Transfrm and Mains Freq

[Tom Phillips]

... we still use the old generating system left by the old colonial power that ruled us, and that is a 100v 25Hz plant.

That is exactly what is needed to power the Christmas lights in Whoville.

[J M Fahey]

Well, *there is* such a Country as Enzonia, although it's more easily known as "Brazil".
Half of it 220V ; half of it nominal 110V (it *used* to be real 110V) but now for fun (life would be boring otherwise) it's actually 127V .
All those old 110V Giannini amps, straight copies of blackface Fenders and factory "upgraded" to 500/530V +V rails with real 110V (or lower) wall voltage, explode like grenades when plugged into modern 127V.
Old electrolytics add to the fun, of course.
All amps sold there customarily carry "110/220V" switches *in the front panel* , so you don't forget to switch them.
Big Sao Paulo city is mostly 110V (127) while Santos, it's port suburb, is 220.
There actualy are streets where one side is 110, the houses in front are 220.
The "frontier" must lie *somewhere*, after all.
That's nothing, I visit every year their ExpoMusic, a big NAMM type event, inside a 3 or 4 block convention center.
Which sits right in the middle of 110V Sao Paulo .... but is internally wired for 220 .
Now you'd expect a monster poster or billboard at every Expo door stating "you are entering a 220V place" or something .... not a hint.
Musicians who go there in a hurry to demonstrate something often blow their own equipment .
It happened to a friend of mine while I was there, a helpful stage hand plugged and turned on his Vox Valvetronix 100 amp to "save him some setup time"
The booth owner refused all responsibility because "the wall jacks were properly labelled 220V" with thumbnail sized (no kidding) black and white computer printed labels stating so.
I was just visiting, 2000Km away from my bench, so I had to go buy a cheap soldering iron, $20 multimeter, etc. to repair it for free (hey, I was staying at his home).

EDIT: to add to the fun, (hey, Brazil is famous for being fun), they have the largest overseas community of "Nisei", Japanese residents outside Japan proper, and it's customary for kids to travel to Japan for one year after they finish High School, to kep tradition alive, practice the language, visit Family, and of course work for 1 year at Toyota, Nissan, Sony or whatever, which looks good on the resume afterwards.
And what do kids buy when they have a few U$ (ok, Yen) free?
Yes, they buy a guitar and a Marshall.
A 100V one by the way, with no extra taps, not even 110 or 120V ones.
And when back home, many plug it straight into the 127V wall outlet.
Just imagine the fireworks, doubly so because some "repair" them with penny nail fuses or similar.
Oh well.

[Alan0354]

There should be loss in efficiency when going from 60 to 50Hz on top of everything said here. It might not be important for most people.

[trobbins]

efficiency of ???

[trem]

JMF.....thanks for the stories.
Brazil sounds like an Electricians Nightmare.
Never a dull moment I guess.
Interesting fact(s) you bring up about The Japanese population.
Learn something new everyday I guess :-)
thanks again

[Alan0354]

efficiency of ???

Output power vs input power. Same idea about the iron core you guys talking about. Lower frequency result in drawing more current from the primary. That's the reason you need more winding and/or bigger core. Like you guys said, the transformer gets warmer with lower freq. More power is loss to heat means less efficiency.

Actually, even at 60Hz, the efficiency is quite low, I believe it's only about 30% to 40%. But don't quote me on this. That's the whole thing about switching power supply, they work at 40KHz, they need a lot less turns, using a much much smaller core and average efficiency of a switcher is at 75% or so. Just the transformer alone is even higher.

We design high voltage power converters, we get about 5V per turn!!! That is for a 24V input transformer, there is only 5 turns in the primary!!!! A 100W transformer is about 1.5"X2". And that was only because the secondary is 5KV and we need a lot of turns.

[trobbins]

For a transformer that is designed to operate with 50/60Hz, and at the operating voltage, the extra loss operating at 50Hz (versus say 60Hz) is I would suggest insignificant compared to say changing the bias on the output stage by a few percent (Watts dissipation wise), or raising the heater voltage by say 0.1V.

I think the value here is to be more aware of when operating conditions go outside of design intent, and to have a general appreciation of how the mains input current waveform can start to move away from sinusoidal, with a growing peakiness at 'top-dead-centre' of the voltage waveform as the core BH operating point moves ever more in to and out of the saturation region each half cycle.

On a tangent, if it was the OT starting to overdrive in to the OT core saturation region, then some guitarists would call that extra mojo. If the PT was starting to do it, but in a non-destructive manner, then I'd be thinking some guitarists would be also hearing extra mojo - probably in the same way as guitar amp listening tests have often preferred the influence of power supply ripple additives in their sound, compared to a perfectly filtered power rail.

[R.G.]

I have a 1948 Hammond catalog that lists some transformers as designed for 25 "Cycles". The 25 Cy transformers weigh quite a bit more, like 15 pounds vs 9 pounds for the 50/60 Cy unit.

The frequency put out by an alternator (that what makes our electric power these days in almost all cases) is completely determined by the number of poles and rotational speed. I hated learning all this back in my sophomore "motors" class, but it's turned out to be really useful over the years. In the early days of electric power when Edison was sending out DC, George Westinghouse set up an AC system. I believe that the Westinghouse system was 25Hz because that was what was practical to build the (for then) huge rotating alternators to generate it.

25Hz was dropped when it was realized that you could make your loads, motors and transformers, half as heavy if you'd double the frequency. Making alternators was better by then, so they did, and saved a lot of iron and copper. But 25Hz hung around in isolated pockets for a long time, before the entire country was connected into two power grids. That's where the 25Hz stuff came from. It was an ugly older sibling.

Many times a transformer that has only a 120V primary will be 60Hz only but one with dual primaries will specify 50/60Hz. This is a cost cutting move by the manufacturer to maximize profit in North America.

There should be loss in efficiency when going from 60 to 50Hz on top of everything said here. It might not be important for most people.

That gets slippery. There may be an efficiency loss in the transformer, maybe not. It depends on the design and the iron. If the design is conservative (that is, designed to not get into the edge of saturation) at 50Hz, the loss will be minimal. If it's a 60Hz design on the edge of saturation and the manufacturer simply decides to let it run hot, yeah, there's a loss. There are other things to consider for overall amp efficiency with frequency changes.

Actually, even at 60Hz, the efficiency is quite low, I believe it's only about 30% to 40%. But don't quote me on this.

I won't - it's incorrect. A well designed 50/60Hz mains transformer usually runs well over 95% efficient, and they can easily run 99%. Calculation of iron and copper losses is a stock item for mains power transformers, and almost as well known as V = I*R, if a little more tedious to compute.

That's the whole thing about switching power supply, they work at 40KHz, they need a lot less turns, using a much much smaller core and average efficiency of a switcher is at 75% or so. Just the transformer alone is even higher.

We design high voltage power converters, we get about 5V per turn!!! That is for a 24V input transformer, there is only 5 turns in the primary!!!! A 100W transformer is about 1.5"X2". And that was only because the secondary is 5KV and we need a lot of turns.

Well, there are quite a few things contained in the whole thing about switching power supplies. The basic efficiency of a transformer - at any frequency - is power out divided by power in. The difference is the wire loss, core loss, and radiation losses. Wire losses are pure I²R, but you simply must do the math to get the true RMS currents in both primary and secondary to compute this. The secondary current depends on the load, rectification and filtering. The primary current is the reflected version of the secondary, plus core excitation losses and radiation losses.

The excitation losses are the price you pay for transforming. To have a transformer, you simply have to pump up the magnetic field in the core, and this has to be AC. The core has losses depending on the material and the amount of its B-H curves being used by the excitation. "Dumb" 50/60 Hz cores have a dead simply relation of core loss to excitation because they're always being driven symmetrically about 0, so the core losses depend purely on (1) the shape of the B-H curve for the material, (2) how close to the edge of the saturation lip the primary volt-turns drive the core, and (3) the physical configuration of the core material in resisting eddy current losses.

Radiation losses are the unavoidable losses of input current spent in projecting M-field leakage out of the core, capacitive leakages through the coils, and any capacitive leakage out of the coils to the rest of the world. These are tiny with low frequency transformers, but get big in switching power supplies.

Switching power supplies generally have to go to ferrite cores because eddy current losses would melt iron cores at those power levels and frequencies. Induction heaters are devices which melt metals and other things by forcing an alternating magnetic field through them and melting them with the eddy current losses. The only good way around eddy current losses is to reduce the size of the conducting regions inside the core. This is why mains transformers are laminated - it makes eddy current losses smaller. At higher frequencies you need thinner laminations, which is why output trannies sometimes use much thinner lams than mains transformers. The next step up is sintered iron dust, then bound iron dust. Then you have to step to ferrites to get higher resistivity to avoid crushing core losses.

Ferrites have about 1/10 to 1/5 the "magnetic goodness" of silicon-steel laminations, but you're forced to use them because the iron would be unavoidably lossy at higher frequencies. This forces other consequences too.

High frequency transformer design has its own set of gotchas, and so does high voltage transformer design.

All this is easy - if tedious - to compute.

On OT Saturation:
I suspect that this almost never happens in a musical amplifier. Transformer saturation might be a nice symmetrical soft compression, but transformers saturate from the low end first. Unless an OT is marginal for the lowest frequencies fed to it, it doesn't saturate. Even if it's designed to be marginal to saturate, this is a bass effect that you can't get at higher frequencies, and that you probably can't through the preamp/PI stages to the output stage.

What saturates a transformer core is the size of the volt-time integral. If the voltage is fixed, as it is by the power supply in an amplifier, then the only way to increase volt-time is to lower frequency. Saturation edge is reached almost linearly with the volt-time integral (the "almost" being from the curvature of the B-H characteristic), so if you have the ability to just saturate with a low E, then it takes twice the signal voltage to saturate an octave up, and four times to saturate with two octaves up. the power supply doesn't stretch that far.

If an OT is designed to saturate at say, middle E (162Hz, about), then the output tubes are in trouble with low Es, unless you've done funny stuff in the preamp to try to match the size of the incoming signal to the linear falloff in saturation point with declining frequency. While this is possible, I've never seen an amp design that does it intentionally. And most preamps deliberately cut lows in the path from input to PI. This makes OT saturation even harder if the OT would saturate at the bottom notes.

As an aside, just to start some thinking, you cannot saturate a transformer from the secondary. No combination of secondary loading will cause saturation in a transformer that is being fed a non-saturating volt-time on the primary. You can cause similar incindiary effects by pure overheating/overloading, but it's not saturation.

[Alan0354]

Yeh, I meant to come back and correct the 30% efficiency of the 60Hz transformer, that's the number of a linear regulated power supply, the transformer is much better. I am not an expert of transformer and I am not trying to be. I was just talking about when you lower the freq from 60 to 50Hz, you need a little bigger size so you are not running to the edge.

[R.G.]

I understand. Yes, linear regulated power supplies are often very inefficient. You are correct - the lower the frequency, the bigger the transformer has to be for equal primary voltages.

[Tom Phillips]

Another fact to consider about transformer efficiency is that it depends upon the loading. For example, if the current being drawn from the secondary is light then the magnetization current flowing in the primary could be a significant portion of the overall primary current draw. Taken to the limit of zero secondary current such as occurs in a wall wart transformer when the device it powers is turned off we have power being drawn but none being delivered. That's 0% efficiency. There are other factors that affect the efficiency of course.

My point is that the efficiency number published by a transformer manufacturer is usually the maximum efficiency for the device and it only applies at one point in the operating curve. This is not usually a factor in guitar amplifier design. I mention it to illustrate that specifications are often dependent upon other factors. It sticks in my mind because of a time when I was evaluating the efficiency of a standard EI lamination transformer vs. a toroidal transformer for a project. The two transformers were those available from the junk box. In general a toroid would be more efficient if both were designed for the same application. It turned out that the standard transformer was more efficient for the particular application because the toroidal has such a larger VA rating that it would have been running at such a smaller fraction of it's capability that it would not achieve its efficiency potential.

Cheers,
Tom

[loudthud]

All amps sold there customarily carry "110/220V" switches *in the front panel* , so you don't forget to switch them.

In the US there are some odd plug/recepticles that are sometimes used for 240V. They are configured so that a 120V plug won't fit into a 240V outlet. This worked when power cords were always attached to the appliance. Then came the IEC standard that permitted detachable power cords with a universal connector on the back of the appliance. Are the outlets in Brazil configured so that a 120V cord can be plugged into a 240V outlet?

[R.G.]

Another fact to consider about transformer efficiency is that it depends upon the loading. For example, if the current being drawn from the secondary is light then the magnetization current flowing in the primary could be a significant portion of the overall primary current draw. Taken to the limit of zero secondary current such as occurs in a wall wart transformer when the device it powers is turned off we have power being drawn but none being delivered. That's 0% efficiency. There are other factors that affect the efficiency of course.

Good point. Efficiency is often quoted only at full load, where it looks best. There is some rationale behind that, as full load is often the place where the most power is wasted. The "80 plus" specs on computer power supplies take this into account and require over 80% at all powers over ... um 10%?? Have to go look.

Transformers design actually amounts to setting a nearly-fixed core loss for the given primary Volt-time point, +/- input voltage variation. The iron is worked at near saturation all the time, and eddy currents are pretty much fixed. The wire losses are variable, and start at the minimum of no load, only the magnetizing current running in the primary; the wire losses go up from there to the design maximums at full current.

The "full current", by the way, is the current which over time produces the design-maximum temperature at the hottest spot inside the trannie for transformers which deal primarily with power. Transformers are run as hot as the insulation will (practically, reasonably, and reliably) run and still have a reasonable lifetime.

My point is that the efficiency number published by a transformer manufacturer is usually the maximum efficiency for the device and it only applies at one point in the operating curve. This is not usually a factor in guitar amplifier design. I mention it to illustrate that specifications are often dependent upon other factors. It sticks in my mind because of a time when I was evaluating the efficiency of a standard EI lamination transformer vs. a toroidal transformer for a project. The two transformers were those available from the junk box. In general a toroid would be more efficient if both were designed for the same application. It turned out that the standard transformer was more efficient for the particular application because the toroidal has such a larger VA rating that it would have been running at such a smaller fraction of it's capability that it would not achieve its efficiency potential.

Absolutely correct.

It's common for the no-load losses in a conservatively designed power transformer with Class A/105C insulation to be from 1% to 5% of full load power.

[trem]

120v is the standard house current in the USA, but most homes do have 240 coming in. The 120v is one side of the 240. Your breaker box alternates one phase or the other. Half your breakers are off one phase and half off the other.

That is something I never even considered.
I imagine there is reasoning behind the decision somewhere.
I hope this is not a dumb question, but.....
Does it help...further back the line, or something...to spread current use somewhat evenly across both halves of the 240 Volt supply.?
Thank You

[R.G.]

That is something I never even considered.
I imagine there is reasoning behind the decision somewhere.
I hope this is not a dumb question, but.....
Does it help...further back the line, or something...to spread current use somewhat evenly across both halves of the 240 Volt supply.?

There is a long thread of reasoning behind it, and the quirks are often buried.

It turns out that (thank you, Mr. Tesla!) three-phase is the lowest number of AC phases that's economically advantageous to distribute over long distances. Most power distribution is 3PHase except for some oddities and some very recent very high voltage DC distribution. So it gets out to the world as 3-phase. Three-phase can be star (three windings which all share one middle "neutral" point and three unconnected outputs) or "delta" - only three windings, with no "middle neutral". Delta is more economical to distribute because you can leave out the fourth wire, but you have to be really careful about balance, because the neutal in star is what carries any differences in current in the phases. Delta needs close balance, but is one wire cheaper.

Most neighborhoods have 3-phase, and from that make three 240Vac isolated phases. Each 240 phase is centertapped, so you get in effect, six phases of 120V with a "neutral". The neutral is grounded at the transformer, and also at the house. Each house gets both sides of the same 240Vac and the neutral. Each end of the 240 is balanced around the neutral, and from end to end is 240Vac. You can now power *both* 120Vac low power stuff and also the higherr power 240V air conditioners, electric driers, ovens and cooktops. When houses are wired, one of the things the electrician is supposed to do is to add up the loading on the red and black phases (those are the colors; neutral is white) and for "typical" scenarios provide close-to-balance. In addition, all the power on both phases comes from the 240V back at the distribution transformer, so imbalances may move the effective centering of the neutral, but cannot imbalance the distribution transformer. Generally, equal numbers of houses of about equal expected power are connected to each phase of the distribution transformer, often three houses per transformer.

So - yes, you're right, it was to get lower power limits in the wall wiring with 15A-per-breaker 120V, and high power 15A-per-breaker 240V, and to spread the loading around.

At least that's what I remember of the Motors course and my sometimes brushes with the wiring codes.