Greetings, friends. Today I was hoping to spark a brief discussion on proper heater wiring. This idea came from an Instagram poster known as Amptech74 who repairs guitars and tube amps and probably seen more violently explosive tube behavior than most of us in Hi-Fi land, with the notable exception of @Tubelab_com.
Anyway, here's the idea: don't use the Center Tap that's hardwired into the power transformer. When a power tube fails, the plate voltage is going to find the easiest path to ground. If the plate arcs to the cathode, it still has to go the cathode resistor to get home, but if it arcs to the heater, it only has to fight the resistance of the copper wire. If that voltage happens to pass thru the power transformer, well, there's gonna be damage. But if you are using an artificial center tap, the HV can only cook those 2 resistors, saving your expensive iron.
So. Waddya think?
w
Anyway, here's the idea: don't use the Center Tap that's hardwired into the power transformer. When a power tube fails, the plate voltage is going to find the easiest path to ground. If the plate arcs to the cathode, it still has to go the cathode resistor to get home, but if it arcs to the heater, it only has to fight the resistance of the copper wire. If that voltage happens to pass thru the power transformer, well, there's gonna be damage. But if you are using an artificial center tap, the HV can only cook those 2 resistors, saving your expensive iron.
So. Waddya think?
w
Not at all.
The heater wires can conduct many amps, the HT flashover is milliamps and has to go through the cathode resistor, if used before it gets to the heaters. So lots of obstacles for those angry pixies.
Use a centre tap either physical or through 100R resistors.
Check and change your output valves regularly to avoid this scenario in the first place.
If you are paranoid, fit a zener diode across the cathode resistor. That will fail short circuit when your old valve commits suicide in your amplifier.
The heater wires can conduct many amps, the HT flashover is milliamps and has to go through the cathode resistor, if used before it gets to the heaters. So lots of obstacles for those angry pixies.
Use a centre tap either physical or through 100R resistors.
Check and change your output valves regularly to avoid this scenario in the first place.
If you are paranoid, fit a zener diode across the cathode resistor. That will fail short circuit when your old valve commits suicide in your amplifier.
Maybe test them? I'm guessing this is mostly a guitar/PA amp issue anyway, as they would have a rough life compared to your typical hifi amp.
jeff
Don't epiphanies feel so cool until you put 99% of them into real context. ....pooooph. It's more likely that someone would slip with a meter probe and short the plate pin to a heater pin. With all the heaters in parallel, you get B+ on all at once and hope that some other tube doesn't fail H-K to gnd.
Current from the HV rail is limited by the series DCR of PT and OPT.
Heater winding is less than 100 milli ohm, no HV can build up across it.
Surge current from reservoir cap is taken care of by thermal inertia of thick wire.
The heater filament itself is by far the weakest element in the circuit ...
Heater winding is less than 100 milli ohm, no HV can build up across it.
Surge current from reservoir cap is taken care of by thermal inertia of thick wire.
The heater filament itself is by far the weakest element in the circuit ...
I have seen lots of dead guitar amps in my lifetime. Let's explore some of the possible scenarios. Johnny B Metalhead is jumping all over the stage and accidently jams the guitar into the back of the amp smashing one or more output tubes shorting random bits of tube internals together. Mr. Fixit, the expert amp tech has previously replaced whatever fuses that were inside the amp with some newfangled No-Blo fuses to stop those annoying blown fuses when Mr. Roadie plugs the 120 volt amp into 240 volts.
Tube plate shorts to G3 (beam formers) which may be grounded or connected to the cathode. If grounded the plate is shorted to ground with B+ at the other end of the OPT. Result, fried OPT, Fried rectifier tube, or both.
Tube plate shorts to G2. Limited fireworks here, but amp will have no output or very low output. If continuously played something else may blow.
Tube plate shorts to G1. G1 has a short unhappy life, but the tube was already broken, so the likely casualties are similar to the next scenario.
Tube plate shorts to cathode. Here the cathode resistor (if equipped) will be very unhappy and blow open, followed shortly by an exploding bypass cap. Excess current could damage the OPT, blow the rectifier tube, or both.
Tube plate shorts to the heater. Oddly, this is a common failure in a cranked guitar amp, especially big ones. The short is not due to Johnny Metalhead's antics, but usually a mismatched load or no load due to speakers blowing and failing to an open voice coil. An amp switching from saturation to cutoff into an unloaded OPT will results in a LOT of voltage being generated in that OPT, THOUSANDS of VOLTS. In a perfect world the voltage will rise to infinity, but in the real world it will rise until there is an ARC. The arc will occur wherever an opportunity presents itself. I have seen three common places, the OPT itself, leading to destruction, from pin 3 (plate) of the tube socket to pin 2 (heater) of the tube socket, or from pin 3 to pin 2 inside the base of the tube itself. Now we have kilovolt level energy jumping to the heater circuit looking for a return path to ground. Remember that center tapped heater winding? If it is indeed tied directly to ground, the failure will usually be contained to the tube and socket involved in the arc. Black plastic tube sockets will carbonize which will lead to burning and possibly a fire. That's due to Mr. Fixit's No-Blo fuse. A proper fuse should blow. If that CT heater was grounded through a resistor or two, or a resistor in parallel with a capacitor, those parts will fail. Resistors usually blow open, caps usually short. Now we have lots of voltage running through the heater circuit. Lots of ugly stuff happens here.
When I was about 17 years old I was at an outdoor concert and a popular local Miami band was rocking a county park when I heard the unmistakable sound of a bass guitar speaker failing. First it sounded raspy, but it got worse. Before the band could finish that song, the sound deteriorated into noise, followed by silence and a panic on stage. The Ampeg SVT was on FIRE! Several weeks later I got to autopsy that SVT. It had been running at full crank when the mismatched speaker load went open. The OPT was fried a couple of the 6550's were bad, a 12AX7 and a few other parts were bad as well.
As vinylkid58 stated this is most likely a guitar amp issue. HiFi amps rarely see clipping. Guitar amps in the hands of some players ALWAYS see clipping. The fire ball in the OPT, tube or socket is due to an open or severely mismatched load. Many think that the voltage seen on the plate of an output tube only goes to twice the B+ voltage. I have seen peaks of 2400 volts on the plate of a 430 volt B+ guitar amp that was driven about 10 dB into clipping into a properly matched guitar speaker AT RESONANCE. The resonance of most guitar speakers lies within the range of standard guitar.
Tube plate shorts to G3 (beam formers) which may be grounded or connected to the cathode. If grounded the plate is shorted to ground with B+ at the other end of the OPT. Result, fried OPT, Fried rectifier tube, or both.
Tube plate shorts to G2. Limited fireworks here, but amp will have no output or very low output. If continuously played something else may blow.
Tube plate shorts to G1. G1 has a short unhappy life, but the tube was already broken, so the likely casualties are similar to the next scenario.
Tube plate shorts to cathode. Here the cathode resistor (if equipped) will be very unhappy and blow open, followed shortly by an exploding bypass cap. Excess current could damage the OPT, blow the rectifier tube, or both.
Tube plate shorts to the heater. Oddly, this is a common failure in a cranked guitar amp, especially big ones. The short is not due to Johnny Metalhead's antics, but usually a mismatched load or no load due to speakers blowing and failing to an open voice coil. An amp switching from saturation to cutoff into an unloaded OPT will results in a LOT of voltage being generated in that OPT, THOUSANDS of VOLTS. In a perfect world the voltage will rise to infinity, but in the real world it will rise until there is an ARC. The arc will occur wherever an opportunity presents itself. I have seen three common places, the OPT itself, leading to destruction, from pin 3 (plate) of the tube socket to pin 2 (heater) of the tube socket, or from pin 3 to pin 2 inside the base of the tube itself. Now we have kilovolt level energy jumping to the heater circuit looking for a return path to ground. Remember that center tapped heater winding? If it is indeed tied directly to ground, the failure will usually be contained to the tube and socket involved in the arc. Black plastic tube sockets will carbonize which will lead to burning and possibly a fire. That's due to Mr. Fixit's No-Blo fuse. A proper fuse should blow. If that CT heater was grounded through a resistor or two, or a resistor in parallel with a capacitor, those parts will fail. Resistors usually blow open, caps usually short. Now we have lots of voltage running through the heater circuit. Lots of ugly stuff happens here.
When I was about 17 years old I was at an outdoor concert and a popular local Miami band was rocking a county park when I heard the unmistakable sound of a bass guitar speaker failing. First it sounded raspy, but it got worse. Before the band could finish that song, the sound deteriorated into noise, followed by silence and a panic on stage. The Ampeg SVT was on FIRE! Several weeks later I got to autopsy that SVT. It had been running at full crank when the mismatched speaker load went open. The OPT was fried a couple of the 6550's were bad, a 12AX7 and a few other parts were bad as well.
As vinylkid58 stated this is most likely a guitar amp issue. HiFi amps rarely see clipping. Guitar amps in the hands of some players ALWAYS see clipping. The fire ball in the OPT, tube or socket is due to an open or severely mismatched load. Many think that the voltage seen on the plate of an output tube only goes to twice the B+ voltage. I have seen peaks of 2400 volts on the plate of a 430 volt B+ guitar amp that was driven about 10 dB into clipping into a properly matched guitar speaker AT RESONANCE. The resonance of most guitar speakers lies within the range of standard guitar.
Does that say it is a good idea to solder a large value resistor across the speaker outputs somewhere inside the amplifier if it is one regularly run turned up to eleven?
kind regards
Marek
One of those white ceramic "cement" resistors rated at 10 watts and 100 ohms wired across the speaker terminals allowed me to unplug the speaker at full crank on a test amp without any fireworks. This amp used TV sweep tubes for output and a generic 6600 ohm guitar amp quality OPT that was wired for 3300 ohms. I set an identical OPT on fire several years ago when my DIY test load made with Radio Shack parts failed to an open. The same large sweep tubes were used in both tests. Some would argue that the resistor consumes a few watts of power and affects the sound. These are true statements, but losing a couple watts on a 100 watt amp would not be heard, and the sound IS changed according to the scope pictures.....that nasty spike on the trailing edge of a clipped square wave gets rounded off a bit. To me it sounds better.
No, just an adapter or a UK, or EU power cord.
We had several LTX automated stations at work (late 1970's). All used similar minicomputers, but some of the CRT monitors were 120 volts and some were 240 volts. Both plugged into the side of the computer rack. When the setup operators swapped CRT's around "stuff" happened. Oddly the 120 volt monitor would live for an hour or more on 240 volts until the main filter cap in its power supply vented its stink......while the monitor kept on working!
Attachments
The nasty spike on the trailing edge is high frequency content. A zobel network of 0.1 uF and 10-20 ohms does pretty much the same thing without as much waste power with typical music.
The spikes will go away completely with free wheeling diodes. Clipping is flat top.
The spikes will go away completely with free wheeling diodes. Clipping is flat top.
A zoebel network works great on a HiFi amp that rarely sees clipping. The resistor smokes and the cap gets hot on a cranked guitar amp, as it must dissipate all that HF energy. The diodes do work, but some don't like the sound, and they are also prone to failure. Parts added to reduce failures may actually increase or move them. These effects are highly player and speaker dependent.
A 10 ohm zobel actually runs cooler than 100 ohms that gets full range. Even when driven to 11, or even 15. You just need the same 5 or 10 watt resistor, not a 1/2 watt like you would see on a hi fi amp. All the big transistor PA amps oversize them. As far as the cap getting hot, you just need a low tan delta cap, to keep the dissipation where it belongs. In this application it’s not really a “zobel” anymore, it’s a snubber - just like the one used to prevent HF ringing on the power supply transformer. I’ve seen it across the primary, but the only ones I’d trust there are class 2 3KV+ ceramics, as they would see 1200V on a 600V B+. Those are quite nonlinear, and not low DF either. So on the secondary it stays when I build one.
The diodes are absolutely required with transistor amps, as even a “benign” 2X overvoltage is instant death. They are never what goes out, even on big iron pigs like the CS800 that DJ’s often crank to 15. When they do fail, they fail short - but their current is limited to the amount that was stored in the inductance - it can’t spike higher. Pretty hard to kill a 3A rectifier when the cathode can only emit an amp of current to start the reaction. It has to tolerate the voltage anyway, even in normal operation.
The diodes are absolutely required with transistor amps, as even a “benign” 2X overvoltage is instant death. They are never what goes out, even on big iron pigs like the CS800 that DJ’s often crank to 15. When they do fail, they fail short - but their current is limited to the amount that was stored in the inductance - it can’t spike higher. Pretty hard to kill a 3A rectifier when the cathode can only emit an amp of current to start the reaction. It has to tolerate the voltage anyway, even in normal operation.