Hello all,
I've been reading along with you, and I've learned a lot myself thanks to your posts! I don't mean to interrupt the flow of ideas - I just wanted to ask a question to wg_ski regarding his statement a few posts ago:
Just for our understanding and learning, can you describe what you mean by a power supply flywheeling (or what happens when a power supply flywheels)?
I've been reading along with you, and I've learned a lot myself thanks to your posts! I don't mean to interrupt the flow of ideas - I just wanted to ask a question to wg_ski regarding his statement a few posts ago:
Just for our understanding and learning, can you describe what you mean by a power supply flywheeling (or what happens when a power supply flywheels)?
rtarbell said:
The leakage inductance in the trafo will act as a filter at very heavy load. This is what REALLY causes the voltage to drop. At light load, the filter cap dominates, charging to the peak voltage in short pulses and you get a bit of sag and some ripple. At heavy load (several times the VA rating, usually) the inductance starts to be significant and what you get is a choke-input filter. The "flywheel effect" causes the current through the rectifiers to become more constant instead of short pulses. Ripple actually decreases with load. Recall that inductors don't like changes in current, and will generate back EMF to try to keep it that way. When the load is sufficiently high (or the inductance high enough), the output voltage drops to 0.637 X the peak value (the time average value) and doesn't drop beyond that except for the effects of parasitic resistances (which are small). The end result is that a DC supply will drop from 1.4x the RMS value of the trafo at no load to 0.9x the RMS value at severe overload. The transition in between can be rather rapid, depending on the trafo and how much capacitance you have.
Sure?
The leakage inductance is on the AC side of the rectifier so won't work as the common inductive filter supply. It will make the conduction angle wider though as you say. Output can drop below rectified average.
Instead of thinking of the leakage inductance, replace it with an ideal transformer with inductance in series with the primary.
If you short out the output of the rectifier you will essentially have an inductor across power line, so current is still limited because of its inductance.
But still, current will increase with decreasing output voltage and conduction angle gets wider too. Probably that's why it doesn't drop too much.
The leakage inductance is on the AC side of the rectifier so won't work as the common inductive filter supply. It will make the conduction angle wider though as you say. Output can drop below rectified average.
Instead of thinking of the leakage inductance, replace it with an ideal transformer with inductance in series with the primary.
If you short out the output of the rectifier you will essentially have an inductor across power line, so current is still limited because of its inductance.
But still, current will increase with decreasing output voltage and conduction angle gets wider too. Probably that's why it doesn't drop too much.
megajocke said:
On something poorly wound, or with intentionally poor coupling. With a primary wound closer to the core and evenly wound, primary side reactance is a small portion of it.
So-called "inherently limited" trafos or these split-bobbin EI units that they use for low capacitance, the primary side reactance is quite high. They drop like stones under any sort of load. A 1-amp unit could flywheel at 100mA. But we don't use those to power big audio amps. I've yet to load a 60 Hz toroid heavy enough to get it to drop below rectified average. Except maybe the one time I melted a potted 225VA unit making hydrogen from salt water back in undergrad school.
I agree that in the center tapped full wave rectifier the distribution primary/secondary will matter and there will be an effect similar to the one of inductive filters.
But in the circuit used in audio amps we draw AC from both ends of winding of the transformer so it doesn't matter if all leakage inductance is transferred to the primary.
The drop isn't proportional to current though, the output resistance decreases the more current you draw as the conduction angle widens so it fits with your observations.
But in the circuit used in audio amps we draw AC from both ends of winding of the transformer so it doesn't matter if all leakage inductance is transferred to the primary.
The drop isn't proportional to current though, the output resistance decreases the more current you draw as the conduction angle widens so it fits with your observations.
See attachment, first 20% or so drop fast but after that it gets much stiffer. Also, ripple doesn't increase much with current.
This was with 1mH leakage inductance referred to secondary, bridge rectifier and 100V peak (70V RMS) input. 30mF smoothing and neglible resistive loss. This leakage inductance is on the high side compared to what would be found in a power amp transformer but the only real difference would be the scaling of currents. Red = output voltage. Green = load current
This was with 1mH leakage inductance referred to secondary, bridge rectifier and 100V peak (70V RMS) input. 30mF smoothing and neglible resistive loss. This leakage inductance is on the high side compared to what would be found in a power amp transformer but the only real difference would be the scaling of currents. Red = output voltage. Green = load current
Attachments
AndrewT said:
To discharge the caps after the amp is turned off...
take a look at ESP's original drawing...
An externally hosted image should be here but it was not working when we last tested it.
im going to follow the schematic very closely, because like this i can supply a higher voltage to the drivers
and the beauty is that 1.5KVA unit i'm looking at has 4 secondaries ( 2x 68V and 1x 12V ) which allows me run ~130VDC rails and ~150VDC drivers
will be interesting to get zapped by that for sure...
I have felt 220V 50Hz... i wonder whats 150VDC gonna feel like
In that case it will be 95V idle on main rail and 105V on drivers side.
Still plenty of juice if you ask me, and lets not forget that's just for 1 channel !
Both power supply's will be ties together in parallel using a separate 4x5000uF bank that will do the load/current balancing while not starving either supply of current/voltage.
Still plenty of juice if you ask me, and lets not forget that's just for 1 channel !
Both power supply's will be ties together in parallel using a separate 4x5000uF bank that will do the load/current balancing while not starving either supply of current/voltage.
The core doesn't enter into transformer power transfer at all, only how many turns you need and how much wire fits. Bigger core means less number of turns and more room for wire. The wire will be shorter but there is also room for thicker wire which decreases resistance even more and gives you a higher powered transformer.
It's the impedance of the mains, wire resistance in the transformer and its leakage inductance that limit the the short circuit current - which means it can get very high.
It's the impedance of the mains, wire resistance in the transformer and its leakage inductance that limit the the short circuit current - which means it can get very high.
That seems to be a very common misconception and is not how it works.
The power is not transfered "through" the core. Without the core there would still be power transfer but magnetizing inductance would be in the same order as leakage inductance so it would be very inefficcient (in normal applications at least).
Flux = Integral((Applied voltage per turn)*dt)
Flux density = Flux / Area
(SI units)
Current does not matter.
Actually, if you look closer at the transformer, current drawn from secondary actually *decreases* flux because of the winding resistance of the primary, bringing the transformer *further* from saturation.
If you still don't believe it, it's easy to see that current flowing in secondary and primary produce opposing magnetomotive forces - sum is 0 so the core doesn't see the power transfer. But this isn't really the driving force, it's just a consequence of the close coupling between primary and secondary.
The power is not transfered "through" the core. Without the core there would still be power transfer but magnetizing inductance would be in the same order as leakage inductance so it would be very inefficcient (in normal applications at least).
Flux = Integral((Applied voltage per turn)*dt)
Flux density = Flux / Area
(SI units)
Current does not matter.
Actually, if you look closer at the transformer, current drawn from secondary actually *decreases* flux because of the winding resistance of the primary, bringing the transformer *further* from saturation.
If you still don't believe it, it's easy to see that current flowing in secondary and primary produce opposing magnetomotive forces - sum is 0 so the core doesn't see the power transfer. But this isn't really the driving force, it's just a consequence of the close coupling between primary and secondary.
Small core = thinner wire and more of it. Higher resistance and reactance so it drops faster per ampere.
Transformers *do* make more noise under load, which may lead some to believe it's saturating. Saturation moves the core (makes it buzz, becasue it's being magnetized harder on each cylce). At high currents the WIRES make noise due to magnetostriction. The two coils get magnetized opposite to one another harder, and try to move with respect to one another. It can make a loud buzzing, too. Listen to a shorted pole pig some time.
Transformers *do* make more noise under load, which may lead some to believe it's saturating. Saturation moves the core (makes it buzz, becasue it's being magnetized harder on each cylce). At high currents the WIRES make noise due to magnetostriction. The two coils get magnetized opposite to one another harder, and try to move with respect to one another. It can make a loud buzzing, too. Listen to a shorted pole pig some time
.Adrculda said:
An amp running on +/-95 sure makes a lot more sense for a first timer than one running off +/-130. What you'll end up with is a souped up USA1310 which can *really* put out 400/8R, 650/4R and 1k/2R. Two ohm power may even be better than that, but I wouldn't expect more than 1300W.
And 100 volt caps will save you a bundle compared to 150's.
DON'T go putting them in series unless you use entire power supplies in series (which fixes the maximum voltage across each cap to a maximum value). Caps don't necessarily "age" at the same rates.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.