"MJ15XXX something, or the good old BCV64/65 or TIP35/36 and push them to the limit, your slope circuit would be a must. Those devices have no power dissipation tolerance at just slightly over 30V .. "
The MJ15015/16 have 180W SOA at 60V
The MJ15022/23 have 160W SOA at 80V
The MJ21195/96 have 240W SOA at 80V
The only MJ series I am aware of that has a SOA curve that bends at 30V are the old MJ802/4502
The MJ15015/16 have 180W SOA at 60V
The MJ15022/23 have 160W SOA at 80V
The MJ21195/96 have 240W SOA at 80V
The only MJ series I am aware of that has a SOA curve that bends at 30V are the old MJ802/4502
Lars Clausen said:Mikeks: I´d say lets take your amp and my amp, crank them up to full power and short the outputs. Then let's see which one last longer
I gather you have given a lot of thought to this issue, but as i see it there is one particular reason that close SOA monitoring on the reactive load side is simply unnecessary, and a waste of time. But on the other hand you don't seem too open towards other people's suggestions, so i'll just let you work this one out yourself ....
At least i can say that any circuit that limits the current in the output stage - like the one that limits the current in the output stages, according to the SOA slopes - will always influence the sound quality. Enough for me to not even consider it.
It limits the current, but you don't accept that people call it a current limiter ???
Have a nice weekend
We will feel glad, if you also compare the short circuit with our mosfet amp also, which upon short circuit will go into latch mode for indefinate period, till the user resets it.
Regarding the influence of protection circuit on sound quality, i dont agree with you , since we have set the current limit of our to be 3 times the RMS current at max voltage and thats the worth of simply using mosfets. Bipolars wont stand in such conditions.
meanwhile your Zeta will do the right work for you..........
Hi Workhorse,
I don't see the usefullness of comparing an amp that will shut down on a short in this test. Whether you use mosfets or bipolar transistors in this situation is immaterial since the amp has shut down. No SOA issues since you are now talking about short protection. As you have over built your amp, the same could be done with bipolars. Big deal.
Your point about SOA issues with mosfets was made a while ago. Now back to other useful input from others.
-Chris
I don't see the usefullness of comparing an amp that will shut down on a short in this test. Whether you use mosfets or bipolar transistors in this situation is immaterial since the amp has shut down. No SOA issues since you are now talking about short protection. As you have over built your amp, the same could be done with bipolars. Big deal.
Your point about SOA issues with mosfets was made a while ago. Now back to other useful input from others.
-Chris
I sense a certain amount of semantic confusion. VI-limiting and short circuit protection are not synonomous. They may, however, intersect to a greater or lesser degree.
I expect most VI-limiters deal with a short at least briefly. Hopefully, long enough to deal with someone who forgets to turn off the amp while connecting the speakers. While persistent shorts may result in thermal damage after a while, I think that circumstance is relatively rare at least for home the user.
A pure short circuit protection scheme (I saw a schematic recently of one) may do a hard shut down when the load drops below say1 ohm but leave you unprotected from speakers with dips just short of 1 ohm. This might be all that's needed if use many, many parallel output devices beyond what usual for the supply rails. Brute force solutions have the merit of simplicity.
I expect most VI-limiters deal with a short at least briefly. Hopefully, long enough to deal with someone who forgets to turn off the amp while connecting the speakers. While persistent shorts may result in thermal damage after a while, I think that circumstance is relatively rare at least for home the user.
A pure short circuit protection scheme (I saw a schematic recently of one) may do a hard shut down when the load drops below say1 ohm but leave you unprotected from speakers with dips just short of 1 ohm. This might be all that's needed if use many, many parallel output devices beyond what usual for the supply rails. Brute force solutions have the merit of simplicity.
Sam, indeed. If you take enough output devices, you may not need SOA modelling, because of the massive overkill to ANY possible complex load situation. Nelson Pass' designs come to mind. In such a case, a simple short protection may be all that is needed. But it has a price, of course.
Jan Didden
Jan Didden
sam9 said:
True!
VI protection limits current as a function of device voltage..(as Jan pointed out somewhere)....
...while crude short circuit protection provides a constant current limit regardless of device voltage......
...in other words, the short-circuit locus is merely a flat horizontal line across the SOA plot (linear-linear scales assumed)....
...this makes for grossly inefficient use of available SOA, as the flat short-circuit locus HAS to reside below the max. voltage, power dissipation parabola (MOSFETs), or the secondary breakdown parabola (BJTs), leaving more than 90% of available SOA unused.....
Therefore, it should be apparent that simple short-circuit protection is exceedingly crude...and inefficient...and reeks of lack of commitment...laziness.....ineptitude.....cluelessness....rank amateurishness.....(have i missed any choice adjectives here....
Let me know... )mikeks,mikeks said:
...crude...and inefficient...and reeks of lack of commitment...laziness.....ineptitude.....cluelessness....rank amateurishness.....(have i missed any choice adjectives here....
Can you, please, point to few schematics with short circuit protection realizations which may be called as "good from mikeks's
point of view"?
Workhorse said:
It is A way, you can't say it is the best way. For one thing, it certainly is an expensive way. But, you set your priorities, and make your decision.
This is also a cultural thing. Europeans often have the tendency to spend some more time to try to find a low-cost solution to a problem in engineering. North Americans often solve problems quicker by throwing money (or material) at it. I'm not saying one or the other is 'better', it is just something I have often seen during my involvement in industrial and military engineering.
Jan Didden
janneman said:
Hello Jan Didden,
Time always matter in such decisions, as far as professional gear is concerned you cannot compromise with lower cost yet effective solution, here in INDIA people are crazy about how much of stuff is implemented in the product apart from compactness. An overkill always serves as a best option in real time field operation ++ intelligent designing of protection always creates an exceptional powerware solution to delight the professional customers.
The thing is very different in homeaudio or domestic consumerware.
regards,
Kanwar
anli said:
mikeks,
Can you, please, point to few schematics with short circuit protection realizations which may be called as "good from mikeks's
... as an example:
Jan Didden
Edit: I'm not sure this meets Mikeks approval, although I guess he would give an imperceptible nod..
Attachments
More mumblings on output protection philosophy
While I can sympathize with Mikek's approach in the sense that careful and dedicated design pays off in better use of a device capability - in this case absolute maximum voltage / current / power dissipation and in the case of bipolars second breakdown limits -, it is also true this approach not necessarly guarantees the best overall compromise.
The real world is pesky in this respect and we have on occasions found the neat, elegant and efficient engineering solutions we can take pride on, must yield to a far less sexy design that does the job closer to the bottom line. Accountants are good at this, lawyers are inverse transforms.
Except for the most extreme power applications where there is no choice but the available devices to protect, most often it pays off both in cost and simplicity to go the overkill way.
An example that comes to my mind (please do not take this as a subject for further discussion, it does not belong to this thread and may be treated elsewhere), is the active vs. passive crossover decission.
Good engineers with a knack for fine design can and do come out with passive networks that behave fairly, but it must be recognized the problem is not at all trivial given the mess of interactions and impedance fluctuations. The reward, a single point of connection for a single amplifier.
Active crossovers for its part may seem an overkill, devoting an amplifier for each frequency band. But design of a high performance 5 pole two way divider for example, is strightforward and inexpensive, and each amplifier sees a load impedance as good as it gets within it's operating range.
I am not sure whether the active crossover way turns out in the end also to be the less expensive, taking into account the relatively high cost of quality inductors and capacitors, and the comparatively low cost of even discrete amplifiers, once the relative powers for each band are reasonably balanced.
In this case, the overkill way seems to make sense both in performance and cost despite of being less refined from a design point of view.
Rodolfo
While I can sympathize with Mikek's approach in the sense that careful and dedicated design pays off in better use of a device capability - in this case absolute maximum voltage / current / power dissipation and in the case of bipolars second breakdown limits -, it is also true this approach not necessarly guarantees the best overall compromise.
The real world is pesky in this respect and we have on occasions found the neat, elegant and efficient engineering solutions we can take pride on, must yield to a far less sexy design that does the job closer to the bottom line. Accountants are good at this, lawyers are inverse transforms.
Except for the most extreme power applications where there is no choice but the available devices to protect, most often it pays off both in cost and simplicity to go the overkill way.
An example that comes to my mind (please do not take this as a subject for further discussion, it does not belong to this thread and may be treated elsewhere), is the active vs. passive crossover decission.
Good engineers with a knack for fine design can and do come out with passive networks that behave fairly, but it must be recognized the problem is not at all trivial given the mess of interactions and impedance fluctuations. The reward, a single point of connection for a single amplifier.
Active crossovers for its part may seem an overkill, devoting an amplifier for each frequency band. But design of a high performance 5 pole two way divider for example, is strightforward and inexpensive, and each amplifier sees a load impedance as good as it gets within it's operating range.
I am not sure whether the active crossover way turns out in the end also to be the less expensive, taking into account the relatively high cost of quality inductors and capacitors, and the comparatively low cost of even discrete amplifiers, once the relative powers for each band are reasonably balanced.
In this case, the overkill way seems to make sense both in performance and cost despite of being less refined from a design point of view.
Rodolfo
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.