Method to Measuring Inductance of Output Transformers

Подписаться 18 тыс.

Просмотров 4,2 тыс.

50% 1

This is one of a couple of methods that have been investigated lately and it seems to be the easiest to implement and gives usable results in determining the actual inductance values of audio output transformers and large iron core inductors for use in modeling an output transformer in LTSpice.
These measurements can become tedious and sometimes frustrating and very far off what is expected. I do not guarantee in any way that you will get the results I get or the results you expect or furthermore, the correct answers. As I said in the video, it is a method that I find reasonable and it is surely better than blindly guessing. Thanks for watching.

Развлечения

Опубликовано:

14 ноя 2017

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 18

@bobh.3705 6 лет назад

I absolutely love the methodical, logical, ease with the way you communicate through your videos. You make it look so easy but I know better than that. I know I have a better understanding of this subject and many more because of your videos. You are the very reason I went back to vacuum tubes after 50 years of electronics. Thank you so much.

@jwl9286 6 лет назад

What a test set! Fantastic job! I've committed XL=2πfl to memory! When you reffered to it my ears resonated. I try to visualize how it works but you make it happen. On a lighter note Herman's Hermit's measured 8 Henry's in the song Henery the eighth I am!

@dennisgouveadeazevedo6238 6 лет назад

Thanks for the video David, I just love it. I guess there's always something cool about building your own test gear.

@davidsoffer3307 6 лет назад

Great information as always! Thank you.

@InssiAjaton 6 лет назад

I have noticed that you show the detector of the GR bridge in wide band position, but it looks like you use the internal 1 kHz oscillator. When in the ancient past I used the same bridge, I concluded that my results were more steady if I selected the 1 kHz narrow band detector with the internal oscillator. Only when I needed some other frequency, I changed the detector to wide band. That relates to this situation, where you could use the 70 Hz to feed the GR bridge. Another issue is the excitation amplitude. As it happens, the hysteresis loop of the core varies by frequency and amplitude. With mu-metal or permalloy cores you could get good results at a wide frequency range and also with minimal amplitude effect. You find those only in very low power transformers, mainly microphone transformers. Output transformers generally never use other than ordinary silicon steel laminations. And there comes a big gotcha. Very low signals don’t “tilt” the hysteresis loop anywhere enough to get the maximum inductance. Of course, when you increase the drive signal, already before the final saturation, the inductance starts dropping. Then another issue. Somebody suggested there might be antenna effects from the other loose wires. Not much with the low 70 Hz frequency. And twisting the loose wires together would actually increase load capacitance. But again, even that would start having effect at the upper end of the audio frequency range. And finally, chokes are what they are, basically just two wires. But the transformers offer typically at least two windings with four wires. That introduces another possibility. You could measure the inductance of the secondary winding. If you also measure the turns ratio (say 50 V RMS at primary producing 50 mV RMS on secondary or the turns ratio 1,000:1), you get the primary inductance by multiplying the secondary inductance by 1,000,000, which is the square of the turns ratio. Actual output transformers are likely to have turns ratios closer to 100:1 and the the multiplier would be 10,000. Anyway, often this would be far easier kind of measurement.

@ElPasoTubeAmps 6 лет назад

I found this particular measurement process pretty flaky and very hard (if not impossible) to get the same results each time. I abandoned this approach for the process in the video I posted right after this one that takes into account the peak values of the sine wave. My results became very good and compare right along with the GR 1650A. However, there are some exceptions that I get from the GR instrument that I think are actually more accurate with my method for iron core inductors be it a choke or transformer. One example is a Chicago Standard output transformer that I can consistently measure the primary L at about 40H. I use low frequencies between 30 and 70 Hz and keep the voltage to 100 mV or less. The GR runs as 1 KHz, as you know, and the results I get from it on the Chicago transformer are between 60 and 70H. I am pretty sure that is not right. That is too high. 40H makes sense. I have had the Ls vs Lp mentioned before and as I am sure you know, Ls is for Q up to 10 and Lp is for Q between 1 and infinity. The Q of these chokes and transformers are often above 10 so I have to use Lp to make the DQ scale dip. As for transformer turns (voltage) ratios, I find the typical 30-50 watt vacuum tube output transformer to have a ratio of around 25:1 or an impedance ratio of 625:1. That is the typical 5000 ohm plate load to 8 ohm secondary. I experimented with placing small capacitors across both the exciting voltage and the coil under test with values like 500 pF and I found no change in any of the measurements, I am sure primarily, because I use low frequencies like 30 and 40 Hz where Xc is so high it doesn't change anything. Thanks for your thoughts and suggestions.

@garytrautman5404 6 лет назад

Great idea, measuring the resistance with an AC signal. Avoids magnetizing the core, too! I would have thought a voltage level much higher than 70mV would be needed to get up into the range where the core permeability significantly kicks in and magnetization losses are swamped - but it looks like that may not be a big deal. Have you tried a higher voltage (5V RMS or so)?

@ElPasoTubeAmps 6 лет назад

This adventure in measuring L in transformers has been a real eye-opener. I still know virtually nothing about transformers but one thing I have learned is that they are much more sensitive to outside influence, when being measured, than I imagined. This is not true under operating conditions when hundreds of mA is flowing but they need to be measured with care and away from any influencing energies including a metal bench. A manual worth looking into is the General Radio Type 1632-A Inductance Bridge that can be found for free at the Bama Boat Anchor website. Page 8 talks about measurement of iron-cored inductors and very low induction levels in what they call the Rayleigh range where a ferromagnetic core has a linear permeability characteristic that is a linear function of excitation level. I know that is a mouth-full and I was copying if from the manual. With that said, it works out in my experiments to keep the voltage as low as possible and with a newer more complex method that I have found out about and used, I can raise the voltage level to a full volt with dependable results. I have raised it to as high as 6.5 volts but measurements start becoming unpredictable and incorrect for my calibration inductors. My "newer" method takes into account all of the vectors and not just scalar values like the so called, square-root-of-three, method does.

@garytrautman5404 6 лет назад

GenRad manuals are worth their weight in gold - as is all the excellent homework you share here. I imagine that all the flying leads act as antennas that pick up stray EMF and transform it onto the winding being measured. It may help to twist each windings leads together, winding by winding (unless it's very old and the copper has become brittle). Thank you seems to fall short of what you deserve - one more outstanding individual to be thankful for this Thanksgiving. Enjoy the Holiday. PS: Have you considered contacting the Discovery Channel? You have the right stuff.

@billmoran3812 6 лет назад

So, I'm wondering if the difference between small coils and iron core transformers is due to the potential core saturation and the reflected impedance of the other winding? Interesting experiment.

@ElPasoTubeAmps 6 лет назад

Hi Bill, There are some procedures in the GR manuals on incremental inductance values where the core is biased with different levels of current but in the method I have found most useful and the way the GR1650-A measures inductance, we do it at the other end of the spectrum. 50 mV is so small and the core current is in the uA range so, again, according to one of the GR manuals on an inductance bridge, this is where the inductance is stable. They sure know a lot more about it than me so I chose to measure it that way. The problem at the low end is having instruments that are consistent and accurate at the mV and uA level and just moving the transformer across the table or making these measurements on a metal workbench can cause havoc. Demagnetizing the core with a couple of volts across the primary and slowly reducing it to zero over about a three minute period is something else General Radio recommends for iron core inductors. When things get really strange and nothing is turning out right, I have done this and it does seem to settle the transformer core down and give reasonable answers.

@thomjmoore 6 лет назад

Have you ever considered doing a tutorial on LTSpice?

@manuelminutello3027 6 лет назад

Could you make a video about plate chokes? How do you determinate the loadline of a plate choke? 73 IU3IRR

@ElPasoTubeAmps 6 лет назад

I have always used the B&W 800 or the National R-175 choke. I found a really interesting article on chokes that might be interesting so the link is below. www.w8ji.com/rf_plate_choke.htm

@EJP286CRSKW 6 лет назад

You're also getting a small error by not taking into account the 600R source resistance of the SG505 ... or are you? Great video, great box.

@ElPasoTubeAmps 6 лет назад

Hi Esmond, I put a 10 ohm resistor across the oscillator to swamp that part and stabilize the voltage. It is attached via one of the Pomona type connectors on the distribution box. Internal resistance of the inductor surely plays a part and then these transformers appear to become slightly magnetized and then measurements start to skew. It's a challenge...

@pa4tim 6 лет назад

What about the real resistance of the transformer ? Is that so low you can forget it ? If I understand it well you measure the impedance, not the reactance. To come closer you can subtract the DC resistance . I think you know the math , something with complex numbers ;-)

@ElPasoTubeAmps 6 лет назад

I think you are absolutely right about the internal series resistance of the inductor - it does play an important part but as you mention, it is a vector and not a scalar and cannot simply be subtracted. I think an improved method might be to actually use the full complex (vector) differential equation of (I hope I can type it correct...) v(t)=Ldi/dt. That is easy to turn around into, as I have used it as a scalar value of, L=V/di/dt. That approach of measuring current was working really well in some cases but I think internal series R was getting in the way. Possibly an approach that would be more accurate would be to pulse the inductor and measure the slope of the current (so we really get the rate of change, di) and then measure time (dt) and go through the little bit of differential calculus which will change sin to cos and model in LTSpice for a better answer. Maybe that is next although as my friend, Bob mentioned, a saw-tooth waveform may excite instabilities in a real inductor that causes other difficulties... Whewww...