# Physically measuring reflected impedance

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### Windcrest77

Disregarding math from published specs and spice model techniques. So in practice, at the breadboard using physical components. Would a cheap impedance meter be a good way to actually measure reflected impedance? Here is a link to a popular meter the TOA. It seems to work fine at measuring the impedance of a coil at 1000Hz. I have several used OPT that I acquired at swap meets, many of which have multiple secondaries. Unclear markings, etc. I understand that when a OPT has multiple secondaries you can mix and match them in series or parallel to make the most use. But when you do that you need to then know what your reflected impedance becomes. Since I'm not a spice wiz, nor a math wiz, I thought there must be a way to simply empirically measure my resulting primary impedance at 1000Hz? I know about turns ratios and the base formulas involved and that is great if you know the specs of your OPT and you are not using multiple secondary windings, etc. I don't currently own an impedance meter, but I think I need one now. Seems like it will make life much easier. My objective is to do this all from an empirical, physical, measurable way, free of spice or specs followed by a little match. How do I actually measure reflected impedance? Basically hook up my secondaries to intended speakers, then measure. Here is the TOA video, will this do it or any of the other inexpensive handheld impedance meters used by audio installers?

Last edited:

#### Osvaldo de Banfield

IMHO, the easiest way to do it is, supposing you have an audio oscillator, inject a signal into the secondary (lower turns), and measure input and output voltage, whence you can easily reckon turns ratio (Vout/Vin). Then, given an output load resistance, multiply it to the vale squared, and you will get the load impedance seen from the anodes.

Example: Vin = 1V, Vout = 10V, then Vout/Vin = 10V/1V = 10
When the trafo is loaded with an 8Ω speaker, the impedance seen from the anode will be 10² * 8 = 800Ω.

You can also do it with line frequency via a variac and a isolation transformer.

Last edited:

#### smoking-amp

You need to know the turns ratios. Measure witn an Ohm-meter all the wires to diagram out what is connected to what and the Ohms across them. This will tell you primaries (Hi Ohms) from secondaries (low Ohms).

Then put some (isolated for safety) AC across the full primary and measure the voltages across the various windings or connected taps (including the primary)
From the voltage ratios, you can directly determine the turns ratios. Impedance ratios are just the square of those V ratios.

If you can guess which winding or taps was the 8 Ohm one, you can determine the impedance of the other windings via the various turns ratios squared.

The more tricky part is estimating the power handling ability and freq. response. Some test equipment needed for this. Although a Variac and ammeter with graphed results of current versus Voltage can give you a rough estimate. When the current starts curving upward rapidly versus applied voltage, you are reaching magnetic saturation. Do not go further. Operation of the xfmr as an audio OT should derate the voltage to about 70% of the saturation voltage. Then you need to convert the line voltage measured to the lowest frequency intended to convey. At half the freq. (from line source), then half the audio voltage. Ie, linear scaling.

Last edited:

#### Osvaldo de Banfield

Then put some (isolated for safety) AC across the full primary and measure the voltages across the various windings or connected taps (including the primary)

As tube output transformers are voltage step down, injecting a low voltage at the primary will give very low secondary voltage with high error measurements. I again suggest to inject the voltage from the secondary, which gives higher voltages easily measure them.

#### smoking-amp

I use a Variac to feed the primary, 120 to 140 max V typically.

I also use a current probe and scope to measure the saturation voltage level by looking for current peaking in the feed line. For this, using a Variac voltage injected into a winding (like 16 Ohm) that can be saturated readily. Carefully increase V from 0 (while watching the ammeter and scope for peaking) so as to not blow a fuse. (if one has a current probe and scope, much better for estimating safe insertion voltages than plotting current graphs)

Note: Magnetizing current and winding resistance will cause inaccuracy in the voltage measurements for turns ratios. I insert the excitation into two different windings initially (two separate test groups), and only record voltage ratios between the zero current windings.

Last edited:

#### Windcrest77

IMHO, the easiest way to do it is, supposing you have an audio oscillator, inject a signal into the secondary (lower turns), and measure input and output voltage, whence you can easily reckon turns ratio (Vout/Vin). Then, given an output load resistance, multiply it to the vale squared, and you will get the load impedance seen from the anodes.

Example: Vin = 1V, Vout = 10V, then Vout/Vin = 10V/1V = 10
When the trafo is loaded with an 8Ω speaker, the impedance seen from the anode will be 10² * 8 = 800Ω.

You can also do it with line frequency via a variac and a isolation transformer.

Thanks Osvaldo, I'll try it with my function generator as that indicates voltage right off and its variable. If the OPT has multiple taps on a single winding I should be able to determine which one is the lowest tap, then just use an ohmmeter to id the other taps up the coil.

Also some more esoteric OPTs have what I assume are multiple separate secondaries, (not simply a tapped single-secondary). On the Sowter site they suggest using all the secondaries in series or parallel for best high frequency transfer by not leaving a "dark" unused secondary there just eating up current. If I come across such an OPT as that... I assume I just wire the secondary coils in series/parallel first, that will set up the secondary as a single coil, then use the same procedure you described to calculate whats reflected?

#### Windcrest77

You need to know the turns ratios. Measure witn an Ohm-meter all the wires to diagram out what is connected to what and the Ohms across them. This will tell you primaries (Hi Ohms) from secondaries (low Ohms).

Then put some (isolated for safety) AC across the full primary and measure the voltages across the various windings or connected taps (including the primary)
From the voltage ratios, you can directly determine the turns ratios. Impedance ratios are just the square of those V ratios.

If you can guess which winding or taps was the 8 Ohm one, you can determine the impedance of the other windings via the various turns ratios squared.

The more tricky part is estimating the power handling ability and freq. response. Some test equipment needed for this. Although a Variac and ammeter with graphed results of current versus Voltage can give you a rough estimate. When the current starts curving upward rapidly versus applied voltage, you are reaching magnetic saturation. Do not go further. Operation of the xfmr as an audio OT should derate the voltage to about 70% of the saturation voltage. Then you need to convert the line voltage measured to the lowest frequency intended to convey. At half the freq. (from line source), then half the audio voltage. Ie, linear scaling.

Thanks for the power measuring method, I haven't gotten that far yet in my needs here because I am just testing a as many flea-watt/headphone circuits as I find time. For a computer desk amp servicing either near-field speakers (with a small SS sub below 80Hz) or headphones for at night. Most transformers I have are minimum 10 watt just based on appearance.

#### Tubelab_com

I use a Variac to feed the primary, 120 to 140 max V typically.

When doing this it is a good idea to place a load resistor across the secondary, whose value is somewhere in the 20 to 100 ohm range. The line voltage has a lot of high frequency noise that can cause ringing in the OPT making the meter read a somewhat higher than expected value.

#### directdriver

When doing this it is a good idea to place a load resistor across the secondary, whose value is somewhere in the 20 to 100 ohm range.

I was wonder how to determine the value on a loading resistor, on either primary or secondary, to get the proper turns ratio. Without a load, everything tends to measure a touch higher.

#### Tubelab_com

Several years ago I reported some rather unbelievable power output numbers from common tubes.....things like 100 watts from a pair of 6L6GA's. Several readers called BS to the point that I doubted my own numbers. Never mind that I had made my career on getting notoriously hard to measure stuff right the first time, this was a 1 KHz sine wave. One by one possible measurement errors or uncertainties were eliminated until the actual primary impedance of the OPT was all that was left.

I took several known brand name OPT's and devised ways to measure them until I had multiple different ways to arrive at the published ratings with low error. Then I tested the junk box OPT's that I had been using to torture these tubes. These transformers were specified to be 6600 ohms to 0-4-8-16, and I was running an 8 ohm load on the 16 ohm tap, which should be a 3300 ohm OPT.....and it was within a few ohms.

The simplest method that gives the right answer was to plug the full primary into isolated line voltage and measure the secondary voltage. A resistor that was close to the expected load gave the most correct reading, but anything close, say 8 to 50 ohms on the 8 ohm tap gave similar readings.

Note that this all happened before CFL and LED lamps along with flat panel TV sets were common. Those tend to pollute the line power worse than anything I saw back then.

The big power numbers were real and the result of driving the tubes deep into AB2 operation. Some tubes like AB2, and some don't.

#### DF96

An unused secondary does not eat up current, and I doubt if Sowter said that it did. It merely avoids the opportunity to have a reduced secondary resistance and so may reduce voltage a little.

#### Osvaldo de Banfield

Note that this all happened before CFL and LED lamps along with flat panel TV sets were common. Those tend to pollute the line power worse than anything I saw back then.
And consider that you live in a country where laws must be respected. Here, doesn't occur the same, and myriads of led lightning including house and street lamps without line filtering, heterodyning between them in the same area, and with variable frequency as they (I suppose) have RC oscillators. Ham HF band are nowadays UNUSABLE.

#### Tubelab_com

you live in a country where laws must be respected.

I'm guessing that our light bulbs and fixtures come from the same place in Asia tha yours do, and cause pretty much the same problems....Yes, there are RF emission regulations, but actual enforcement is rare.

As it is in most places, big money gets it's way.

Somewhere in the mid 1990's the power company put in this set of 765 KV transmission lines right across the street from my house in Florida. As soon as they went live 160, 80 and 40 meters became useless. The "S" meter was pegged on 160 and 80 with only a loud 60 Hz buzz from the speaker. 40 was marginally useful on a cool dry night, which rarely occurs in south Florida. During the summer rainy season the only contacts I could make were from Europe in the early morning using JT-65 on 14.076 MHz. Much of my experimentation was done in the VHF and above region.

AT&T killed all use of 902 and 2304 by planting a cell tower right across the street. It also killed any hopes of EME contacts on the higher bands since the large amounts of AM modulated RF in the air required a filter in front of the LNA, raising my system noise floor by an instant 2 dB or so.

#### Attachments

• 2012-12-08 11.53.24_x.jpg
662.4 KB · Views: 64
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.