500 watts into 1ohm SS amplifier design

[mund24]

pa sub din sir tony! for sure dami matutunan dito!

[tony]

 

we have thus far specified Vceo of our output trannies, so how about Ic?

in post #1,  Io peak = 32A, now this is for a pure resistance load, speaker loads are seldom resistive....

let us say phase angle of 60*, so now our current Ic needs to be 32A/cosine 60* = 32/0.5 = 64 amperes

if i can get a quad of these: http://www.datasheetcatalog.org/datasheet/sanrex/SQD300AA100.pdf

but we will work with the ussual suspects, the 2sc5200, mj21193, etc.....

more later.....

[n0th1n6]

Subscribing muna at dumudugo na ilong ko

[DJFEL]

sub...ok to para sa aken..nice one..

[tony]

It's not just wire size. Even the way you connect the wires (cables) matters a lot.  Every fraction of an ohm due to poor contacts means several watts of wasted power.

yes, i know very well what loose connections can do, our electrical vault room burned down as a result, we had to work 3 days for 24 hours straight, we were all like zombies after that....

[tony]

That made this topic even more interesting. These will teach the readers how good engineering, solid foundation, and attention to details, can overcome difficult engineering problems.

thank you sir, my aim here also is to demonstrate how audio power amps are designed and hopefully built.....

 

[tony]

In addition to power loses due to ohmic drop, another amplifier system parameters that will suffer when wire resistance comes into play is the Damping Factor. How this affect the sonic quality may spark another heated debate, and if it does, lets do it in separate thread para di magulo itong main topic .

One appealing application that comes to mind when toying with amplifiers driving low impedance load is speaker box where the amplifier is housed right inside the speaker box. Powered speaker ba ang tawag duon? Short wire runs, soldered joints, and elimination of connectors solves some of the problem raised my master tintop.

hopefully with valuable insights from you, TTH, labgruppen and others we can see this thru.....please feel free to argue when you feel i am straying...

[tony]

^right on sir..  this amp should be used as near to the speaker as possible to avoid losses while maintaining a reasonable amount of DF.. it also calls for tons of NFB to have an equivalent internal impedance as low as possible..
hmm, how about using a "sense lead" (ala laboratory power supplies)? it might cause instabilities at high frequencies, but if there is a way to apply it to low frequencies only, where DF really matters, then it could help.. some integrated amps have a feature like this, kenwood is one iirc, but i have not seen one in action.

i recall in the 80"s Philips came up with this "motional feedback system", they demonstrated it in a bar wherein a band was playing, they made the speakers appear as though they were part of the performance, then one by one the band members got up and left the stage one by one until all of them were all in the backstage....
at this point, the audience erupted in a wild applause.......

[tony]

we need output transistors with Vceo of 100v and an Ic of 64 amps, these are minimum requirements, nothing stops us from using better transistors in terms of these specs...the only stumbling block is that of cost and availability...

a DIY'er diys not because gusto nyang makatipid, but more so because gusto nya "yung kanya ang masunod", commerical amps are designed and built with profit margin in mind....

a DIY'er does not think of profit but rather of satisfaction in having built something to his liking......enough of my rantings....

let us look at what seem to be a favorite power trannie here, the 2SC5200 and its compliment the 2SA1943, http://pdf1.alldatasheet.com/datasheet-pdf/view/30356/TOSHIBA/2SC5200.html

if you look at the specs for Ic you will see, it says 15Amps, so that you may think that 64/15 = 4.26 or 5 trannies are good......there are other considerations......

we want to use the part of the hFE characteristic where beta does not droop too much, and i choose 4 amps, so now our ouput trannie count becomes 64/4 = 16 trannies or 32 trannies including the pnp side.......

but wait, there is still more....

[tony]

can we use a smaller number of power trannies? let's find out......

the answer to that is SOA, transistor safe operating area, by definition it is the area on the trannies' operating or active region wherein secondary breakdown is less likely to happen....

secondary breakdown happens with a simultaneous occurrence of high voltage and current, it is a hotspot on the die surface where localized heating happens and destroys the chip....we do not want that to happen to our amplifier....

here are some good readings on the topic: http://sound.westhost.com/soa.htm , http://en.wikipedia.org/wiki/Safe_operating_area

some discussion here: http://www.diyaudio.com/forums/solid-state/158520-output-transistor-safe-operating-area.html

and here: http://www.diyaudio.com/forums/solid-state/4091-redesign-leach-amp-pcb-integrated-247-output-devices.html

[tony]

other references: http://home.alphalink.com.au/~cambie/ETI466Web.htm#Output , http://www.diyaudio.com/forums/solid-state/49255-safe-operating-area.html

[mund24]

^

nice info with those links! thanks sir tony +1

[tony]

to continue......

yesterday, i said that using 2SC5200/2SA1943 combo, we will need 16pairs of these trannies....

what about the MJL21193 and MJL21194?

looking at the datasheet: http://pdf1.alldatasheet.com/datasheet-pdf/view/172457/ONSEMI/MJL21193.html

if we were to follow the same logic that we used in selecting for 2sc5200, we will find that beta droop is not pronounced until about 8amps, but in this case i will use 6amps, so therefore, we need 64/6, or 11 pairs of these trannies.....

and if we were to use TO-3 style package then we will need 11/1.5 or 8pairs of these MJ21193/MJ21194 trannies...

then what about TIP35/TIP36?
let us look at the datasheet: http://pdf1.alldatasheet.com/datasheet-pdf/view/12674/ONSEMI/TIP35C.html

here we see that 3amp is good for this device, so that we will need 64/3 or 22 pairs for this amp......

no one is preventing you to use more pairs, this is diy, and there is no accountant looking behind our backs to slap our hands if we go over budget.....

[TinTopHack]

no one is preventing you to use more pairs, this is diy, and there is no accountant looking behind our backs to slap our hands if we go over budget.....

Right... no accountant pero nan dyan naman si Mrs... Sobra pa sa accountant.

[tony]

to summarize:

2SC5200/2SA1943 ------ 16 pairs

MJL21193/MJL21194 ----- 11 pairs

MJ21193/MJ21194 -------  8 pairs

TIP35C/TIP36C ----------  22 pairs

notice that up to this point i have not mentioned about the device dissipation Pd max, i have thus far tackled, Vceo and Ic....
The reason we do not use this spec as a figure of merit as it has not much use to us in designing amps....

looking at 2SC5200, at dc operation and if case is held to a temp of 25*C it can dissipate 150 watts, so that if heatsinking is such that case temperature reaches 50*C it can still dissipate 75 watts......but we do not want to operate our trannies this way, we want to have it to operate at much lower temperature as heat is the number one killer of solid state devices....

the MJL21193/94 otoh while rated at 200watts at 25*C case temp has a soa of 150watts at 25*C....
  
the TIP35C/36C is rated at 125 watts  at 25*C case temp has only 50watts soa at case temp of 25*C....

SOA were taken at dc curves, 50 volts, except for the MJL21193/94 which has only a 1sec curve provided....

[tony]

Right... no accountant pero nan dyan naman si Mrs... Sobra pa sa accountant.

sinabi mo pa 

[Johannah Aleksandria]

Hintay ako ng update sa makapag-assemble ng ganito. Sa ngayon malabo pa sa akin ang benefits nito, maybe more readings.
Go go lang sir.

[tony]

what else do we need to consider in designing an output stage?

for me, hFE and Ft,

hFE is  dc current gain, while Ft is the transition frequency, it is the frequency wherein the trannie gain becomes unity.....


how do the candidate trannies fare in this aspect?

2SC5200, Ft is 30mhz, while hFE is almost 100 at 4amps, but for our purpuses we will just use datasheet spec of 55min

Mj21194, Ft is 4mhz and hFE is about 45 at 6amps....

TIP35/36, Ft is 3mhz, and hFE is about 105 at 3amps.....

so for purposes of this design, i choose the 2SC5200  with its compliment, the 2SA1943....

[tony]

so we have the 2SC5200 as output trannies, what is the drive required? http://pdf1.alldatasheet.com/datasheet-pdf/view/408787/TOSHIBA/2SC5200_04.html

taking the hFE at 4amps as 100, peak current input to base required then equals 64/100 or 640mA, this is the current that needs to be delivered to the bases of our output trannies....

let us look for a driver...remember that 640mA peak current needs to be delivered to the bases to push the output trannies hard...bear in mind that our output trannies have an Ft of 30mhz, using a driver such as the TIP41C/42C is a waste,(http://pdf1.alldatasheet.com/datasheet-pdf/view/427094/ONSEMI/TIP41CG.html) these devices have lower Ft's and are therefore slow.....

therefore, as driver i will also go for the 2SC5200...http://pdf1.alldatasheet.com/datasheet-pdf/view/408787/TOSHIBA/2SC5200_04.html

what is the required drive to the base of this driver so it can put out 640mA?

looking at the datasheet, hFE at 640mA is also 100, so that 6.4mA is required at the base of this power trannie used as driver....

at this level, it seems that there is no need for a triple darlington type output configuration.....

here we see how this japanese trannie has very desireable traits....

more later...

[Min]

Cge lng sir tony tuloy mo lang pag se share nito kahit na naka nose bleed intindihin, basa basa lang baka sakali makayanan.

+1 po sa pag se share.