My understanding is that the Peak Inverse Voltage for a tube rectifier will be
2 x TX V (RMS) x 1.4
so for the LL1665 for example, which is 530-0-530, PIV = 1484v
I'm happy working on that basis to select a rectifier, but the WE 274B page says
Max TX input to choke filter = 660v RMS per plate
Peak Inverse Voltage = 1500v
I've mailed WE, but if anyone could help clarify that would be good. The PIV rating was under "Design Centre Values" section, is that the issue ?
2 x TX V (RMS) x 1.4
so for the LL1665 for example, which is 530-0-530, PIV = 1484v
I'm happy working on that basis to select a rectifier, but the WE 274B page says
Max TX input to choke filter = 660v RMS per plate
Peak Inverse Voltage = 1500v
I've mailed WE, but if anyone could help clarify that would be good. The PIV rating was under "Design Centre Values" section, is that the issue ?
adding an 1n4007 in series with each plates makes you forget piv worries...
i can not see what your problem is, just use them, and be mindful of the first cap
in case you use cap input filter..
i can not see what your problem is, just use them, and be mindful of the first cap
in case you use cap input filter..
Take a choke input power supply, with the choke meeting or exceeding the critical inductance requirement, and the required current load on the power supply.
That power supply is transformer secondary, rectifier, choke, and then filter cap.
The rectifier has a voltage drop due to the current. Suppose it is 50V.
That first capacitor after the choke will only see (Secondary RMS x 0.9) - the rectifier drop).
Given a 660Vrms secondary, that is 594V.
594 - 50 = 544V.
The primary to the plate is 660V x 1.414 = 933V
933V + 544V = 1477V. That meets the 1500V spec.
Solved! . . . Well Sort of:
Do not rely on this, because the choke input supply does not act like a choke input supply while the vacuum tube loads are warming up. The current is too low to meet the critical inductance versus the load current. So it acts like a capacitor input power supply during the rest of the amp's warm up, and the unloaded rectifier has a lower voltage drop, and the first cap will charge up to almost rms x 1.414.
That power supply is transformer secondary, rectifier, choke, and then filter cap.
The rectifier has a voltage drop due to the current. Suppose it is 50V.
That first capacitor after the choke will only see (Secondary RMS x 0.9) - the rectifier drop).
Given a 660Vrms secondary, that is 594V.
594 - 50 = 544V.
The primary to the plate is 660V x 1.414 = 933V
933V + 544V = 1477V. That meets the 1500V spec.
Solved! . . . Well Sort of:
Do not rely on this, because the choke input supply does not act like a choke input supply while the vacuum tube loads are warming up. The current is too low to meet the critical inductance versus the load current. So it acts like a capacitor input power supply during the rest of the amp's warm up, and the unloaded rectifier has a lower voltage drop, and the first cap will charge up to almost rms x 1.414.
i will also pay attention to the total impedance of the voltage source to plates,
they give at 100 ohms per leg, so that if your traffo secondary is at 80 ohms per leg,
then a 20 or 22 ohms resistor in series will take away some of your fears...
they give at 100 ohms per leg, so that if your traffo secondary is at 80 ohms per leg,
then a 20 or 22 ohms resistor in series will take away some of your fears...
Also note that mains ACV will vary, as will the diode peak voltage drop with peak current, so there is some variation around the 'numbers'.
The PIV of any ss diode added to a plate should really be at least the PIV of the anode - so two 1N4007 in series would be appropriate for each anode. These ss diodes seem to be the cheapest modern insurance policy available for a valve diode power supply, closely followed by fusing the CT.
The PIV of any ss diode added to a plate should really be at least the PIV of the anode - so two 1N4007 in series would be appropriate for each anode. These ss diodes seem to be the cheapest modern insurance policy available for a valve diode power supply, closely followed by fusing the CT.
Minus the forward drop of the other rectifier.
In the sim below, input both sides is 100V Peak 70.7V RMS. The rectifier reverse voltage is 199.2V (forward voltage with 1N400x rects is ~~0.7V). This is 2*Vrms*1.4 minus forward drop. Expected output is 70.7V*0.9 or 63.63V, simmed output is 62.9V (again ~~0.7V forward drop).
I agree: the 660V and 1500V limits do NOT agree (even with large vacuum drop), and so much so that I wonder if there is a typo.
Attachments
I guess my post #4 was misunderstood. Here goes again:
1. The choke input supply has the choke first, and the filter cap second (and let us use a perfect rectifier, 0V drop). A choke input supply charges the cap that follows the choke to the Average voltage of the alternation.
The Average of a sine alternation is 0.9 x the Vrms.
The voltage to the filter cap is 660Vrms x 0.9 = 594V, the cap charges to + 594V; it does not charge to 660V x 1.414, because this is a properly designed and properly loaded choke input supply.
The number to remember is + 594V, the voltage on the cap.
Please note that when the rectifier is back biased, the cathode will have +594V on it, it will not have +933V on the cathode.
And back biased means one plate is back biased, while the other plate is forward biased (full wave rectification).
2. The secondary first goes positive to 660V x 1.414 = +933V
The secondary then goes negative, -933V cutting off the diode.
The number to remember is -933V on the diode anode.
4. The diode has - 933V on the anode, and + 594V on the cathode from the capacitor.
The differential voltage from anode to cathode is -933V to +594V = 1527V.
Now we are close to the 1500V rating peak inverse voltage for the choke input supply.
4. If the rectifier has a voltage drop when in forward conduction (it does), that voltage drop reduces the voltage on the cap by that amount. Let the tube rectifier drop be 30V under load.
+ 594V - 30V drop = 564V on the capacitor.
The differential voltage across the diode is -933V to +564V = 1497V
I vote for the fact that the old Western Electric engineers did know a thing or two about their 274 rectifier.
By the way, when the 274 rectifier is used in a capacitor input power supply, the maximum recommended capacitor is 4uF.
How many schematics on the web are there that use more than 4uf, in search of less hum (and reduced rectifier life)???
I have seen the same kind of improper selection of input capacitor values for other types of tube rectifiers maximum capacitance rating.
5U4, 5AR4, 5Y3, . . . etc.
Arcing?, tube life?, hot rectifiers, anybody?
1. The choke input supply has the choke first, and the filter cap second (and let us use a perfect rectifier, 0V drop). A choke input supply charges the cap that follows the choke to the Average voltage of the alternation.
The Average of a sine alternation is 0.9 x the Vrms.
The voltage to the filter cap is 660Vrms x 0.9 = 594V, the cap charges to + 594V; it does not charge to 660V x 1.414, because this is a properly designed and properly loaded choke input supply.
The number to remember is + 594V, the voltage on the cap.
Please note that when the rectifier is back biased, the cathode will have +594V on it, it will not have +933V on the cathode.
And back biased means one plate is back biased, while the other plate is forward biased (full wave rectification).
2. The secondary first goes positive to 660V x 1.414 = +933V
The secondary then goes negative, -933V cutting off the diode.
The number to remember is -933V on the diode anode.
4. The diode has - 933V on the anode, and + 594V on the cathode from the capacitor.
The differential voltage from anode to cathode is -933V to +594V = 1527V.
Now we are close to the 1500V rating peak inverse voltage for the choke input supply.
4. If the rectifier has a voltage drop when in forward conduction (it does), that voltage drop reduces the voltage on the cap by that amount. Let the tube rectifier drop be 30V under load.
+ 594V - 30V drop = 564V on the capacitor.
The differential voltage across the diode is -933V to +564V = 1497V
I vote for the fact that the old Western Electric engineers did know a thing or two about their 274 rectifier.
By the way, when the 274 rectifier is used in a capacitor input power supply, the maximum recommended capacitor is 4uF.
How many schematics on the web are there that use more than 4uf, in search of less hum (and reduced rectifier life)???
I have seen the same kind of improper selection of input capacitor values for other types of tube rectifiers maximum capacitance rating.
5U4, 5AR4, 5Y3, . . . etc.
Arcing?, tube life?, hot rectifiers, anybody?
Last edited:
Nah, I'm not keen on that explanation for choke input.
Each diode PIV is twice the peak level of one side of the secondary winding, so for 660-0-660Vrms secondary each diode sustains a PIV of 660x1.4x2=1850V. At the peak of the mains sine waveform, the cathode is at the same level as the conducting anode, which is 1850V away from the other anode that is not conducting - the choke and output dc level are not dictating PIV.
The datasheet refers to 1500V as a design centre maximum PIV, and I'd suggest is lower than 1850V as it would relate to a continuous applied voltage condition.
The datasheet refers to 440VAC for capacitor input filter condition (ie. 1250V PIV), which is much less than choke input but is more related to peak cathode current conduction limit that occurs at the peak of the voltage sinewave given a certain effective resistance.
Each diode PIV is twice the peak level of one side of the secondary winding, so for 660-0-660Vrms secondary each diode sustains a PIV of 660x1.4x2=1850V. At the peak of the mains sine waveform, the cathode is at the same level as the conducting anode, which is 1850V away from the other anode that is not conducting - the choke and output dc level are not dictating PIV.
The datasheet refers to 1500V as a design centre maximum PIV, and I'd suggest is lower than 1850V as it would relate to a continuous applied voltage condition.
The datasheet refers to 440VAC for capacitor input filter condition (ie. 1250V PIV), which is much less than choke input but is more related to peak cathode current conduction limit that occurs at the peak of the voltage sinewave given a certain effective resistance.
Last edited:
trobbins,
You were right. You had a good point that the plate with the other phase causes the cathode to be at the peak voltage minus the rectifier drop. I missed that fact.
Perhaps the different voltage ratings are related to the tube heating caused by the current of cap input, versus the tube heating caused by the current of choke input. (the integral of I squared x the voltage drop).
This integral of I squared x DCR is also used for some transformers. The maximum current rating of the transformer is two-fold. One current rating for capacitor input, and another current rating for choke input.
There have been cases when the filament power was removed, but the tube kept rectifying. That was due to the high filament to plate current causing the filament to remain in a sustained hot condition. (don't operate the tube this way).
The Tungsol 5U4G data sheet did have a note that only applied to choke input operation.
It said the secondary max rms voltage rating was with the secondary unloaded.
But it did not have that same requirement for the secondary max rms voltage rating of cap input operation.
You were right. You had a good point that the plate with the other phase causes the cathode to be at the peak voltage minus the rectifier drop. I missed that fact.
Perhaps the different voltage ratings are related to the tube heating caused by the current of cap input, versus the tube heating caused by the current of choke input. (the integral of I squared x the voltage drop).
This integral of I squared x DCR is also used for some transformers. The maximum current rating of the transformer is two-fold. One current rating for capacitor input, and another current rating for choke input.
There have been cases when the filament power was removed, but the tube kept rectifying. That was due to the high filament to plate current causing the filament to remain in a sustained hot condition. (don't operate the tube this way).
The Tungsol 5U4G data sheet did have a note that only applied to choke input operation.
It said the secondary max rms voltage rating was with the secondary unloaded.
But it did not have that same requirement for the secondary max rms voltage rating of cap input operation.
There is a very straightforward way to see it must be the RMS end-to-end voltage times 1.414 (then minus drop).
But we don't normally draw this affair to emphasize the back-to-back diode series connection across the end-to-end AC wave from the transformer.
Drawn like this, we see that it hardly matters "what" is connected to winding CT or diode CT. The two diodes together MUST block the entire winding peak voltage. Since one diode is conducting with low drop, the other diode must take essentially full peak voltage.
Any "rational" rectifier will have "low" drop. In days before large crystals, our vacuum bottle had drops that seem "large" now: 50V even 100V. But these were scaled for 300V-600V power supplies. The useful DC was larger than the diode drop. If drop is not "small", your electric bill goes to the rectifier not the actual load. Yes, 5-tube radios did lose a lot of voltage and power to the rectifier; larger hungrier systems usually did better.
But we don't normally draw this affair to emphasize the back-to-back diode series connection across the end-to-end AC wave from the transformer.
Drawn like this, we see that it hardly matters "what" is connected to winding CT or diode CT. The two diodes together MUST block the entire winding peak voltage. Since one diode is conducting with low drop, the other diode must take essentially full peak voltage.
Any "rational" rectifier will have "low" drop. In days before large crystals, our vacuum bottle had drops that seem "large" now: 50V even 100V. But these were scaled for 300V-600V power supplies. The useful DC was larger than the diode drop. If drop is not "small", your electric bill goes to the rectifier not the actual load. Yes, 5-tube radios did lose a lot of voltage and power to the rectifier; larger hungrier systems usually did better.
Attachments
if you have 2 x 660 vac = 1320 volts ac.... i will choose dioders that has 3.1 x 1320 volts of piv rating, or 4100 volts, 5 x in4007's paralleled with 1 meg 1 watt equalising resistors should do it....these diodes are dirt cheap...
this is one reason i dislike ht with ct psu traffos...
this is one reason i dislike ht with ct psu traffos...
Tony, ss diodes used in series do indeed need some derating.
A 660-0-660V secondary does indicate 1850V peak nominal, which would I suggest be covered by three 1N4007 in series for each valve anode.
I wouldn't use voltage sharing resistors or capacitors across each ss diode for the reasons discussed in link, but some may want to.
https://www.dalmura.com.au/static/Power%20supply%20issues%20for%20tube%20amps.pdf
A 660-0-660V secondary does indicate 1850V peak nominal, which would I suggest be covered by three 1N4007 in series for each valve anode.
I wouldn't use voltage sharing resistors or capacitors across each ss diode for the reasons discussed in link, but some may want to.
https://www.dalmura.com.au/static/Power%20supply%20issues%20for%20tube%20amps.pdf
i would rather be on the safe conservative side....diodes are dirt cheap,
i buy 1N4007 by the thousand....
i buy 1N4007 by the thousand....
- Status
- Not open for further replies.