Fast or slo blo?

[lowell]

I'm trying to iron out why the following seems true:

So tube amps typically have slo blo mains fuses and fast blo HT fuses. Also it seems that solid state amps tend to have ONLY fast blo fuses.

Tube amps have an initial startup surge because of the heaters correct? Therefore a slo blo is appropriate, as a fast blo would blow from the cold heaters "heating" up.

The HT fuse is fast blo because there is no startup surgein the HT line (heaters are cold) and any surge in the HT line is a bad thing: shorted tubes, bleeders etc...

I'm thinking that solid state amps have fast blo fuses for the mains because solid state amps tend to have turn-on surge suppression designed into them. (thermistors, relays etc...) Is this correct? It's one of those things I've always thought I knew but when I really thought about it I wasn't completely sure.

[teemuk]

Tube amps have an initial startup surge because of the heaters correct?

The heaters are at least partial reason, yes. A cold filament will draw more current but the current is limited when the filament begins to heat up and due to that increase in resistance. Basically, the operation of it is very much like that of a light bulb limiter.

BUT who says solid-state amps have only fast blow fuses...? I've seen plenty with slo-blo and for the good reason that the inrush surge can be ludicrously high in them as well. They don't have filaments alright, but the reservoir capacitances are often substantially larger than in tube amps and the rectifiers are almost without question solid-state and "instantly conducting". Also, throw in a toroid with high VA rating and the stage is set for massive inrush current draw.

It's not so much a problem of < 50W amps but once you start to get past that the inrush is a very real phenomenon and will eat conservatively rated "fast" fuses for breakfast. Either you throw in a fast fuse rated for a very high current, or a slow blow fuse with more effective rating, closer to limits of real current draw of the amp when voltages have settled.

Some SS amps do have inrush current limiters, mainly PA amps and alike with substantially high output power ratings. I think the rule of thumb recommendation is to start fitting those things in once the VA rating of the power transformer goes past 300. But it's somewhat rare to see amps in 50W - 200W range using such. There are thermistors in eg. Fender and Crate amps, but then again eg. Marshall SS amps often seem to be without inrush limiters and are fused with slo-blo. If it says something like "T1A" or "T2A" in the schematic it refers to a slo-blo. (T = time lag). And it's quite common to see that "T" in there.

I would think that without the thermistor or some other kind of inrush limiter almost all of those SS amps with fast-acting fuses would indeed require a slo-blo mains fuse, since the rail-to-rail voltage is usually in the range of 60 volts or more, total rail capacitance can easily be in the range of 10000uF, VA rating can be close to 200 and the rectifiers are semiconductor diodes.

I think some tube amps with tube rectified PS could infact cope with fast blow mains fuse since the rectifier limits the inrush and sets a limit for the reservoir capacitance.

[R.G.]

As much as anything else, it can be a matter of transformer size. All power line transformers have a turn-on surge all by themselves, whether they're loaded or not. The size of the surge depends on the size of the transformer and exactly where in the AC cycle they are turned on compared to where in the AC cycle they ran down to when turned off (that is, the remanent flux in the iron from turn off).

This can be big. We used to have a 15A variac in the power supply lab that regularly popped wall breakers at turn on. Measurement of the surge showed occasional peaks of over 200A. But this was a very big transformer. As Teemuk notes, SS amps in the smallish range (up to 100-200W) tend to not have inrush limiters. Beyond that, the inrush is so big that you start needing some kind of slow start.

Charging surge to filter caps and/or heaters is on top of the transformer inrush.

It's helpful to remember that AC power fuses are not there to save components; they're put in to stop fires when the amp is already so damaged that it will start melting and burning. If an AC power fuse happens to save your amp, that's nice, but it's not what it's there for. Other fuses are put in amps to (try to) save components. If you really want to save components, do current limiting with electronics, not fuses. Semiconductors in particular are notorious for giving their lives to save fuses.

[Jazz P Bass]

Semiconductors in particular are notorious for giving their lives to save fuses.

Good observation, RG
"Ooh ooh that smell. Cant't you smell that smell"

[R.G.]

And correspondingly, semiconductors are fast enough to stop sub-microsecond overloads, if they're used properly.

I once made what I subtly called a "Tube Saver". It watched tube current through a 1 ohm resistor, lowpass filtered that, and shut off the current with a high voltage MOSFET in the cathode circuits. There was various stuff lighting LEDs to say what went overcurrent, options to let it stay off for a while then come back on, stay off til power cycled, etc. The big problem was slowing it down enough so that it didn't over-eagerly trip on transients. I could cut the bias wire to the output tubes and ... nothing happened except the overcurrent light came on. Same with turning the bias pot too far.

I did that with analog stuff and CMOS logic, but it could be done better these days with a single $1 microcontroller, a MOSFET and a few other parts. The uC would do the lowpass filtering and watching for fast transients, slow overcurrents, timing how long to let an OC go on, looking at transformer temperature, etc.

Real tube diehards are nearly always offended by the idea, but it worked well. No, it did not audibly change the tone of the amp, as demonstrated by a switch which shorted the current-watching resistors. But a real tube diehard would say it really sounded worse, it just can't be measured, or heard as worse when you're testing it. 8-)

[jpfamps]

And correspondingly, semiconductors are fast enough to stop sub-microsecond overloads, if they're used properly.

I once made what I subtly called a "Tube Saver". It watched tube current through a 1 ohm resistor, lowpass filtered that, and shut off the current with a high voltage MOSFET in the cathode circuits. There was various stuff lighting LEDs to say what went overcurrent, options to let it stay off for a while then come back on, stay off til power cycled, etc. The big problem was slowing it down enough so that it didn't over-eagerly trip on transients. I could cut the bias wire to the output tubes and ... nothing happened except the overcurrent light came on. Same with turning the bias pot too far.

I did that with analog stuff and CMOS logic, but it could be done better these days with a single $1 microcontroller, a MOSFET and a few other parts. The uC would do the lowpass filtering and watching for fast transients, slow overcurrents, timing how long to let an OC go on, looking at transformer temperature, etc.

Real tube diehards are nearly always offended by the idea, but it worked well. No, it did not audibly change the tone of the amp, as demonstrated by a switch which shorted the current-watching resistors. But a real tube diehard would say it really sounded worse, it just can't be measured, or heard as worse when you're testing it. 8-)

Tubesync have a system that seems to work like this.

TubeSync - Guitar Tube Amp Biasing, EL34 s valves / tubes and more

It's amazing what some people can hear, especially as many builders already include current sense resistors to allow easy bias adjustment (the MOSFET will have no significant resistance when fully on).

By only reservation with this system is that in the event of a valve short, lifting the cathodes may cause the cathode/heater insulation to break down and the HT imposed on the filament supply and subsequent collateral damage. I've seen this happen.

[R.G.]

Tubesync have a system that seems to work like this.

TubeSync - Guitar Tube Amp Biasing, EL34 s valves / tubes and more

Hey - someone else caught up! I did the tube saver back in 1999 as I remember. They include auto biasing, looks like. I did an auto bias for the Workhorse line of amps, then decided not to put it in because of tube drift at startup, and the ease of biasing with the red-green light indicators.

By only reservation with this system is that in the event of a valve short, lifting the cathodes may cause the cathode/heater insulation to break down and the HT imposed on the filament supply and subsequent collateral damage. I've seen this happen.

I worried about that on the first proto. As a practical matter, I never had one do that in the small number of tests I ran. Just as a preventer for that, I put in a 100K power resistor paralleled with the MOSFETs. This kept things in bounds when the MOSFETs turned off.

Once you make the step to actively sensing what happens and taking steps to prevent ugliness once it does, it's simple enough to include an optocoupler and triac on the primary side or HT winding or a MOSFET in the return side of the HT winding too. That's not fast enough to save a tube in a real meltdown because of the filter caps, but it would stop a majority of possible collateral damage. There are other things I'd include as a full amp saver system. I'm sure the guys you did the pointer to will have thought of some of those as well.

[Steve Conner]

But a real tube diehard would say it really sounded worse, it just can't be measured, or heard as worse when you're testing it. 8-)

I think a real tube diehard would say "F*** in' thing keeps cutting me off in the middle of my solo"

[R.G.]

I think a real tube diehard would say "F*** in' thing keeps cutting me off in the middle of my solo"

Yep. Good solos are always done in conditions where you're not giving your life for your music, but your amp is.

My friend the amp tech says he can't count the number of times someone has brought in a melted-down amp and said "it was really sounding GREAT just before it blew up. Can you make it always sound like that?" The classic illustration to that is an EL34 he handed me with a conical depression in one side where **the glass melted** and sucked down into a cone into the tube.

Hey... here's an idea: a device that fades in a "boost channel" to another power section as you get louder and louder. Maybe something like a compressor that begins compressing signal to the "tone amp" and feeding more to the "loud amp" or PA at some power level.

[Steve Conner]

Well, I experimented with these protection circuits too, and that's what they did, cut off the amp just as it started sounding good. When I investigated in more detail, it turned out that all the classic amps were destroying their tubes by design. In the light of this it was impossible even to define what "protection" meant, so I gave up.

[R.G.]

Well, I experimented with these protection circuits too, and that's what they did, cut off the amp just as it started sounding good. When I investigated in more detail, it turned out that all the classic amps were destroying their tubes by design. In the light of this it was impossible even to define what "protection" meant, so I gave up.

Good observation. Destroying tubes by design is another way to look at it. I suspect the actual designers of the classic amps would be horrified if they heard the uses to which their amps are being put, but that's another issue too.

Hmmm.... I wonder if it's possible to design a tube specifically for musical amplifier use that is beefed up to withstand the normal (ab)use that musicians do these days, and preserve the internal electron ballistics that account for how the tube sounds. For instance, making the plates and grids out of a tungsten core, plated with iron, plated with (whatever) to make them essentially un-meltable. Making the glass envelope out of high temperature borosilicate glass that's much thicker. Making the heaters either heavier, thicker, more heat and evaporation-resistant and with more heater/cathode insulation for voltage issues; that kind of thing. It'll never be done of course, but I suspect even mild redesign effort for longevity in the specific application would do a lot. It's a different optimization point from the designs which went into electronics in the Golden Age.

[teemuk]

^Maybe. I already know that it doesn't sound as good as "vintage" and "NOS", though.

[jmaf]

So, I guess the OP's points about filaments, RG's about transformer core filling up, and cap surge account for 100% of the fast-blo slo-blo debate? A switched mode supply with tiny input transformers and caps don't ever need slo blo?

[R.G.]

So, I guess the OP's points about filaments, RG's about transformer core filling up, and cap surge account for 100% of the fast-blo slo-blo debate? A switched mode supply with tiny input transformers and caps don't ever need slo blo?

It happens a different way in switching power supplies. They uniformly - even the tiny ones - have the AC power line connected to a diode bridge and that dumps right into a filter cap. In this case, the surge is both big and fast. Unless a power supply has some kind of slow start designed into it, it's going to pull a power on surge.

[Steve Conner]

Yes! They are nasty little buggers as far as inrush current goes. In a piece of gear with a transformer, the resistance and leakage reactance of the windings limits the inrush, but with a switcher, you're basically connecting a big, low ESR, discharged capacitor straight to the mains. If you read the datasheet for an OEM switcher carefully, you'll see a specified inrush current of something like 80 amps.

Here in 240V land the surge is at its absolute worst, and it's not uncommon to hear a pop and see a sizeable blue flash from the outlet when plugging them into the wall. My Apple laptop power adaptor does it every time.