Any over-voltage protection device needs to constrain the plate voltage(s) to less than the likely insulation or working voltage rating of the winding, but not interact with normal signal voltages.
If a device is placed across each primary half winding (plate to B+ CT) in a PP amp, then signal voltages on the device happily reach B+ level. Even a vintage OT is likely to have insulation designed for about 3x the B+ level (eg. 1500Vpk for a 500VDC B+), and modern OT's should be much more if due care was being applied.
So for a MOV, its minimum 1mA DC conduction voltage should be safely well above the max expected B+. A review of the MOV conduction curves would then likely indicate that the worstcase MOV voltage at some relatively high current (perhaps an amp or so) is sufficiently below say 1500V.
How much current can be forced through a MOV in the actual application depends on an assessment of the conditions leading up to an over-voltage, and the circuit loop for the transient current. I had a fair look at all that in the link, and so far can't really identify anything deficient with a well selected MOV.
http://dalmura.com.au/projects/Output%20transformer%20protection.pdf
If a device is placed across each primary half winding (plate to B+ CT) in a PP amp, then signal voltages on the device happily reach B+ level. Even a vintage OT is likely to have insulation designed for about 3x the B+ level (eg. 1500Vpk for a 500VDC B+), and modern OT's should be much more if due care was being applied.
So for a MOV, its minimum 1mA DC conduction voltage should be safely well above the max expected B+. A review of the MOV conduction curves would then likely indicate that the worstcase MOV voltage at some relatively high current (perhaps an amp or so) is sufficiently below say 1500V.
How much current can be forced through a MOV in the actual application depends on an assessment of the conditions leading up to an over-voltage, and the circuit loop for the transient current. I had a fair look at all that in the link, and so far can't really identify anything deficient with a well selected MOV.
http://dalmura.com.au/projects/Output%20transformer%20protection.pdf
"Did I already mention TV damper diodes in this thread?"
I assume those are SS damper diodes, so there is no cathode to heater issue.
"to design the power supply correctly so it could accept the wall voltage range."
Its not just line over-voltage that's an issue here. Someone tripping over the speaker wires, or forgetting to connect them, or a tube arcing over internally, a tube pulled out of it's socket during operation or a bad socket, a parasitic oscillation, NFBK instability, capacitive speaker wires. All of these could generate a voltage spike or burst of oscillation. The OT is resonant at some frequency. The leakage L in a P-P OT or the air gap in a SET OT store magnetic energy that can be released as a V spike if tube conduction suddenly fails. Especially SET OTs, totally built in self destruction, ready and armed by the DC current. I don't think 400 Ohms across the 8 Ohm output will save you here.
I assume those are SS damper diodes, so there is no cathode to heater issue.
"to design the power supply correctly so it could accept the wall voltage range."
Its not just line over-voltage that's an issue here. Someone tripping over the speaker wires, or forgetting to connect them, or a tube arcing over internally, a tube pulled out of it's socket during operation or a bad socket, a parasitic oscillation, NFBK instability, capacitive speaker wires. All of these could generate a voltage spike or burst of oscillation. The OT is resonant at some frequency. The leakage L in a P-P OT or the air gap in a SET OT store magnetic energy that can be released as a V spike if tube conduction suddenly fails. Especially SET OTs, totally built in self destruction, ready and armed by the DC current. I don't think 400 Ohms across the 8 Ohm output will save you here.
Last edited:
Maybe not in this thread, but I remember this as an option as well, so you must have suggested them as well. I think its also an interesting option - especially since I have a lot of them...
How do they fare when B+ rises beyond tolerance?
800V goes beyond 3kV? Never.
I mean P-P amp only.
It's not a power supply; it is a primary inductance of an output transformer that can generate kilivolts from 250V B+ when an amp is clipping, but a load is disconnected.
Last edited:
So by doing this, the guitar amp is protected with no speaker connected?
Any explanation as to why such a high resistance is used? -- big difference between 400 ohms and an 8 ohm speaker.
Also, the 12W rating -- is this not dependent on the output of the amp?
Finally, what is the best way to deal with a parallel speaker jacks? Can one resistor protect against either/both in use, or does this need two resistors, one for each jack?
I would just like a simple form of protection for my tube guitar amp against the speaker becoming accidentally disconnected.
Any help appreciated.
Making a Tube amp that will not fail under any and all improper or unintended applications, or with parts that fail in various ways, . . . is like making a car safe for a very bad driver (or for a very good driver surrounded by very bad drivers).
I will not build a push pull pentode wired amp (unless I am sure the negative feedback will keep it stable and gain limited under no load conditions to the extent it will prevent self-destruction).
The same goes for a SE pentode wired amp.
When you are connecting a pentode wired amplifier into a system, and you don't hear anything (because one speaker wire disconnected), you turn the volume up, and . . . Snap!
Of course I could use a fairly low resistance load that is internally located, and in parallel with the loudspeaker, but that effectively defeats one of the purposes of using pentodes (high plate resistance that does not rob power from the load).
Just my preference.
I will not build a push pull pentode wired amp (unless I am sure the negative feedback will keep it stable and gain limited under no load conditions to the extent it will prevent self-destruction).
The same goes for a SE pentode wired amp.
When you are connecting a pentode wired amplifier into a system, and you don't hear anything (because one speaker wire disconnected), you turn the volume up, and . . . Snap!
Of course I could use a fairly low resistance load that is internally located, and in parallel with the loudspeaker, but that effectively defeats one of the purposes of using pentodes (high plate resistance that does not rob power from the load).
Just my preference.
On the PP 6v6 guitar amp I built for my grandson I hardwired a power resistor across the secondary for just such a reason. I don't remember if it was 150 or 250 ohm.
What is the safety factor for the amp (a guitar amp, stereo amp; different designs)?
With amp gain turned up full (I can't hear anything, because the speaker is not connected) can the load be:
2X the output tap rating?
4X the output tap rating?
10X the output tap rating?
It depends on the topology of the amp (triodes and triode wired tubes adhere to an output stage gain of u; pentodes have MUCH higher gain under open load conditions).
Although different, the problem can exist for SE and for PP amps.
Parallel loads suck up power.
Each amp topology has its own best methods for protection, including how the amp will be used and abused.
With amp gain turned up full (I can't hear anything, because the speaker is not connected) can the load be:
2X the output tap rating?
4X the output tap rating?
10X the output tap rating?
It depends on the topology of the amp (triodes and triode wired tubes adhere to an output stage gain of u; pentodes have MUCH higher gain under open load conditions).
Although different, the problem can exist for SE and for PP amps.
Parallel loads suck up power.
Each amp topology has its own best methods for protection, including how the amp will be used and abused.
@6A3sUMMER -- thanks for trying to help.
Some of that detail is my post above.
"I would just like a simple form of protection for my tube guitar amp against the speaker becoming accidentally disconnected."
I read the post a few pages back from DavesNotHere, which I quoted in my question. Probably easier to direct you to my earlier post #65 than to post the details again.
Difficult for me to answer some of your questions -- some of the technical details are above my pay grade. I'm not an amp designer, or expert, and I have more than one amp. Typically, these are fairly low wattage (15 to 20 watts) PP guitar combo amps, using 12AX7s in preamp/PI and either 6V6 or EL84 output tubes. They are mostly cathode biased, and mostly do not have NFB. IE, they are simple guitar amp circuits, not highly-engineered Hi-Fi applications.
I am careful with the use of my amps but, given that they are continually carted from gig to gig, my long experience has shown the potential for solder joints to fail and sockets to lose connectivity -- etc etc. I can fix that stuff -- I can solder reasonably well, I can make simple mods, and I understand tube amp safety implications. For me, the solution cannot lie in the design of the amp, it has to be an add-on or modification.
If I turn the amp on, and the speaker connectivity is lost, the amp could sustain damage before I have any indication that something is wrong
I was just looking for a simple solution for a user who is not an expert. Perhaps that's not possible?
not an expert. Perhaps that's not possible?
Some of that detail is my post above.
"I would just like a simple form of protection for my tube guitar amp against the speaker becoming accidentally disconnected."
I read the post a few pages back from DavesNotHere, which I quoted in my question. Probably easier to direct you to my earlier post #65 than to post the details again.
Difficult for me to answer some of your questions -- some of the technical details are above my pay grade. I'm not an amp designer, or expert, and I have more than one amp. Typically, these are fairly low wattage (15 to 20 watts) PP guitar combo amps, using 12AX7s in preamp/PI and either 6V6 or EL84 output tubes. They are mostly cathode biased, and mostly do not have NFB. IE, they are simple guitar amp circuits, not highly-engineered Hi-Fi applications.
I am careful with the use of my amps but, given that they are continually carted from gig to gig, my long experience has shown the potential for solder joints to fail and sockets to lose connectivity -- etc etc. I can fix that stuff -- I can solder reasonably well, I can make simple mods, and I understand tube amp safety implications. For me, the solution cannot lie in the design of the amp, it has to be an add-on or modification.
If I turn the amp on, and the speaker connectivity is lost, the amp could sustain damage before I have any indication that something is wrong
I was just looking for a simple solution for a user who is not an expert. Perhaps that's not possible?
not an expert. Perhaps that's not possible?
Horses for courses then - solder a power resistor across the secondary winding at the output transformer - perhaps aim for 5-10% of the output power rating (cranked) of the amp. Eg. a 50W 8 ohm speaker load could have a 2.5W max cranked loss in to the added resistor without anyone really noticing. Perhaps use a 5W ceramic resistor (to give power rating margin).
P = 50W = V.V/R where R = 8 ohm. So output voltage is 20Vrms. That is then the max voltage across your added resistor, so that resistance will be R(add) = 20V x 20V / 2.5W = 160 ohm. So could use a 150 or 180 ohm 5W resistor.
As discussed, there are better ways to protect, but as far as simplicity, it goes some way to keeping a load on the OT primary windings.
P = 50W = V.V/R where R = 8 ohm. So output voltage is 20Vrms. That is then the max voltage across your added resistor, so that resistance will be R(add) = 20V x 20V / 2.5W = 160 ohm. So could use a 150 or 180 ohm 5W resistor.
As discussed, there are better ways to protect, but as far as simplicity, it goes some way to keeping a load on the OT primary windings.
Interesting about just using a resistor to protect all guitar amps.
As was mentioned, some simple SE and PP guitar amps use pentode (pentode wired) output stages, and do not have feedback.
The pentodes without feedback are essentially constant current devices.
Constant current into 8 Ohms versus 160 Ohms dictates a large voltage difference.
Take such an amp that puts out 20V into an 8 Ohm load, remove the 8 Ohm speaker, but do have a 160 Ohm load that is always there inside the amp to protect it (inside it won't disconnect).
Now, without the 8 Ohm speaker connected, we can potentially have the amp producing up to 20V * (160 Ohms/8 Ohms) = 400V at the output.
We will not get there, but the primary will be quite unloaded, and we still may get Arc, Arc, Arc.
As I said earlier, the topology of the amp often dictates what are acceptable and not acceptable protection methods.
I would have to think long and hard for a good protection method for pentode un-fedback guitar amps.
I think it may take a combination of parts and circuits.
As was mentioned, some simple SE and PP guitar amps use pentode (pentode wired) output stages, and do not have feedback.
The pentodes without feedback are essentially constant current devices.
Constant current into 8 Ohms versus 160 Ohms dictates a large voltage difference.
Take such an amp that puts out 20V into an 8 Ohm load, remove the 8 Ohm speaker, but do have a 160 Ohm load that is always there inside the amp to protect it (inside it won't disconnect).
Now, without the 8 Ohm speaker connected, we can potentially have the amp producing up to 20V * (160 Ohms/8 Ohms) = 400V at the output.
We will not get there, but the primary will be quite unloaded, and we still may get Arc, Arc, Arc.
As I said earlier, the topology of the amp often dictates what are acceptable and not acceptable protection methods.
I would have to think long and hard for a good protection method for pentode un-fedback guitar amps.
I think it may take a combination of parts and circuits.
The pentodes are current sources up until they run out of voltage headroom (B+). But OT resonance could boost the voltage higher. I think a reasonable protective R could prevent the OT resonance (critical damping maybe). One might want to figure the primary inductance, convert that to reactance at the OT resonant frequency, then load the OT with that equivalent R somewhere. A calculation for a typical OT would be interesting to see what the R comes out as.
Guitar amps often use 1/4 inch Input jacks.
If you were to use a 1/4 inch Output jack, the danger would be to mistakenly
connect the guitar plug to the Output jack instead of the Input jack (like in a gig
in a darkened club room). That could destroy the guitar circuits.
But, if you can be sure to prevent that, you could use a 1/4 inch Output jack
that has an 'auto' switch as part of the jack. That is a certain style of 1/4 inch jack.
This is all inside the amp box.
With an 8 Ohm 25 Watt load resistor (inside) wired to the switch, it will be connected whenever there is no 1/4 inch Plug inserted in the Output jack. So, there will be the 8 Ohm resistor loading the amp.
When you plug in the speaker plug, now the 'auto'switch does not make contact with the 8 Ohm resistor. Now the speaker is the load.
Caution: if the speaker cable fails, or if the speaker fails, and the 1/4 inch plug is still inserted, there will be no load on the amp.
I hope I made it understandable.
If you were to use a 1/4 inch Output jack, the danger would be to mistakenly
connect the guitar plug to the Output jack instead of the Input jack (like in a gig
in a darkened club room). That could destroy the guitar circuits.
But, if you can be sure to prevent that, you could use a 1/4 inch Output jack
that has an 'auto' switch as part of the jack. That is a certain style of 1/4 inch jack.
This is all inside the amp box.
With an 8 Ohm 25 Watt load resistor (inside) wired to the switch, it will be connected whenever there is no 1/4 inch Plug inserted in the Output jack. So, there will be the 8 Ohm resistor loading the amp.
When you plug in the speaker plug, now the 'auto'switch does not make contact with the 8 Ohm resistor. Now the speaker is the load.
Caution: if the speaker cable fails, or if the speaker fails, and the 1/4 inch plug is still inserted, there will be no load on the amp.
I hope I made it understandable.
The switched in protection gives me an idea. If a circuit monitored the output tube drive voltage and cathode current, it could tell if the load is missing. Maybe just use a single transistor, like the protection in SS amplifiers, to shut down the screen voltage for some time window, whenever drive voltage exceeds expected cathode current by some threshold. Maybe need to ignore HF, since OT leakage L could cause an issue at HF. Would be nice to have an LED warning indication too.
The pentode push pull amps does seem to be less likely to arc than a pentode SE amp when the speaker load is disconnected.
PP amp in AB operation - one of the tube's plate can go to near 0 volts, while the other tube's plate goes to near 2X B+. If there is enough leakage reactance between the 2 halves of the primary, and if the one tube is completely cut off, then it can rise to more than 2X B+.
SE amp - when the only output tube cuts off, then the plate can rise to much more than 2X B+.
It is too bad that it seems that there is no single simple and sure way to make all the guitar amps safe under no load conditions, which was the original request of this thread.
There have been many good ideas, but it still may require different solutions for each amp.
Anything that causes both output tubes to no longer draw current, will require a study of whether the power supply capacitors will be able to handle the rise in B+ voltage.
PP amp in AB operation - one of the tube's plate can go to near 0 volts, while the other tube's plate goes to near 2X B+. If there is enough leakage reactance between the 2 halves of the primary, and if the one tube is completely cut off, then it can rise to more than 2X B+.
SE amp - when the only output tube cuts off, then the plate can rise to much more than 2X B+.
It is too bad that it seems that there is no single simple and sure way to make all the guitar amps safe under no load conditions, which was the original request of this thread.
There have been many good ideas, but it still may require different solutions for each amp.
Anything that causes both output tubes to no longer draw current, will require a study of whether the power supply capacitors will be able to handle the rise in B+ voltage.
6A3summer, a simple way is related to who is doing the modification.
I always put a MOV across each primary winding/half-winding, as discussed at start of thread, but I'm happy to work through those technicals, and am ok with messing around in an amp. That form of protection is valid for all amp OT configurations I reckon.
I always put a MOV across each primary winding/half-winding, as discussed at start of thread, but I'm happy to work through those technicals, and am ok with messing around in an amp. That form of protection is valid for all amp OT configurations I reckon.
You need a "load" which basically stays out of the way when operating normally but clamps/absorbs excess inductive peak voltages; both safety measures can be reasonably covered by a combination of reverse biased plate diodes and MOVs across primary windings.
Not 100% safe or fool proof, but think 90% or more effective.
Of course a fool or destructive genius will insist until even this protected transformer may short, why not? .... but at least it´s many steps in theb right direction.
Snubbers, while a great idea in power supply or power electromagnetic loads protection (think motors, electromagnets, power transformers, etc.) *will* change sound signature in an audible way, if "strong" enough to protect and be powerless if designed not to interfere.
Not 100% safe or fool proof, but think 90% or more effective.
Of course a fool or destructive genius will insist until even this protected transformer may short, why not? .... but at least it´s many steps in theb right direction.
Snubbers, while a great idea in power supply or power electromagnetic loads protection (think motors, electromagnets, power transformers, etc.) *will* change sound signature in an audible way, if "strong" enough to protect and be powerless if designed not to interfere.