Over the years I've noticed several postings related to sparking rectifier tubes. I thought I would bring the subject up again to perhaps help some folks contemplating new builds. According to the tube data for the GE 5AR4 tube available here
http://tubedata.milbert.com/sheets/093/5/5AR4.pdf
for capacitor input power supplies, like the SSE/TSE/SPP, the 5AR4 needs a minimum amount of resistance in the plate supply feeding each plate of the rectifier tube. RATING TABLE III on page 3 shows that as the transformer voltage increases, more resistance is needed in the transformer circuit. In my SSE build, I used a 400V Antek toroid power transformer which has only 28 ohms resistance and 50 ohms impedance (according to PSUD software) in the secondary circuit. So I added a pair of 100 ohm wirewound resistors in series between the transformer HV leads and the PCB to comply with RATING TABLE III. I had 25 watters which I used although 5 watt rating would probably be plenty.
As far as I can tell when playing around with PSUD, there are two times when the 5AR4's current draw I(D1) rating is approached or exceeded; during inital startup to charge the capacitors, and after startup if using power hungry tubes like KT-88s/6550s in the SSE.
For the parts I used and operating conditions, adding the resistors made the error messages go away on PSUD power supply simulator under both scenarios (startup and after 11 seconds or so).
Adding the recommended optional in-rush limiter thermistors (on negative leg of B+ and in series with power transformer primary) will help for startup, but I'm not sure if they have enough resistance after warming up to limit the 5AR4 current surges when driving the bigger tubes in the SSE under all conditions as the tubes age when using low impedance power transformers.
Although most folks don't have problems, I've read about the fireworks some experience and decided to play it safe and spend a few pennies on resistors to potentially save expensive transformers and tubes.
Hopefully this will help. I'm still pretty new to all this, so feel free to chime in if I'm leading anybody astray. Thanks.
http://tubedata.milbert.com/sheets/093/5/5AR4.pdf
for capacitor input power supplies, like the SSE/TSE/SPP, the 5AR4 needs a minimum amount of resistance in the plate supply feeding each plate of the rectifier tube. RATING TABLE III on page 3 shows that as the transformer voltage increases, more resistance is needed in the transformer circuit. In my SSE build, I used a 400V Antek toroid power transformer which has only 28 ohms resistance and 50 ohms impedance (according to PSUD software) in the secondary circuit. So I added a pair of 100 ohm wirewound resistors in series between the transformer HV leads and the PCB to comply with RATING TABLE III. I had 25 watters which I used although 5 watt rating would probably be plenty.
As far as I can tell when playing around with PSUD, there are two times when the 5AR4's current draw I(D1) rating is approached or exceeded; during inital startup to charge the capacitors, and after startup if using power hungry tubes like KT-88s/6550s in the SSE.
For the parts I used and operating conditions, adding the resistors made the error messages go away on PSUD power supply simulator under both scenarios (startup and after 11 seconds or so).
Adding the recommended optional in-rush limiter thermistors (on negative leg of B+ and in series with power transformer primary) will help for startup, but I'm not sure if they have enough resistance after warming up to limit the 5AR4 current surges when driving the bigger tubes in the SSE under all conditions as the tubes age when using low impedance power transformers.
Although most folks don't have problems, I've read about the fireworks some experience and decided to play it safe and spend a few pennies on resistors to potentially save expensive transformers and tubes.
Hopefully this will help. I'm still pretty new to all this, so feel free to chime in if I'm leading anybody astray. Thanks.
This is known to help tube life especially with new production 5AR4's too, but should help NOS too.
UF4007 diodes will be a bit quieter for a few more cents.
tube rectifier diode mod
Randy
UF4007 diodes will be a bit quieter for a few more cents.
tube rectifier diode mod
Randy
Last edited:
Technically,You don't. But it might be nice to keep the soft-start quality of a tube rectifier,or perhaps the voltage drop it provides. Or if you just want to keep the equipment somewhat 'original'.
I really can't see how this improves a tube rectified PS, unless you feel that a SS rectifier is an improvement on a tube rectifier, since that is what you end up with. Which is valid. I guess I would argue that if you want a SS rectifier, use one, if you want a tube, use one.
Adding a silicon diode in series with each plate of a tube rectifier does offer some added protection against rectifier arcing. Actually it helps quench the arc once it starts.
The rectifier tube has a tough life. It must fill the empty filter caps as its cathode is just starting to emit electrons. The filter caps look like a nearly short circuit load on the rectifier tube as it is warming up and at a fraction of its emission capacity. The tubes produced 50 years ago have a cathode coating that is nearly uniform and all of it heats at the same rate. Some of the tubes sold today have an uneven coating which heats up faster in the thin spots forcing all of the start up current to take place in a small area. The total current demands on the tube may be in spec, but the current density may be very high in one small area of the cathode. This can lead to an arc if the current density is too great.
Once an arc starts it is hard to stop. If the input filter cap has partially charged the voltage across the arc will be higher when the plate swings negative. Adding the solid state diode in series with the plate removes that negative half cycle stress from the rectifier tube, and it halves the total time of voltage stress that is feeding the arc. This 16mS of no voltage across the arc can be enough to stop it.
Adding a silicon diode in series with the plate should not cause the harsh harmonic hash that a pure silicon rectifier does since the tube is still doing the work. In a pure silicon rectifier the diode switches on rather abruptly as soon as the secondary voltage rises 0.7 to 1.0 volts above the voltage stored in the filter cap. This abrupt switch - on causes a current spike that can find its way into other circuits and into your audio.
With the tube / diode combination the diode will turn on as the secondary voltage rises 1 volt or so above the capacitors voltage, but the current will still flow until according to rectifier tubes VI characteristic as it does with the tube alone. You will just lose about 1 volt of B+.
This mod is relatively easy to do on the SSE. Simply disconnect the red transformer wires from the PC board, solder a 1N4007 or UF4007 diode on the end of each wire with the stripe on the end not soldered to the red wire, cover the diode with heat shrink tubing, and insert the striped end of the diode back inth the connector on the board where the red wire was.
This mod will also work on the Simple P-P board but DO NOT do it to the Tubelab SE board or you will not have any negative bias and your output tubes will fry!
The rectifier tube has a tough life. It must fill the empty filter caps as its cathode is just starting to emit electrons. The filter caps look like a nearly short circuit load on the rectifier tube as it is warming up and at a fraction of its emission capacity. The tubes produced 50 years ago have a cathode coating that is nearly uniform and all of it heats at the same rate. Some of the tubes sold today have an uneven coating which heats up faster in the thin spots forcing all of the start up current to take place in a small area. The total current demands on the tube may be in spec, but the current density may be very high in one small area of the cathode. This can lead to an arc if the current density is too great.
Once an arc starts it is hard to stop. If the input filter cap has partially charged the voltage across the arc will be higher when the plate swings negative. Adding the solid state diode in series with the plate removes that negative half cycle stress from the rectifier tube, and it halves the total time of voltage stress that is feeding the arc. This 16mS of no voltage across the arc can be enough to stop it.
Adding a silicon diode in series with the plate should not cause the harsh harmonic hash that a pure silicon rectifier does since the tube is still doing the work. In a pure silicon rectifier the diode switches on rather abruptly as soon as the secondary voltage rises 0.7 to 1.0 volts above the voltage stored in the filter cap. This abrupt switch - on causes a current spike that can find its way into other circuits and into your audio.
With the tube / diode combination the diode will turn on as the secondary voltage rises 1 volt or so above the capacitors voltage, but the current will still flow until according to rectifier tubes VI characteristic as it does with the tube alone. You will just lose about 1 volt of B+.
This mod is relatively easy to do on the SSE. Simply disconnect the red transformer wires from the PC board, solder a 1N4007 or UF4007 diode on the end of each wire with the stripe on the end not soldered to the red wire, cover the diode with heat shrink tubing, and insert the striped end of the diode back inth the connector on the board where the red wire was.
This mod will also work on the Simple P-P board but DO NOT do it to the Tubelab SE board or you will not have any negative bias and your output tubes will fry!
Hi George,
There are several reasons why rectifier diodes shouldn't be used to replace rectifier tubes in older gear. You know this, but many people don't. For similar reasons, using a #83 rectifier (or similar) as an "upgrade" causes very similar problems, being a mercury rectifier with an almost constant 15 V drop with varying current draw.
You can use silicon diodes if you also include enough series resistance to duplicate the tube versions I-V curve. I guess a purist could also install a series zener for the static voltage drop, but that's way out there. The added resistance will help with generated noise on switching, and a parallel R-C snubber across the diode will silence the beast. Snubber circuits can be seen across tube rectifiers in some designs, although this may just be a capacitor without the resistor.
The point is that there is no reason why using a silicon diode means you have to live with HF noise bursts at the line frequency. Just think a bit when the design is finalized.
One thing I didn't see mentioned is a common reason why tube rectifiers may arc. Increased capacitance used that is well beyond the maximum rated values. There is a certain designer that sold equipment that makes the twin Cetron full wave rectifiers arc. Now, that's a crime - and it's ignorant. Increased capacitance will also create more switching noise against the inductance of even simple wire (never mind the power transformer). Don't do this 40uF input capacitance is enough to get the job done.
-Chris
There are several reasons why rectifier diodes shouldn't be used to replace rectifier tubes in older gear. You know this, but many people don't. For similar reasons, using a #83 rectifier (or similar) as an "upgrade" causes very similar problems, being a mercury rectifier with an almost constant 15 V drop with varying current draw.
You can use silicon diodes if you also include enough series resistance to duplicate the tube versions I-V curve. I guess a purist could also install a series zener for the static voltage drop, but that's way out there. The added resistance will help with generated noise on switching, and a parallel R-C snubber across the diode will silence the beast. Snubber circuits can be seen across tube rectifiers in some designs, although this may just be a capacitor without the resistor.
The point is that there is no reason why using a silicon diode means you have to live with HF noise bursts at the line frequency. Just think a bit when the design is finalized.
One thing I didn't see mentioned is a common reason why tube rectifiers may arc. Increased capacitance used that is well beyond the maximum rated values. There is a certain designer that sold equipment that makes the twin Cetron full wave rectifiers arc. Now, that's a crime - and it's ignorant. Increased capacitance will also create more switching noise against the inductance of even simple wire (never mind the power transformer). Don't do this 40uF input capacitance is enough to get the job done.
-Chris
The Quicksilver amplifiers were built using a first cap of 320 mfd. Now this helped cause the demise of the two or at least one of the 5AR4 rectifier tubes within say about 3 months of re-tubing the rectifiers.
The mod on my Quicksilver amp has kept the same rectifier tubes running in the Quicksilver amp since 2002. Since 2002 I have mentioned it a few times here in the forum. While not the perfect solution it does work quite well.
The mod on my Quicksilver amp has kept the same rectifier tubes running in the Quicksilver amp since 2002. Since 2002 I have mentioned it a few times here in the forum. While not the perfect solution it does work quite well.
A tube that has manufacturing defects bad enough that it is going to spark out, will usually do so the very first time you turn it on. If they have lived this long, they should be OK.
As the emission dies off with age the possibility of a sparking death increases. Again an uneven cathode coating or improper alignment will cause uneven cathode wear.
As the emission dies off with age the possibility of a sparking death increases. Again an uneven cathode coating or improper alignment will cause uneven cathode wear.
Absolutely!
I have to say that it's been a very long time since I have seen any that bad. The situation where the cathode has been unevenly coated is another real quality problem. You may also see the cathode oxide material flaking off or maybe even rolling around in the tube. Pitch those in the bin. If the quality of the tube is that poor, I would expect all kinds of problems. Low vacuum and solder failure in the base pins are other signs of shoddy workmanship. Even the appearance of the internal elements may not be parallel with others too. Never use stuff like that.
I'm agreeing with George that your rectifier tube was / is very probably a good one. Arcing and other forms of discharging within a good tube can wreck it in no time. There is a possibility that your tube was damaged. Do you have another piece of equipment you can plug this tube into to test it?
I have seen more than my fill of badly designed power supplies in equipment with low secondary winding impedance and silly large filter caps. Using a high capacitance filter that is above the maximum rating in the tube manual is just plain wrong. This doesn't even buy you anything in the way of performance.
-Chris
Can't find the mod on site I linked to anymore so try this.
http://www.diyaudio.com/forums/tubes-valves/121057-diodes-rectifier-question.html?postid=1481237#post1481237
http://www.diyaudio.com/forums/tubes-valves/121057-diodes-rectifier-question.html?postid=1481237#post1481237
Last edited:
As George says they take the load off the tube helping older or new production last. It should help if the amp is also short cycled also!
Some amps (big ones) load the tube rectifier near or to the max so it would of course help there the most such as for example it is a popular mod on the forms for the Dynaco ST-70
Some amps (big ones) load the tube rectifier near or to the max so it would of course help there the most such as for example it is a popular mod on the forms for the Dynaco ST-70
Last edited:
Hi rmyauck,
The resistors reduce the peak currents to safe levels. They also reduce the "hot switching current" as the tube runs normally.
Know how to replace a tube rectifier? Use a diode, Zener(s) and resistance in series. The diode does what you would expect it to do. The zener provides the voltage drop seen across a tube in normal operation, and the resistors replicate the resistance of the rectifier tube. You should therefore end up with the same B+ voltages and close to the same performance. You're almost there now, why even leave the tube in circuit?
The delay in B+ the tube offers only helps the capacitors with their peak voltages, and if you are using a filament type rectifier (5U4 - say), there is no advantage to using a tube.
-Chris
The resistors reduce the peak currents to safe levels. They also reduce the "hot switching current" as the tube runs normally.
Know how to replace a tube rectifier? Use a diode, Zener(s) and resistance in series. The diode does what you would expect it to do. The zener provides the voltage drop seen across a tube in normal operation, and the resistors replicate the resistance of the rectifier tube. You should therefore end up with the same B+ voltages and close to the same performance. You're almost there now, why even leave the tube in circuit?
The delay in B+ the tube offers only helps the capacitors with their peak voltages, and if you are using a filament type rectifier (5U4 - say), there is no advantage to using a tube.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.