Can someone enlighten me about what happens to a transformer when the secondary is heavily loaded (short term duration load only, so over-heating should not be a factor). I understand that the voltage may drop and regulation will be poor.
I am wondering how the magnetic circuit will behave under brief high loading - any potential problems? For instance, let's say the transformer has 50V dual secondaries each rated for 2A. What happens if you try to draw 4A in short bursts? Is this doing anything bad to the transformer, or operating it in an "unsafe area"?
I am wondering how the magnetic circuit will behave under brief high loading - any potential problems? For instance, let's say the transformer has 50V dual secondaries each rated for 2A. What happens if you try to draw 4A in short bursts? Is this doing anything bad to the transformer, or operating it in an "unsafe area"?
Nothing bad at all for the magnetics in my understanding. Drawing current out reduces the B in the core - that's at its highest off-load. I reckon your issues are thermal ones - copper losses.
Oddly enough, something similar is what happens all the time in capacitor filtered rectifier supplies, only the big current pulse happens 120 times a second. If you were to gate the 2X rated load at 50% duty cycle, everything would be quite ordinary. Of course any secondary overload causes primary overload, so if pulse power gets too extreme and long enough duration it might be hard to protect the primary with a fuse and of course losses go up. At some point a larger transformer is required, but for short periods transformers can be massively overloaded without breaking anything. That's one of the reasons simple rectifier supplies are so reliable.
Heating the copper of both primary and secondary are the only issues. Note that there are two thermal limits. The normal steady-state one is set by the rate at which heat can diffuse away from the windings. The pulse one is set by the thermal capacity of whatever you are heating. The time boundary between the two is set by thermal time constants but you are unlikely to stray too far away from the steady-state situation unless you are doing something unusual.
What about PS noise and VA stage clipping from voltage drop, agree that temp may not be an issue because of short duration ...
The amplifier is at the other end of the supply cables.
The amp will have it's own local decoupling. This provides all the transient current demand and meets the short term non transient demand.
The PSU simply recharges the decoupling back up to normal voltage.
It's the enormous overload capability of a linear PSU that completely outclasses SMPS.
The amp will have it's own local decoupling. This provides all the transient current demand and meets the short term non transient demand.
The PSU simply recharges the decoupling back up to normal voltage.
It's the enormous overload capability of a linear PSU that completely outclasses SMPS.
The OP was asking about the effect on the transformer. Of course, voltage droop will have other well-known effects on the circuit.a.wayne said:
.
Actually, heat is pretty much the main issue.
What happens to the magnet circuit is that the magnetic field grows stronger with increased current flow, and affects a wider area. There are horribly complicated formulas to for calculating this, but the bottom line is that more current = more magnet effect. Which in this case doesn't matter.
The problem here is the same as with any conductor, whether straight or wound into a coil. Current flow heats wire, simple as that.
With transformers it's purely a mechanical problem. The turns on the inside of a coil are insulated by the turns on the outside of the same coil. Sure the material used is copper wire, which conducts heat [and electrical energy] well, but the outside turns are being heated just as much as the inside turns, so cooling is a balancing act. The engineers who design transformers worry a lot about this.
The result of excessive current flow is excessive heat, which reduces the life not of the wire itself--except in extreme cases--but of the insulation on the wire. All of which results in guaranteed early failure.
The rule of thumb with all electrical devices is 80%. A rating is always stated, but is never used. Multiply the rated load by 80%, and that's the safe continuous working load.
For instance, a 50 volt transformer rated at 2 amps (100 watt transformer) would be specified to work continuously at no more than 80 watts, which at 50 volts would be 1.6 amps. Work it any harder and you're running maxed out, which always means you're taking your chances.
.
Actually, heat is pretty much the main issue.
What happens to the magnet circuit is that the magnetic field grows stronger with increased current flow, and affects a wider area. There are horribly complicated formulas to for calculating this, but the bottom line is that more current = more magnet effect. Which in this case doesn't matter.
The problem here is the same as with any conductor, whether straight or wound into a coil. Current flow heats wire, simple as that.
With transformers it's purely a mechanical problem. The turns on the inside of a coil are insulated by the turns on the outside of the same coil. Sure the material used is copper wire, which conducts heat [and electrical energy] well, but the outside turns are being heated just as much as the inside turns, so cooling is a balancing act. The engineers who design transformers worry a lot about this.
The result of excessive current flow is excessive heat, which reduces the life not of the wire itself--except in extreme cases--but of the insulation on the wire. All of which results in guaranteed early failure.
The rule of thumb with all electrical devices is 80%. A rating is always stated, but is never used. Multiply the rated load by 80%, and that's the safe continuous working load.
For instance, a 50 volt transformer rated at 2 amps (100 watt transformer) would be specified to work continuously at no more than 80 watts, which at 50 volts would be 1.6 amps. Work it any harder and you're running maxed out, which always means you're taking your chances.
.
.
Hold everything. I just realized what I think you might really be asking.
With audio circuits, a too-small power supply results in distortion. How much, or whether-at-all, is unpredictable, it depends on the circuit and the components.
.
Hold everything. I just realized what I think you might really be asking.
With audio circuits, a too-small power supply results in distortion. How much, or whether-at-all, is unpredictable, it depends on the circuit and the components.
.
With audio circuits it is the peak current available to drive transients that is important. Most full range audio signals are 1/5 to 1/7 peak to average. Some supplies are undersized deliberately as a cheap way to protect the output stage from being overdriven into failure.
I have seen some commercial "undersized" supplies described as being constructed that way to provide a lot of "dynamic headroom" for short durations. For instance early NAD amps.
Undersized power supplies are often the result of economic pressures, marketing palaver makes the resulting performance limitations seem like an advantage. (I worked for a large audio company for many years.) An undersized power supply may also allow you to reduce the size of the heat sink for further savings. You get lots of dynamic headroom, but of course the secret is that that headroom power rating could well have been the rated continuous power with a larger supply and sufficient heat sinking. (Assuming adequate SOA in the output devices and drivers, another area where corners may be cut to save money.)
If I read your response correctly, you are inferring that the manufacturer did a "cop out" by cutting corners and reducing costs. However, real music signals (or at least natural uncompressed music) has a large crest factor, and this is best reproduced by a system with significant headroom of 20dB or more over average levels. Peaks are short in duration, and this is a very good match for the "undersized power supply" and an amp that could have been IMO "over supplied" for continuous duty.
I see it as a natural fit to music playback.
I agree with Kevin. Saying a soft PSU would give headroom seems cynical to me.
A stiffer PSU would not cut any headroom, it provides a stable platform for the
circuit to work as intended. A soft PSU is instable, voltage will vary with load.
A not constant PSU makes the claim "headrom" ad absurdum as in fact it reduces
headroom.
Certainly not what you want. Still the only arguements for smaller trannies are cost
and weight, which in the end is almost the same as cost. 😉
I have a 30 VAC 2amp dual secondary transformer I use in an amp test rig.
At this current any faults on new amps just make the transformer hum but don't blow up components like mosfet output transistors.
I attach a new amp then power up. If there is a loud hum from transformer then there is a fault and I turn off immediately. No blown components, no blown fuses and no fuss.
At this current any faults on new amps just make the transformer hum but don't blow up components like mosfet output transistors.
I attach a new amp then power up. If there is a loud hum from transformer then there is a fault and I turn off immediately. No blown components, no blown fuses and no fuss.
I don't agree with your negative feelings about a "soft" PSU. The salient question is what is the duration of the peaks? Music has brief high power transients and a much lower constant power demand. This is why the average power for in-home stereos is often on the order of 1W only. Why "overdesign" a power supply so that the amplifier can supply over 100W continuously when it will never be needed (continuously). The transformer needs to accommodate the average current demand AND have sufficient voltage for peaks. This means VA is sized for the average power plus a little extra and the rail voltage is sized for the peak power needs. If the rails sag it is only for a brief moment and then they recover - this will not impact the performance of the amplifier in terms of THD, bandwidth, etc.
Hi,
There is no point IMO in saying a "soft" PSU is poor compared to a "hard"
PSU that turns a music amplifier into a laboratory amplifier and costs more.
The extra expense is best spent on improving the music amplifier.
A typical case would be a multichannel receiver that the PSU hasn't
a hope in hell in driving all channels continuously but work fine
most of the time. A big supply would be a minor improvement.
An uber supply and uber heatsinking, pretty pointless for purpose.
e.g. the classic NAD3020 would do 50W per channel at clipping
level 8R (say 3% distortion) music programme, but no chance
of about 60% (peak dissipation) of that continuous in a soak
test, hence rated at 25W 8R and hence soak tested at 15W.
FWIW a music amplifier on decent music programme at full volume
(mild clipping) spends about 80% of the time under 20% full power.
Overdesign of PSU's is typical in many DIY amplifiers, which makes
the also typical underdesign of heatsinking an even bigger problem.
rgds, sreten.
There is no point IMO in saying a "soft" PSU is poor compared to a "hard"
PSU that turns a music amplifier into a laboratory amplifier and costs more.
The extra expense is best spent on improving the music amplifier.
A typical case would be a multichannel receiver that the PSU hasn't
a hope in hell in driving all channels continuously but work fine
most of the time. A big supply would be a minor improvement.
An uber supply and uber heatsinking, pretty pointless for purpose.
e.g. the classic NAD3020 would do 50W per channel at clipping
level 8R (say 3% distortion) music programme, but no chance
of about 60% (peak dissipation) of that continuous in a soak
test, hence rated at 25W 8R and hence soak tested at 15W.
FWIW a music amplifier on decent music programme at full volume
(mild clipping) spends about 80% of the time under 20% full power.
Overdesign of PSU's is typical in many DIY amplifiers, which makes
the also typical underdesign of heatsinking an even bigger problem.
rgds, sreten.
Last edited:
Charlie, you take some false estimations here. You don't need to repeat the peaky
character of music. This is true, but anyone here does already know it. Even more
this doesn't mean a soft PSU (= to small transformer) giving reasonable results. No
offense also, I just don't agree with you. 🙂
BTW, in real life when you listen with an average 1 W you'd rarely need 100 W at
peaks, but that's not my point though.
How will you know the time span of those peaks? Why accept PSU voltage sagging?
This IS a compromise soundwise. And do you know the current drwan by the loud-
speakers? What does happen if impedance is lower? The PSU needs to supply double
the current at half the load (say @ 4 Ohm instead of 8 Ohm).
The transformer has to deliver all the current needs. Ideally the PSU caps have just
to smooth the power, not to supply any current. If caps have supply quick power,
it's the job of the smaller ones close to the current sucking transistors. A capacitor
that is far away cannot deliver anything quickly.
For example for a 100/8 Ohm amp I would want it 200/4 Ohm, suggesting a 400 VA
transformer per channel. No less. I f you don't that power, why going for that high
voltage at all? If 25 W is sufficient, why not just put a 20 V PSU in it?
Howrever, anyone here has do decide on his own. If you want a "soft" PSU with a
weak transformer - do it. You have to realize, though, that it is still a compromise.
😉
Ok, let's just say it's "inferior". 😉 The PSU is defintitely the place where spending
money makes sense if music is what you aim for.
Are we talking about multichannel or Hifi?
True, this is a typical example of a budget amp with a very poor PSU. It sounds
accetable a lower levels, but no at higher listening levels and/or with current
demanding speakers. It's a compromise and you hear that it is. 😉
Probably true. But why should it sound awful at the other 20 %. Don't get me
wrong, I don't want to be too polemic or cynical, but whe it's hard to find words
things are getting difficult.
I don't think most overdesign their PSUs. They just do it right while most com-
mercial amps do compromise their PSUs. I also don't think many have to small
heatsinks, at least not me. It's a different question anyway.
depends on the capacity of your traffo in the first place...
overloading traffos is not necessarily a bad thing....
the question you have to answer for yourself is are you
willing to accept the added temperature rise without getting scared?
a lot depends on the insulation whether it can take the rise in heating without melting...
as to your coppers, the worst thing is voltage sagging...
but if you are overloading your traffo listening to music,
i doubt if that is something to be worried about....😉
over the years of repairing amps, i have seen a lot of commercial
amps whose traffos makes me wonder...
how are those amplifiers able to meet their power specs on such skimpy traffos?
and yet they do, considering that the FTC required only 5 minutes of full power testing,
transformers have to be abused at much longer time than this to burn-out...
and still most of the times it is those semicon devices that gave up instead of the traffo
in case of amplifier abuse.....
transformers have way much bigger thermal mass than semicon devices...
growing up as child, i would put candles atop the power traffo of our 21 inch Admiral BnW tv set,
i delight in seeing those candles melt away.....yet, that tv set served us continuously for 14 years
without serious breakdowns, save for occasional tube replacements....
As answered already, no harm will be done for short term overloads, even 1000% overload, although the transformer may become physically audible.
Regarding suitable ratings, for older pop/rock music with high peak factor, an under-specified transformer causing voltage sag, is of no great consequence.
When playing modern highly compressed/limited music, or the likes of Bach, Toccata and Fugue in D minor, collapsing supplies can become audible.
'Dynamic Headroom' is really a measure of how good, or lousy the supplies are...serious amps measure less than 1dB.
Dan.
- Status
- Not open for further replies.