Hello All,
Is their a practical limit to the inductance of the choke used in a choke input filter after a rectifier?
I am trying to understand the attached graph for the max rating of an EZ81 using a choke input filter.
It seems to suggest that the larger the coke (in henries) used the lower the max output current rating is at a given voltage. I have never seen this mentioned before.
Am I just misinterpreting the graph?
Thank in advance for your advice.
Best regards,
John
Is their a practical limit to the inductance of the choke used in a choke input filter after a rectifier?
I am trying to understand the attached graph for the max rating of an EZ81 using a choke input filter.
It seems to suggest that the larger the coke (in henries) used the lower the max output current rating is at a given voltage. I have never seen this mentioned before.
Am I just misinterpreting the graph?
Thank in advance for your advice.
Best regards,
John
> practical limit to the inductance of the choke used in a choke input filter after a rectifier?
Cost, weight, size, I*R drop.
> It seems to suggest that the larger the coke (in henries) used the lower the max output current rating is at a given voltage.
If the current in a choke-input filter is too low, it works as a CAP-input filter, with much higher output voltage.
Cost, weight, size, I*R drop.
> It seems to suggest that the larger the coke (in henries) used the lower the max output current rating is at a given voltage.
If the current in a choke-input filter is too low, it works as a CAP-input filter, with much higher output voltage.
The max. O/P current of a PSU is ultimately linked to the capabilities of the power trafo.
A good approximation for choke I/P filter critical current, in mA., is given by V/L. Notice that smaller critical currents are associated with larger inductances.
A bleeder resistor in parallel with the 1st filter cap. is a common technique for ensuring the critical current is drawn. It works out that 1000 Ω of resistance/H. of inductance is needed in the bleeder. Larger inductances waste less of the total current in the bleeder. Unlike the situation with cap. I/P filters, the full RMS current capability of the rectifier winding is accessible, when choke I/P filtration is employed.
A good approximation for choke I/P filter critical current, in mA., is given by V/L. Notice that smaller critical currents are associated with larger inductances.
A bleeder resistor in parallel with the 1st filter cap. is a common technique for ensuring the critical current is drawn. It works out that 1000 Ω of resistance/H. of inductance is needed in the bleeder. Larger inductances waste less of the total current in the bleeder. Unlike the situation with cap. I/P filters, the full RMS current capability of the rectifier winding is accessible, when choke I/P filtration is employed.
Last edited:
A true choke input filter has these characteristics:
1. The Secondary RMS voltage in x 0.9 = DC Volts out (of course, you have to calculate for the additional DC voltage drops from the Rectifier's voltage drop, and the Choke DCR).
2. The current drawn from the secondary and the rectifiers is the Average current (not peak current of a cap input filter). Cooler secondaries, cooler rectifiers (no peak transient Isquared losses that a cap input filter has; the loss is average Isquared loss).
3. All other things being equal, the voltage regulation with load current variations is good (has to have the critical inductance per the minimum load current). A cap input filter does not have good voltage regulations with load current variations.
Critical Inductance, you ask?
There is a formula for Critical Inductance:
60Hz power mains, and full wave rectification (120Hz full wave):
Critical Inductance = 350/Load mA.
Example, 50mA load. 350/50mA = 6 Henry choke
You have to use a 6 Henry choke or more (like 6, 7, or 10 Henry)
50Hz power mains, and full wave rectification (100Hz full wave):
Critical Inductance = 420/Load mA.
Example, 50mA load. 420/50mA = 8.4 Henry choke
Use a 10, 12, or 15 Henry choke or more.
I use choke input filters for B+, whenever I can.
Just my preference.
The graph that you posted is both for the safe operation of the rectifier, and to show the predicted DC out with a given RMS in, and choke inductance.
It does not show the required Critical Inductance to make the filter be a true choke input filter. You will get more than Secondary RMS x 0.9DCV output with a choke that is less than critical inductance (again, you have to take the additional voltage drops due to the rectifier drop, and the voltage drop of the choke DCR).
1. The Secondary RMS voltage in x 0.9 = DC Volts out (of course, you have to calculate for the additional DC voltage drops from the Rectifier's voltage drop, and the Choke DCR).
2. The current drawn from the secondary and the rectifiers is the Average current (not peak current of a cap input filter). Cooler secondaries, cooler rectifiers (no peak transient Isquared losses that a cap input filter has; the loss is average Isquared loss).
3. All other things being equal, the voltage regulation with load current variations is good (has to have the critical inductance per the minimum load current). A cap input filter does not have good voltage regulations with load current variations.
Critical Inductance, you ask?
There is a formula for Critical Inductance:
60Hz power mains, and full wave rectification (120Hz full wave):
Critical Inductance = 350/Load mA.
Example, 50mA load. 350/50mA = 6 Henry choke
You have to use a 6 Henry choke or more (like 6, 7, or 10 Henry)
50Hz power mains, and full wave rectification (100Hz full wave):
Critical Inductance = 420/Load mA.
Example, 50mA load. 420/50mA = 8.4 Henry choke
Use a 10, 12, or 15 Henry choke or more.
I use choke input filters for B+, whenever I can.
Just my preference.
The graph that you posted is both for the safe operation of the rectifier, and to show the predicted DC out with a given RMS in, and choke inductance.
It does not show the required Critical Inductance to make the filter be a true choke input filter. You will get more than Secondary RMS x 0.9DCV output with a choke that is less than critical inductance (again, you have to take the additional voltage drops due to the rectifier drop, and the voltage drop of the choke DCR).
Last edited:
I think the graph's lines showing choke inductance values is meant to represent the minimum inductance required for a particular output voltage-current capability on the given transformer voltage characteristic curve.
Eg. when Vtr=2x250V, at least 1H is needed to allow 200Vdc - 180mAdc loading capability; where a lower loading current would not maintain critical current, and a higher loading current would exceed the diode capability. Whereas 10H would be needed if loading was down at 25mA; where a lower loading current would not maintain critical current, and a higher loading current up to 180mA could be applied without exceeding the diode capability.
And note that is for 50Hz.
Eg. when Vtr=2x250V, at least 1H is needed to allow 200Vdc - 180mAdc loading capability; where a lower loading current would not maintain critical current, and a higher loading current would exceed the diode capability. Whereas 10H would be needed if loading was down at 25mA; where a lower loading current would not maintain critical current, and a higher loading current up to 180mA could be applied without exceeding the diode capability.
And note that is for 50Hz.
trobbins,
50Hz, 100 Alternations/sec, full wave rectification:
420/180mA = 2.33 Henry, Minimum for true choke input filter.
420/25mA = 16.8 Henry, Minimum inductance for a true choke input filter.
Or, those are not True Choke Input Filters.
Putting just any random value of choke in front of a cap does not make it a true choke input filter.
60Hz, 120 Alternations/sec, full wave rectification:
350/180mA = 1.94 Henry, Minimum for true choke input filter.
350/25mA = 14 Henry, Minimum inductance for a true choke input filter.
Or, those are not Choke Input Filters.
Putting just any random value of choke in front of a cap does not make it a true choke input filter.
You can use a choke that is smaller than the Critical Inductance value.
But the currents will be Transient currents, very similar to Cap input filters; not Average currents that true choke input filters are.
Undersized choke input filters will give DC Volts output of Larger than 0.9 x RMS Voltage, and Smaller than 1.414 x RMS Voltage (when you take into account the additional DC voltage loss due to the Rectifier Drop, and the additional DC voltage drop due to the choke DCR).
A 2x250VRMS secondary that has 0 Ohms DCR, will blow up the rectifier if you use less than 1 Henry choke, and use more than 180mA load current.
Poof, as they say!
50Hz, 100 Alternations/sec, full wave rectification:
420/180mA = 2.33 Henry, Minimum for true choke input filter.
420/25mA = 16.8 Henry, Minimum inductance for a true choke input filter.
Or, those are not True Choke Input Filters.
Putting just any random value of choke in front of a cap does not make it a true choke input filter.
60Hz, 120 Alternations/sec, full wave rectification:
350/180mA = 1.94 Henry, Minimum for true choke input filter.
350/25mA = 14 Henry, Minimum inductance for a true choke input filter.
Or, those are not Choke Input Filters.
Putting just any random value of choke in front of a cap does not make it a true choke input filter.
You can use a choke that is smaller than the Critical Inductance value.
But the currents will be Transient currents, very similar to Cap input filters; not Average currents that true choke input filters are.
Undersized choke input filters will give DC Volts output of Larger than 0.9 x RMS Voltage, and Smaller than 1.414 x RMS Voltage (when you take into account the additional DC voltage loss due to the Rectifier Drop, and the additional DC voltage drop due to the choke DCR).
A 2x250VRMS secondary that has 0 Ohms DCR, will blow up the rectifier if you use less than 1 Henry choke, and use more than 180mA load current.
Poof, as they say!
Last edited:
The minimum inductance is not for protecting the diode , for that you must have the nominal series resistance , but to achieve continuous mode operation , V medium = 0,9xVrms and "regulation" . Dependent on current and voltage like in the graph .
So for an PP amplifier you should use high inductance for the voltage to remain stable from no load idle current to max power .
So for an PP amplifier you should use high inductance for the voltage to remain stable from no load idle current to max power .
For most 5V rectifier tubes, and 6.3V consumer rectifier tubes, the requirement for series resistance is generally only for Capacitor Input Filters.
There generally is no requirement for series resistance when the choke is at the data sheet spec, on the tube data sheets and on the tube data sheet choke input graph.
For capacitor input filters, the required series resistance includes the DCR of the Secondary; the DCR of primary x primary to secondary step up ratio; and any series resistors that are installed to make the total resistance meet the spec.
Generally, If you calculate (and install) the Critical Inductance for the lowest current the amplifier will have, then no series resistor is necessary.
Rectifier Tubes Love to have choke input filters that have at least critical inductance.
They can put out higher DC current than a cap input filter can (unless the cap input filter has a very large series resistor, and that resistance causes B+ droop when the current goes up).
So much for power supply software. All you need is a 1950s era Radio Amateur Handbook.
There generally is no requirement for series resistance when the choke is at the data sheet spec, on the tube data sheets and on the tube data sheet choke input graph.
For capacitor input filters, the required series resistance includes the DCR of the Secondary; the DCR of primary x primary to secondary step up ratio; and any series resistors that are installed to make the total resistance meet the spec.
Generally, If you calculate (and install) the Critical Inductance for the lowest current the amplifier will have, then no series resistor is necessary.
Rectifier Tubes Love to have choke input filters that have at least critical inductance.
They can put out higher DC current than a cap input filter can (unless the cap input filter has a very large series resistor, and that resistance causes B+ droop when the current goes up).
So much for power supply software. All you need is a 1950s era Radio Amateur Handbook.
... I assumed that the choke has no significant ohmic resistance and the cap is big in our days . The reactance of the choke could be or be not enough at startup .If it is saturating at that surge current certainly not .By the way, the graph did not show the critical inductance values for the currents and voltages on the graphs.
They showed the Maximum Safe values of inductance, versus voltage and current, when there is no series resistance.
Series resistance in either a choke input filter, or in a cap input filter causes droop, when the current goes up.
Now you know another reason I like choke input filters, that have at least critical inductance.
Have you ever seen a rectifier spark?
I have. Live and learn.
From the graph:
250V RMS = 353.6V peak
1 Henry @ 100Hz (50Hz mains x 2 = full wave rectification) = 2 x pi x f x L
1 Henry is 628.3 Ohms of inductive reactance (at 100Hz full wave).
353.6V peak / 628.3 Ohms = 0.563 Amps peak (a little less than that, because the waveform has higher frequency harmonic structure, so a little higher total effective inductive reactance than 628 Ohms).
After the choke starts charging the first cap, the peak currents go down very fast.
Page 5 of the Philips EZ81 data sheet shows a schematic of the necessary resistance of the primary, secondary, and added resistors in each plate circuit (and shows a capacitor input filter).
There is no spec and no diagram that shows resistance for a choke input filter.
A big clue for using a choke input filter is to use one that has a current rating much larger than the DC current will be at maximum amplifier running current. Prevent saturation, use a husky choke.
They showed the Maximum Safe values of inductance, versus voltage and current, when there is no series resistance.
Series resistance in either a choke input filter, or in a cap input filter causes droop, when the current goes up.
Now you know another reason I like choke input filters, that have at least critical inductance.
Have you ever seen a rectifier spark?
I have. Live and learn.
From the graph:
250V RMS = 353.6V peak
1 Henry @ 100Hz (50Hz mains x 2 = full wave rectification) = 2 x pi x f x L
1 Henry is 628.3 Ohms of inductive reactance (at 100Hz full wave).
353.6V peak / 628.3 Ohms = 0.563 Amps peak (a little less than that, because the waveform has higher frequency harmonic structure, so a little higher total effective inductive reactance than 628 Ohms).
After the choke starts charging the first cap, the peak currents go down very fast.
Page 5 of the Philips EZ81 data sheet shows a schematic of the necessary resistance of the primary, secondary, and added resistors in each plate circuit (and shows a capacitor input filter).
There is no spec and no diagram that shows resistance for a choke input filter.
A big clue for using a choke input filter is to use one that has a current rating much larger than the DC current will be at maximum amplifier running current. Prevent saturation, use a husky choke.
Last edited:
There are a few simplifications for calculating critical inductance. The equation that Philips used for the EZ81 datasheet is L(crit) = Vo / (942 x Io), where Vo and Io are the DC load conditions, and 942 = 2 x pi x 100Hz x 0.667. And after checking the maths, I can confirm that the Philips datasheet does show critical inductance 'lines'.
Choke input filtering may be a concern for a hot turn-on event, depending on the situation.
For a normal power on situation, the resistance of the valve diodes fall from a high value, and the choke current builds up from zero typically without any gross overshoot. In simulation land, PSUD2 manages that with a 0.5 sec ramp up of supply levels. I haven't done, or seen, any examples of ramp up conditions for a cold start-up with a choke input filter.
For a hot start-up, after a few cycles the choke current can peak well above rated DC current level, and the affect of choke saturation on inductance dynamically changes, and the diode surge current may exceed its datasheet rating. PSUD2 can provide some insight to what may be experienced by changing the inductance, but that requires that choke inductance is measured for different levels of DC current and at DC current levels that significantly exceed the rating. I haven't seen this topic discussed in any vintage book afaik.
Choke input filtering may be a concern for a hot turn-on event, depending on the situation.
For a normal power on situation, the resistance of the valve diodes fall from a high value, and the choke current builds up from zero typically without any gross overshoot. In simulation land, PSUD2 manages that with a 0.5 sec ramp up of supply levels. I haven't done, or seen, any examples of ramp up conditions for a cold start-up with a choke input filter.
For a hot start-up, after a few cycles the choke current can peak well above rated DC current level, and the affect of choke saturation on inductance dynamically changes, and the diode surge current may exceed its datasheet rating. PSUD2 can provide some insight to what may be experienced by changing the inductance, but that requires that choke inductance is measured for different levels of DC current and at DC current levels that significantly exceed the rating. I haven't seen this topic discussed in any vintage book afaik.
Last edited:
trobbins,
Interesting.
But I am not sure we are there yet.
The derivation for the number 0.667 is 2/pi
That is the term for the average voltage value of a pair of full wave rectified alternations.
2 x pi x 100Hz x 0.667 = 2 x pi x 100Hz x 2/pi = 400.
It does Not equal 942.
Earlier, I said the number for 100Hz full wave was 420.
That is a lot closer to 400 than 942 is.
Have we come full circle?
(pun intended).
Interesting.
But I am not sure we are there yet.
The derivation for the number 0.667 is 2/pi
That is the term for the average voltage value of a pair of full wave rectified alternations.
2 x pi x 100Hz x 0.667 = 2 x pi x 100Hz x 2/pi = 400.
It does Not equal 942.
Earlier, I said the number for 100Hz full wave was 420.
That is a lot closer to 400 than 942 is.
Have we come full circle?
(pun intended).
Last edited:
trobbins,
If you use 942/mA, and 190mA, that is 4.96 Henry.
If you use 420/mA, and 190mA, that is 2.21 Henry.
I still am not satisfied with the interpretation of the graph.
The graph in Post # 1 shows 1 Henry, 190mA, 250V in, and 200V out.
250V x 0.9 = 225V. 225V - 25V rectifier drop = 200V
So that means that the rectifier has only a 25V drop when the DC current is 190mA . . . I doubt the EZ81 can do that.
I think that 1 Henry is less than Critical Inductance on that graph.
I use solid state diodes, and choke input. If I account for the voltage drops caused by the 'primary DCR x turns step up ratio', and the 'secondary DCR', and the 'Choke DCR', and the 0.6V - 1.0V diode drop, it works for me.
60Hz power mains (120Hz full wave), 350/mA, = critical inductance.
With a 350V-0-350V secondary, and critical inductance choke input, 350 x 0.9 = 315VDC.
The output is 315V - the drops caused by the primary x step up ratio, secondary, and choke DCRs.
It works out good.
Actual circuits prove the numbers.
What am I missing in my calculations and measurements?
If you use 942/mA, and 190mA, that is 4.96 Henry.
If you use 420/mA, and 190mA, that is 2.21 Henry.
I still am not satisfied with the interpretation of the graph.
The graph in Post # 1 shows 1 Henry, 190mA, 250V in, and 200V out.
250V x 0.9 = 225V. 225V - 25V rectifier drop = 200V
So that means that the rectifier has only a 25V drop when the DC current is 190mA . . . I doubt the EZ81 can do that.
I think that 1 Henry is less than Critical Inductance on that graph.
I use solid state diodes, and choke input. If I account for the voltage drops caused by the 'primary DCR x turns step up ratio', and the 'secondary DCR', and the 'Choke DCR', and the 0.6V - 1.0V diode drop, it works for me.
60Hz power mains (120Hz full wave), 350/mA, = critical inductance.
With a 350V-0-350V secondary, and critical inductance choke input, 350 x 0.9 = 315VDC.
The output is 315V - the drops caused by the primary x step up ratio, secondary, and choke DCRs.
It works out good.
Actual circuits prove the numbers.
What am I missing in my calculations and measurements?
Last edited:
So, from Post # 1 graph: 200Vo / (942 x Io), Where Io is in Amps (not mA).
200 / (942 x 0.19) = 200 / (179) = 1.1 Henry
Now we are getting there.
That still does not account for the differences in the critical inductance formulas.
50Hz (100Hz full wave) and 60Hz (120Hz full wave) should only disagree by 5/6, or 6/5, depending on which frequency you start 'with', and which frequency you compensate 'to'.
All I know is that for 60 Hz mains (120Hz full wave), 350/mA always works.
And for 50Hz (100Hz F W ) 420/mA works.
For 50 Hz mains (100Hz full wave) 420 /190mA = 2.2 Henry.
That is a 2:1 disagreement of the two formulas.
. . . Perhaps it has a very large margin built in.
200 / (942 x 0.19) = 200 / (179) = 1.1 Henry
Now we are getting there.
That still does not account for the differences in the critical inductance formulas.
50Hz (100Hz full wave) and 60Hz (120Hz full wave) should only disagree by 5/6, or 6/5, depending on which frequency you start 'with', and which frequency you compensate 'to'.
All I know is that for 60 Hz mains (120Hz full wave), 350/mA always works.
And for 50Hz (100Hz F W ) 420/mA works.
For 50 Hz mains (100Hz full wave) 420 /190mA = 2.2 Henry.
That is a 2:1 disagreement of the two formulas.
. . . Perhaps it has a very large margin built in.
The difference between critical inductance expressions certainly does come down to the simplifications made on the way to an expression, and even then the starting point is nearly always a sinewave for voltage.
For example the post #13 expression is via Schure (1958) with the condition that the peak value of the ac ripple current through the choke being less than the dc output current, and assuming the ripple voltage is 0.667 x dc output voltage Vo.
I think Schade's 1947 IRE paper was the first to progress mathematical expressions for rectifier operation that included the effective source resistance of the transformer and diode, and used the simplification of only assessing the mains rectified 2nd harmonic (and not higher harmonics).
Schade notes that for the typical situation where choke impedance is much more than the following capacitor impedance, then a further simplification allows L(crit) = RL/1000, an expression that was first used by Dellenbaugh and Quinby in QST in 1932, and for 50Hz and where RL = Vo / Io is the DC load resistance.
For example the post #13 expression is via Schure (1958) with the condition that the peak value of the ac ripple current through the choke being less than the dc output current, and assuming the ripple voltage is 0.667 x dc output voltage Vo.
I think Schade's 1947 IRE paper was the first to progress mathematical expressions for rectifier operation that included the effective source resistance of the transformer and diode, and used the simplification of only assessing the mains rectified 2nd harmonic (and not higher harmonics).
Schade notes that for the typical situation where choke impedance is much more than the following capacitor impedance, then a further simplification allows L(crit) = RL/1000, an expression that was first used by Dellenbaugh and Quinby in QST in 1932, and for 50Hz and where RL = Vo / Io is the DC load resistance.
Last edited:
Use whatever formula works for your amplifiers.
I designed a vacuum tube amplifier with the B+ using solid state diodes, A 20H choke input, followed by 200uF, a 180 Ohm series resistor, 200uF, a 1k series resistor, and 200uF.
The push pull outputs B+ comes after the 180 Ohm series resistor, and the input stage B+ comes after the 1k resistor.
Hum at the amplifier output is less than 100uV.
It has a power factor of 0.94 (about as resistive a load as you can get).
That means the choke input is really working properly.
How many of you can and do measure your amplifier's power factor?
I bet it would pass the World EMI standards, but perhaps not World power efficiency standards.
But since it is not a commercial amplifier, it does not have to (at least not yet, but laws change for the "better" all the time).
Now where are those power usage police today?
I designed a vacuum tube amplifier with the B+ using solid state diodes, A 20H choke input, followed by 200uF, a 180 Ohm series resistor, 200uF, a 1k series resistor, and 200uF.
The push pull outputs B+ comes after the 180 Ohm series resistor, and the input stage B+ comes after the 1k resistor.
Hum at the amplifier output is less than 100uV.
It has a power factor of 0.94 (about as resistive a load as you can get).
That means the choke input is really working properly.
How many of you can and do measure your amplifier's power factor?
I bet it would pass the World EMI standards, but perhaps not World power efficiency standards.
But since it is not a commercial amplifier, it does not have to (at least not yet, but laws change for the "better" all the time).
Now where are those power usage police today?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.