Fusing the B+ on your home builds ?

[SteelerFan88]

Curious as to the general consensus on fusing the B+ and/or filaments on your amp builds.

1. a good idea or not really necessary ?
2. why do some builders do it and some do not ?
3. your own personal opinions on the matter ?

I am just full of interesting questions lately but hey......the only dumb question is the one not asked...right ?

 

[pdf64]

See http://www.geofex.com/tubeampfaq/TUBEFAQ.htm#ampmods and http://www.valvewizard.co.uk/fuses.html
It seems a good idea to fuse the B+ winding at least (in addition to the primary).
Existing amps may include fusible components, eg 1 watt screen grid resistors.
If metal oxide flame retardant types are used, these can often quickly blow in response to a tube short, thereby greatly reducing the exposure of transformer windings to fault currents, whilst minimising the collateral damage.
Pete

 

[twangbanger]

See http://www.geofex.com/tubeampfaq/TUBEFAQ.htm#ampmods and http://www.valvewizard.co.uk/fuses.html
It seems a good idea to fuse the B+ winding at least (in addition to the primary).
Existing amps may include fusible components, eg 1 watt screen grid resistors.
If metal oxide flame retardant types are used, these can often quickly blow in response to a tube short, thereby greatly reducing the exposure of transformer windings to fault currents, whilst minimising the collateral damage.
Pete

Great info , thanks for sharing !

 

[Ht @ Clipper Amps]

Hard for me to tell someone not to add measures to protect their amp. If you are grounding your filaments with two 100 ohm 1/2 watt resistors, you are good there as the resistors will act as fuses. As far as the B+ goes, at worst it won't hurt anything to fuse it.

 

[SteelerFan88]

Hard for me to tell someone not to add measures to protect their amp. If you are grounding your filaments with two 100 ohm 1/2 watt resistors, you are good there as the resistors will act as fuses. As far as the B+ goes, at worst it won't hurt anything to fuse it.

Yeah my filaments use an artificial CT (2x100R resistors) so yes....they are good to go but I think adding an inline fuse after rectification can not hurt anything and add a measure of protection there but if the amp is built correctly and tested thoroughly, the chances that a failure will occur are minimal I think. I may throw a fuse there just to do it though....certainly does not hurt anything

 

[pdf64]

If you are grounding your filaments with two 100 ohm 1/2 watt resistors, you are good there as the resistors will act as fuses

I can't think of any benefit of those ground reference resistors fusing.
Rather it may expose the tubes to significant harm, eg if a power tube short sends B+ into the heater such that the ground reference is lost, then the entire heater chain may get pulled up pulled up to a high voltage, far above the tubes' Vh-k rating. The result being that every tube suffers damaged insulation between the heater and cathode, and heater current will leak into the signal chain.
Due to the above, my view is that ground reference resistors should have a suitably large power rating to withstand such fault currents for as long as it takes for the B+ or primary fusing to take effect, eg >3 watts. Such a rating may not protect the ground ref resistors from heat damage but they should help to retain the ground reference and mitigate the risk of damage to other tubes.
Pete

 

[VacuumVoodoo]

+10
Resistors are not fuses. Using them as such is simply asking for trouble.
As for "fusible resistors" - these are not general purpose fuses. There's one in some Fender amps used incorrectly. When this one blows it causes thermal runaway on power tubes.

 

[trobbins]

A common fault scenario is flashover from output tube plate to heater. In that scenario, saving the OT is often the top priority, with tube damage a lower concern. Once the flashover extinguishes (eg. both resistors in the humdinger turns to smoke and carbon), then the heaters don't remain at plate voltage level.

 

[pdf64]

I agree that resistors are not fuses and that using them as such is poor practice.
The main reason for this seems to be that when a resistor is subject to significant overdissipation, its fusing characteristics are not specified, defined or characterised, and so when used as such, their protective function can't be relied on. Additionally, when blown, a different type / brand of resistor may be fitted than was originally used, that whilst having similar 'headline' specs, behaves very differently under heavy overdrissipation
However, whilst being mindful of the above, there are a lot of amps out there, eg most BF and SF Fenders, in which the 1 watt rating of their screen grid resistors can be beneficial in terms of a protective function. This may be especially so if flame retardent types are used, eg most MOX.
Due to this, my feeling is that the 1 watt rating of these resistors should be retained when they're replaced or for amps of a similar design, with only their PT primary fused. To increase the power rating would expose the amp's most expensive components (the transformers) to prolonged fault currents, in the event of common failures such as some types of tube short. Slow blow / time delay fuses, as generally specified for power transformer primaries, can a take a long time to blow, especially if the fault current is limited, and prolonged exposure to such currents may result in damage to the insulation of transformer windings.

As for "fusible resistors" - these are not general purpose fuses. There's one in some Fender amps used incorrectly. When this one blows it causes thermal runaway on power tubes

Please could you provide further detail about this, eg which amp model/s and component reference or location?
Thanks - Pete

 

[SteelerFan88]

I always use 2w rating on my 100R resistors for my artificial filament center tap and 3w rating on my grid/screens. I use 5 or 10w rated resistors for cathode bias and 1 watt for the rest of the circuit....I like things to be a little "overkill".....I sleep better at night

 

[Jerry Glass]

Personally, I don't fuse the B+ on my builds but I have really developed an appreciation of the "save the show" fuses like the ones used on 100 watt JCM900s. They don't necessarily protect anything but in the event of a tube failure in the middle of a performance, they drop the offending pair and allow the show to continue.

 

[VacuumVoodoo]

Re: "Fusible resistors"

Please could you provide further detail about this, eg which amp model/s and component reference or location?
Thanks - Pete

Pete, here it goes:
65 Deluxe Reverb drwg No: 041285
R69 22 Ohm
Faulty (or bad choice of cap's ESR) C36 draws excessive current at power on, R69 opens, this zeroes the bias, output tubes go bright orange but the plate current overload is not enough to blow the primary fuse. It was just enough to activate thermal fuse embedded in PT primary winding. So there you have it. An improperly applied 5 cent component kills a PT because someone was too cheap to install a fuse in the B+ line.

 

[pdf64]

Ha, that was a bad idea.
How strange to fuse (kinda) the C- but not the B+.
Thanks for pointing it out.
Pete

 

[trobbins]

I always use 2w rating on my 100R resistors for my artificial filament center tap .....I sleep better at night

For higher power amps, with plate terminals next to heater terminals, this scenario can be difficult to protect against:

The fault current goes through one half OT PP winding, and may be limited by circuit resistances (PT and OT windings, valve diodes, humdinger resistance) such that a PT winding fuse (if used) may take a while to blow if the level actually gets to the fuse rating. I often use a 0.6W trimpot humdinger or small fixed humdinger resistors, which I hope will fail first in that situation.

But I also try and go to the root cause, with MOV limiting of peak plate voltage - which should ease the chance of that kind of flashover in 600V+ amps, although grungy environments are still a long term risk.

 

[donnyjaguar]

HT fuses can blow if there's a momentary interruption to the AC voltage. Very inconvenient.

 

[trobbins]

Do you think that is due to the recharge of the main HT filter capacitor, plus the idle current, causing a higher level of fuse current than in a cold start up? Or would this be more for cathode biased output stages that probably lose a lot of the bias (compared to a typical fixed bias supply that would likely stay up in level)?

 

[donnyjaguar]

If the bias voltage falls faster than the HT, the current in the output tubes can draw a lot of current while the filaments are still hot. A good reason not to "over-cap" an amplifier.

 

[trobbins]

If the bias voltage falls faster than the HT, the current in the output tubes can draw a lot of current while the filaments are still hot. A good reason not to "over-cap" an amplifier.

I'd suggest that there is little if any risk is moderately over-capping an amp, with respect to the turn-off scenario.

For a cathode biased PP example, the bias is proportional to the sum of anode and screen currents. So drawing a lot more anode current would mean a higher bias - so the anode current doesn't increase 'a lot'.

For a fixed biased PP example, the bias is completely dependant on the bias supply circuit, and typically over-capping that supply may cause stress at a following turn-on, not turn-off. Even if bias went quickly to 0V, the anode current can still only go to 0V Vgk characteristic anode current-voltage curve.

 

[donnyjaguar]

I'm talking fixed bias, I guess I shoulda made that clear. Both sets of "joules" have to be considered here.

 

[trobbins]

One way to look at the stress that could be applied to the anode (or with a little more difficulty a screen) is how long would it take to discharge the main filter cap.

So lets assume a tube could conduct 200mA for Vgk=0V for the operating screen voltage. Lets assume a pentode PP output stage with the valves at idle. If the mains turned off, and Vgk fell to 0V instantaneously, then simplistically anode current could quickly ramp to 200mA, where it would remain fairly constant whilst the main B+ discharged (assuming screen voltage stayed up).

For a 500V B+, and 100uF cap, it would simplistically take dt = 500V x 100u / 0.2A = 250ms to discharge the cap. The power level being dissipated in the anode during that time is ramping down, as B+ is ramping down. The screen voltage will also in practise be falling, so the anode current will also be falling and not nominally constant.

The other aspect to note is that the cap discharge energy is dissipated in two valves, and the peak power level would likely be less than a comparison with a cranked amp where one valve is effectively carrying that peak current level with nearly B+ on it for a 50% duty cycle.

From a joule perspective, that cap has 12.5J to discharge, which would occur over about half a second. The mass of an anode is relatively large, which allows it to handle quite high power levels for short periods of time before the temperature rise gets ugly. And note that many anodes have happily changed to red for a short while and the valve soldiered on, and millions of valve amps have happily soldiered on even when turned off many times.