Question regarding fusing the B+ rail

[Retrovert]

I just recently started using mosfets to save power switches
http://audiokarma.org/forums/index....orth-5-and-an-hour-or-so-of-your-time.504673/

Yes, I saw that. Elegant and simple. I've been using external switches on a socket strip to save mechanical wear and tear beyond the contact damage.

I was thinking a mechanical relay in concept. I didn't think there'd be an arc because contacts are shunted by fuse. Mosfet takes it out of picture.

The arc occurs wherever the contacts break a circuit. So if you had a switch in the B+ the arc would occur at that switch. The fuse is just a piece of wire somewhere in the feed. All the fuse can do is limit current by overheating and opening, assuming it does not itself arc, and it takes a while to do that.

Here's a video which demonstrates the problem: www.youtube.com/watch?v=Zez2r1RPpWY

Notice how the DC arc (a mere 220 VDC) sustains itself across a large distance, far larger than the distance between contacts in a relay.

 

[triode17]

Brice - This is what I spec. into my own designs or existing vintage equipment. I break the B+ line going to the Output Trans. C.T. Insert a DC ammeter. Run the amp to full sine wave power and measure the current. I then multiply the current by 150% and use the next highest value fast blow fuse. Of course, this only protects the output trans. from a power tube short.
To go all the way, I add two fuses to the H.V. secondaries. This protects the rectifier if a filter cap. shorts and protects the power transformer if either a cap. shorts or the rectifier(s) short. But this measurement is of AC current.
Ultimately, use a ceramic body type fuse as they have an arc quenching powder inside to prevent arc overs which can re-conduct, if a glass fuse element gets vaporized onto the glass from end to end.

 

[nj pheonix]

I understand the arc when opening a circuit. What I'm not following is, if the fuse and the relay contact are in parallel, the circuit isn't opening. Only a small fraction of a milliamp would be broken due to the very low resistance path of the fuse

 

[Retrovert]

Contacts don't cleanly close on relays. There's slight bounce so the two contacts may not instantly and fully close. Irregularities build up on the surface from wear and tear, but also from pitting resulting from wear and tear and arcing.

Anyway, an arc formed on open or close can weld the contacts shut. This is why speaker protection systems can't use relays to open the circuit. Well, they can, but it doesn't always work as expected...

You need to check the DC rating for any relay. This is usually a fraction of the AC rating. Why smaller? The reason is arcing. In AC every 16 mSec the zero crossing extinguishes any arc that might be speculating about the possibility of occurring. In DC? No such luck. The arc just goes and goes and goes until you turn the juice off. Watch the video I above linked.

If the power supply has chokes and capacitors, these will still be energized when the relay coil de-energizes and the contacts open. As that circuit tries to break the flow of current, their voltage rises and they'll discharge through the contacts causing an arc, either damaging the contacts or welding them together. Read about flyback diodes and how they prevent destructive arcs from flux decay.

Interrupting high voltage with a relay is not a trivial thing to do, and I'd suggest careful thought is required before using an electromechanical relay. Relays are oft used and oft misunderstood.

 

[triode17]

That's a perfectly legitimate solution if it were for, say, low-voltage heaters.

I above described that solution as using a time-delayed NC relay to switch out a resistor which current limited B+. The problem is that when the relay closes the contacts may arc because of the voltage across them (contact closing jitter) and will become pitted or weld shut. Switching high voltage is tricky.

I'm also not sure what happens as the relay coil energizing stops when the unit is turned off if any B+ remains. As the contacts open with B+ across them this will cause a arc, damaging the contact and potentially welding them shut. It depends on how fast B+ declines. I would not want to base a circuit on that.

Instead of a mechanical relay it would need to be a high-voltage MOSFET. Which allows the conduction of the MOSFET to be varied with an RC constant.

So, yes, one may use a bypass, and its a perfectly good concept, but the implementation cannot be electromechanical.

 

[triode17]

Retrovert- I disagree with your statement, "The problem is that when the relay closes the contacts may arc because of the voltage across them..."
The voltage across the resistor starts out high, because it's absorbing the power needed to charge the caps. quickly. The longer the time delay, the lower the series voltage across the resistor becomes. After a short time, the voltage across the caps. becomes much larger than the drop across the series resistor. At that time, you can safely short the resistor out with the relay.
If anything, the open relay contacts could arc if they can't withstand the initial resistor voltage.

 

[Retrovert]

Retrovert- I disagree with your statement, "The problem is that when the relay closes the contacts may arc because of the voltage across them..."
The voltage across the resistor starts out high, because it's absorbing the power needed to charge the caps. quickly. The longer the time delay, the lower the series voltage across the resistor becomes. After a short time, the voltage across the caps. becomes much larger than the drop across the series resistor. At that time, you can safely short the resistor out with the relay.
If anything, the open relay contacts could arc if they can't withstand the initial resistor voltage.

First, consider powering off which is where the big problem happens. At this point the system is fully energized and the closed contacts (relay must be energized to close) are now opening (normal state of the relay is open), interrupting the flow of B+. I'd written:

If the power supply has chokes and capacitors, these will still be energized when the relay coil de-energizes and the contacts open. As that circuit tries to break the flow of current, their voltage rises and they'll discharge through the contacts causing an arc, either damaging the contacts or welding them together. Read about flyback diodes and how they prevent destructive arcs from flux decay.​

Do you agree with this?

Second, consider the system when on and the relay is closing to take the current-limiting resistor out of the loop. There's plenty of energy stored in the inductors and capacitors. When the relay closes the jitter in the contacts moves them slightly apart during the close. That tiny open/close jitter causes micro-arcing which damages the contacts.

Relay contacts must be rated to withstand the voltage being switched in or out. It is tricky to interrupt high voltage, which is what happens when the relay de-energizes. As proof the system contains energy, think about hot-switching the amp. For up to tens of seconds that energy is retained, unless bleeder resistors are used for the capacitors and flyback diodes for the chokes. The transformer also has some residual flux in it which must go somewhere.

If the choke cannot dump its flux as current the voltage will rise (dv/dt is huge) until it can. This is why an unloaded output transformer arcs over: the flux collapses and the voltage rises until current can flow. Same thing in an inductor.

So there is much potential for trouble with a relay.

Make sense?

 

[Retrovert]

Wait, there's an easy solution to this. Talk about being primed by the question instead of seeing the forest.

The relay and the resistor are placed in line with the AC B+ secondary winding before the rectifier. That way AC is being switched, not DC, and there's no arcing problem. So at startup the maximum B+ winding current is limited by the resistor, and whatever it puts out is rectified and filtered, and that switches out letting the voltage rise to its proper level. The fuse won't blow because the maximum current is below its peak current.

Don't switch the problematic DC, switch the easy AC. Still need a high-voltage AC relay, of course, but that's far more easily obtained than a HV DC relay.

I think that solves the issue if one wants to use a relay instead of a MOSFET.

 

[Brice]

Thanks Andy for your suggestion.

Maybe the idea of fusing the VDC B+ is just wrong, unless spending $30+ for a high voltage fuse...

All vintage McIntosh had a fuse after the SS rectification, and before the choke, a 2A slo-blo littlefuse 250VAC which never ended up as a mini Tchernobyl as far as I know.

Maybe the current we have is so low compare to your video Retrovert.

 

[Retrovert]

Maybe the idea of fusing the VDC B+ is just wrong, unless spending $30+ for a high voltage fuse...

Wow. Where you buying your $30 fuses? I want to sell fuses like that!

I just did a quick check. A 500 VDC @ 0.125 A ceramic cartridge fuse is $5 from Mouser, not the lowest-cost vendor around, but it has an decent search tool.

My dream of a multi-continent $30 audiophile fuse empire collapses. Sadness...

All vintage McIntosh had a fuse after the SS rectification, and before the choke, a 2A slo-blo littlefuse 250VAC which never ended up as a mini Tchernobyl as far as I know.

Well, those are the wrong fuses to use, for the reasons I above explained.

AC fuses are not suitable for DC use, particularly high voltage.

Don't believe me, believe LittleFuse:

www.littelfuse.com/technical-resources/frequently-asked-questions.aspx#question_0
Can fuses with an AC voltage rating be used in a DC applications?
Fuses must be rated for the voltage AC or DC in which they will be used. Generally, fuses have a DC voltage rating that is half of the maximum AC voltage rating. (Example: LLSRK_ID - 600 VAC, 300VDC) Consult the factory for specific DC voltage ratings. ​

Every other fuse vendor will clearly state the same limitations. Those warnings are not to sell you more expensive parts, its because AC fuses do not identically function in DC applications; some do not function at all in DC applications.

I am amazed that using an AC fuse on a DC circuit (for which it is NOT rated) made it past UL.

Oh, wait a minute! Does it have a UL rating. Not all products do. This turkey fryer, for example, does not: www.youtube.com/watch?v=yObDuYTfudY

Using 2A fuse on the B+ is very large, btw. That would only fail if something melted down into slag.

Maybe the current we have is so low compare to your video Retrovert.

Um, no. Not at all. This is just basic physics. Low current destructively arcs. Ever see a Jacob's ladder? That is low current but high voltage. Once the arc gets the air ionized it is easy to maintain, even as the electrode gap widens.

Voltage is what initiates the arc (must be high enough to exceed the insulator's breakdown voltage, in this case an air gap), and the current flow maintains it and does the damage. But low current arcs are still very damaging. Here's an example. Walk across the carpet, scuffing your feet. Now get close to a grounded switchplate. That discharge is high voltage, low current.

That video was just for illustrative purposes, BTW, to show that once started a DC arc does not readily stop unlike an AC arc. I was hoping the concept of zero-crossing in AC vs. no zero-crossing in DC would be understood.

 

[gadget73]

Unfortunately using the wrong part just because its cheap is not the most uncommon thing in the world. I'm sort of curious if McIntosh specified that fuse in the service literature, or if it just ended up being the one that was commonly installed because "fuses is fuses".

Definitely would agree that 2 amps seems like a whole lot of current for a B+ fuse. It won't pop from charging surges though.

 

[Retrovert]

Unfortunately using the wrong part just because its cheap is not the most uncommon thing in the world. I'm sort of curious if McIntosh specified that fuse in the service literature, or if it just ended up being the one that was commonly installed because "fuses is fuses".

Ok. I located the service manual. (How did we live without the internet?)

"The MC240 uses a 3.2 ampere sloblo type fuse." The schematic shows this as a fuse on the transformer primary, not on the B+.

I don't know if different McIntosh models differently fused things, and am not about to go on an easter egg hunt to find out. I'd suspect not. Just a traditional primary fuse of dubious value for anything other than seriously adverse conditions.

So this is a very different sort of arrangement than fusing B+.

Definitely would agree that 2 amps seems like a whole lot of current for a B+ fuse. It won't pop from charging surges though.

Exactly. It won't pop from anything short of the End of Days. So what does it do? This must have been the discussion:

Marketing Drone: (knocks on open door frame) Hi! I'm from Marketing. (fake smile) You're the lead engineer for the new amplifier? Just wanted to pop over and say we need some sort of whatchamacallit, umm, you know, one of them, errr, fusey things on the new amp. (nods and looks expectant)

Engineer: Ahhh, yes. I discussed this with your boss and he was dead set against that. Told me there was no budget for safety features. (rubs hands) Nice you've come around. (smiles, very excited) Well, we'd need to do some engineering for a time-delay relay with a resistor and bypass to reduce inrush load and ensure accuracy in fusing, and would likely need to install fuses on the transformer, power supply, and each output transformer, just to be sure. I can work that out for you as an engineering change order so your boss can approve it. Be done by the end of the day. (takes out sliderule starts moving it, looks up, excited) This will be a great safety feature, you now.

Marketing Drone: (visibly dismayed, gives frown) Riiiight... Look, we, uh, just need some sort of fusey thing on this. (lowers voice conspiratorially) You know, for the stores. (fake smile) You can do that, right? (nods, looks expectant)

Engineer: (looks up) What??? What???

Marketing Drone: Okaaaay. (speaks as if to child) Ummm, let's be a bit clearer. All rightey. Budget like, oh, say, maybe, uh, twenty cents? (raises eyebrows, looks expectant) Yeah? Can do?

Engineer: (incredulous, shakes head, speaks as if to a moron) A fuse holder and the fuse is thirty-five cents. (waves hands, visibly annoyed) All adding that might do is keep the thing from bursting into fire if something went horribly wrong, but the power transformer could be fried, the rectifier and tubes could be slag, and the output transformers could be in flames. That would be pretty much useless as a protection measure. The owner would likely need to buy a new amp. (folds arms, nods, looks annoyed and dubious) And let me guess. You want a slow-blow fuse, too, so it never blows and never needs replacing, right? (shakes head and rolls eyes in disgust)

Marketing Drone: (face lights up with happiness) Thirty-five cents? Never blows? Yes! (pumps hand in air) Got it in one! Knew you could handle this! Right, then, we're all done here! Excellent. Let me know when the fusey thing is installed so we can take photos for the ads! (turns around, leaves office with spring in step)

Engineer: (grumbles, puts away sliderule, takes out briefcase, removes newspaper, turns to "Help Wanted" section, begins reading)​

 

[Brice]

Found one suitable: LITTELFUSE 0508.315MXP Fuse, Cartridge, 508 Series, 315 mA, 1 kV

 

[Retrovert]

And there was much rejoicing.

But that's 1 kV and you're paying for that extra capacity. Might be over specifying the part.

Ok, it's only $7, but you'll need six fuses which adds up:

(2) for the transformer secondary halves before rectifier
(1) after rectifier but before filter capacitors
(1) after the power supply filter bank (capacitors and inductors)
(2) for the output transformer center taps. Roughly half that current capacity for each of the output transformers.​

Also should fuse the heater supply and the rectifier supply. Protects against shorts.

As long as you're doing this, you might want to look into flyback diodes for the output transformers, chokes, and power supply transformer. Prevents arcing from flux collapsing.

Also bleeder resistors for the capacitors.

Belt, suspenders, and a single-action Colt .45. Just in case. Never know what trouble turns up in the saloon when you're playing a piano concerto.

 

[Brice]

All right, thanks all for your advice.

 

[Steve O]

Ok. I located the service manual. (How did we live without the internet?)

"The MC240 uses a 3.2 ampere sloblo type fuse." The schematic shows this as a fuse on the transformer primary, not on the B+.

I don't know if different McIntosh models differently fused things, and am not about to go on an easter egg hunt to find out. I'd suspect not. Just a traditional primary fuse of dubious value for anything other than seriously adverse conditions.

So this is a very different sort of arrangement than fusing B+.



Exactly. It won't pop from anything short of the End of Days. So what does it do? This must have been the discussion:

<snip>

​

And now for the rest of the (Mac) story:

Most vintage Mac models employing SS rectification utilized a voltage doubler circuit (MC225, MC40, MC240, and MC75). Since the B+ was in the vicinity of 450VDC, a high current non center tapped secondary @ ~ 170VAC was utilized. Additionally, most of these models placed a 250VAC rated fuse on one leg of the sec. of appropriate current rating prior to the doubler circuit. Thus the fuse sees the 170 VAC secondary and not the 450VDC post rectifier and filter. I think it's pretty safe to assume that this arrangement is comfortably within the capabilities of a 250 VAC rated fuse. Curiously, this secondary fuse did not show up in any of Mac's schematics or other documentation that I'm aware of. But it's in the physical product if one looks for it.

In addition to the secondary fuse, Mac also employed the traditional primary fuse and in many versions, an NTC inrush limiter in the primary circuit. From a this it should be evident that in Mac's later consumer tube amps (1960s), safety and long term reliability were important design condiderations.

Edit: Just located a SAMS folder for the 240 dated 1961 that does indeed call out the secondary fuse on the schematic and in the parts list. Speced as 3A 250V AGC type. Interestingly, I usually find a 2A SB in that position. Undocumented production change or???

 

[trobbins]

Steve O, that location of secondary fuse for a doubler is normal and expected. The littlefuse app guide (link in post #8) in the Voltage rating section indicates that a 250VAC is ok for use in secondaries with voltages well above 250VAC, so not an issue for typical valve amps (even perhaps for those with secondaries exceeding 500VAC).

Doubler's incur much higher initial rms current levels, such that a slow-blow T fuse on the secondary is highly likely to provide better discrimination, as a significantly lower fuse rating can be used (compared if you were choosing a fast blow F). Although an NTC can generally be of benefit, it can be a pain for fuse selection, as the thermal cooling time constant of the NTC can be quite slow, whereas the main DC filter caps typically discharge in a second or so - and hence any trigger happy person on the mains AC switch can stress the fuse if the fuse had its rating/type reduced 'because of the NTC".

 

[Retrovert]

Although an NTC can generally be of benefit, it can be a pain for fuse selection, as the thermal cooling time constant of the NTC can be quite slow, whereas the main DC filter caps typically discharge in a second or so - and hence any trigger happy person on the mains AC switch can stress the fuse if the fuse had its rating/type reduced 'because of the NTC".

Yes.

A better approach uses a relay to switch the (now hot) thermistor out of circuit so it may cool. This is the classic issue with the Amperite delay relay (a bi-metallic switch is heated for the delay), which should not be used to carry the entirety of the current when hot although this is commonly, and mistakenly, done. Since the thermistor is placed on the primary or secondary, an AC relay can be used without issue.

Being thermal devices both must be allowed to cool. so that hot-switching (admittedly a bad idea) goes through the same delay cycle. The relay allows such cooling, but it does, of course, have some fixed time and rapid switching before it has fully cooled would have the same problem.

The relay further eliminates the heat from the thermistor and prolongs the life of the (sealed) delay relay.

A MOSFET is the ideal solution.

 

[Retrovert]

Additionally, most of these models placed a 250VAC rated fuse on one leg of the sec. of appropriate current rating prior to the doubler circuit. Thus the fuse sees the 170 VAC secondary and not the 450VDC post rectifier and filter. I think it's pretty safe to assume that this arrangement is comfortably within the capabilities of a 250 VAC rated fuse. Curiously, this secondary fuse did not show up in any of Mac's schematics or other documentation that I'm aware of. But it's in the physical product if one looks for it.

Interesting. I looked at a different service manual and no such fuse is listed. The only one is the rear panel. This must have perplexed customers and resulted in professional service. Which may have been the goal. Because that blown fuse might indicate a bad tube or output transformer.

Yes, a 250 VAC fuse on an 170 VAC transformer winding is, for certain, exactly within specifications and intended use.

Thanks for that bit of arcana.

 

[Brice]

It was done intentionally, to prevent an average user to change the fuse with a bigger one "just to make it work". The internal would blow and that's it. Service call and the amp wasn't been damaged.

Anyway I looked at the Acrosound schematics, and they added a fuse on the cathode side 1/8 A for each tubes. That's funny as for 2 tubes, it's exactly where I am: 250mA.

If I add a 125mA fuse at each cathode, this would work at protecting the OT of a suicidal tube.
Do they need to be high voltage here? I do not believe so as the internal resistance of a conducting tube never goes to 0. Unless we have a catastrophic short of course.
What do you guys think?