To fuse or not to fuse (on the secondary side of the transformer)?

[rock soderstrom]

Hi folks,

I have already seen a lot of variations on this topic. What is your opinion?

Edit: The mains (primary) side of the transformer is of course always fused.

If you vote for the fuse on the secondary side, where would you implement this fuse?

Before or after the rectifier?

For a bipolar power supply, see attachment, is one fuse in the ground path sufficient or do I need two fuses for the +/- voltages?

 

[beatnik]

I have seen similar supplies with fuses on +V / -V lines after the regulator, but I did not understand if they were actually useful since the LM317 / LM337 has inbuilt overload protection, and extra diodes can be added for short circuit protection like on your schematic.

I guess a fuse before the rectifier would also protect from shorts in the rectifier itself, but this fault might be already covered by the mains fuse.

 

[JohnRoberts]

It depends, what is the purpose for the fuse?

UL sometimes adds fuses to products (after the transformer) to prevent fires caused by overheating from some expected component failure. Many solid state devices (like diodes) fail as short circuits that can then pass excessive current.

These internal fuses that UL requests fall under the category of "no user serviceable parts".. They require printed stickers so service technicians know what value to replace with in case of fusing.

JR

PS; The last big console power supply I did had fuses and two color LEDs on every rail. The LEDs indicated Green for all good, Red for fuse open, and dark for PS rail down.

 

[rock soderstrom]

It depends, what is the purpose for the fuse?

UL sometimes adds fuses to products (after the transformer) to prevent fires caused by overheating from some expected component failure. Many solid state devices (like diodes) fail as short circuits that can then pass excessive current.

In my opinion, the fuse should protect against fire. I think the transformer is the biggest potential danger if it gets too hot due to overload.

I experienced this myself once. Over 30 years ago, I worked in a large and old theater. Once, during a show, a lighting control desk actually smoked heavily and then burned! The smoke was so strong that the show had to be stopped, a fire extinguisher was used, the transformer had caught fire due to massive overload (shorted rectifier??).

I am currently building a Studio Spring Reverb in a wooden box (similar to AKG BX style). The electronics and transformer are mounted on a metal plate, a situation similar to a guitar amp, hence the question.

 

[JohnRoberts]

In my opinion, the fuse should protect against fire. I think the transformer is the biggest potential danger if it gets too hot due to overload.

Many modern transformers already have internal thermal fuses. This is common failure in commercial wall warts.

I experienced this myself once. Over 30 years ago, I worked in a large and old theater. Once, during a show, a lighting control desk actually smoked heavily and then burned! The smoke was so strong that the show had to be stopped, a fire extinguisher was used, the transformer had caught fire due to massive overload (shorted rectifier??).

I've never had any of my designs catch fire, but some of Peavey's early class D amps were jokingly called "fire starters"....

I am currently building a Studio Spring Reverb in a wooden box (similar to AKG BX style). The electronics and transformer are mounted on a metal plate, a situation similar to a guitar amp, hence the question.

It doesn't take much energy to start a fire (think small incandescent light bulbs). Wood doesn't ignite until approx 450 F (270C ? check my math). Usually you can smell if components in a prototype are getting too hot... If you can't hold a finger on one it may be too hot. A simple heat test is to put a drop of spit on the hot item. If it sizzles and boils away the temperature is >100'C/212'F

JR

 

[rock soderstrom]

Many modern transformers already have internal thermal fuses. This is common failure in commercial wall warts.

Yes, I've seen that too but I suspect that not all manufacturers indicate that on the label. Is there any way to find out? How is this realized by thermal fuse or bimetal switch?

It doesn't take much energy to start a fire (think small incandescent light bulbs). Wood doesn't ignite until approx 450 F (270C ? check my math). Usually you can smell if components in a prototype are getting too hot... If you can't hold a finger on one it may be too hot. A simple heat test is to put a drop of spit on the hot item. If it sizzles and boils away the temperature is >100'C/212'F

I use MDF as a veneer substrate, that has even a higher flash point but eventually that burns too.

I always check my builds for too warm components in operation, what is too warm will be exchanged for higher wattage components.

I am more afraid of fatal errors that can lead to unforeseen situations.

 

[JohnRoberts]

Yes, I've seen that too but I suspect that not all manufacturers indicate that on the label. Is there any way to find out? How is this realized by thermal fuse or bimetal switch?

If the transformer is UL approved there is a file number with specifications for the transformer build. I suspect VDE has similar regulations while I have not had any dealing with EU safety agencies this century.

I use MDF as a veneer substrate, that has even a higher flash point but eventually that burns too.

I always check my builds for too warm components in operation, what is too warm will be exchanged for higher wattage components.

I am more afraid of fatal errors that can lead to unforeseen situations.

That's why we build prototypes. DIY gear used inside your living space needs extra attention. Low power stuff is generally less dangerous but as I already shared it doesn't take much power to make something glow.

JR

 

[zamproject]

Hi

I am currently building a Studio Spring Reverb in a wooden box (similar to AKG BX style).

what is similar ? the wooden box or the spring reverb design ? interested to know more

(unfortunately my beloved BX20 have more or less fatal issue...as most of them)

Cheers
Zam

 

[CJ]

We used polyswitches on the secondary, their resistance goes up when current exceeds their rating, but, we had to use twice the rating we were trying to protect as self heating would cause them to shut off by themselves. After the fault is removed, they reset by themselves. This was to satisfy UL without having to add 4 extra fuse holders and fuses, the circuit board we were using did not have enough room. The polyswitch looks like a tantalum cap.

 

[rock soderstrom]

Hi

what is similar ? the wooden box or the spring reverb design ? interested to know more

(unfortunately my beloved BX20 have more or less fatal issue...as most of them)

Cheers
Zam

No, my Reverb just has a similar form factor, which is an upright "wooden" box (veneer on MDF). Additionally, as a special feature, it has a simple delay integrated.

The main problem for a BX clone are the bidirectional transducers. Unfortunately, no idea where you could get something like that...

 

[rock soderstrom]

We used polyswitches on the secondary, their resistance goes up when current exceeds their rating, but, we had to use twice the rating we were trying to protect as self heating would cause them to shut off by themselves. After the fault is removed, they reset by themselves. This was to satisfy UL without having to add 4 extra fuse holders and fuses, the circuit board we were using did not have enough room. The polyswitch looks like a tantalum cap.

Now that you mention it, I remember those things. PPTC devices seem like the right component to do the job for transformers that don't have a built-in fuse. The "auto reset" feature is quite sexy.

In fact, some transformers come with them out of the box.
Interesting, I'll take a closer look!

https://en.wikipedia.org/wiki/Resettable_fuse

 

[gyraf]

(unfortunately my beloved BX20 have more or less fatal issue...as most of them)

Are you aware that the pickups have been reconstructed recently?

 

[warpie]

Wouldn't a thermal fuse on the transformer itself avoid most of these scenarios?

 

[PermO]

I've build a lot of projects for the Dutch Broadcasting Lab.

These are all designed and prototyped and tested inhouse, a bunch of parts gets ordered and I start building them.

All the designs I have build for them had those polyswitch self restoring fuses on every power rail, standard practice.

Not a bad idea to copy I guess.

 

[JohnRoberts]

We used polyswitches on the secondary, their resistance goes up when current exceeds their rating, but, we had to use twice the rating we were trying to protect as self heating would cause them to shut off by themselves. After the fault is removed, they reset by themselves. This was to satisfy UL without having to add 4 extra fuse holders and fuses, the circuit board we were using did not have enough room. The polyswitch looks like a tantalum cap.

The PTC fuses are nice but more expensive than conventional fuses.

Back in the 90s I designed a trick power amp (PMA70+) with a cap doubler in the rails so momentarily it could put out 2x the voltage, 4x the power. This small 35W power amp could put out over 100W peak. Of course it can't do that continuously without 100W heatsinks and power transformer, so I used a PTC device in series with the rail boost circuit. That allowed the amp to put out 100W+ for short transients, 60W for something like 15 seconds until the PTC got hot, then collapsing to the 35w continuous. Playing dynamic music it just sounded like a 100W amp.

JR

 

[zamproject]

The main problem for a BX clone are the bidirectional transducers. Unfortunately, no idea where you could get something like that...

As Jacob say one guy in Berlin redo them

Are you aware that the pickups have been reconstructed recently?

Yep, but beside one or two dead coil, mine also suffer from substrate degradation in the kind of TO39 package that hold the micro wire. I'll try one day...

About the actual topic, poly-fuse are very convenient, I use them intensively in the automation system for protection over stall condition.
Just be careful if ordering, about current rating...some basic descriptions state the hold current, other the trip current, you better double check the datasheet to be sure about calmed spec. (I do the mistake once )

Best
Zam

 

[jeffbro]

Are you aware that the pickups have been reconstructed recently?

Are they available for purchase? Where?

 

[rock soderstrom]

As Jacob say one guy in Berlin redo them

Huh, that was new for me. In my neighborhood! Sometimes the good is so close.

 

[rock soderstrom]

Wouldn't a thermal fuse on the transformer itself avoid most of these scenarios?

Just to understand, do you mean on the primary or secondary side?

 

[Michael Tibes]

is one fuse in the ground path sufficient or do I need two fuses for the +/- voltages?

I would use fuses in the +/- voltages. You could still have current flowing between the rails without gnd. And you'd change the impedance of the gnd path.

SSL used special 10R resistors on each pcb in the +/- voltages. These behave like a fuse, they open the circuit but don't burn. This practice makes sense to me for low current applications.

Michael