Secondary Fuse Value

[trobbins]

You could estimate the fuse current by simulation - that requires a few measurements made on the power transformer for its effective resistance and voltage parameters, as well as an understanding of the power supply and amp circuitry. I use PSUD2 program for that assessment. That can also indicate the crest factor that a meter would need to cope with. The link is to an article that shows how that can be done:
https://dalmura.com.au/static/Valve%20amp%20fusing.pdf

[Gaz]

Only if your meter allows for a higher crest factor with voltage measurements than with current. Check your meter manual. If I had to guess I would think that a crest factor of 5 should be sufficient. Maybe someone knows more exactly?

The manual for my Fluke reads:

"AC crest factor can be up to 3 at full scale, 6 at half scale. For non-sinusoidal wave forms add -(2% Rdg + 2% full scale) typical, for a crest factor up to 3."

Would that explain my high readings?

[Gaz]

You could estimate the fuse current by simulation - that requires a few measurements made on the power transformer for its effective resistance and voltage parameters, as well as an understanding of the power supply and amp circuitry. I use PSUD2 program for that assessment. That can also indicate the crest factor that a meter would need to cope with. The link is to an article that shows how that can be done:
https://dalmura.com.au/static/Valve%20amp%20fusing.pdf

Thanks, Tim. I really appreciate that article - it's the best and most detailed I've found by far. I was actually going over and over it before getting frustrated and posting here. I just simply don't have enough electronics background to put it all together. I've simulated my own PS, but am having trouble choosing a fuse value based on the results. Perhaps you can help me. Here are the screenshots of my sim results. Thanks!

[Helmholtz]

"AC crest factor can be up to 3 at full scale, 6 at half scale. For non-sinusoidal wave forms add -(2% Rdg + 2% full scale) typical, for a crest factor up to 3."

If I interpret that statement correctly, it doesn't give any information about accuracy for crest factors above 3.

[loudthud]

Isn't crest factor just Peak/RMS ? In the first example it looks like the crest factor is just less than 3. T1 peak current 1.7697A, RMS .61188A . The assumption is that the numbers for the transformer are correct. You need primary and secondary resistance and the no-load secondary Voltage. Is 50/60 Hz set correctly ? Two Ohms for ESR of C1 is the default number and is probably on the high side.

[trobbins]

Can you confirm which parts of the model you have measurements for? Do you measure 340V on the secondary unloaded? Did you use the secondary winding effective resistance calculator (buried in PSUD2) to get the 14.5 ohm value (ie. insert the primary mains voltage and the primary and secondary winding resistances)?

Do you effectively have 265mA max of loading on B+ (ie. are you able to measure and sum the various currents drawn from B+, such as the Output transformer or output stage cathodes, and any screen and preamp feeds)? And what is the B+ DC voltage at 'full volume'?

Do you have similar measurements for idle conditions (sum of currents, and B+ DC voltage)?

Measurements for the above questions allow the simulation model to be verified to get some confidence that the model is adequate for the task at hand, given all the other assumptions and margins that are being applied.

If your model was adequate, then fuse continuous current = 612mArms with a crest factor of 2.9. The following estimated levels indicate a 0.75A T type IEC fuse would be fine - a 0.75A fast blow would not. The fuse current rating is fairly high, so the steps between ratings is fairly large (given that an IEC fuse would have to be at least 0.62A).



Assessing a UL248-14 model fuse, it would have to be at least 0.83A rating, so likely a 1A Slow blow. There are no minimum multiplier requirements, only maximum, so not much help for design. You may find UL248-14 compliant fuses with typical characteristic curves - which you could use, but you don't have a guaranteed minimum performance to design with.

[Helmholtz]

The total DC current I(R1) demand of a 100W amp with a B+ of 400V is typically around 500mA at full power. So actual load resistance must be much lower than 1.7k. Total DC current at full power can be found by inserting a 1R resistor right after the reservoir cap and measuring DC voltage drop.

[Helmholtz]

I took some measurements with my Fluke meter using the Max function. Here are my readings:

With 1kHz sine wave signal at full volume: 1.3A

What do you get at full output power (just before clipping into rated load) in normal meter mode (not using the Max function)?
With a 100W Marshall I'd expect around 1.0Arms.

[nickb]

In case you know the PT's rated secondary current, Merlin states that the fuse is typically rated for twice to 3 times the rated transformer current.

I've become very skeptical recently of this rule of thumb after meeting the Masters of Doom Metal (fake name) and the Incendiary Power Transformers. They play their amps at full tilt. A 100W amp came in from them with a blown power transformer. 1A fuse was fitted to HT secondary for a 420mA winding. Seemed reasonable enough. Fitted new PT, checked it out and sent it on it's way. Comes back a few weeks later with a burnt out power transformer again, and I mean burnt. The heater winding was being used at it's limit so I added an auxiliary heater transformer and off it went. Came back again, albeit after a few years, with the same "flaming" problem. Questions as to mode of use followed. Reproducing those conditions on the bench gave a 950mA RMS secondary current. A 1A fuse would never blow and so the transformer would over heat. Not a good situation.

So, if you fit a fuse that is bigger than the transformer rating you risk a fire under the right conditions. Therefore, how can we in all good conscience advise using a 2 x or even 3x rating ? The theory is the mains fuse blows to prevent a fire, and it would, but not until it's too late for the poor transformer.

A time delay fuse ran at 1.5 times it's rating will take something like 60 mins to fail. This seems to me a perfectly reasonable arrangement given the long thermal time constant of a transformer and so allows for irregular peak loudness of more normal use. If that fuse keeps blowing the transformer is not going to last very long.

On the other hand, there is the surge current when you close the standby switch. I suspect that the reason for the 2x-3x rule of thumb. So we have opposing requirements which leads me to conclude that the fuse rating should be close to the transformer rating and that the issue of surge current needs to be dealt with by other means for example a negative temperature thermistor.

[Helmholtz]

I agree and and need to clarifiy that Merlin's statement seems to be an observation rather than a recommendation.

http://www.valvewizard.co.uk/fuses.html

(Actually PT manufacturers should know best and I think they should be bound to specify the fuse ratings.)

[uneumann]

1A fuse was fitted to HT secondary for a 420mA winding. ..... Reproducing those conditions on the bench gave a 950mA RMS secondary current. A 1A fuse would never blow and so the transformer would over heat. Not a good situation.

Yikes - Agreed - but the transformer rating is the problem. No fuse can solve that. That amp simply can not be played safely.

[pdf64]

Maybe add some protecting resistance in series with the winding, in order to limit the max current?
Similar to what would be done with a tube rectifier, should the effective HT winding resistance be too low.

[nickb]

Yikes - Agreed - but the transformer rating is the problem. No fuse can solve that. That amp simply can not be played safely.

..as is self evident But perhaps you should tell that to Hammond. It's their JCM800 100W transformer 290HX. I'll leave it at that to keep this thread on topic (fuse ratings).

[SoulFetish]

But perhaps you should tell that to Hammond. It's their JCM800 100W transformer 290HX.

You know, might not be a bad idea. The model has failed in this fashion enough times to determine a pattern. If I were Hammond, I would want to know.
I have a question(s) as it relates to this topic. I can use this hammond transformer as an example.
This is essentially a problem of heat, with the breakdown of the insulation causing the failure, correct?
Could hammmond keep the same transformer specs, as it’s currently designed, and change only the class of insulation to an appropriately higher temperature, and have the transformer work normally at the load you mentioned?

[Helmholtz]

I checked the specs of a few more 100W amp PTs. All of them have a rated B+ winding current below 50% of the amps' current demand at full power (using PSUD2).
It seems that these PT ratings/specifications are based on the assumption that max power is typically only demanded for a period of time that is short compared the thermal heat-up time of the PT with ample low power cooling periods between power peaks.