Output Transformer Question

[John Fisher]

In my finite understanding I know that with an OT, The impedance is determined by the turns ratio of the primary and the secondary windings. I was wondering, what effect is there with the impedance remaining the same supposedly matching the tubes in a Push pull situation only with fewer turns in the primary and fewer in the secondary. I figured that you would at least need thicker wire but how would the performance be with fewer windings.
Thanks for your response.

[Steve Conner]

The turns ratio determines the impedance ratio. If you want to match a 100 ohm source to a 1 ohm load, you need a turns ratio of 10:1.

If this 10:1 transformer were ideal, then you could also use it to match a 10 ohm source to a 0.1 ohm load, or a 1000 ohm source to a 10 ohm load, and so on.

But real-world transformers are not ideal, they have annoying characteristics like leakage inductance, magnetizing inductance and DC resistance. The number of turns and the size of the iron core are two important variables that the transformer designer can tune to get the compromise he wants.

Example: let's take an OT that was well designed, and halve the number of turns in both windings. The magnetizing inductance, leakage inductance and DC resistance all go down by a factor of 4. The volt-second product for saturation is halved.

If the transformer was rated to cover 20Hz-20kHz before, then it is now better suited to the range 80Hz-80kHz.

You could get the bass performance back by using the transformer at a lower impedance level. Say it was a 50 watt 3k:8ohm OT before. After halving the turns it is a 660 ohm:2 ohm one. If you used it with a 660 ohm source and 2 ohm load, the bandwidth would go back to 20Hz-20kHz.

You could also get the bass performance back by using a bigger core. You have now scaled up the whole OT, making it appropriate for a higher powered amp.

Transformer design is a craft like pickup winding, but at least it is a 100+ year old craft and very well documented. The RDH is still the best source I've seen for output transformer design.

Simplifying grossly, if you know the lowest frequency you want to pass at full power, the universal transformer EMF equation will tell you how many turns you need for a given cross-sectional area of core.

[R.G.]

Good write up Steve!

[John Fisher]

Thanks Steve! Very helpful reply.

The turns ratio determines the impedance ratio. If you want to match a 100 ohm source to a 1 ohm load, you need a turns ratio of 10:1.

If this 10:1 transformer were ideal, then you could also use it to match a 10 ohm source to a 0.1 ohm load, or a 1000 ohm source to a 10 ohm load, and so on.

But real-world transformers are not ideal, they have annoying characteristics like leakage inductance, magnetizing inductance and DC resistance. The number of turns and the size of the iron core are two important variables that the transformer designer can tune to get the compromise he wants.

Example: let's take an OT that was well designed, and halve the number of turns in both windings. The magnetizing inductance, leakage inductance and DC resistance all go down by a factor of 4. The volt-second product for saturation is halved.

If the transformer was rated to cover 20Hz-20kHz before, then it is now better suited to the range 80Hz-80kHz.

You could get the bass performance back by using the transformer at a lower impedance level. Say it was a 50 watt 3k:8ohm OT before. After halving the turns it is a 660 ohm:2 ohm one. If you used it with a 660 ohm source and 2 ohm load, the bandwidth would go back to 20Hz-20kHz.

You could also get the bass performance back by using a bigger core. You have now scaled up the whole OT, making it appropriate for a higher powered amp.

Transformer design is a craft like pickup winding, but at least it is a 100+ year old craft and very well documented. The RDH is still the best source I've seen for output transformer design.

Simplifying grossly, if you know the lowest frequency you want to pass at full power, the universal transformer EMF equation will tell you how many turns you need for a given cross-sectional area of core.

[Enzo]

And remember that the transformer has no impedance of its own, just ratios.

[Steve Conner]

Well, I just tried to explain that it does have an impedance of its own. I guess not so much an impedance, as a range of impedances that it's best suited for.

I've heard it said that transformers are like gearboxes. A F1 car might travel at 200mph with the engine spinning at 10,000rpm. And a mining dump truck might do 20mph with its huge Caterpillar engine thundering at 1000rpm.

So it follows that the transmission from the F1 car ought to work in the dump truck, right? Same gear ratios.

[Chuck H]

It's probably worth mentioning that WRT guitar amps, fudging is usually OK. For example, it's common practice to pull a pair of tubes from a four tube amp and then double the speaker load WRT the impedance setting to "correct" the impedance. This can be done because transformers work on ratios.

And my dump truck can go faster than 20. But my race car can't go 200. Of course I used the same transmission for both.

[Enzo]

Oh, not arguing with you. Obviously the ratio is not the whole equation. Just pointing out that there is nothing inherently 4k ohm impedance about a transformer winding. What MAKES it a 4k impedance is the ratio and the load on the other end.


Just as that gearbox doesn't demand a car go 43 miles per hour. It only goes 43 miles per hour if the engine speed is appropriate.



I remember some 40 years ago or more, they were holding road rallies. And there was some sort of formula for handicapping the cars, and it involved something like car weight and horsepower and maybe gas mileage. I forget the details: ton-miles per gallon maybe. SO it behooved the racers to use small lightweight cars like little MGs or something to take best advantage of the formula. Then one year someone figured out that the other end of the formula could be used, so they showed up in a cement mixer. It was really heavy, got poor mileage, but had lots of power - or something. That's it, got really low mileage, but moved a lot of weight in the proces. They won the race. The rules were amended by the next season, but just goes to show you...

[kg]

Well, I just tried to explain that it does have an impedance of its own. I guess not so much an impedance, as a range of impedances that it's best suited for.

I've heard it said that transformers are like gearboxes. A F1 car might travel at 200mph with the engine spinning at 10,000rpm. And a mining dump truck might do 20mph with its huge Caterpillar engine thundering at 1000rpm.

So it follows that the transmission from the F1 car ought to work in the dump truck, right? Same gear ratios.

i actually like this analogy.

gearboxes don't care a whit about horsepower. they only care about torque.

you might say that the "torque" represents magnetic flux.

it could also be the winding current density (amps per cross sectional area).

those are your two real limiting factors... resulting in saturation and shorted/open windings respectively.

[Chuck H]

That clarifies my point with the dump truck/race car re-reference joke. You could make a transmission that would "work" for both. But it wouldn't be at all ideal for both. This also is often done with off the shelf transformers (not unlike what was used in many vintage guitar amps). The same transformer used for a pair of 6V6's at 16 watts with an 8k primary may have been used for a pair of 6l6's at 40 watts with a 4k primary.

For the record, I'm not poohing the importance of the details. I just wouldn't want the OP to go from this thread with the thought that everything needs to be just so or an amp will sound bad or explode. There is the occasional post here where someone building a guitar amp is trying to calculate very finite OT parameters based on the operating conditions, And it really isn't necessary. Close is fine. Maybe better sometimes. Hard to know before you hear a result. Always within safety margins though.

[Steve Conner]

Yes. A guitar amp OT is more like your trusty pickup truck that gets used as a dump truck during the week and a race car at weekends.

[Chuck H]

I had one. A '71 Ford F100 with a 302. I loved that truck. Sold it cheap and bought a van when I moved to the PNW where it rains all the time. I hate driving a van.

[J M Fahey]

halve the number of turns in both windings. The magnetizing inductance, leakage inductance and DC resistance all go down by a factor of 4

Sorry but inductances are quadratic and will go down by 2 squared (=4) ; resistance is linear and will go down by a factor of 2.
Obviously, if all you do is "halve turns".
No mention was given of increasing wire diameter.
jm2c

[kg]

nice catch JM!

keepin' us honest...

[Steve Conner]

I thought it was only fair to pick a heavier gauge of wire that would make the same use of the winding window as before.

So twice the cross-section, half the length, one quarter the resistance.