I must have had too much time up my sleeve over the xmas break - link to an article that tries to cover the design and awareness needed for fusing in valve amps, and in particular the power transformer high voltage secondary. I used to just choose a value that I thought would be appropriate, without doing any real assessment - so thought there must be a better way...
http://dalmura.com.au/projects/Valve amp fusing.pdf
http://dalmura.com.au/projects/Valve amp fusing.pdf
I'm happy to take critique and highlight errors and ways to improve. Unfortunately it doesn't help much when staring at a large pile of assorted fuses with a smattering of markings of uncertain origin - so perhaps of more practical use by those who buy fuses from a quality tech company who have specs and datasheets for the fuses they sell, and even then you will need to keep them separated and identifiable.
Ciao, Tim
Ciao, Tim
Very good write up.
One might want to take into account the ambient temperature and it's effects on the fuse as well as on the thermistor. Unfortunately, fuse suppliers do not supply temperature related performance data.
I have measured differences in performance of fuses in fuse holders and leaded fuses of the same series/rating (Littlefuse 3AG 312 series vs 318 series, UL248-14) due to heat entrapment in the fuse holder.
Does it make more sense to fuse the anode supply to the tube using a DC Fast Acting type such as the KLK D series? This would seem to provide better protection of the output transformer as the stored energy in the filter cap would be available to blow the fuse. Fusing the transformer still leaves the stored charge in the filter caps to do damage to the OPT and other circuits.
Of course that does not protect the screen if driven from a separate supply. Always trade-offs.
One might want to take into account the ambient temperature and it's effects on the fuse as well as on the thermistor. Unfortunately, fuse suppliers do not supply temperature related performance data.
I have measured differences in performance of fuses in fuse holders and leaded fuses of the same series/rating (Littlefuse 3AG 312 series vs 318 series, UL248-14) due to heat entrapment in the fuse holder.
Does it make more sense to fuse the anode supply to the tube using a DC Fast Acting type such as the KLK D series? This would seem to provide better protection of the output transformer as the stored energy in the filter cap would be available to blow the fuse. Fusing the transformer still leaves the stored charge in the filter caps to do damage to the OPT and other circuits.
Of course that does not protect the screen if driven from a separate supply. Always trade-offs.
Yes, 'ambient' temp is an interesting topic in a valve amp.
Some fuses are lucky enough to be in open style holders under the chassis in a 'cool' section. Other hotter situations could include a fuse in an enclosed holder, or placed next to a power resistor, or exposed to valve heat, or placed in a spaghetti sleeve or heatshrink. And when a fuse is enclosed, then thermal dissipation via terminal wiring may or may not be an influence.
The general datasheet deratings I've seen are typically down to about 90% at 60-80C ambient.
Datasheets can give max voltage drop specs at rated current, and Schurter even give a typical voltage drop, which is often a lot lower than the max spec. Of interest is that the power dissipation by the fuse cartridge is somewhat low for typical amplifier current ratings used in HT supplies, but certainly increases for current levels greater than a few amp. Which would imply that self heating in a fuse holder is not too big a derating issue for fuse current rating less than a few amps.
If a fuse selection method is using the minimum I-t limit curve then there is inherently a very large margin to what may be the single line 'characteristic' I-t curve provided by a manufacturer.
I guess one hurdle to face when collating measurements for such an issue is that individual fuse tolerance is likely to be quite wide, and so large batches of fuses would be needed as controls versus being subject to a set change in conditions in order to give good statistics.
Fusing the B+ is also a very interesting topic. I can see that three hurdles should really be jumped first - a suitably DC rated fuse; protection of the output transformer from L.di/dt; and making the fuse de-energise the plate and screen supplies (and not just the plate).
Protecting a PP output transformer would typically mean the fuse being placed between the main B+ filter cap, and a substantial filter cap(s) on the output transformer and screen supply side that would restrict any voltage rise on the output transformer CT terminal or screens.
I would suggest that it is easier and probably better to fuse the output valve cathodes. But it may be easier to discuss benefits of one or other method by focussing on what happens in a particular fault situation, which is why I tried to describe some common fault situations in order to appreciate the prospective fault current level that could pass through a fuse.
A filter cap can certainly assist the opening of a fuse by contributing a level of I2t in addition to the prospective current available from the mains supply source of power. I don't think there is a simple accurate equation to take the capacitor CV2/2 energy and pass it through a fuse to derive a specific I2t level. Some simple approximations could go a long way I guess, such as assuming the filter capacitor forces a constant current of some value through the fuse, and assuming the capacitor voltage then drops at a certain dV/dt from B+ to zero, and hence get an approximate time t to estimate the I2t contribution, and then check the I2t rating of the particular fuse to see how they compare.
Hmmm.
Some fuses are lucky enough to be in open style holders under the chassis in a 'cool' section. Other hotter situations could include a fuse in an enclosed holder, or placed next to a power resistor, or exposed to valve heat, or placed in a spaghetti sleeve or heatshrink. And when a fuse is enclosed, then thermal dissipation via terminal wiring may or may not be an influence.
The general datasheet deratings I've seen are typically down to about 90% at 60-80C ambient.
Datasheets can give max voltage drop specs at rated current, and Schurter even give a typical voltage drop, which is often a lot lower than the max spec. Of interest is that the power dissipation by the fuse cartridge is somewhat low for typical amplifier current ratings used in HT supplies, but certainly increases for current levels greater than a few amp. Which would imply that self heating in a fuse holder is not too big a derating issue for fuse current rating less than a few amps.
If a fuse selection method is using the minimum I-t limit curve then there is inherently a very large margin to what may be the single line 'characteristic' I-t curve provided by a manufacturer.
I guess one hurdle to face when collating measurements for such an issue is that individual fuse tolerance is likely to be quite wide, and so large batches of fuses would be needed as controls versus being subject to a set change in conditions in order to give good statistics.
Fusing the B+ is also a very interesting topic. I can see that three hurdles should really be jumped first - a suitably DC rated fuse; protection of the output transformer from L.di/dt; and making the fuse de-energise the plate and screen supplies (and not just the plate).
Protecting a PP output transformer would typically mean the fuse being placed between the main B+ filter cap, and a substantial filter cap(s) on the output transformer and screen supply side that would restrict any voltage rise on the output transformer CT terminal or screens.
I would suggest that it is easier and probably better to fuse the output valve cathodes. But it may be easier to discuss benefits of one or other method by focussing on what happens in a particular fault situation, which is why I tried to describe some common fault situations in order to appreciate the prospective fault current level that could pass through a fuse.
A filter cap can certainly assist the opening of a fuse by contributing a level of I2t in addition to the prospective current available from the mains supply source of power. I don't think there is a simple accurate equation to take the capacitor CV2/2 energy and pass it through a fuse to derive a specific I2t level. Some simple approximations could go a long way I guess, such as assuming the filter capacitor forces a constant current of some value through the fuse, and assuming the capacitor voltage then drops at a certain dV/dt from B+ to zero, and hence get an approximate time t to estimate the I2t contribution, and then check the I2t rating of the particular fuse to see how they compare.
Hmmm.
Last edited:
I've also been thinking about how to fuse an amp using ~600V plate and ~300V screen voltages. I did find the KLK fuses, and also the Littlefuse 508 - the 508 is standard 1/4" size and rated to 1000VDC. Expensive though!
As for protecting the screen, in this case (a transmitting tube that can take a bit of screen dissipation) I think I can get away with a series resistor (bypassed to filament) to limit the current.
I think the key thing for people to understand is that you cannot use a normal AC fuse and expect it to behave with DC current, especially at higher voltages. The fuse will turn into a brilliant plasma arc in the event of overcurrent. It will probably open eventually, but it might start a fire in the process.
Pete
As for protecting the screen, in this case (a transmitting tube that can take a bit of screen dissipation) I think I can get away with a series resistor (bypassed to filament) to limit the current.
I think the key thing for people to understand is that you cannot use a normal AC fuse and expect it to behave with DC current, especially at higher voltages. The fuse will turn into a brilliant plasma arc in the event of overcurrent. It will probably open eventually, but it might start a fire in the process.
Pete
Yes that is the intent. However positioning the fuse in that location does not manage a diode short circuit fault. Also note that as the fault current increases, then the diodes conduct for an increasing time up to each mains half cycle (rather than conducting for just a small duty cycle).
The intent is to use a fuse with an AC voltage rating that is equal to or exceeds the secondary winding AC voltage.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.