Mr Murphy says that a $10 transistor always blows before a 10 cent fuse.
What about fuse @ the output ?
I have made recently a short circuit protection that works perfect and is very simple. Can disconnect the speakers in 14us. I just have to made some fine tuning to finish it up. The whole PSU is 10cm x 11.5cm. Protection circuit total current consumption is approx 5mA only.
The movie is bellow (I do not know how to tag it in the post).
https://www.youtube.com/watch?v=R4u4BiysseA&feature=youtu.be
For now sorry for the language ( I will prepare a new movie in English in next few days).
The yellow trace is output voltage and the blue is current.
Regards
The movie is bellow (I do not know how to tag it in the post).
https://www.youtube.com/watch?v=R4u4BiysseA&feature=youtu.be
For now sorry for the language ( I will prepare a new movie in English in next few days).
The yellow trace is output voltage and the blue is current.
Regards
Attachments
Leach omits the V part of the IV detection. This allows the SOA to be exceeded when Vce is higher. Devices can still fail even though the Leach I slope detection is operating properly. This is particularly the issue when reactive loads impose a high Vce while high Ic passes.
15A devices typically have a much reduced current capability when Vce is at half rail to rail voltage. See datasheet's SOA graph.
Typical high power BJTs show max continuous Ic ~ 1A when Vce=50V and even less if the Tc is ~80°C
Short duration one shot pulses are shown on the datasheet SOA graph. See what is shown for the 25°C 100ms and 10ms pulses and then de-rate for temperature.
Last edited:
But we are talking about short circuit. The event last for a second before the fuse blows. If you have multiple pairs of transistors, say just 4 pairs, each conduct 2.5A for 1 sec before the fuse blows.
Attached is the datasheet of the transistor I chose. It is supposed to be a weaker transistor. The SOA shows about 2.5A at 50V ( I assume 1sec duration). If you look at the MJW3281/1302 they are 3.5A.
This is a once in a life time event ( I hope!!), not a regular event. My interpretation of the graph is the transistor is safe if you stay within the SOA graph even if it happens repeatedly.
Last edited:
This is my guess, series resistance and series inductance. It's a very thin wire even though it's short. I read a lot of people are against this also.
I disagree.
resistance is not bad and inductance too since there is no core..(even good)
Also think it would save the transistors unlike fuses at the rails.
(here im dilеmmed and dont know actually)
one think im sure -> its not bad, it could be neutral or positive.
glad to see another opinions, i like this idea.
Last edited:
Because to be effective as protection, a wire fuse must be rated very close to the maximum power rating and at that point, it will be heating and cooling rapidly, distorting the output. See "fuse distortion" https://books.google.com.au/books?i...6AEIUDAL#v=onepage&q=fuse distortion"&f=false
'Seems mosfet, you couldn't resist coming back for more.
Short circuit current will be very much higher than 10Apk. If your output stage and output cabling has a resistance of 1r0 and the PSU capacitors are charged to 50Vdc, then the short circuit current will be 50Apk.
If, as is more likely in a well built high power amplifier that, the total short circuit resistance is <0r2, then the fault current could be >250Apk
This example of 2.5A @ 50Vce is more like the current driven into an 8ohm test load.
I assume for normal amplifier operation that the peak transient current into the reactive speaker load is HALF that at maximum power into the rated resistive test load and the Vce is equal to the supply rail voltage, then the 100ms SOA after temperature de-rating gives a good indication of amplifier reliability.
i.e. for a 100W into 8ohms amplifier running on ±50Vdc supplies, then the maximum power current is 5Apk (Pmax = 5Apk² * 8r0 / 2 = 100W) A reasonable SOA check value is therefore 2.5A @ 50Vce when driving a speaker load. This would be a 1pr 250W devices output stage, or a 2pair 130W devices output stage.
If you want to design for a different maximum output power or for a different speaker load, then do the new numbers based on that half maximum current into your loading.
Last edited:
Short circuit current protection is not easy to design for.
The current rise time can be enormous (>>500A/us). The time delay until the "switch" opens could allow the current to rise to damaging levels.
eg an F4A fuse passing 100A could continue conducting for 1ms. Will an output device survive longer than that and not suffer any damage? An F10A fuse may pass that 100A fault current for 10ms. Try to find some fuse datasheets to see how long fuses take to rupture and for the arc to extinguish.
The current rise time can be enormous (>>500A/us). The time delay until the "switch" opens could allow the current to rise to damaging levels.
eg an F4A fuse passing 100A could continue conducting for 1ms. Will an output device survive longer than that and not suffer any damage? An F10A fuse may pass that 100A fault current for 10ms. Try to find some fuse datasheets to see how long fuses take to rupture and for the arc to extinguish.
Last edited:
I put the fuse on the pcb after the reservoir cap, you won't have to discharge the cap to blow the fuse.
Whether it's driving a 8 ohm, or 4 ohm is irrelevant. The SOA spec is for a given Vce. You just have to make sure you have enough transistors pairs so the total SOA is higher than the fuse.
I have a suspicion people blow transistor instead of rail fuse because they don't have enough pairs of transistors. if you have 2 pairs, then the max current for 50V is only 5A. If you put a 10A fuse, you got a problem. But if you have 4 pairs and you use a 8A fuse, that is within the spec.
Last edited:
Transistors or diodes can pass huge current if the duration is very short too.
I still think the key is to have more pairs with at least 0.25ohm emitter resistor to guaranty current spread evenly.
I think when you say the current limiting in Leach amp doen't work very well mainly the amp doesn't have enough pairs of transistors. The limiting circuit should prevent big in rush of current like you described. Problem is the circuit is not very accurate in when the limiting started. It might be too high when it start already. You need both the rail fuse and limiting circuit to be safe. The circuit limit the current within uS time. Then the fuse blown within 1 sec.
That's what I have, both the Leach and the rail fuse with 9 pairs in planning!!!.
I think when you say the current limiting in Leach amp doen't work very well mainly the amp doesn't have enough pairs of transistors. The limiting circuit should prevent big in rush of current like you described. Problem is the circuit is not very accurate in when the limiting started. It might be too high when it start already. You need both the rail fuse and limiting circuit to be safe. The circuit limit the current within uS time. Then the fuse blown within 1 sec.
That's what I have, both the Leach and the rail fuse with 9 pairs in planning!!!.