I've got another wacky question. Is there any such thing as "too much" inductance in a power supply choke? I ask because I have been winding my own power and output transformers for some time now with success, and just recently got around to trying an air-gapped choke. I made two revisions of it, one using a full-stack of iron from a microwave oven transformer. Original primary an two secondaries completely removed first. The second is slightly smaller. It uses a half-stack of iron out of an old APC UPS for enterprise/server use. Still pretty big though.
On the APC iron, I managed to wind 3294 turns of 28 awg. 115 ohms DC resistance. On the microwave iron, I managed 3030 turns of #28. DC resistance 147 ohms (its circumference is larger.)
My cheap LCR meter thinks they're both mosfets for whatever reason, even though it'll usually measure reasonable amounts of inductance. I tried measuring sine wave voltage drop across them, but there was no change in voltage at any frequency that I tried. So I tried injecting pulses into the circuit below and then measuring the frequency that the circuit oscillated at after the pulses. Then plug that frequency along with the known capacitance value into a LCF calculator. I don't know how accurate that was, but...
The microwave iron version returned a reading of 39.518 H. The APC iron version returned a reading of 19.225 H. That was staggering. FAR more that I am used to seeing in a power supply choke. Only time I've seen anything close to this was in a very early Atwater Kent console radio. It used a large choke along with two smallish film caps for filtering. Likely because caps were still very primitive at the time. Aside from the risk of inductive kickback (and the physical size), is there any inherent disadvantage of using a choke of that size? A choke like this would be inherently expensive to produce, but I've got a stack of free iron and a virtually infinite supply of #28 magnet wire at my disposal.
In any event, I name all my projects for record keeping. This one is called "Legato."
On the APC iron, I managed to wind 3294 turns of 28 awg. 115 ohms DC resistance. On the microwave iron, I managed 3030 turns of #28. DC resistance 147 ohms (its circumference is larger.)
My cheap LCR meter thinks they're both mosfets for whatever reason, even though it'll usually measure reasonable amounts of inductance. I tried measuring sine wave voltage drop across them, but there was no change in voltage at any frequency that I tried. So I tried injecting pulses into the circuit below and then measuring the frequency that the circuit oscillated at after the pulses. Then plug that frequency along with the known capacitance value into a LCF calculator. I don't know how accurate that was, but...
The microwave iron version returned a reading of 39.518 H. The APC iron version returned a reading of 19.225 H. That was staggering. FAR more that I am used to seeing in a power supply choke. Only time I've seen anything close to this was in a very early Atwater Kent console radio. It used a large choke along with two smallish film caps for filtering. Likely because caps were still very primitive at the time. Aside from the risk of inductive kickback (and the physical size), is there any inherent disadvantage of using a choke of that size? A choke like this would be inherently expensive to produce, but I've got a stack of free iron and a virtually infinite supply of #28 magnet wire at my disposal.
In any event, I name all my projects for record keeping. This one is called "Legato."
Yes. The supply voltage will collapse after a sudden increase in load current. The inductor should be just big enough to remove 120Hz ripple.JasonWatkins said:
Ed
In choke input supply, there is critical inductance of choke, at and above which the current in power transformer secondary becomes uninterrupted sine wave. The rough approximation of critical inductance in henries is load voltage in volts divided by load current in milliamps. For example, if the load draws 100 mA at 400 V, the critical choke inductance is 4 H.
Inductance above critical will not affect behavior of power supply at nominal current in any way other than voltage drop in added ohmic resistance of the choke. However, it will make the supply more flexible - the choke input behavior will be maintained at lower output current. In the above example, a 8 H choke will not significantly change behavior at 100 mA, but extend choke input operation down to 50 mA.
Measuring choke inductance is meaningless unless it is done at exactly same conditions at which the choke operates in your supply, DC current and AC voltage across the choke. Both of these parameters dramatically affect choke inductance.
Inductance above critical will not affect behavior of power supply at nominal current in any way other than voltage drop in added ohmic resistance of the choke. However, it will make the supply more flexible - the choke input behavior will be maintained at lower output current. In the above example, a 8 H choke will not significantly change behavior at 100 mA, but extend choke input operation down to 50 mA.
Measuring choke inductance is meaningless unless it is done at exactly same conditions at which the choke operates in your supply, DC current and AC voltage across the choke. Both of these parameters dramatically affect choke inductance.
Okay. I didn't know how much I was going to get exactly, so I just wound on as much as I could get with the intention of checking afterward. I assumed that more is more. I am going to be adapting the Mullard 520, and it calls for a large choke. I figure it's better to make one if possible. I'll try again, and test under more ideal conditions. Thanks for the clarification.
The reference capacitors you used are suited to RF coils, not LF ones. The pF values should be nF at least, maybe low µF. The distributed capacitance of your windings is significantly larger than the test caps, and this causes an artificial increase in the computed inductance.
Shortly , more inductance and low resistance is better , no surprise that would consume more materials ... of course after a point it is useless to increase more .
As can be seen from many rectifier valve datasheets, a choke input power supply provides better regulation than a capacitor input type, but the load needs to be above a critical level for that to occur, and a higher inductance is a benefit for that application.
I was just characterising a Parmeko P480 choke rated at 20H, 120mAdc, with DCR=240 ohm for choke input application. With 20Vac across the choke, the inductance was about 30H at 5mA, falling to 22H at 130mAdc. But inductance would be higher for a choke input application as the ac voltage across the choke is much higher.
You could look through Patrick Turner's archived website for how to design practical chokes, and how to adjust the gap for a given outcome.
There are also other simple ways to measure choke inductance for a power supply application: https://dalmura.com.au/static/Choke measurement.pdf
I was just characterising a Parmeko P480 choke rated at 20H, 120mAdc, with DCR=240 ohm for choke input application. With 20Vac across the choke, the inductance was about 30H at 5mA, falling to 22H at 130mAdc. But inductance would be higher for a choke input application as the ac voltage across the choke is much higher.
You could look through Patrick Turner's archived website for how to design practical chokes, and how to adjust the gap for a given outcome.
There are also other simple ways to measure choke inductance for a power supply application: https://dalmura.com.au/static/Choke measurement.pdf
That is a great resource that I didn't know about. Thank you for sharing. Unfortunate to hear about his passing. He was clearly very talented and passionate about what he did.
The link, if anyone is interested in reading it:
https://turneraudio.com.au/
Fortunately, I didn't have to modify a lathe to wind mine, nor do I have to fabricate bobbins. I picked up a cheap coil winder off of Ebay. The NZ-1. They run about $35 USD. They're hand-cranked, and geared to roughly 1:8. I can't keep turns neat at that speed though, so I moved the hand crank to the main axle for 1:1 winding. Takes all day to wind an OPT, but the results are worth it to me. Also has a counter on it. Manually counting turns STINKS. 3d-printing in Nylon can knock out bobbins all day long, too.
I'm still reading through his site, but thank you. And thanks to everyone else for sharing. Given me several things to consider for the next iteration(s).
You might do some simulations in PSUD2. It's interesting to watch the behavior of a supply to step changes in the load current and see how it responds. Changes in inductances and capacitors has noticeable effect on the smoothness of the voltage when it step changes. I try to get a nice smooth, and hopefully quick, transition of the voltage between current levels. Too high an inductance or capacitance can really muck up the behavior. Now, I'm not sure this is a valid test for an audio amp but it seems like a good idea.
I would imagine the intra-winding stray capacitance is quite large, so 390pF seems very low for characterizing this - try more like 10nF and you'll be less likely to be getting error due to stray capacitance.