I'm trying to replace rail fuses with a current sensing circuit and a low RDS mosfet. This should eliminate the possibility of fuse distortion while better protecting the amplifier on overcurrent as well.
This will consist of two small daughter cards soldered in place of the existing fuse holders that are interconnected so if one protection circuit activates, it will signal the other circuit to activate, shutting down both rails.
I'd like to add provisions for interface with other protection circuitry in the amplifier to activate these as well. If a speaker DC detector activates, it would be nice to disconnect both rails from the amplifier before the supply caps dump through the shorting output device, for example.
This will consist of two small daughter cards soldered in place of the existing fuse holders that are interconnected so if one protection circuit activates, it will signal the other circuit to activate, shutting down both rails.
I'd like to add provisions for interface with other protection circuitry in the amplifier to activate these as well. If a speaker DC detector activates, it would be nice to disconnect both rails from the amplifier before the supply caps dump through the shorting output device, for example.
Last edited:
Sorry, but this sounds a pretty crazy thing to try and do - making the amplifier a LOT less safe, and more likely increasing any imaginary distortion. A simple fuse is a VERY reliable and VERY safe component, and probably has no more effect on the sound than a piece of wire (which is all it is).
Many are building amplifiers without rail fuses. This is a safer alternative to that.
Standard slow blow fuses properly sized to the current needed in an amplifier will start to melt on peaks of a sine wave under maximum load. You can see it as discolouration in the center of the fuse element. You can also see it on a scope. It looks like clipping. Normal amplifiers have too large of a fuse in them to properly protect output devices. How many times have you heard of output devices blowing without taking out a rain fuse?
I've already got a very reliable protection circuit running that will shut down my amplifier and open speaker connections faster than a fuse can blow on a short circuited output. This will enhance that protection circuit by opening connection to the supply caps on overcurrent or any other type of fault.
Standard slow blow fuses properly sized to the current needed in an amplifier will start to melt on peaks of a sine wave under maximum load. You can see it as discolouration in the center of the fuse element. You can also see it on a scope. It looks like clipping. Normal amplifiers have too large of a fuse in them to properly protect output devices. How many times have you heard of output devices blowing without taking out a rain fuse?
I've already got a very reliable protection circuit running that will shut down my amplifier and open speaker connections faster than a fuse can blow on a short circuited output. This will enhance that protection circuit by opening connection to the supply caps on overcurrent or any other type of fault.
Hi,
I do that but I using a microprocessor to control /regulate the rails voltage and a hall sensor to protect the speakers. If the current reached the set current high limit the rails voltage are shutdown to protect the speakers. I used it to control the rails voltage for an LM3886 amplifier. Attached it is a block diagram showing how it is done.
I do that but I using a microprocessor to control /regulate the rails voltage and a hall sensor to protect the speakers. If the current reached the set current high limit the rails voltage are shutdown to protect the speakers. I used it to control the rails voltage for an LM3886 amplifier. Attached it is a block diagram showing how it is done.
Hi Tauro0221
I remember your circuit from the 21 Century Protection thread. I believe it was you who recommended this current sensor I am using.
That's NOT what the fuses are for, they are safety devices (mostly to protect the expensive mains transformer, and subsequent potential fire damage), NOT a crude attempt to save the output devices from been killed by misuse.
Your idea of 'properly sized' is most probably FAR lower than the fuses should be - which could explain why you imagine they might have some effect on the sound.
Interesting that you consider a tiny (and almost certainly inaudible) distortion as more serious that your method of 100% distortion - which killing the sound totally is
That's NOT what the fuses are for, they are safety devices (mostly to protect the expensive mains transformer, and subsequent potential fire damage), NOT a crude attempt to save the output devices from been killed by misuse.
Your idea of 'properly sized' is most probably FAR lower than the fuses should be - which could explain why you imagine they might have some effect on the sound.
Interesting that you consider a tiny (and almost certainly inaudible) distortion as more serious that your method of 100% distortion - which killing the sound totally is
I put a fuse on the mains side of my transformer to protect it. I put protection after the supply to protect the rest of my amplifier. There are many more parts to an amplifier than a transformer.
It seems I have offended you by trying to improve on an antiquated safety system. I use a microcontroller to control safety circuits as well. Some of us have left the stone age.
Now , you comment. But , did not do the reading.
It is a valid patent - Patent US6587027 - Solid state fuse - Google Patents
I can see where your "crazy" statement comes from never being in an
industrial process environment.
They were using solid state breakers in the 70-80's for industrial
circuit protection. Those old pieces of silicon are most likely still working.
Now , with uOHM MOSFETS and better hall sensors , you get this-
Solid State Breakers | Solid State Relays | Perfect Switch, LLC .
These breakers protect "life and limb" out at the plant site , military
compliant - built to last 100 years. None of the old fuse's drawbacks.
"Increasing distortion" OMG - Don't you read ( u OHM !!! ) like a bloody
14ga wire with NO fuse.
A standard fuse is much more of a nichrome resistor element than
a 14 ga copper conductor - more reading needed !
At the 30 second overload rating of a 15A fuse , over 500mV drop. as it
gets incandescent , many ohms before it goes. the SS fuse goes
from uOHM to infinity instantly.
OS
I read a little and read the hall sensor...........That's a awfully fancy way to do a simple job. Just the sensor must be more expensive than any of the power devices used.
In my book a simple rail fuse should do it IF you use the money to put in more pairs of power transistors. I am sure you save money having more power transistors. More transistors in parallel, the much better chance they survive long enough to burn the fuse first.
Looking at the circuit, I am not sure you even save pcb space compares to just putting more power pairs. AND not to mention the more pairs you use, the more forgiving is the choice of transistors, less requirements of constant Hfe at high current( as you have more pairs in parallel), Less requirement on SOA stuff. You can buy cheaper transistor instead of "the" transistor.
That's one thing I don't understand. reading a lot of the schematics here. Poeple try to get a lot of power out of one or two pairs of power transistor, then in turn, worry about the SOA, beta droop, resorting to "the" transistor. I started out with 5 pairs of output transistors. Don't tell me the rail fuse can out last all these transistors. You get better and cheaper transistors when you put less requirements on them.
AND, who cares about 1/10 of an ohm through the fuse? For one, this is at the collector, there is a lot of common mode rejection. These are secondary effect. Put by pass caps right at the collector to smooth out the voltage. I use a 10uF in parallel with a 0.1uF right on the collector of each of the power BJT. If needed to, I can put more.
Worry more on the layout. All the talks, then when I look at a lot of the layout posted, long power and ground trace, no bypass cap any where close. resistance is proportional to length/width. with long trace, the voltage drop can be worst than a fuse that is so short. I don't see anyone worry about this very obvious problem.....Then laugh at me using ground and power planes.
In my book a simple rail fuse should do it IF you use the money to put in more pairs of power transistors. I am sure you save money having more power transistors. More transistors in parallel, the much better chance they survive long enough to burn the fuse first.
Looking at the circuit, I am not sure you even save pcb space compares to just putting more power pairs. AND not to mention the more pairs you use, the more forgiving is the choice of transistors, less requirements of constant Hfe at high current( as you have more pairs in parallel), Less requirement on SOA stuff. You can buy cheaper transistor instead of "the" transistor.
That's one thing I don't understand. reading a lot of the schematics here. Poeple try to get a lot of power out of one or two pairs of power transistor, then in turn, worry about the SOA, beta droop, resorting to "the" transistor. I started out with 5 pairs of output transistors. Don't tell me the rail fuse can out last all these transistors. You get better and cheaper transistors when you put less requirements on them.
AND, who cares about 1/10 of an ohm through the fuse? For one, this is at the collector, there is a lot of common mode rejection. These are secondary effect. Put by pass caps right at the collector to smooth out the voltage. I use a 10uF in parallel with a 0.1uF right on the collector of each of the power BJT. If needed to, I can put more.
Worry more on the layout. All the talks, then when I look at a lot of the layout posted, long power and ground trace, no bypass cap any where close. resistance is proportional to length/width. with long trace, the voltage drop can be worst than a fuse that is so short. I don't see anyone worry about this very obvious problem.....Then laugh at me using ground and power planes.
Last edited:
Hi Alan.
OS and I both have amplifiers with output devices spread out over 16'' heat sinks. No more room for output devices.
What I'm thinking of would be around 1 1/4'' square board standing up off the main amp board on leads from where the rail fuse holder would normally be soldered. There's around $20 worth of parts there. It's considerably more than a fuse and holder cost, but this is DIY. Who cares?
Wow , I guess they are not ready
. Reading the comments , I guess no one has seen the greatly increased
ripple that a high Rser fuse will cause to be locally decoupled.
And to say eliminating that increases distortion .... I wonder if it
was this hard to go from tube to transistor ? Or from phono to CD ...
Oh wait , I remember - it WAS !
OS
Hi Alan.
OS and I both have amplifiers with output devices spread out over 16'' heat sinks. No more room for output devices.
What I'm thinking of would be around 1 1/4'' square board standing up off the main amp board on leads from where the rail fuse holder would normally be soldered. There's around $20 worth of parts there. It's considerably more than a fuse and holder cost, but this is DIY. Who cares?
I don't mean to be personal attack or anything. I am just hot on the trod in the speaker protection and read a lot about protection and all. I just join in and talk. I read through almost 60 pages on the protection thread and I impression I got was people got way carried away with the protection. As DIY, fixing the amp is the easiest part, when it fails, whether it blow 10 or blow one transistor, you have to open it up and fix it. Only thing to me that needs protection is the expensive speaker that I assume people don't know how to fix.
If you already have a lot of pairs of transistors, I don't think any problem will burn all the transistors, it just take out the fuse. I am just talking from my observation on a lot of the schematics posted here that use two pairs for a 150W amp!!! I won't even consider less than 4 pairs for over 100W.
We didn't just design the protection board to adequately protect the amplifier. It's more fun to overbuild things and take them to the extreme. Much like you did with your output stage in your amplifier. I'm not in this hobby to just assemble some parts and make it work. I'm in it for the challenge of innovating new ideas and designs. I'm a hot rod builder for a living, so this is a way of life for me. The more people tell me it's a stupid idea and it won't work, the more I push to make it work.
To me saying it's overbuilt and unnecessary is like saying her boobs are too bouncy. Never!
In that light, I am sure your circuit kick the butt of the fuse, it's like the Roll Royce of protection.We didn't just design the protection board to adequately protect the amplifier. It's more fun to overbuild things and take them to the extreme. Much like you did with your output stage in your amplifier. I'm not in this hobby to just assemble some parts and make it work. I'm in it for the challenge of innovating new ideas and designs. I'm a hot rod builder for a living, so this is a way of life for me. The more people tell me it's a stupid idea and it won't work, the more I push to make it work.
To me saying it's overbuilt and unnecessary is like saying her boobs are too bouncy. Never!
In that light, I am sure your circuit kick the butt of the fuse, it's like the Roll Royce of protection.
That's what I'm hoping for.
It's not the cost of the power devices, it is the time finding the other 117 blown devices when the 18 blown output transistors let rail voltage run around. The problem with rail fuses, one blows, the other doesn't. Even aside from any real or mythical distortion. The blown PA amp I bought and worked on, the 10000 uf rail caps took out the PWB trace to the crowbar SCR instead of blowing the AC breaker. I trial repaired it, another weak old OT went, blowing an 18 ga buss wire I repaired the PWB trace from rail to the crowbar with. You do the math, that is a massive melting current. As far as the cost of the FET, the cost of one 15" black widow woofer is about $138 + $10 freight and 7% use tax, a $8 hall effect sensor is a lot cheaper than that. It is apparant that this amp was connected to speakers over and over again after it failed. I can't see myself doing much better in an actual public performance unless I have a red light on the front panel to tell me to quit trying, even though the roadie/helper found the short in the speaker wiring and corrected it. My speaker carrying yard man is not real bright, but saves my back for not much money.
A rail FET circuit can take out both rails, if control is properly coupled. In terms of the actual circuit shown, I see one fet, perhaps to take out the "ground"? Rail voltages can run from plus to minus without involving the ground, I don't see the point. My PV1.3k has only a 4000 ohm resistor between speaker and ground anyway.
One quibble with this circuit, I haven't downloaded the IC datasheets, but I don't see a latch to keep the circuit out after an overcurrent event. My circuit also has a front panel LED to show that the protection circuit is red or green. I also don't like three power supplies, I'm having enough trouble cramming in another wall transformer to produce one additional voltage. Putting this control circuit on the audio +-15 v circuit strikes me as asking for trouble, so I went to a separate supply.
Instead of hall sensors, I'm looking at using the voltage developed across the rail FET to set the defect latch. Optoisolators can shift the voltage developed around to other reference points like digital ground, as long as the voltage is >.6 v on the opto input diode. My target trip current is 25-30 A so this voltage is not out of the question. Right now my circuit latch is tripped only by the DC detection circuit previously installed to trip the crowbar SCR. I'll look up this IF8025 FET, currently I'm intending to use Fairchild FDP52n20 in the TO220 package. I'd like the same die in a TO247 package (long leads) but the bigger heat sinks seem to come only in surface mount parts with clipped leads.
I don't see the problem with inrush, my rail FET stopper is between the 10000 uf caps and the output transistors. I have only .1 uf ceramics at the output transistor board, and I added those myself, Peavey had no local rail capacitors on the transistor board at all. I stop inrush anyway with a CL101 thermistor on the AC line after the breaker, my PV-1.3k was dimming the room lights when I turned it on.
Edit, datasheetcatalog never heard of a IF8025.
Last edited:
The + and - on the schematic are for orientation. + goes to positive rail input on one rail. - goes to the negative rail input. Two of these boards interconnect through the mosfet driver module, so if one triggers, both are shut down.
This is in the early stages of testing. I haven't bothered with a latching circuit to keep things shut down yet. My own finished products will likely interface with a microcontroller. The first round is pretty much a proof of concept and something to hack to a breadboard when I'm designing something like this. I don't expect it to be a usable product usually.
Yes, supply caps can make a mess when things go south. We're building amps with upwards of 100000uF. That's a big whack when something shorts! That's the main reason I'm experimenting with this. I initially started experimenting with mosfets on the output rails of my power supplies, but the supply boards ware getting very large quickly. OS suggested this fuse idea and it seems to make sense.
This is in the early stages of testing. I haven't bothered with a latching circuit to keep things shut down yet. My own finished products will likely interface with a microcontroller. The first round is pretty much a proof of concept and something to hack to a breadboard when I'm designing something like this. I don't expect it to be a usable product usually.
Yes, supply caps can make a mess when things go south. We're building amps with upwards of 100000uF. That's a big whack when something shorts! That's the main reason I'm experimenting with this. I initially started experimenting with mosfets on the output rails of my power supplies, but the supply boards ware getting very large quickly. OS suggested this fuse idea and it seems to make sense.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.