Correct. Cap. I/P filters conduct only during brief portions of the 360o cycle (conduction angle). The larger the value of the filter's I/P cap., the shorter the amount of time conduction occurs. Also, as the I/P capacitance increases, the current in those charging pulses increases. The ripple current waveform becomes more and more "triangular" as the conduction angle gets smaller. Apply Fourier's Theorem to that ripple waveform and it should be obvious why HF crud appears.
Scan the archives here and over at AA for my posts on "hash" filtration. If you want clean/quiet B+ from large valued cap. I/P filter's, "hash" filtration is (IMO) very much in order.
Another consideration with cap. I/P filters is I2R heating in the power transformer's rectifier winding. Since heating varies with square of the ripple current and inversely linear with the duration of the charging pulses, large valued cap. I/P filters impose a substantial thermal burden on the power transformer. Don't try to draw B+ current in excess of 1/2 of the rectifier winding's AC RMS current capability, when large valued cap. I/P filtration is employed.
Cap. I/P filters are voltage rich and current poor. Choke I/P filters are current rich and voltage poor. TANSTAAFL!
This keeps coming up.
We don't WANT "true RMS" when measuring for cap-input rectifiers. We want Peak or Peak-to-Peak.
I don't see why you expect 360V; 250V*1.414 is 353.5V. ALL power supplies sag. Smallish supplies sag 10% to 15%. 315V is 11%sag, spot-on realistic expectation.
12.5% drop from NO load to FULL load (you are using a Class A amp so idle IS full load) for a non regulated supply is VERY GOOD performance.
Typical is 15/20% , go figure.
Like Jon Snell said, you will get peak calculated voltage *only* in an unloaded supply.
Remember you will lose voltage not only because of power transformer internal DCR but also because of the narrow duty cycle charging pulses (so peak currents are much higer, also substract ripple voltage from what´s available
.
As I mention above, your power supply is actually quite good.
Thanks for all the answers, this cleared up alot of things. Really cool to see people willing to help so much.
But if he measures the 250 at the bridge, under load, the sag has already be taken into account.
And yes, that 250 will have flatted tops due to the rectifier load, but all AC meters, true RMS or otherwise, really measure peak values and are calibrated in RMS, assuming a perfect sine. So a 1V peak will be indicated as 0.7Vrms. A flat topped 250 will be indicated low if anything.
In cases like this you can't do without a scope; there may be much more ripple than expected which messes up the meter measurement.
Jan
Jan
Agreed. If he measures honest-to-goodness 250Vac under load, then the DC should be higher(I'd expect about 340Vdc). Something is not right.
Does the amp play OK or does it hum? A faulty reservoir cap could account for a low average DC reading, but the ripple would be enormous, meaning the amp would noticeably hum.
Jan.didden,
There are 3 completely different kinds of AC meters:
1. Peak Responding
Old VTVM, or old TVM meters are peak responding.
They usually have 2 scales. A scale that reads out in Peak AC; and a scale that reads out the rms of a pure sine wave.
2. Average Responding
Old Simpson or Triplett VOMs are average responding.
But the scale is calibrated for rms of a pure sine wave.
3. True RMS Responding
These meters integrate the AC wave over time, and read out the True RMS of sine waves, square waves, triangle waves, etc.
Start with a triangle wave, measure with all 3 types of meters, you will get 3 different answers.
Start with a square wave, measure with all 3 types of meters, you will get 3 different answers.
Some DMMs are True RMS meters; Some are not.
Know Thy Test Equipment
There are 3 completely different kinds of AC meters:
1. Peak Responding
Old VTVM, or old TVM meters are peak responding.
They usually have 2 scales. A scale that reads out in Peak AC; and a scale that reads out the rms of a pure sine wave.
2. Average Responding
Old Simpson or Triplett VOMs are average responding.
But the scale is calibrated for rms of a pure sine wave.
3. True RMS Responding
These meters integrate the AC wave over time, and read out the True RMS of sine waves, square waves, triangle waves, etc.
Start with a triangle wave, measure with all 3 types of meters, you will get 3 different answers.
Start with a square wave, measure with all 3 types of meters, you will get 3 different answers.
Some DMMs are True RMS meters; Some are not.
Know Thy Test Equipment
So now why do we become more “realistic” all of a sudden when we switch our attention to tubes - when over in the solid state section the advice given is much different? That if the supply drops more than a couple percent under load, and if the amp doesn’t double its 8 ohm power at 4 the transformer is a POS? I can’t ever seem to get the point across that you can only get 5% regulation figures if the trafo is oversized to some ungodly amount - like 1000 VA for every 100 watts. Come here and expectations seem to be much more realistic.
The 12% is spot on what I get from the Antek transformers across the board regardless of size, 88% no load to full load - full load being when the DC output power -volts times amps, under load- is 60% of the VA rating (which is to assume 60% power factor). This is consistent with the advice to use a DC load current about half the RMS rating. EI core trafo regulation figures are often much poorer, and vary much more with transformer size and construction. Little ones can be pretty bad - 25-30% is pretty common among <100 VA EI’s. Their regulation can suck even with AC loading - especially split bobbin types.
How many amps do you have going to the heaters? The heater winding is only 2.5A, and a couple og 6L6GC's will require 1.8A, plus whatever other tubes you have.
"Honest to goodness" does not apply much when waveform is NOT a perfect sinewave but a flat topped one.
And as mentioned earlier, common hobby type meters measure *average* value and multiply by the proper factor to show "RMS"; that holds true, again, only for perfect sinewaves.
Also capacitors are being charged only for a small time, say 15% , and are being discharged by a constant and heavy load (single ended tube amp) the rest of the time (85%?).
A scope will show the true picture, a meter will display "a single number" which to boot is calculated making a lot of assumptions which here do not hold true.
I trust cheap meters only for mains and transformer tap voltages, anything more complex than a sinewave is taken with a grain of salt.
FWIW my rule of thumb is:
Average commercial transformers/supplies: about 20% drop from theory.
POS/junk: 30%, sometimes worse in real sh*t stuff.
Usually associated to hand burning transformers if used loud.
Audiophile cost-is-no-object type: there you can go from overbuilt to silly, from 2X or 3X rated transformers and 4 times larger caps to crazy heavy boat anchors, magnetizer class cap banks, silver wire, etc.
And still you will have some % drop.
Law of diminishing returns they call it.
Power Mains typically have dominant 3rd Harmonic distortion.
That means that they are slightly 'clipped' at the top and bottom, versus the shape of a pure sine wave.
So a True RMS meter will give a higher voltage reading; versus a peak responding meter on the same mains.
A 'clipped' sine wave that reads 120VAC on a True RMS meter, will give a lower B+ voltage; Versus an un-clipped sine wave that reads 120VAC on that same True RMS meter.
And even with a pure sine wave from the power mains, after it goes through the amplifier's power transformer, the B+ secondary will have dominant 3rd harmonic distortion.
The 'clipped' sine wave will result in lower B+ than you expect if you use a cap input filter.
The 1.414 multiplier factor only works on pure sine waves.
I suspect for the same conditions of power mains and amplifier power transformer, a choke input filter will give a B+ Voltage that is closer to what you are expecting.
The 0.9 multiplier factor will probably be more accurate, because the choke input circuit is average responding,
Versus a cap input filter that is peak responding.
Do any of those power software programs take account of the fact that power mains sometimes have 3% or more 3rd Harmonic distortion?
Do any of those power software programs take account of the fact that amplifier power transformers also often have 3rd Harmonic distortion?
Comments, anybody?
That means that they are slightly 'clipped' at the top and bottom, versus the shape of a pure sine wave.
So a True RMS meter will give a higher voltage reading; versus a peak responding meter on the same mains.
A 'clipped' sine wave that reads 120VAC on a True RMS meter, will give a lower B+ voltage; Versus an un-clipped sine wave that reads 120VAC on that same True RMS meter.
And even with a pure sine wave from the power mains, after it goes through the amplifier's power transformer, the B+ secondary will have dominant 3rd harmonic distortion.
The 'clipped' sine wave will result in lower B+ than you expect if you use a cap input filter.
The 1.414 multiplier factor only works on pure sine waves.
I suspect for the same conditions of power mains and amplifier power transformer, a choke input filter will give a B+ Voltage that is closer to what you are expecting.
The 0.9 multiplier factor will probably be more accurate, because the choke input circuit is average responding,
Versus a cap input filter that is peak responding.
Do any of those power software programs take account of the fact that power mains sometimes have 3% or more 3rd Harmonic distortion?
Do any of those power software programs take account of the fact that amplifier power transformers also often have 3rd Harmonic distortion?
Comments, anybody?
Last edited:
nothing unusual about this...if your finals tube are 60mA each, then your psu is being drawn 120ma dc, within your power traffo rating....no need to overthink this...
Last edited:
For a capacitor input filter, the transient current to the first capacitor is many times the DC current.
65mA + 5mA = 70mA DC.
If the transient current is 3x or 4x the DC current, those transients are 210mA or 280mA.
That accounts for lots of voltage drop in the DCRs of the power transformer's primary and secondary windings.
By the way, it sounds like you might not have a bleeder resistor on the B+.
Put a bleeder resistor across the B+ to ground.
Safety First!
Prevent the "Surviving Spouse Syndrome"
65mA + 5mA = 70mA DC.
If the transient current is 3x or 4x the DC current, those transients are 210mA or 280mA.
That accounts for lots of voltage drop in the DCRs of the power transformer's primary and secondary windings.
By the way, it sounds like you might not have a bleeder resistor on the B+.
Put a bleeder resistor across the B+ to ground.
Safety First!
Prevent the "Surviving Spouse Syndrome"
Juuzpo,
Your "Primary was 55 Ohms and secondary 260 Ohms".
With a 230 to 250V turns ratio, that is about 300 Ohms effective total DCR seen by the bridge rectifier.
Your 100uF input filter cap has a capacitive reactance of 15.9 Ohms at 100Hz (Full wave bridge rectification of 50 Hz).
You are observing a lower B+ voltage than you expected.
Most of that is probably due to the DCR of the transformers windings, versus the low impedance of the 100uF cap; so the transient currents are causing large voltage drops across the transformers windings.
Looks OK to me, as long as the transformer is not getting hot.
Live Long and Prosper . . .
Your "Primary was 55 Ohms and secondary 260 Ohms".
With a 230 to 250V turns ratio, that is about 300 Ohms effective total DCR seen by the bridge rectifier.
Your 100uF input filter cap has a capacitive reactance of 15.9 Ohms at 100Hz (Full wave bridge rectification of 50 Hz).
You are observing a lower B+ voltage than you expected.
Most of that is probably due to the DCR of the transformers windings, versus the low impedance of the 100uF cap; so the transient currents are causing large voltage drops across the transformers windings.
Looks OK to me, as long as the transformer is not getting hot.
Live Long and Prosper . . .
Last edited:
Probably more, as I'm quite sure the transformer has other secondaries, which also affect the reflected primary DCR.
But the whole problem is not about DCR voltage drop, as the AC voltage was measured under real load.
It is indeed.
/cdn.vox-cdn.com/uploads/chorus_image/image/65890483/iStock_1176125291.9.jpg)