High Current Leach/leach super amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi i was just wandering if anyone has or knows how to modify the leach/leach super amp to run really low impedences like 1 ohm. I was just thinking of adding more output transistors and also changing the protection circuit so it doesnt kick in until about 0.5 ohm or some thing(but dont know exactly what parts i need to change). Please let me know what you guys think.Please help if you can.

regards

Bowdown
 
I just wanted a high current amp as i have a sub that needs about 2000wrms and is 2 ohms(dual 4 ohm coils) so i was going to build another pair of leach/superamps as i have heaps and heaps of parts for them(I mean heaps). was going to build 2 high current modules that are stable down to 1 ohm then run them bridged at 2ohm. i dont currently have a 4WD.

Cheers

Bowdown
 
Im usually in Bunbury but go to perth(joondalup) every couple of weeks. Yeah see i bought all the stuff in big quantity to get a really good price on the stuff as i was planning on making a few leach amps for my home theater stuff probably end up with 3 or so 6 ch amps and a couple of BIG 1ch amps to run the subs. and i have to build a few amps for my brother too. thats why i didnt worry about over buying. whats p&q mean? sorry. You in Australia? yeah some info on power supply would be good. was thinking of using two or three 2.5-3 kva toroidals per monoblock. but i would have to run 2-3 lots of power sockets to the amps as the sockets can only handle 10 amps. might get a higher current one put in.

cheers

Bowdown
 
Hi bowdown,

OK, lets do some engineering for free ;)

2000 watts @ 2 ohms results in a peak current of

sqrt(2 * 2000watts / 2ohms) = 45 amps

as well as an peak voltage of

sqrt(2 * 2000watts * 2 ohms) = 89 volts

With some losses you'll need about 100 volts on the rails. To make this amp save on a phase angle of let's say 45°, each transistor bank must handle about

1/sqrt(2) * 100 volts * 45 amps = 3100 watts peak.

I suggest, that the transistors have this SOAR capabilities for at least 10ms. When using a MJL21193/MJL21194 for example one transistor is able to handle about 1.5 amps @ 100V at 25°C. Because the junction temperature is higher you'll have to decrease the power handling of each transistor to about 1 amp.

In this case you have to parallel about 30 devices (each bank, positive and negative) for reliable operation.

The efficiency of a power amp is about 65%. This means that you have to dissipate about

0.35 * 2000watts = 700watts.

Although when you don't need this as permanent load requirement, a heatsink of about 0.05K/W is necessary. Forced cooling for this low thermal resistance is mandatory. In this case the heatsing gets about 35°C warmer under full load condition.

Power supply:

1 / efficency * output power = power of supplies

1 / 0.65 * 2000watts = 3100 watts is necessary for your power supply. In this case you'll need a transformer, which power rating is about 1.5 timers higher.

A 4500 watts transformer is needed, when permanent load is required.

To obtain an rail voltage of 100 volts under full load condition and about 10% voltage drop from unloaded to loaded condition a transformer with

100volts * 1.1 / sqrt(2) = 78volts

is required.

Although these are only quick calculations, there shouldn't severe mistakes in.
 
Gees bocka thanx for all this good info. Hmmm dont know what i should do now. This is what i have at the moment at home. I got approx 130 pairs of mjl21193/94, approx 200 pairs of mj15003/4, 5000 pairs of mpsa06/56,250 pairs of mje15030/1,1000 pairs of 2n5416/2n3439 and all the resistors,diodes and caps in the hundreds, 4 2kva torodals giving me +/- 96Vdc and another 4 on their way. i Have 3 meters of raw heatsink (300mm = about 0.3c/w not fan cooled) i can get it cut into any length. Please advise me what other amps i can build with this stuff.

Jaycee i dont mind running it at 8ohm but dunno how to get 2000watts rms at that impedence.

Gees thanx guys for all your help. i really appreciate it.

Regards

Bowdown
 
I would look into building one of Anthony Holton's amps, MOSFET output based, if you really want this kind of power. I would say 2kW for a sub is very, very, very overkill when a 400W is enough to make my house shake.

You're definately going to have to have lots of output devices, a high rail voltage, and probably fan cooling for this. This is an ideal application for a Class D amp but these aren't really practical for the DIY constructor unless he has a lot of experience.
 
I'd build the leach amp. Stick with the MJL21193/21194 as outputs. For the driver I'd recommend another MJL21193/21194. And for the predriver I'll recommend a MJE15032/MJE15033.

The MJ15003/15004, the MJE150031/150032, MPSA05/56 are not applicable for this high power design because of the too low Vce ratings. I haven't a datasheet for the 2n5416/2n3439 on the hand. I'd look for transistors with a Vce of 300V at least. Maybe there are some 2SA/2SC devices that fit these requirements.

The transformers should be sufficiant. Maybe you'll get a little bit less power than 2000 watts.

I think forced cooling with a 0.3K/W heatsink can obtain 0.05K/W. Give it a try. Maybe you measure the temperature under full load conditions.
 
Hi bowdown...

The Bipolar Transistor amps are not suitable for driving ultra low impedance loads....because their SOA is extremely small at high voltages and they also suffer from Second breakdown voltages....they are poor performers when you are in need of large voltage signal along with the same magnitude of current rating..
Ultra low impedance load driving capability falls under the kingdom of only N-channel large die mosfets.... BJT just sags the power even if you parallel them in great numbers...

We had a similar situation like yours but our customer wants to drive his 4 X 800W 18" Selenium 8 ohm Subwoofers in parallel resulting in 3200WRMS@2ohms and it needs at least 4000WRMS of sheer power to do the right justice for the Ricter scale to move upwards.
We implemented 2 bridge amplifiers comprising 8 pairs of APT20M18LVR large die mosfets from APT with + - 90 Volt rails and 4 toroids transformer of 755VA each with total of 16 X 10000MFD/100V Filter caps.....8 X 1000V/35A Bridge Rectifiers
Some specs...
Power output....4000WRMS at 2 ohms
Frequency response....10hz to 50khz
Slew rate.....85V/us
Damping Factor >1000 @ 4 OHMS


bowdown you have so much stock under your command, beefing a leach amp is not worth because they are not designed to handle these type of ultra low impedances...try to find some other bipolar design on this forum ......

K a n w a r
 
The Bipolar Transistor amps are not suitable for driving ultra low impedance loads....because their SOA is extremely small at high voltages

Hi Kanwar,

this is not true, look to the MJL21193 datasheet. 150 watts @100 volts is not so bad for a TO247 package. It's true that bipolars have a second breakdown. You have to take this into account when using them.

About 50-100 watts output power for each transistor on a bank is a typical rating for high-power bipolars. Not neccessarilly every design needs to be a (N)MOSFET design.

Because bowdown has the MJL21193/MJL21194 on the hand he can use them when he designs his amp carefully. And the MJL21193/21194 is cheap.

4 toroids transformer of 755VA each

This is definitely not enough for 4000 watts continous power. Look to my calculations.

beefing a leach amp is not worth because they are not designed to handle these type of ultra low impedances...try to find some other bipolar design on this forum

No, just parallel enough output transistors for low impedances and use a BIG BIG heatsink. The leach amp is intentionally designed or 8 ohms. For 2 ohms the number of output devices is simply 4 times higher. The problem is, that the leach amp is designed for a +/- 58V power supply, not for +/-100V. The leach amps needs to be modified carefully in this case.

just run your amps bridged, [snip]

I think its not impossible to drive it 2000 watts at 8 ohms

Bridging and modifiing the leach amp for 2000 watts @ 8 ohms is nearly the same as modifiing the leach amp to 2000 watts @ 2 ohms. Same number of ouput devices, same power supply...
 
Hello BOCKA......,

this is not true, look to the MJL21193 datasheet. 150 watts @100 volts is not so bad for a TO247 package. It's true that bipolars have a second breakdown. You have to take this into account when using them.

Why not you take a look at the datasheet of APT20M18LVR [www.advancedpower.com] and feel the massive SOA under your command......and imagine what type of amp could be constructed by using these monsters....

About 50-100 watts output power for each transistor on a bank is a typical rating for high-power bipolars. Not neccessarilly every design needs to be a (N)MOSFET design.

Yeah its not necessary for every design to be n-channel mosfet one...but if you want real ++KW of power levels only N-channel mosfet Rules...... that domain

This is definitely not enough for 4000 watts continous power. Look to my calculations.

Do you listen to continuous subsonic Music content on subwoofers or just listen to 10Hz---20HZ continuous Sinewaves only......meanwhile hefty Capacitor bank takes great care of musical dynamics....

No, just parallel enough output transistors for low impedances and use a BIG BIG heatsink.

These would consume lot and lots of space on PCB as well as heatsink and still what you get is power SAG when drving reactive loads... why not use high power devices and force cooling .....

K a n w a r
 
The Leach Super Amp was once called the Double Barrel Amp.

I believe the reason was that the design called for Paralleled as well as SERIES connected output devices. This halves the voltage the ouput transistors sees thus allowing it to operate well within it's SOAR even at higher current and rail voltages.
 
Gees guys thank you all so much for all your help and info.

Bocka, so your saying that the mj15003/4 to be changed to mjl21193/4 but 4 times as many(16 pairs)
the MJE150031/32 to be changed to mjl21193/4
and the mpsa06/56 to be changed to MJE15032/33
is this correct???

the 2n3439/5416 have a Vceo of 350V

http://www.semelab.co.uk/pdf/bipolar/2N3439_40.pdf

http://perso.wanadoo.fr/jm.plantefeve/2N5416.pdf


Cant i use the double barreled (super amp) version of this amp but use mjl21193/4 outputs x 4 and change the mje150031/32 to mjl21193/4. i currently have a double barrelled amp running on +/-96 Vdc and has been running beautifully for just over a year now. Obviously id have to make a new pcb layout, but shouldnt be too much of a hassle i think.

http://users.ece.gatech.edu/~mleach/superamp/
Thanx workhorse for your input, but id rather try and build a leach super amp on steroids as i dont want all or most of the stuff i have on hand to go to waste especially the mjl21193/4 and the mj15003/4(cant really use em for this project anyway unless i want to series them.)
It would be a real pity to see so much good gear go to waste thats all.

Gees guys i really appreciate all your help.

Thanx again

Bowdown
 
Hi,
you have two 4ohm voice coils.
I see three options
1. Run them in parallel requiring 2000W into 2ohms. Ultra currents!
2. Run them in series requiring 2000w into 8ohms. High volts even when bridged .
3. Two amps into separate voice coils requiring 1000W + 1000w into 4ohms. 4 core speaker cable. Two bridged amps avoids High volts.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.