power supply was normally limited to about a 40µf input capacitor value. However, the real ratings are the laid out in the ratings charts for the various conditions.
Short answer is you're over thinking it. You got a bad tube. Tubes plug into sockets because they're designed to be replaced when they fail. Unplug it, throw it in the trash can, and plug in a new rectifier tube.
Not as much info as an RCA sheet but they only mention 4uF here. And I'm under the minimum current by ~100mA, if that matters. Prob does.
There are two important ratings that the input capacitance affects. One is the maximum surge current (non-repetitive), and "hot switching current". Your average rectified current is important, but so are these other specifications. The dynamic characteristics are where most people run into trouble. They don't think about them.
I'm sorry, your response is confusing. A swinging choke is specially constructed to provide a wide range of inductance and handle the unfiltered ripple from the rectifier. I've used standard chokes as input chokes. Some handle it better than others, but in the worst case they "dance" or vibrate excessively, meaning they cannot handle the AC ripple.
After all the amplifier is off, and all the capacitors have been discharged.
Measure the resistance from the power transformer's B+ secondary center tap to each rectifier plate.
If it is at 67 Ohms, then there is enough series resistance for a good working 5U4, to drive a 32uF first cap up to 500VDC (with appropriate secondary voltage).
If it is less than 67 Ohms, use some series resistance from each secondary lead to its plate.
If that does not do the trick . . . find a better 5U4.
(A real 5U4 is a husky device).
If your secondary is 350VAC (350VAC x 1.414 is 495Vpeak), be sure to use B+ caps that are rated for at least 500V.
An unloaded supply might put out 500V, even with the rectifier voltage drop, because an unloaded 350VAC secondary puts out more than 350VAC.
350-0-350VAC is normally rated With load present.
Let Tubelab_com burn up tubes, he does the fireworks for us, so we do not have to.
Thanks Tubelab!
Measure the resistance from the power transformer's B+ secondary center tap to each rectifier plate.
If it is at 67 Ohms, then there is enough series resistance for a good working 5U4, to drive a 32uF first cap up to 500VDC (with appropriate secondary voltage).
If it is less than 67 Ohms, use some series resistance from each secondary lead to its plate.
If that does not do the trick . . . find a better 5U4.
(A real 5U4 is a husky device).
If your secondary is 350VAC (350VAC x 1.414 is 495Vpeak), be sure to use B+ caps that are rated for at least 500V.
An unloaded supply might put out 500V, even with the rectifier voltage drop, because an unloaded 350VAC secondary puts out more than 350VAC.
350-0-350VAC is normally rated With load present.
Let Tubelab_com burn up tubes, he does the fireworks for us, so we do not have to.
Thanks Tubelab!
Last edited:
Just curious what that construction is. Gap? Varnish impregnated windings? It's information that..actually wont do me any good in this lifetime, but a fella can be curious about random things that are encountered by happenstance. Kinda like fan bracing on a guitar vs "ladder". I'll never build a guitar from scratch, but it's still interesting to know how things work and why you get the results you get.
http://www.vias.org/crowhurstba/crowhurst_basic_audio_vol3_025.html
For sure is better impregnated that a normal one , but not universal , you have to calculate and use the right one for that regulation effect
For sure is better impregnated that a normal one , but not universal , you have to calculate and use the right one for that regulation effect
A swinging choke has a specially design air gap to control the core saturation characteristics.
In an inductor input filter (i.e. where the first component after the rectifier is an inductor) it is important that the current through the choke does not fall to zero at any point during the rectification cycle. If this happens then the voltage across the choke will rapidly increase and voltage regulation will suffer greatly. In some cases component and rectifier damage can also result.
Normally, in a closed core design the inductance of a choke will vary with current. With a very small air gap (i.e. no core saturation control) the variance can be quite large. With a very large gap, the variance is very little but the inductance is low. The goal is to choose a gap which provides good saturation control, but still provides sufficient inductance. This is usually done with a controlled air gap. The curves of relative inductance verses current and air gap look like this:
A "swinging choke" is one in which the inductance varies with current in a well behaved and very predictable manner. Like the "medium gap" example above.
Now this is important In choke input filters because there is a limit to the inductance below which the conduction of the rectifier will fall to zero. To combat this, as the current falls you want the inductance to increase to help maintain the current. But at high current, the inductance still needs to be large enough to maintain conduction and provide appropriate smoothing. There is a critical minimum inductance value typically give by the following relation:
In the "large gap" example of figure 14-18 above, the amount of metal necessary to maintain the required inductance with a large air gap is prohibitive. And in the "very small air gap" example, the inductance falls to far too low a value at high load. The controlled gap example allows for both load variation and maintaining the minimum inductance without an overly large and expensive inductor.
The use of inductor input power supply filters was done in answer to technology limitations in the early vacuum tube era. Typically inductors were less expensive and easier to construct than electrolytic capacitors. This is why you see so many filter capacitors in early tube equipment between 1µf and 10µf. Larger capacitors were simply too expensive and difficult to produce. So for high voltage regulation, choke input filters were preferred. The development of the swinging choke was in direct answer to this situation.
However, in the modern era the situation is almost exactly reversed. For us, large value, high quality electrolytic capacitors are inexpensive and plentiful. Chokes however are significantly more expensive. And since, in most respects, properly designed capacitor input filters provide better overall performance than choke input filters, there is little need to ever design a power supply today with a choke input filter.
Thank you for that very thorough explanation! Now I have a question: Why do some regular chokes balk at being used in a choke-input supply? I've certainly experienced this. My understanding was that they cannot handle the amount of AC ripple from the rectifier. If that is the case, what is the mechanism that enables a swinging choke to do so? Maybe it's contained in the explanation above but I need to have it spelled out. ;-)
Compared to a CLC designed inductor, a competently designed inductor for LC choke-input B+ application would typically include more turns, a larger core size, enhanced enamel wire insulation and winding-to-core insulation, and closer tolerance of actual inductance value (ie. gap width tolerance). They are all aspects of design that relate to the additional requirement of managing a high AC voltage (and hence AC related flux swing) as well as the nominal DC related flux level. A swinging choke may also use two gaps of different length, where one gap was designed to saturate at a dc current level approaching max rated.
It is just the construction , how good are impregnated with lacquer . The same a home made transformer can rattle like hell if you are not careful .
There are a number of factors to address in the design. First, the choke must exceed the critical value for the continuous conduction or the rectifier current stops flowing and the voltage spikes. This can manifest as buzzing, ringing, core lamination separation, arcing, etc. Second, the core must not saturate as the current rises. It is important to remember that the instaneous current in the first choke is NOT the average power supply current. It is a combination of the primary ripple frequency and harmonics as the various filter stages draw current. Peak ripple current cannot exceed the saturation point of the choke. Also, at maximum draw the choke should exceed the critical but at minimum load the choke should exceed twice the critical value.
So, if the air gap is too small, the choke swings too wildly and there are conduction problems. If the air gap is too big, the choke doesn't swing and it has to be designed for twice the critical value at minimum load. It also depends obviously on the variation of the load. If it's for a single ended class A stage, then the load is fairly stable and not much variation is required. If it's for a push-pull class AB stage, then the difference between minimum and maximumm is gretaer. If it's for a class B or class C power stage (as used in a transmitter) then the load variation is very large and there can be substantial load variation requiring a carefully designed choke or a more complex filter architecture. The bottom line is if you get the power supply design wrong, or don't get the right parameters set in your choke design, then there will likely be problems.
Yes, thank-you. Even in computer power supplies, where these relatively itty-bitty cores are used, one of the tests we did was how the inductance changes with increasing current. That was part of a "trust but verify" relationship we had with the manufacturers. The graph shown explains a lot. My apologies for my previous ill-advice.
Choke input B+ filter:
I have had very good luck with the Hammond reactor 5H 200mA.
I do not know if it was designed to work as a choke input filter or not, but it works for me.
I hope the quality does not drop.
Your Mileage May Vary.
I thought computer power supplies typically were high frequency switchers;
not the decades old brute force 50/100Hz 60/120 Hz line frequency/full wave rectified frequency chokes.
I have had very good luck with the Hammond reactor 5H 200mA.
I do not know if it was designed to work as a choke input filter or not, but it works for me.
I hope the quality does not drop.
Your Mileage May Vary.
I thought computer power supplies typically were high frequency switchers;
not the decades old brute force 50/100Hz 60/120 Hz line frequency/full wave rectified frequency chokes.
They were, but they used swinging chokes on the secondary side of the converter. Load currents could vary a lot depending on how much was stuffed into a PC. Adding hard drives, memory, upgraded video. The yellow toroid cores had a variable mu which was high at no bias and went down in a controlled way. The green ones were higher mu but not designed for ”swinging” and the red ones were low mu and linear. They still use those for class D filtering.
@6A3sUMMER Was it you that said something about liking SS-rectified choke-input power supplies? Certainly an interesting approach, and I believe I have a Hammond 158Q (5H 150mA) in my stash. How would one determine capacitance for such a supply?