Hey guys,
I've been slowly assembling parts for a large power amplifier,based on the LME49810 chip. The amp will be mono,for subwoofer duty.
I'm looking for input,and suggestions..Any pointers are welcomed!
My current plan is to marry the '49810 chip to the Leach output stage.
LME49810 Datasheet
Leach amp schematic
So far,here's what I've got:
-8pairs of MJ15011 and MJ15012 on a heatsink.
I'm thinking I'll use 7 pairs for the output,and one pair for the drivers. (Q16,Q17 in the Leach amp schematic.)
MJ15011/12 datasheet
-The emitter resistors are 0.698 ohm 1% 10W.
(maybe a tad large,but thats okay for now. I can change them later if everything seems well balanced/matched.)
-For Q14 and Q15 on the Leach schematic,I was thinking about using some 2SD845/2SB755 transistors (or similar) that I have in the parts bin. (got a few to choose from)
-Rail voltage will be ~+/-70 (no load/idle) I'm assuming it will drop to ~+/-60V under a decent load.. Eventually I'll find a bigger power transformer,but for now this one is okay.
-Power trans is kind of 'unknown'.It came from a junked SS amp.
Primary (120Vac) is fused at 8A. Each 50Vac tap had a 5A fuse.(There are 25Vac taps also,which were fused at 10A each.)
I think it was some kind of dual-rail design (class H?)..it had big rects,and filter caps for each set of rails. (~ +/-35V and 70V)
I won't be using the 25Vac taps.
-Main filtering is currently 8200ufx2 and 680ufx2,But I will be adding (alot) more,plus some at the output transistors.
The PSU is snubberized (1ohm and 0.047uf?),and bypassed with 0.33uf film caps,along with a 0.22uf cap across the +/- rails.
I've used a large 35A 600V bridge rectifier,with 0.1uf snubber caps across the pins.
So far in testing the noise and ripple is quite low under a moderate load.
(scope shows <5mV sawtooth at idle)
I wanted to over-build the output stage by a fair amount.
It will probably only ever be used with 4 and 8 ohm loads.
Assuming that the rails hold up,thats a good 400+W into 8ohms-That's Plenty,and should be easy for 7pairs of transistors.
-I need suggestions for the Vbe circuit in general.
I'm thinking of a transistor in a TO220/TO126 package,so I can screw it to the main heatsink easily.
I've got quite a few types to choose from in the bin,and can always order/buy something if I need to.
What transistors have others used here?
Any suggestion as to what works well with the LME48910 chip?
Any other suggestions or tips are appreciated!
I've been slowly assembling parts for a large power amplifier,based on the LME49810 chip. The amp will be mono,for subwoofer duty.
I'm looking for input,and suggestions..Any pointers are welcomed!
My current plan is to marry the '49810 chip to the Leach output stage.
LME49810 Datasheet
Leach amp schematic
So far,here's what I've got:
-8pairs of MJ15011 and MJ15012 on a heatsink.
I'm thinking I'll use 7 pairs for the output,and one pair for the drivers. (Q16,Q17 in the Leach amp schematic.)
MJ15011/12 datasheet
-The emitter resistors are 0.698 ohm 1% 10W.
(maybe a tad large,but thats okay for now. I can change them later if everything seems well balanced/matched.)
-For Q14 and Q15 on the Leach schematic,I was thinking about using some 2SD845/2SB755 transistors (or similar) that I have in the parts bin. (got a few to choose from)
-Rail voltage will be ~+/-70 (no load/idle) I'm assuming it will drop to ~+/-60V under a decent load.. Eventually I'll find a bigger power transformer,but for now this one is okay.
-Power trans is kind of 'unknown'.It came from a junked SS amp.
Primary (120Vac) is fused at 8A. Each 50Vac tap had a 5A fuse.(There are 25Vac taps also,which were fused at 10A each.)
I think it was some kind of dual-rail design (class H?)..it had big rects,and filter caps for each set of rails. (~ +/-35V and 70V)
I won't be using the 25Vac taps.
-Main filtering is currently 8200ufx2 and 680ufx2,But I will be adding (alot) more,plus some at the output transistors.
The PSU is snubberized (1ohm and 0.047uf?),and bypassed with 0.33uf film caps,along with a 0.22uf cap across the +/- rails.
I've used a large 35A 600V bridge rectifier,with 0.1uf snubber caps across the pins.
So far in testing the noise and ripple is quite low under a moderate load.
(scope shows <5mV sawtooth at idle)I wanted to over-build the output stage by a fair amount.
It will probably only ever be used with 4 and 8 ohm loads.
Assuming that the rails hold up,thats a good 400+W into 8ohms-That's Plenty,and should be easy for 7pairs of transistors.
-I need suggestions for the Vbe circuit in general.
I'm thinking of a transistor in a TO220/TO126 package,so I can screw it to the main heatsink easily.
I've got quite a few types to choose from in the bin,and can always order/buy something if I need to.
What transistors have others used here?
Any suggestion as to what works well with the LME48910 chip?
Any other suggestions or tips are appreciated!
hi, im happy to see another LME amplifier being built.
im a be looking forward to your progress, since ill be needing a subwoofer amplifier soon.
im currently working on version 3 of my LME amp project, my amp is no where near as large as your 8 pair output stage, mine uses 2 pairs and the other version uses 3 pairs, hopefully ill learn something from your build so i can upgrade mines. lol.
one suggestion is to implement a transistor protection into your pcb in case of a output shorting. since the LME doesn't protect for shorts.
i didnt implement it on my pcb, so im going to have the protection circuit added on laters, but its really a must.
looking forward to you build.
laters
im a be looking forward to your progress, since ill be needing a subwoofer amplifier soon.
im currently working on version 3 of my LME amp project, my amp is no where near as large as your 8 pair output stage, mine uses 2 pairs and the other version uses 3 pairs, hopefully ill learn something from your build so i can upgrade mines. lol.
one suggestion is to implement a transistor protection into your pcb in case of a output shorting. since the LME doesn't protect for shorts.
i didnt implement it on my pcb, so im going to have the protection circuit added on laters, but its really a must.
looking forward to you build.
laters
if you haven't already seen these.
they're commercially available LME49810 amps, here is a link where they include the schematics, the board, silk and measurements for tree different LME amps.
ax1300
ax1400
ax1600
at the bottom of the page they have the links to the schematics PCB etc...
he also has power supply schematics for them.
just though it may help.
laters
they're commercially available LME49810 amps, here is a link where they include the schematics, the board, silk and measurements for tree different LME amps.
ax1300
ax1400
ax1600
at the bottom of the page they have the links to the schematics PCB etc...
he also has power supply schematics for them.
just though it may help.
laters
It will have protection circuitry for sure. 
I have a big relay that I'm going to use for the output. Some say it won't do much good on a high power/voltage design,but I figure it's better than nothing at all!
Fuses,DC protection,thermal,overload,etc. circuitry will also be used. (Based on the circuits on ESP,Project #33,etc.)
I have a couple fans mounted in a piece of plexi that I'm going to use for cooling.They even have the open-collector "third wire" for tacho,or whatever. Maybe I'll make some kind of swanky 'smart' fan controller later on,but for now it's gonna be simple.
I probably WILL use the lower rail voltage,for initial testing,and perhaps when experimenting with lower impedance loads.
The soft-start,and Aux-PSU are done and tested as well. I've got ~15.5V raw,which is regulated down to +12V and +5V for the LME49810,and whatnot.
The soft-start and output relay are operated from the raw +15V supply.
The soft-start is just two BIG 33ohm resistors in parallel,bypassed by a relay after an R/C delay. (150ohms and 1000uf on the relay coil) It works nicely,and the resistors don't even warm up at all. We'll see if adding more capacitance changes that by a significant amount.
I have a big relay that I'm going to use for the output. Some say it won't do much good on a high power/voltage design,but I figure it's better than nothing at all!
Fuses,DC protection,thermal,overload,etc. circuitry will also be used. (Based on the circuits on ESP,Project #33,etc.)
I have a couple fans mounted in a piece of plexi that I'm going to use for cooling.They even have the open-collector "third wire" for tacho,or whatever. Maybe I'll make some kind of swanky 'smart' fan controller later on,but for now it's gonna be simple.
I probably WILL use the lower rail voltage,for initial testing,and perhaps when experimenting with lower impedance loads.
The soft-start,and Aux-PSU are done and tested as well. I've got ~15.5V raw,which is regulated down to +12V and +5V for the LME49810,and whatnot.
The soft-start and output relay are operated from the raw +15V supply.
The soft-start is just two BIG 33ohm resistors in parallel,bypassed by a relay after an R/C delay. (150ohms and 1000uf on the relay coil) It works nicely,and the resistors don't even warm up at all. We'll see if adding more capacitance changes that by a significant amount.
DigitalJunkie said:
my LME has gotten too hot to touch before, now i use a PC Memory Heatsinks "Ramsinks".
you can use the double sided tape that comes with it, or you can remove it and use thermal paste, drill two holes and attach it to the LME chip, i have one of these but in aluminum and it works great.
i think the one on the picture above may work better then mine, since it has longer fins.
if your driver stage draws the full current that the LME has to offer it will get pretty hot. but as long as it has a heatsink it'll be fine.
Use a separate heat sink for the LME49810 than the output stage. The output stage will get plenty hot but the driver will be more a consistent temperature. You don't want the two to interact thermally. All the thermal compensation for biasing is done outside of the chip so no need to interaction and it will cause some degradation in THD performance.
If driving lots of output devices then it will take lots of work on the drive but only if you don't have an buffer or intermediate driver stage. So you just have to work the numbers to see if you need this current gain stage or not. Most likely, you only need a small heat sink on the driver. Watch National's web pages for more info.
-SL
If driving lots of output devices then it will take lots of work on the drive but only if you don't have an buffer or intermediate driver stage. So you just have to work the numbers to see if you need this current gain stage or not. Most likely, you only need a small heat sink on the driver. Watch National's web pages for more info.
-SL
Yea,since I'm going to be driving so many output pairs,I will probably (have to) use some kind of driver stage. The exact parts,design,etc. are still kind of up in the air though.
OK,I will attach the LME chip to it's own separate heatsink,of a good size. I will have to change my layout a little bit,but I can make it work.
I was wondering if attaching it to the main heatsink would be beneficial,due to the chips built-in thermal protection,but it sounds as if the separate heatsink is the better option.
I'll mount the Vbe parts on a small board,and mount it to the main heatsink,as usual.
OK,I will attach the LME chip to it's own separate heatsink,of a good size. I will have to change my layout a little bit,but I can make it work.
I was wondering if attaching it to the main heatsink would be beneficial,due to the chips built-in thermal protection,but it sounds as if the separate heatsink is the better option.
I'll mount the Vbe parts on a small board,and mount it to the main heatsink,as usual.
Re: The Vbe..
A quick dig through the bag of transistors popped up a handful of 2SC2911's
2SC2911 Datasheet
They look like a possible candidate for the Vbe multiplier to me,but I'm no expert! Shall I give it a whirl?
Edit:
For anyone interested,here's some crude phone-pics of the heatsink right after I drilled it.
It's a big hunk of aluminum I-beam.
A quick dig through the bag of transistors popped up a handful of 2SC2911's
2SC2911 Datasheet
They look like a possible candidate for the Vbe multiplier to me,but I'm no expert! Shall I give it a whirl?
Edit:
For anyone interested,here's some crude phone-pics of the heatsink right after I drilled it.
It's a big hunk of aluminum I-beam.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The Leach is missing a resistor in each half of the protection circuit. Look at Jens' thread on the small Leach clone.
7pair may need a high power device as a driver, equaling your 8pair output stage. The pre-driver will then need an SOAR to suit.
The Leach output needs ~20mV across the emitter resistors. This will make the main heatsink run very hot. Don't put the LME on this hot sink. Use a separate sink and aim to keep Tc < [ambient+20Cdegrees]
400W into 8r0 needs 80Vpk at the output terminal. You can't get that from +-70Vdc.
your 50Vac transformer will give a lot more than +-70Vdc. On no load it is likely to be around +-80Vdc and dropping to +-75Vdc when biased to 200mA. You'll probably need 100V caps for worst case supply conditions.
Can you remove some of the thin, 5A, wire to reduce the open circuit output voltage to around 40Vac to 45Vac.
Alternatively remove half of the thin wire and rewind it in parallel to the half that remains. This will give ~ 25 + 25/2 = 37.5Vac at full VA of the transformer.
7pair may need a high power device as a driver, equaling your 8pair output stage. The pre-driver will then need an SOAR to suit.
The Leach output needs ~20mV across the emitter resistors. This will make the main heatsink run very hot. Don't put the LME on this hot sink. Use a separate sink and aim to keep Tc < [ambient+20Cdegrees]
400W into 8r0 needs 80Vpk at the output terminal. You can't get that from +-70Vdc.
your 50Vac transformer will give a lot more than +-70Vdc. On no load it is likely to be around +-80Vdc and dropping to +-75Vdc when biased to 200mA. You'll probably need 100V caps for worst case supply conditions.
Can you remove some of the thin, 5A, wire to reduce the open circuit output voltage to around 40Vac to 45Vac.
Alternatively remove half of the thin wire and rewind it in parallel to the half that remains. This will give ~ 25 + 25/2 = 37.5Vac at full VA of the transformer.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.