suzyj said:
I disagree. I don’t think that any directly coupled amplifier (unless of very modest power) should be used without a DC output protection scheme – i.e. a relay that disconnects the speakers when a few volts or so DC is detected at the output. This relay can also be conveniently driven from a power-up timer to eliminate turn-on thump.
With rail fuses and DC protection as my absolute minimum of protection circuitry, in 99.5% cases, output device failures won’t be very messy at all.
Cheers,
Glen
Relays work OK, IF you have enough devices in parallel as to not have instantaneous second breakdown in bipolar transistors.
Ideally, I would not use an output relay, but it works and I still can achieve a class A rating on my best power amp, even with an output relay and +/- 80V+ supplies. The output relay is fast enough to work, because 10mS should be tolerated by virtually any amp, before protection cuts in.
Most of you on this website, do not use enough complementary pairs of output devices for a given supply voltage. This allows the potential problem of second breakdown, as well as limited peak output current. For the record I use 9pairs for 400W into 8 ohms and typically 6pairs for 250W, and maybe 2pairs for 125W. We buy in Asia, and the parts must be cheaper there, because we can afford to use them.
Ideally, I would not use an output relay, but it works and I still can achieve a class A rating on my best power amp, even with an output relay and +/- 80V+ supplies. The output relay is fast enough to work, because 10mS should be tolerated by virtually any amp, before protection cuts in.
Most of you on this website, do not use enough complementary pairs of output devices for a given supply voltage. This allows the potential problem of second breakdown, as well as limited peak output current. For the record I use 9pairs for 400W into 8 ohms and typically 6pairs for 250W, and maybe 2pairs for 125W. We buy in Asia, and the parts must be cheaper there, because we can afford to use them.
G.Kleinschmidt said:
Just to elaborate further on the worth of rail fuses: Even without DC fault protection, rail fuses are still a good idea.
In a Class AB power amplifier, each rail fuse only has to be rated to pass the average half wave speaker current, which is much less than the peak speaker current.
Consider a typical HiFi amplifier rated at 100W rms into 8 ohms with a continuous power rating of 1/3 Pmax.
At 1/3 pmax (33W) Vpeak = SQRT(33W*8R*2) = 23V
Ipeak = Vpeak/R = 23/8 = 2.9A
Each rail only has to supply load current for each half cycle, and the average load current drawn from each rail is rather accurately equal to Ipeak/pi, giving 2.9A/3.14 = 0.9A.
A pair of fast blow rail fuses rated at 1.2A each would adequately protect this amplifier for it's given power rating.
1.2A into 8 ohms gives 9.6V. In reality, a speaker with a rated impedance of 8 ohms may have a resistance of say 6 ohms or less.
With 6 ohms and a fuse limited current of 1.2A, the output DC fault voltage is limited to 1.2*6 = 7.2V.
This equates to only a tiny 8.6 watts of speaker dissipation before the fuse rating is exceeded. Any 8 ohm speaker rated to work with a 100W rms power amplifier should be able of cope with that without a problem at all.
Consider the scenario without the rail fuses. A fault condition such as an output transistor with a C-E short may cause the output to swing and sit at anything from 30-40V, depending on the capacity of the power supply. 40V gives 266W into 6R. If you happen to be unable to turn off the amplifier rather quickly, you can almost surely kiss your speakers goodbye.
Cheers,
Glen
john curl said:
G'day John.
The fuse rating I gave in my example is more than adequate for a domestic HiFi amplifier. The only way they're going to blow is if the amplifier is seriously abused (i.e. driven well into clipping).
Anyone who continually needs to crank their volume control up that high should have bought himself or herself a bigger amplifier to start with!
john curl said:
Lighten up john.
My rail fuse rating example was only what I would consider adequate for a domestic 100W into 8ohm rated HiFi amplifier that doesn’t rely on any other form of protection, in response to Suzy’s comment on the utility of rail fuses.
I simply offered it to show that rail fuses can indeed provide adequate speaker protection, and still be sufficiently rated to allow a high continuous power output, providing that the amplifier is not excessively stressed beyond it’s ratings.
Form a DIY perspective (the theme of this board) I think this is relevant.
From a commercial perspective, I don’t think that any 100W amplifier should be designed this way, and I’ve already stated my opinion that even a DIY amplifier of modest power rating, at a bare minimum, requires rail fuses and a DC fault protection relay. For a commercial amplifier, these two forms of protection alone would be less than a bare minimum requirement.
Commercial designers have much less control over whatever dummy may get their hands on the volume control of their designs than the average DIY'er !
Cheers,
Glen
john curl said:
This is unlikely to be true in the presence of a dead short.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=84733&highlight=
Mikeks, I have 1000's of amplifiers in the audio world that use this system to protect them. It works for us. We also have rail fuses, and input fuses. Still, the relays work fast enough for our needs. We once had a probem with our protection because of a manufacturing mistake. We noticed it right away, and I was called in to fix it. Removing a resistor fixed the problem.
john curl said:
Mr Curl,
output devices like the ones you've chosen for the big Parasounds can nowadays be bought for 1/4-1/5th of the rate they went for 20 years ago, even by a diyA fool.
The average EE on this forum apparantly has no different views with regard to output stages than the non-audio EEs i've conversed power amplifier issues with decades ago.
The EE seems to hold the view that reduced dissipation thanks to class AB operation justifies using a minimal output device number, the non-EE ampy constructor here is hoping for the biggest bang for the least invested dollars and spends as little as he can on the parts that don't come NOS.
I've tried dead shorts on relay protected output stages, both deliberate and accidental.
Maybe it's because the relays i use are guaranteed to switch off in 7mS or less, my preference for tossing in even more output devices than Mr Curl or both, but it definitely works.
Michael,
It would surprise me a great deal if Mr Curl hasn't tried a dead short test in +35 years, why don't you ask the big guy ?
I do agree with you that quite a lot of commercial power amps make you raise an eyebrow after a glance on the number of devices in the output stage and the relay model used.
imo, the JC1s have a tremendous bang for the buck, Parasound must indeed buy very cheap in Asia.
And i really doubt that ripping the relays out would make a difference in how the Halo's sound.
Of course, some people ocassionally have maintenance done on their car and change the tires, some don't.
Miss Jackson should remove the brakes on her bike.
Saves weight, and when you've got to go you go.
Shame of such a pretty lady though, same goes for the nice power amplifiers she constructs.
Hi Bob & John,
these various power output stages and the fuses to protect the outside world.
for each of 125W, 250W and 400W into 8ohm arrangements and assuming each are designed to drive 4r , can you specify your preference for rail fuses?
I would always put in a close rated T (delay) mains fuse and I/V protection (probably 2slope) and my preference would then be rail fuses to take care of the longer term excess current draw. A DC detect combined with either rail fuse crowbars or output relay would be the final stage of long term protection.
Taking G's example for a 100W 8r stereo amplifier. Mains T1.6A(220Vac) or T3.1(110Vac), Rail F2.5A. These fuse values allow soft startup and continuous draw of 300W and peak output currents of 40V/8r=5Apk.
these various power output stages and the fuses to protect the outside world.
for each of 125W, 250W and 400W into 8ohm arrangements and assuming each are designed to drive 4r , can you specify your preference for rail fuses?
I would always put in a close rated T (delay) mains fuse and I/V protection (probably 2slope) and my preference would then be rail fuses to take care of the longer term excess current draw. A DC detect combined with either rail fuse crowbars or output relay would be the final stage of long term protection.
Taking G's example for a 100W 8r stereo amplifier. Mains T1.6A(220Vac) or T3.1(110Vac), Rail F2.5A. These fuse values allow soft startup and continuous draw of 300W and peak output currents of 40V/8r=5Apk.
John,
I agree with your view totally. If you have two devices the emitter resitors doubles, four device the it is four time that of a single device and so on. The likelyhood to find yourself in secondary breakdown is less and less possible.
In a commerical design their is no way of telling what the customer will do to your product and while under warrantee it is your problem. Many parrallel devices is expensive one may think. but cheap in the long run.
Cusrtomers are unpredictable and you have to cater for almost any thing that can happen.
I agree with your view totally. If you have two devices the emitter resitors doubles, four device the it is four time that of a single device and so on. The likelyhood to find yourself in secondary breakdown is less and less possible.
In a commerical design their is no way of telling what the customer will do to your product and while under warrantee it is your problem. Many parrallel devices is expensive one may think. but cheap in the long run.
Cusrtomers are unpredictable and you have to cater for almost any thing that can happen.
jacco vermeulen said:
It's Mrs, and it's not a particularly useful analogy.
I use brakes on my bikes regularly, so they're useful. My amplifier has been sitting in my living room connected to my stereo for the last couple of months. Nobody disconnects the speakers, and nobody abuses it.
It's one of the advantages of making something for yourself, rather than making something to sell to others. You can spend your money on things that increase the performance of your amp, rather than having to design something that survives the lowest common denominator consumer. This is DIY Audio, after all.
I don't see the ability to survive a short as being of importance, as the likelihood of that happening to my amp is extremely low. Hence I make different design decisions than I would if (for example) I was designing an amp to be used by musicians.
Regards,
Suzy
PS: For what it's worth, three of my bicycles have no brakes. They're track bikes - for riding on closed velodromes, where brakes are not allowed.
jacco vermeulen said:
Hi Jacco,
I would be pleasantly surprised if your NO relays had an opening time constant of 7mS: it's usually of the order of 15mS, if you're lucky.
Even assuming your relay opened in no more than 15mS when new, this fortuitous state of affairs is unlikely to last; we're talking about an electro-mechanical device after all.
This why i consider my objections here are valid.
suzyj said:
You've just given a reason that the analogy does apply.
I can think of some other stuff to skip if you know exactly what you're doing and the power amp is matched to the load.
H..., i've done a tribe of open frame, life heatsink and protection free amps myself, including russian roulette tube amps for a birthday surprise switch-on in the dark.
Doesn't work anymore with kids around, partners, and the usual guest with grabby fingers.
Regards back to you, Mrs J.
Mike,
SDS is a pleasant surprise, the option of relay exchange servicing is a diy privilege.
PS: like the flag and i'm skipping skippy for Xmas dinner this year again.