Audio amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Darksog

At best this amplifier will deliver about 200 watts RMS into 8 ohms. I would be worried about driving a 4 ohm load with the output stage shown, it might be ok for hi-fi use but for PA the output transistors will die. This does depend on the power supply used but it sounds like you're after some power.

There are other parts of the cct that could be improved also.

Cheers
 
Yes thats it, but you will need to use the 10 FET board if you use IRFP250/350/450 or similar. If you find FETS rated at 250 watts or more then you can use the 6 FET board.

For 500 watts into 4 ohms you will need an 800va centre tapped transformer with AC voltages of 56v 0v 56v. This will give you rails of around +/- 78 volts.

I built my amp almost entirely out of second hand parts and it cost me only $80 US.

It is difficult to put a number on how much it will cost you, it depends on what standard you build it to. Decisions you make about the power supply can vary just that cost by a hundred dollars or so.

If you decide to build this amp but are unsure what a good complete setup would look like, take the time to look at the pictures I posted thoughout the thread and drop me a line (or join that thread) describing what you want to achieve.

Or if you aren't in a hurry you can track / join this thread
http://www.diyaudio.com/forums/showthread.php?postid=759082#post759082
where a new (and more powerful) amplifier is being developed, again with the contributions of other members.


Cheers
 
Darksog said:
hi quasi
i think i use IRFP460 N-FET 500V 20A 280W
and add 2 fets more then it will be 8 fets in amp and 2 ohm load stabil? 75/2/4=9.375A per fet =)


Hi Darksog,

The IRFP460 can be used in the amp, and 8 FETs might be enough but only just. Let me explain;

Whilst a speaker load is rated at say "2 ohms" impedance it's actual impedance could dip under 1 ohm. This is because speaker loads are not resistive but reactive and in reactive loads the current is not in phase with the voltage. This causes real problems (and device failures).

The graph attached shows the possible (worst case) current draw for 8, 4 and 2 ohm speaker loads with 75 volt amp rails, as well as the SOAR of 4 IRFP460s at DC and at 10mS non-repetitive pulse. The graph calculations assume a heatsink temperature of 50 degC.

As you can see a "2 ohm load" can draw 44 amps peak or 11 amps per FET. The DC SOAR curve (blue) is well inside this curve and the 10mS SOAR curve is just outside.

Without going into too much detail this chart tells me that for 2 ohm loads normal music signals that drive the amp into full power only intermittently, 4 x IRFP460s per rail will survive. Extended low frequency signals at continuous full power will destroy the output stage.

Remember also (a lot I know) that the FETs will not share the current equally. So while the average peak current is 11 amps per FET, a few may be passing more.

For me there is not enough room for error with 8 FETs so if I were you, I would use 10 FETs per board as a minimum if the amp was to be driven into 2 ohms.

The IRFP460 also has a much higher gate capacitance (4.2 nF). This means that the driver transistors will have to work harder to cope with the extra charge & discharge current. Please make sure that the driver transistors heatsinking is adequate.

Cheers
 

Attachments

  • soar 4 x irfp460.jpg
    soar 4 x irfp460.jpg
    68.3 KB · Views: 755
Hi Quasi,
For 2 ohms stable operation into reactive or mismatched Loads for sustained periods with low frequency content.. upto16 IRFP460 FETs or 8 Pairs of FETs are required.....when driven from +- 100Vrails....

For +- 75V rails upto 6 pairs or 12 FETs are required..for safe 2 ohm operation..

Gate capacitance doesnot plays an important role in gate drive current requirements...The Foremost important parameter is Total Gate Charge and highest frequency of operation...

Calculate Ig from...this equation below...
Gate Drive Current per FET = Total Gate Charge Qg in nC X Highest Frequency of operation....in Hz

gate resistors must be lowered in value and current drain through gate to source bleeder resistor should be at least 5 times of total gate current of FETs


K a n w a r
 
Thanks Kanwar,

So there you go. With Kanwar's experience in demanding PA applications his view is 12. I say 10 he says 12, in any case you get the picture.

But why the discrepancy??

It gets down to the amplifiers application and cost. If I was installing amplifiers into a concert venue, where the conditions are demanding and loud then Kanwar’s view is valid.

In a home theatre or hi-fi setup where full volume is out of place but the dynamic range is still needed then 8 might be enough.

The power supply capability holds some of the answer. A power supply sized for home use will collapse sufficiently to offer some self-protection. A power supply in a PA amp is usually stronger and likely to hold voltage better, good for the sound but bad for the FETs. In my calculations I factor in a modest 5% power supply droop. Most power supplies (even good ones) would droop more at full power.

As a perspective, most domestic amplifiers I've seen to date would fail the SOAR calculations I have used here.


Cheers
 
quasi said:
The power supply capability holds some of the answer. A power supply sized for home use will collapse sufficiently to offer some self-protection. A power supply in a PA amp is usually stronger and likely to hold voltage better, good for the sound but bad for the FETs.

Cheers


The Present Trend of PA amp Power Supply isn't stronger one..
The PA amp manufactures uses under powered supplies...In the event of 8 ohm load for a typical 1200@2ohm amp the rails are at +-110..gradually upon loading with 4 ohms these sag to +-85V then again to +-70V at 2 ohms load..this type of setup has an advantage of 2 times the peak power for a given rms watts for short periods of time brusts you can say..and close spacing of power amp wattage at different load impedances..
lets say in present day scenario in PA amps..
1200W--2ohms
850W--4ohms
600W--8ohms...
not like 1200W--2 ohms, 600W at 4ohms...350W --8 ohms...with stiff power supplies...time has changed...less weight of amps..highest peak power of music signal brust...close power specs with impedance variations...alike
 
jacco vermeulen said:
Q, slack as in reserve.
How much more favorable would the SOA curve be if not square waves would have been used but sinus waves ?
Who will test a domestic amplifier with a square wave input and a 2 ohms dummy load ?


Hmm...I don't know.

I suspect a manufacturer would rather test into a 2 ohm resistive load with a square wave than a sine wave...
 
jacco vermeulen said:
Q, slack as in reserve.
How much more favorable would the SOA curve be if not square waves would have been used but sinus waves ?
Who will test a domestic amplifier with a square wave input and a 2 ohms dummy load ?


Hi Jacco ,

Square wave testing would not let you to test the sustainance of amp in 2 ohms load..its a very light load...
For Real Hardcore Testing Use a Highly Reactive 2 ohms load and drive the amp with Sinewaves....
A 1mH in series and 22MFD in shunt with 2 ohms resistor will do the fine job....

K a n w a r
 
Hi Jacco...
Yes I have done these test...on my NVMOS amps..

There is a Story behind it....
In my early explorations..When I first designed an amp with 500WRMS@4ohms...Set the current trip shutdown threshold at 22amps peak...Then we took it in a DJ party.. when operating in a program with the usual heavy BASS & Treble CD music ..there wasnot any tripping ..suddenly the main singer came to the stage and started singing...and within few seconds the amp was tripping due to excess current..demand because of a vocal singer voice in live arena.....Everything was out of my understanding...what was happened...Then i remove the trip sensor and the amp functions effortlessly till after 1 hour it gets somked...into fire...due to this singer's voice ...

next day i invited the singer into my place and again repeated the program excatly but with intruments attached for monitoring what was happened....
I was amazed that the peak current demand into typical 2 X 15"4 ohms speaker was 35amperes peak..much greater than what i thought during the vocal sessions....then dismantles the speaker,,a normal crossover was found which poses the highly reactive load when only the mid frequency were fed during the singer singing the vocal....

The amp was observing both Highly peak current drive and highly reactive drive simuntaneously ..the reactive load forces the OFF FET to pass as much as 10 amperes at its max VCC which was 70V ...instant dissipation was 10X70=700W....which insist that paralleling must be done inorder to sustain the operation....
Then to check this validity I run the same procedure on the another amp which was Studiomaster 2000E....upon encountering that load this amp limits it current heavily and distorts the sound very much..... hwereas our prototype does the job fine with more parallelled output fets..

From That Day I constructed a Reference Reactive load which was to be used as a testing measure for my amps.....
During Highly reactive loads...The SOA must be observed with FULL VCC accross the FET Drain to Source and its ability to pass the max current at that point must be considered when driving it with reactive loads....I choose the value of max reactive current equals One Third of Peak current into pure resistive load...per FET at Given VCC....
Lets say if peak current capability is 10 amperes then reactive current capability of the output stage must be at least 3.33 amperes...or it would not survive...Some users prefer VI limiting...I dont like it because it heavily distorts the sound....


K a n w a r
 
Hi, Kanwar,

.. when operating in a program with the usual heavy BASS & Treble CD music ..there wasnot any tripping ..suddenly the main singer came to the stage and started singing...and within few seconds the amp was tripping due to excess current..demand because of a vocal singer voice in live arena.....Everything was out of my understanding...what was happened...

I experienced the same thing here in my workdesk. I've noticed it for a long-long time, but I'm not sure what is happening until you write that. I think it has to do with resonance with passive Xover or/and the speaker, but I haven't research it.

I notice that when I play Disco music with full bass, the whole system is not sucking current as much as if the music played is only a female voice.

Female voice is much more a "sucker" than bass tones :D
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.