1500 Watts RMS - 2 Ohms, to come soon

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Amplifier is ready, made by hurricane, one of my friends, tomorrow we gonna evaluate the Melter, making comparisons.

He made it awfull, awfull assembling, he planned it since first component, and is modern, has voltage regulation in the differential, and power supply is stabilized electronically with 8 hi power Motorola transistors. Each channel will use around 30 transistors, more than 20 will be power units.

He said that could measure 1.5 Kilowatts over 2 ohms, i will check it.... in multisim it can go more, but simulation is not guaranteed.

Will be my birthday gift for you all, he said that i can give them to everyone, you gonna like, because modern.

I am thinking in use 4 Computer power Supplies.... will try first... Computer 1 KW AC mains stabilizer from 220 to 110 volts.

Titanic well be alive soon

Carlos
 

Attachments

  • titanic.jpg
    titanic.jpg
    64 KB · Views: 1,710
Commentable Thoughts

Hi Carlos glad to hear about ur amp.

BUT... one comment indeed for u.

At Higher Power Amplification BJT output Stage stands nothing in the FRONT of N-Channel Quasi complementary Mosfets output Stages.

This is not criticising, nor it is a mistaken entity or any advertisement,
BUT IT IS INSANE REALITY THAT BJT's HAVE NO MATCH FOR N-CHANNEL MOSFETS WHEN IT COMES TO HIGH POWER LEVELS.

Believe it or not ! , u can ask i think ANTHONY HOLTON [www.aussieamplifiers.com]ofcourse about the performance of mosfets.:smash:

High Pressure Regards,
AMPMAN

BTW : GODDLUCK FOR UR TITANIC
 
Yes, Amp Man and Filipe, really strong this one

This one put me afraid to touch... to much voltage on it.

Amp man, no doubts about that. this kind of schematic i know very well not the better, but the idea is the young ones, they become crazy with those things, this way, as i have a friend that donne it and is using a long time, a good way to make boys happy.

And i feel good too.

The boys can use it to demolition, old houses Filipe, those ones we have here and your country too.... just put 10 hertz and go volume up in many speakers.... maybe can be helpfull.

The problem is because the kids can have some electrical problem, because of supply, this way, i will make this advise to beginners, do not construct this one, better to use plus and minus 40 volts to be safe.

Now i will let here the schematic, i will change some, because found bad square wave response around 20 hertz....and this cannot still this way.

Also not satisfied with square response at 60 kilohertz, is becaming rounded... only 40 Kilohertz response... also power is too big, around 1700, i think in reduce it to around 1K2.

Now starting to understant the supply he made, and try to increase a little, because his supply loosing 20 percent.... i will try 10 percent maximum.

Here is the schematic, will change a little, i can see faulty condensers, also i can see small error voltages, and 10 percent differences in stand by current...gonna check all things, and in 15 days i will put the titanic completed, with supply, and stabilized supply.

To the ones crazy to try.... go ahead and help me with informs.

I am afraid the voltage, because the young boys crazy for those things.

I am satisfied with clean 30 watts each channel, normally i use 2 watts each channel.

regards all....like schwartznegger said...."I'LL BE BACK!"

I think good idea go to states and represent speakers there... we have Selenium, also Bravox, some Arlen, Also Phillips and many others.... i like Bravox, but Selenium is good too, all them has their own qualities, depending on use. Also a lot of special speakers made for powerholic competition , with speaker installed on automobiles too.

The transistor already used are not the last Motorola units, have to try others and hear again, also try computer stabilizer, the ones not connected direct to mains, better primary and secondary transformers. Those autotransformers are dangerous.

ATTENTION BOYS, I CANNOT BE RESPONSIBLE IF YOU ELETROCUTED YOUSELF, THIS ONE DANGEROUS!

The amplifier will use 12 pairs of the biggest Motorola transistor available, and suply will use 4 pairs.

Carlos
 

Attachments

  • titanic schem.jpg
    titanic schem.jpg
    50.1 KB · Views: 1,669
Now i will put regulator schematic

The supply used by friend hurricane, the one that constructed Titanic, was loosing too much voltage....more than 40 percent... this way no good power achieved.... around 600 watts maximum in 2 ohms.

The amplifier was considered, alike the DVD movie film "A Knights Tale":

Bull and hurricane, you was measured, you was weigthed, and considered insufficient.

The amplifier Titanic, had 12 position related to 13 amplifiers, he win only one...people agreed that may be supply problem, but they do not like the sound when tested in 30 watts rms level too, they perceived loss of deep bass too. And i also perceived this way.

So, to make this monster, may construct a very heavy supply, one separated 50 pounds transformer i suppose.... or bigger than that.

Now i will finish to try improve, and loose interest in improve circuit, because supply will be a very big problem!

But can help if you want to do it, i am at your service friends...just call me, and i will be there!

Transistors used in supply and amplifier was:

First stage, the diferential stage , BC556/546 pairs.
Second stage 2SA1321/2SC3344.
Third stage 2SC4753/2SA1837.
Output uses MJ15022 and MJ15023, 12 pairs

Supply use same output transistors, 4 pairs, feedback was 2SA1321/2SC3344 and suply driver was MJ15023 and MJ15024

Bias current is 10 miliamperes each transistor, emitter resistors are 0.2 ohms... total bias around 140 mA each rail, ... 15 watts plus 15 watts iddle power dissipation....

Needing more help to construct, i may be helpfull and glad to help.

Carlos
 

Attachments

  • supply titanic.jpg
    supply titanic.jpg
    63.7 KB · Views: 1,337
Hi Carlos,

I must admit I have not studied this thread as much as I should, short of time, but I'd like to look at the current issues raised by Amp_Man. Forgive me if I'm dredging up old knowledge, but I believe he is right; you need more current amplification.

First up, let's work backwards for voltage swing peak to peak into 2R.

1500 watts into 2R (nominal) is 155Vpp, or 54.8Vrms. Since the peak value of this output is 77.5V, into 2R a current of 38.75A is required, assuming the load is precisely 2R and NOT reactive. A reactive load causes difficulties because the current and the voltage will be out of step; for a capacitive load, current is a maximum at zero output voltage, and this puts the entire rail voltage across the output devices just as they pass their highest current. Needless to say, the SOAR of almost any transistor on planet earth has the odd difficulty with this.......:hot:

Your plan to use 12 output pairs means, with proper current sharing (I'd use 0R47 emitter resistors to force current sharing an minimize mismatch influences), each output device would pass a peak current of 3.22A. Once we select our rail voltage at around 84V (a 60-0-60Vac would be fine here), this should be sufficient for SOAR limitations into real world loads, but that's a can of worms I won't address here. 12 pairs should be fine, particularly if you use current limiting for protection.

I haven't got the specs on the 15003/4 device here, but I believe beta at 3.2A is around 30. Let's assume 30; if I'm wrong someone please correct me. At 38.75A peak, we therefore have a drive current requirement of 1.29A for the entire output stage, call it 1.3A peak. This current can only come from the driver.

I do have specs on the 2SC4793/2SA1837, and you can expect a current gain of around 100 in operation. So, for 1.3A peak, we need a drive at the base around 13mA. This is a minimum figure, and probably around 15-16mA would be a better estimate.

This is too much for your voltage amp to supply, as I believe you want to run it at 10mA. Correct? You need extra current through the stage to keep the VAS and the bias generator ticking over, and peak positive current tells us that at least 13mA will flow straight off the CCS, into the base of the driver, but a little extra should remain to flow down to the VAS (assuming an NPN VAS on the negative rail).

If you use a bipolar VAS, and I suspect you might, you still need an excess of current through the VAS to permit the inactive VAS to pass current, and in truth the current variation should not be too great so as linearity does not suffer by overworking the NFB loop.

In either the SE or balanced VAS topology, therefore, you really do need at least 25mA flowing. This is a pretty stiff VAS stage current, and it could be problematic to drive with the input stage, unless you move to darlington VAS, which I don't recommend as their parasitics are too high and thus their speed is lowish.

All this points to the need for another stage between driver and outputs, making a triple. If it were me, I'd simply put in another emitter follower, probably use very high speed transistors around 4mA in TO92 for the VAS (this puts device dissipation at around 83 x 0.004 = 332mW which is just acceptable in a 'tall' TO92), then a couple of very fast TO220s for the predrivers, then run with the 4793/1837 for the drivers.

The 4793/1837 will pass 2A continuous, and 4A transient peak, adequate for this design. They are 100/70MHz devices, very fast, high linearity, low input capacitance, and I'd certainly use them. And they are not too expensive!

Of course it goes without saying that an additional stage will make the amp more difficult to stabilize. But rail voltage is still manageable at +/-84V, the outputs are not that fast, and it should be fine with a combination of miller cap and phase lead from output to feedback node.

Hope this helps,

Cheers,

Hugh
 
I wouldn't design any circuit to use 2SC4793 and 2SA1837 with Ic higher than 300mA in normal operation, this is due to the fact that their current gain starts to drop above 300mA and they are rated at 1A max. These are fast but fragile transistors so prepare to use 6 pairs or so. Another alternative is MJE15030/MJE15031 or MJE15032/15033, these are more rugged and have almost flat gain up to 2.5A but they are also slower and suffer from higher capacitances so their base should be driven from a low impedance [a typical flaw from OnSemi devices]

In some old unfinished prototype I have lying here, I was using 3 pairs of 2SC4793/2SA1837 to drive 6 pairs of 2SC3264/2SA1295. It was intended to run on +-100V rails and be happy with loads above 3 ohms [typical Rdc of the coil of a 4 ohm bass driver or two 8 ohm bass drivers in paralell] and show some realiability down to 1.5 ohms. The output is a triple darlington but the predriver and the preceding stages are made with standard BC550C/BC560C/BD139/BD140 transistors since they are powered from +-15V rails due to the virtual ground approach I'm using. I hope to have the money, the time and the inspiration required to finish it some day [It's a pity to see the transistors required for 4 channels sitting in a box and the 40x20x4cm heatsinks filling with dust over time]
 
Oops!! Eva is quite right, my mistake, apologies Effendi.....:xeye:

Substitute the drivers with MJE15030/31 (one per rail should be sufficient) but now the VAS current will need to come up to about 12mA to drive the higher input capacitance of the ONSEMI devices.

You could use the 4793/1837 in the VAS role, actually. They are certainly fast enough, and at 12mA they will be dissipating 1W apiece on 84V rails, so heatsinks are now required on all devices from VASs forward.

Thanks Eva (is this Guido Tent?)

Cheers,

Hugh
 
Commentable Thoughts

AKSA said:
Hi Carlos,

I'd like to look at the current issues raised by Amp_Man. I believe he is right; you need more current amplification.

Hugh


ThanKs AKSA for clarifying the statement about current amplification to Destroyer X

Carlos , u must insert an extra pre-driver stage to drive ur driver transistors.

With regards
AmPmAn
 
This amplifier was measured, weigthed and considered not enough

Supply voltage reduced a lot, can not hold the amplifier, i sent back to Hurricane and said him:

Here is your hot potato! hohohô

Had not motivation to increase it, because it was considered not good sounding amplifier, also with very low power related all his maximum theorical power... in 30 watts comparison test, this amplifier gonne the almost last place related others.

More details in the Van Melter comparison test made.

Voltage, when hard driven, reduced 40 percent as i already said... this way, cannot be so good, and thats the reason nothing explode for a while.

Thanks all informs, i will give all those to hurricane, also i have some better transistors to give him too.

If you like, as an exercise for you, without doubts much better knowledge than mine, can modified it entirelly, because no owner too, this one belongs to mankind, this way, can be Amp man, Eva's or Aksa's amplifier, i am short circuited brain those last 3 days, cannot think in this earthquaker, because too much surprised with DOUGLAS SELF amplifier, that i am simulating those days.... without believe what Multisim are showing me, perfect square wave in 10 hertz, also 100 kilohertz!

This Multisim 2001 never shown me this performance!, no amplifier simulated did that till now!, i had concluded that was Multisim 2001 bug!, All amplifiers i made,modified , or copied and entered to simulation, no one did that.... this way i will make it real.... now trying to have 2N5401 because could not find them... and this one cannot be substituted, because quality goes down.(this explains why not good simulation in Aksa and perfect sounding amplifier in real world)

Also i never had this experience before, always moving transistors models and having the performance almost the same..... this time not!.... cannot put others, tried 11 models in simulation...no one worked there, hard unbalance... had to change a lot to stabilize and equalize rail voltages.

I have a lot to learn, thats the real thing, sorry the heavy Troy Horse, a Greek gift as we say in my country.... only power...no good quality the way it is in schematic, and with a falling down supply voltage.

Thats the reason i could have good simulation results with Aksa too.... and the sound, real world is wonderfull, magnificent, perfect.... and simulation dennies that....because of 2N5401 again.

This thread amplifier is more to young boys make loud guitar music than to hear good and soft music, is more usefull to teenagers surrounding that he is at home. .. i already run away of this one.

Thank you all, there's a Troy Gift, maybe a challenge for some newcomer enginneer, or a good exercise to someone.

Have you already seen, in simulation, perfect square waves in deep bass?.....Douglas Self could make one circuit that shows that perfect square wave in simulation.... those things makes me feel that is better to find another hobby, because 40 years doing amplifiers, and beeing surprised already...without understanding what's is going on!

bye and thanks

Carlos
 
Good Mr. Teal, if something cracks here in my building i will think you made it shake

You made my building shake too much.

The supply used from hurricane, my friend, was bad, when switch on, voltage reduced a lot, this way, no explosion on transistors.

The first stage, will be more safe using 2N5401 or other with better voltage capability.... also the Output, use MJ15025, to replacement of others, read entire thread, many friends helped a lot with adequate replacements of more adequated drivers.

Thanks Eva, also Hugh, and all others that came here to help.

This amplifier needs a very heavy transformer, and do not look at energy bill, i think you will be surprised. hohohô

Regards, and good luck to you........but your neighboors!!!!

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