Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

SOA, Voltage and Watts question
SOA, Voltage and Watts question
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools Search this Thread
Old 22nd May 2008, 11:53 PM   #1
ricsmuts is offline ricsmuts  South Africa
diyAudio Member
Join Date: Nov 2006
Default SOA, Voltage and Watts question

This one got me thinking.

I was asked by a friend who competes in IASCA SPL competitions to make him a BIG amp (personally I suggestted rather use decent speakers, efficiancy and intelligent design to attain his goal) Some people just don't see it my way.

Nevertheless I am taking the challenge for the sake of it, and I also want to see his face when we promptly melt his so called 1000W voice coils - Hey, he asked, I will deliver, the consequences are his deal.

By BIG we are talking possibly +/- 2000W continuous 1000W per channel (yes that's about 250+A) which is actually another problem to be dealt with but is do-able.

Now to get this kind of power we are going to end up with something that's about 6 feet long and looks like a semiconductor manufacturing plant in the trunk.

The questions that have arisen are:

1) Output stages, what? Triple darlington, quadruple darlington, or something else?

2) Voltage, looking at SOA graphs of MJ15004/MJL4302/2SA5200/2SA5242 and everything else I can think of, at Voltages above 80V per rail it appears the gains obtained by the higher voltages are not very significant at all in relation to the "possibility of blowing up very expensive transistor stuff" factor

At 100V per rail I can run about 0.3A per output transistor, and that's running on the limit. To be sensible, how close to the SOA max should I go?

3) If I use 70V rails, what could be the expected real life max output into 4 Ohm with a good design? (I refuse to let him run 2 Ohm, as doing so, I may as well build a good amp and then just short it out with a pair of car jumper cables) I would expect to still have some kind of damping from the amp at these power levels and 4 Ohm might still give some.

4) Sound quality becomes irrelevant at 2000W and headache pill quality becomes increasingly important. The question is If I use only class B to avoid trying to bias 5000 unmatched output transistors, getting it slightly wrong and exploding his battery, will it make an audiable difference to the bass?

5) Are there any published known working designs of this size out there as I am venturing into awkward territory here?

Before somebody tells me this, Please, no thanks I don't want to go buy pre made amps off the shelf because they are better and cheaper and easier and more powerfull than to make myself and SPL is stupid etc etc,

I honestly would like to build it at totally unreasonable cost, for the sake and the fun and challenge of it, not reason, and will probably have some friction with my buddy when I melt his speakers, and have sleepless nights trying to get it to work, I'm building it, not to listen to it myself, but because that is what engineering and challenges are about.

Thanks for any help and advice in advance.
  Reply With Quote
Old 23rd May 2008, 01:16 AM   #2
Perry Babin is offline Perry Babin  United States
diyAudio Member
Join Date: Nov 2003
Location: Louisiana
Hopefully, wg_ski will help you with the design specifics. I'd suggest the following...

If he can use any speaker he chooses, suggest that he uses a speaker with multiple coils (4 coils). Then you can get the power from lower voltage. This will allow you to do a couple of things that will make the project better (IMO). If you use lower rail voltage, you can use TIP35Cs and TIP36Cs. They are dirt cheap and rugged. With a regulated supply, you'll be able to easily get 800 watts into 2 ohms. If you build four 800 watt boards, that's 2400 watts RMS into a speaker with four 2 ohm coils. You can get more power with lower impedance coils.

I suggest individual boards for several reasons. The most important is ease of replacement if one gets damaged beyond repair. This goes for the power supply boards as well as the audio boards. Having multiple boards also means that the amp can be repaired if one board fails. It will also be easier to build.

If you decide to use high rail voltage, it will probably be easier to keep the transistors cool if you use plastic case transistors. To transfer heat quickly and continuously, you need a lot of aluminum behind the transistor. This is difficult with TO-3 transistors unless you have custom sinks manufactured. Many amps attach the TO-3 transistors to relatively thin (1/4") plates and attach the plates to the sink. The plastic transistors with heavy aluminum behind them is likely to transfer the heat better. TO-3 transistors look cool but they're difficult to work with.

If you use triple darlington, and need to squeeze every possible watt out of it, you'll need to produce a driver supply voltage. There are many ways to do it. Without it, you won't be able to drive the output to the rails.
  Reply With Quote
Old 23rd May 2008, 05:08 AM   #3
1moreamp is offline 1moreamp  United States
diyAudio Member
Join Date: Dec 2005
Location: Northern California
Try looking under the Solid State forum here on DIY, There are several very fertile projects going on over there that will wet your appetite. Just add your on Switch mode power supply and you will be in business.

On a side note class AB in not the most efficient way to go for a sub amp. Class D has its own section also with some very fertile ideas going on in that forum.

My 200 time 4 amp draws 125 amperes at 800 watts, so I think your going to be more into the 300+ ampere draw on an AB design

Here are a few links to start you out below:



Good luck on your project....
  Reply With Quote
Old 23rd May 2008, 12:38 PM   #4
ricsmuts is offline ricsmuts  South Africa
diyAudio Member
Join Date: Nov 2006
Nice designs

That Lynx is a real beaut, enough to sidetrack me into making two for myself
  Reply With Quote
Old 23rd May 2008, 04:10 PM   #5
wg_ski is online now wg_ski  United States
diyAudio Member
Join Date: Oct 2007
Experience has shown me that 80 volt rails are the optimum for a standard class B design. Any lower, and you can't get kW levels effectively. Any more you're up against SOA limits. One of my best amps is a CS800 "clone" that runs at this voltage. It's basically the same capability as the old-school 94 pound boat anchor, but with upgrades to the circuitry. It measures 440W/ch at 20 Hz at 4 ohms, and 650 at 2. Seven MJ21193/4 (T0-3 can) pairs reliably drives real world 2 ohm loads, and I've taken it down to six speaker pairs with no trouble. For the flat-packs on the same load I would use eight to ten. Go to a dozen or more in parallel you can drive a welding rod (or maybe a 1 ohm load). The MJL21193/4 pair is the most cost effective per watt of SOA. The C5200's and clones may be "faster" or more linear but not as rugged - and on a triple darlington it simply does not matter if the outputs have a 30 MHz fT or not. In the end you can violate the DC 25C SOA by about a factor of two, but that's all. In reality, what you do is calculate the effective thermal resistsance for a 50mS half-sine pulse and de-rate for termperature (maybe 70C). It turns out these devices are good for about 400 watts instantaneous dissipation up to the second breakdown point (80V) which is what limits the supply voltage to this. You still calculate the heatsink for average power. Bridge two channels at 4 ohms and you're over 1200W on an unregulated supply. Better regulation will give more power, but if you're running off a 12V supply with a DC/DC, you're going to drop a lot in the wiring and regulation wouldn't be the most brilliant move. It would just keep drawing more and more current till something in the power supply gives out.

DO NOT worry about matching output devices - other than use the same type from the same date code in any parallel bank. From the same lot, the quiescent current values will be plenty close enough, and with any decent value emitter resistors they'll share ok under load. There may be as much as 20% mismatch at quiescent, but that's only a few mA. The amount of thermal feedback is dependent on the current - and that will increase when you start drawing load current.

To go higher than 80V rails you must think about stacking output devices in series/parallel or using multiple rail voltages and going class H. You don't see that done in car amps very much, but it's not impossible. You wanna be the first to DIY it? I have given some consideration to building an EP2500 "clone" for a car system. Requires quad 55V rails and independent power supplies for each channel which makes the power supply the real challenge. 1900 real-world bass watts would make a lot of noise in an enclosed space. For more ideas, see my thread under solid-state, about the 2kW PA amp I'm working on. 3-step class H similar to RMX4050, but using my favorite "guts" in the amplifier itself.
  Reply With Quote
Old 23rd May 2008, 04:22 PM   #6
ricsmuts is offline ricsmuts  South Africa
diyAudio Member
Join Date: Nov 2006

Please remind me to buy you a beer when I am that side.

You always answer my questions properly - thanks
  Reply With Quote


SOA, Voltage and Watts questionHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Decibel Watts VS Watts ? fivestring Solid State 11 1st April 2009 02:08 PM
Easy (newbie) Watts & Speaker Question Lash Car Audio 10 8th May 2008 03:02 PM
quick question about watts and subwoofers tooooslow Subwoofers 3 21st April 2006 03:28 PM
Sound quality of amps at 0.3 watts vs @ typical 5 watts rick57 Solid State 2 26th March 2005 07:25 AM
Voltage question for Zen 4 Buhl Pass Labs 9 24th August 2003 07:54 PM

New To Site? Need Help?

All times are GMT. The time now is 12:53 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 16.67%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio