Blown 1 OHM current-sense resistors

[Leo_Gnardo]

Thanks for the link. What was your original rationale behind choosing 3Ws?

When I started putting in 1 ohm current sense R's, I had @ 100 ½W 5% CF's I'd gotten ultra cheapo from a surplus house. Although none had ever fused on me, when the pile started looking a bit thin, I snooped the Mouser catalog for something that would 1) give better precision, 2) not likely to break and 3) wouldn't break the bank. So I landed on these Dales.

About point 1. Not too many shops have a milli ohm meter, I sure don't. I have to trust the manufacturer to get the value and 1% precision right. So, call it sloppy if you want, but it's a matter of faith that the value is what it says it is. Until I score a milli-ohm meter, I gotta do what I can with what I got.

Later, when I found Ampeg (or Cratepeg as some like to call 'em) was using the 10 ohm version I figured I was on to something good. Whatever other complaints there may be (and there are many) about SVT-CL and similar amps, their use of a 1% WW part hits me in the right spot.

The recent trend towards miniaturizing resistors IMHO is a bad move. For us anyway. Manufacturers & distributors will offer more attractive prices on the mini ones, and boost prices on the proper-size ones to try and drive the buyers to accept the mini versions. Sic semper transit tyrannosaurus rex post propter copter honk. ('Twas ever thus.)

I like the parallel diode idea - BUT keep in mind if a 1 amp spike is enough to kill the resistor, you'd better have a diode that can handle in excess of that current. Drat, round and round we go again, more expense & trouble to save a cheap part...

[woodyc]

...The recent trend towards miniaturizing resistors IMHO is a bad move...

I don't get it either. At any particular power, a smaller part is going to get hotter (heat capacity), and its going to take longer to cool down (surface area.) There's no way around it.

I was wondering if maybe the new small ones have a better alloy or a better way of bonding it to the leads?

Or maybe they are just leaving it to the user to figure out they need a heatsink?

Have you seen these... they're teeny...

http://riedon.com/media/pdf/UB.pdf

[Steve Conner]

When installing an expensive part to save a cheap one, bear in mind that both parts are cheap compared to the cost of a service call to replace either of them. Of course the service cost is borne by the user if you can just get the thing to last through the warranty period.

A 1N4007 might only be rated at 1A continuous, but it can handle a 30 amp peak surge for half a mains cycle. A 1N5408 can handle 200A peak.

I imagine once a tube had had its innards randomised by vibration inside a combo amp for a while, it could quite easily arc between plate and cathode and blow a small cathode resistor to smithereens.

[R.G.]

Dead right Steve, Leo, and Enzo.

If you're trying to keep the resistors going, use a higher-wattage wirewound. Wirewound is tough, and probably more reliable under overload than the diode that would "save" it.

But the fact that there's enough current flowing somewhere to put more than 1W into a 1 ohm resistor means that the resistor failure is a symptom, not a cause. If the resistor rating is increased until the resistor no longer fails, then something else in the chain leading to that resistor will fail next. That might be the cause, might be another victim like the resistor was.

Sometimes that's the only way to get back to what the original problem is, but it's a slow, frustrating way to get there.

Enzo's on to something. B+ will have a hard time generating an amp for any significant time, and it's likely that this might cause other issues as well. A short to the heater supply can easily supply multiple amps.

Power resistors are often rated in energy as well as power. Energy is power times time; you buy energy for your house in kilo-watt-hours. A short pulse of current in a resistor generates power equal to I-squared R. How much heat that generates is I-squared-R times the time it goes on. So a nanosecond blip of huge power won't make it even get warm, but 200% of rated power for an hour may well burn it out. So to pick things for pulse-power applications, look at the energy rating in Joules or sometimes just "I-squared-T" rating. Many fuses also list I²T, to tell you something about their performance under pulses of current.

A momentarily shorted tube might try to dump the main filter cap into the cathode resistor, but it has the resistance of half of the OT primary in series with it to limit the current. But a momentary short from cathode to heater is only limited by the cathode resistor.

Hmmm. If the heater is hard-centertapped to ground, this could kill a cathode resistor. Soft-centertapping with a pair of 100 ohm resistors means that a cathode-heater short would be limited by the 100 ohm centering resistor as well as the cathode resistor. Maybe a good place to dig.

[Gaz]

As always, thanks for the thoughtful replies everyone.

To be clear, my interest in the topic at this point is two-fold: To figure out what the hell happened to the amp in question, and to figure out a better design choice in the future. Don't worry, I'm not looking to throw in some giant resistors until I know for sure what caused the resistors to fuse in the first place.

From everyone's comments, it seems like logically it could be a heater to cathode short because the large current available there. The specific circuit I used for the filament circuit on this amp was a 300R pot for a hum balance/humdinger with he wiper going to an elevated, filtered voltage off the B+ line in the preamp. So, I'm not sure if that minimizes the possibility of that kind of current flow happening.

I have a 1.5A Slow Blow fuse in one leg of the bridge rectifier on the AC side (per Valve Wizard). Under normal operation I have measured about 940mA AC with max signal there with an ammeter, and 604mA DC on the other side of the rectifier. I thought 1.5A would be a good choice on the AC side. Does it make sense this fuse did not blow before the resistors? Normal max cathode current is about 300mA per pair of tubes.


Now from a design perspective, for future builds I'm wondering what the best course of action would be. A 3W resistor allow 1.7A, and I hope the fuse would blow before that blew, but I'm thinking it may be good to toss a 2A diode in there that would turn on well before that limit. That way I would not be relying on the diode to "save" the resistor, but also not on the resistor's brute power alone. Is that a reasonable line of thinking?

Thanks!

[Enzo]

The diode limits the voltage to about a half volt. Half a volt on a 1 ohm resistor is about a 1/4 watt, isn't it?

If this stuff is burning up, you already have a serious problem. We hope a fuse will be blowing. But there is no reason to let parts burn up in the meantime.

As to redesigning things, there just isn't anything in an amp that can;t fail or be burnt up. OK, OK< the chassis won't burn... SOMETHING will always be the weakest link. As Juan said, if we beef this part up, something else will go up in flames. SO just because some part croaks or burns, that doesn't mean it needs to be redesigned.

Seriously, along the highways they would put these large speed limit signs up on posts. When a car skidded off the road, and into these things, it would break off the post and the sign would fall over. Someone decided this was a problem and changed the sign posts to steel posts set in concrete. Sure enough, cars stopped knocking them over, instead cars were now running into immovable steel posts. That increased damage to the cars and increased injury to the occupants. Eventually they came to realize this was not the right approach. Now they are back to 4x4 posts, and at the bottom, they drill a large hole through the post to act like perforation - it makes a predictable weak spot to give way, rather than stop the car dead.

Where do you want YOUR weak spot to be in the amp?

[Gaz]

The diode limits the voltage to about a half volt. Half a volt on a 1 ohm resistor is about a 1/4 watt, isn't it?

If this stuff is burning up, you already have a serious problem. We hope a fuse will be blowing. But there is no reason to let parts burn up in the meantime.

As to redesigning things, there just isn't anything in an amp that can;t fail or be burnt up. OK, OK< the chassis won't burn... SOMETHING will always be the weakest link. As Juan said, if we beef this part up, something else will go up in flames. SO just because some part croaks or burns, that doesn't mean it needs to be redesigned.

Seriously, along the highways they would put these large speed limit signs up on posts. When a car skidded off the road, and into these things, it would break off the post and the sign would fall over. Someone decided this was a problem and changed the sign posts to steel posts set in concrete. Sure enough, cars stopped knocking them over, instead cars were now running into immovable steel posts. That increased damage to the cars and increased injury to the occupants. Eventually they came to realize this was not the right approach. Now they are back to 4x4 posts, and at the bottom, they drill a large hole through the post to act like perforation - it makes a predictable weak spot to give way, rather than stop the car dead.

Where do you want YOUR weak spot to be in the amp?

Thanks, Enzo, I ideally want the weak spot to be the fuse. I emphasized that I'm not trying to 'fix' this amp by shoving in some overrated resistors, but just want to do it smarter next time if possible. The more I think about it, the more I think high wattage wirewound resistors are the way to go. They're simpler, and can be made large enough to withstand a large amount of current. If I play the resistor vs. diode game, I'm tempted to use something as high as a 1N5408 to handle 3A, and a 3W resistor. The diode would turn on at about half the resistors power rating and take over from there, while the resistor alone would handle about 1.7A alone. It seems better in that i can handle higher current and will protect the resistor, but part of me believes that less components is better, and just to stick a couple 4-5W resistors in there. Since this is really up to debate and personal experience to a certain degree, I'm really looking for some good old fashioned opinion here

Now, about the initial failure, I'm still curious about the fuse not blowing, and if my heater arrangement could still lead to such a high current fault without taking out everything in it's path (hum balance pot, heater elevation circuit, tubes).

Thanks again for your thoughts everyone!

[R.G.]

Once again Enzo has the right question - what do you want the weak spot to be?

One obvious answer is to build in a weak spot where you can get at it easily. This is the theory behind fuses. The problem with fuses is that you may have to use a lot of them, and they have tolerances and time delays that may make them difficult to pick for the protection you want to happen.

This philosophy is what is behind electronic protection. In general, you can set an electronic protection to a finer resolution than a fuse. Power supplies in general (although not in tube amps) quit being gee-it-never-did-THAT-before devices when the protections started being integrated to shut down on overcurrent, overvoltage, over temperature, etc. And they became forgiving. Most power supplies these days can't be damaged easily by dead shorts and such because they have an electronic "weak link" that stops the damaging condition before it eats parts and then allows things to come back up when the badness is removed.

Tube amps don't do this for historical and psychological reasons, although B+ fuses and AC primary fuses have been reasonably well accepted.

In this particular amp's case, something is Not Right. Other amps don't commonly have this problem, so this amp is unusual in some way. It's probably not a design issue, it's an implementation or random chance issue.

Some amps use 1/4W resistors in the cathodes of the power tubes deliberately to act as both current sensors and fuses. Ugly, but possible. The owners typically do the equivalent of putting in a bit of aluminum foil over the AC fuse by putting in much bigger-power resistors. They no longer fuse and open, but then the things they were "protecting", the tubes and power transformer, are at risk.

I've done some private work on issues like this by putting together a sensor that looks at per-tube current and sensing too-much-for-too-long currents, and shutting things down when that happens. It worked pretty much as designed.

I don't know if it would help in this case, but maybe or probably. It opened the cathode circuit entirely by turning off a 50-milliohm MOSFET in the cathode. That would catch any tube faults or inside-the-tube shorts, but not problems with shorts directly to the cathode resistor outside the tube.

Again, I don't think this is a design problem - it's a bug in this amp, but a subtle one.

[tubeswell]

Post hoc ergo propter hoc is a logical fallacy. DOn;t lock your mind around a cause and effect relation just because two things happen at the same time. The other day I sneezed, and at that exact moment, a police car siren went off on the street right out front. I am still trying to figure out how my sneeze caused that.

llamavirus

[Gaz]

RG,

I think we may have posted at the same time, but again, I really appreciate you comments. User T. Robbin's once wrote: 'it's tough to know how paranoid to be.' I definitely feel that every time there's the odd failure like this that has me thinking about where I wasn't paranoid enough!

I'd like to think that if I'd made the resistors a wee bit bigger, used a wirewound type, or used the 'diode trick', that the fuse would have popped. I surely didn't intend the resistors to act like fuses. I'm just having trouble determining how much current I could expect to see at the cathodes max. I know at least 1A

P.S. The MOSFET idea sounds brilliant (but we all say that whenever you mention MOSFETs, don't we?... then go right back to our old ways )

[Enzo]

One useful way to look at fuses is in terms of power. With 120v mains and a 3A fuse, that represents 360 watts. To blow that fuse, something inside has to dissipate more that 360 watts. Or a B+ fuse? Got 480v and a 1A fuse? Then 480 watts has to be coming through that fuse to even think of blowing it. Think about that when a 1 watt resistor burns up yet doesn't blow the fuse.

[Leo_Gnardo]

OK, OK< the chassis won't burn... SOMETHING will always be the weakest link.

Plz allow me an off topic moment:

Just for the record, one customer brought me his collection of 3 Burman amps a few years ago. Absolutely excellent amps! One little problem though, very light duty chassis. One started to break apart and someone tried to weld it back together, burnt holes in the corners making it even more unstable. A lost cause. It would have been a major project to make a custom chassis & move the amp lock stock n barrel, no dice on that idea. Great amp, just don't move it anywhere. :P

[Gaz]

One useful way to look at fuses is in terms of power. With 120v mains and a 3A fuse, that represents 360 watts. To blow that fuse, something inside has to dissipate more that 360 watts. Or a B+ fuse? Got 480v and a 1A fuse? Then 480 watts has to be coming through that fuse to even think of blowing it. Think about that when a 1 watt resistor burns up yet doesn't blow the fuse.

Enzo, you, with your Socratic answers That makes perfect sense considering Ohm's Law, but I feel even more clueless because I'm having trouble envisioning what's actually happening in the amp when a fault occurs. It seems to me now, that the fuse would never blow before those cathode resistors!

[km6xz]

Gaz, you will get to the heart of the problem sooner if you forget the distraction of the resistor power rating or size, as mentioned several times above, it is a symptom, not the cause. Logic would have it that there is only one path for enough current, over a long enough time being the heater windings on the transformer. But that assumes the circuit is wired as intended, the socket, chassis, and wiring(or if using a PCB, a path through the board). Was it s fluke or arc, or is there more peak currents than you have read. The meter will not respond to fast risetime/decaytime events so try your scope and see if there are unexpected signal characteristics that are not picked up by the meter.
The recommendation of 1% resistors assumes that the value means that much by itself as to what is dropped across it. If going by the tube manual, the conventional current target might not really be relevant to your particular tubes since we know from brand to brand, or tube to tube off the same assembly line, they can vary widely.
Setting goal for idle current assumes a lot but says little about what the actual performance of the tube is in your circuit. Find the optimum setting for your amp, with your tubes, and record what the value is at idle for future reference. You need to find the best operating point not particular cathode idle current. At would point does the output waveform have characteristics associated with the best sound? Your scope will be handy in that task.

[Enzo]

I agree with Stan, the issue is not the resistor, the issue is what happened. I offered the power thing to explain why a resistor can burn up without taking out a fuse.

Look at the data sheet for a fuse or fuse type. ANy fuse company publishes curves on their fuses.

A 1A fuse does not just pop at the first moment you pass 1A. I looked up a 3A fast fuse. At 3A, it won;t blow. AT 4A it takes two minutes to blow. At 5A it takes a second or two. AT 10A, maybe 1/8 second.

A 3A slow blow fuse will run at 4A all day. At 7A it takes about 10 seconds. 10A takes about 2 seconds.


And that is plenty of time for something to burn up before any fuse blows.