Mono blocks using National LME49830

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

I am planning to start a new mono blocks project using the National LME49830 chips. The circuit employed will be from the application note available at: http://www.national.com/an/AN/AN-1850.pdf. I am having almost all the components readily available with me.

Some of the parameters/requirements are:
1. Each mono block should be able to generate an output power of 150W RMS
2. the speaker load will be 8 Ohms.
3. Extremely silent power supply.

I was not able to find a simple and cost effective supply for this project. I am not so interested in the power supply application note that is available from National as it is a bit bulky and costly. Also, some of the components are not easy to find. But, I am interested in taking the delay and mute sections from that note.

Can someone help me ?

Best regards,
Bins.
 
With the exception of ClassA monsters, the power supply is not the source of humm and hiss and therefore does not have to be extremely silent. The cause for this is then grounding, cabling, distance of the VAS to transformer etc. So I think - especially as these ICs usually have huge PSRR - that you will be happy with a classic linear supply (transformer+ caps).

Or am I missing something? By the way I couldn't find a schematic for a PSU in the referenced application note or anything else that would suggest something special?

Have fun, Hannes
 
Hi,
a dual secondary 40+40Vac transformer will give ~58 to 59Vdc.
With good smoothing and a well built amp you can easily achieve 150W into 8r0.

If you want to have the ability to drive a 4ohm load then you will need to double the smoothing capacitance and double the VA rating of the transformer.

For 150W into 8r0, a 300VA with +-20mF will do.

My 230:40+40Vac gives +-58.5Vdc when fed with 240Vac and the amp bias is 220mA. It gives 170W into 8r0 from a 3pair output stage.
 
Hi Bins,

build the right half of the PSU in the app-note (Fig.1) and instead of all the soft-start/relay fun, put a thermistor, CL60, in series with the primary leads of the transformer.

Works very well.

Have fun, Hannes

PS: by the way, I'm not sure you need soft-start for 2x20000uF at all.
 
Hi ,

Had a brief look at the thermistors and their application in current limiting. The CL-60 solution seems good. How can we provide the mute control if we take off the soft start section ?

Can I reduce the 20,000uF supply reservoir to a lower value around 10,000uF without affecting the amp performance ? What is the main purpose of this ?

Best regards,
Bins.
 
For 150W into 8r0, a 300VA with +-20mF will do.

My 230:40+40Vac gives +-58.5Vdc when fed with 240Vac and the amp bias is 220mA. It gives 170W into 8r0 from a 3pair output stage.

Can you comment on the power supply you had used
1kVA 230:40+40Vac with twin 35A rectifiers (4rectifiers for the two channels). Separate Audio Grounds to the two channels. Both Audio Grounds connected to the Safety Earth with their own Disconnecting Network (6 off 25A rectifiers in total for the two channels).
+-45mF(15//15//15mF, 63V) to each channel. Decoupling 820//820//820//820uF twice on each PCB. Can't make these too big, or it blows the PCB fuses at switch on.
 
Hello AndrewT,

I was a little busy with my light speed project and schroeder tonearm. I am planning to restart my power amplifier project as soon as possible. As stated in the National application note, I am going to use a 24V - 300VA dual secondary transformer along with 700V - 35A bridge rectifier.

But, I would like to use 10,000μF value capacitors instead of the stated 20,000μF capacitors as they are readily available with me. Will there be a performance hit back ? Please suggest.

Will this power supply be able to drive the left and right mono blocks ?

Best regards,
Bins.
 
Hello AndrewT,

OK, then I will use discrete power supply for the left and right channels. What about the capacitor value ? Will a 10,000 μF value be enough ? Can I use them instead of the 20,000 μF ?

Will there be a performance hit back ? Please suggest.

Best regards,
Bins.

they won't be monoblocks.
Two power amps fed from a common PSU must be a 2channel amplifier.
 
Last edited:

opc

Member
Joined 2004
Paid Member
Bins,

For the same price as that Epcos capacitor in your post at the top of this page, you could buy 4 of these:

UNITED CHEMI-CON|ESMH800VNN103MA50T|CAP ALUM ELECT 10000UF 80V SNAP | Farnell India

You'd need to buy at least 4 of the Epcos caps if you wanted to build monoblocks, which will be about Rs. 8408. If you buy the caps in the link above, you could get 20 of them for the same price, which would give you 50,000 uF per rail instead of 10,000 uF.

Don't be afraid of the snap in type, as they're easy to make a simple PCB for, or you can always wire them together with some nice copper buss wire. You're paying an enormous premium for the screw terminal type.

As for the LME49830, it's a superb chip and I highly recommend it. If I were you, I'd run 60 volt rails on the output fets, and run the LME49830 on a separate supply at about 65-70 volts. This lets you swing the fets right to the rail, which means you need less rail voltage to attain a desired output power, and you can bias higher without dissipating as much power. Keep in mind that the designs I've built with this chip really like a lot of bias, I've taken THD+N measurements at 10mA increments from 20mA to 2A, and the performance just keeps getting better the higher you go.

Best of luck!
Owen
 
if you adopt the design philosophy that amplifier passband and NFB along with PSU time constants are inter-related then changing from 20mF to drive an 8ohms speaker will affect the frequency response.

Starting with your +-10mF of smoothing per monoblock.
The PSU RC time constant is 8 * 0.01 = 0.08s = 80ms.
Using that philosophy then the NFB RC <= 0.08 / sqrt(2) <= 56ms.
let's choose a feedback pair of 27k and 1k0. This would require a DC blocking capacitor of <=56uF. Choose 47uF giving RC = 47ms.
Following on from there the input high pass filter should be <= 47 / sqrt(2) <=33ms.
Lets assume the Rin = 27k to match the feedback resistor.
Then it follows that the DC blocking capacitor at the input should be <=1u2F. Choose 1uF.
The input filter of the power amp is 1uF+27k. F-3dB of this single pole passive filter is 5.9Hz.
You may just be able to hear the loss of bass response coming from your speaker with this 1uF blocking cap and that 27k input impedance.

Alternatively you can ignore the possibility of the inter-related RC time constant requirement of the power amplifier and simply set the passband of the amplifier to any value you fancy.

I can only comment on what I find with my systems, I cannot hear your systems.
 
Hi,

The SNAP type is having only 80V whereas the capacitor used in the power supply application note is having 100V rating. Will that be an issue ? If you don't mind, can you send me some power supply schematics for LME49830. My requirement is to develop a 150 watt mono block amplifier to drive 8 Ohms speakers.

Best regards,
Bins.

Bins,

For the same price as that Epcos capacitor in your post at the top of this page, you could buy 4 of these:

UNITED CHEMI-CON|ESMH800VNN103MA50T|CAP ALUM ELECT 10000UF 80V SNAP | Farnell India

You'd need to buy at least 4 of the Epcos caps if you wanted to build monoblocks, which will be about Rs. 8408. If you buy the caps in the link above, you could get 20 of them for the same price, which would give you 50,000 uF per rail instead of 10,000 uF.

Don't be afraid of the snap in type, as they're easy to make a simple PCB for, or you can always wire them together with some nice copper buss wire. You're paying an enormous premium for the screw terminal type.

As for the LME49830, it's a superb chip and I highly recommend it. If I were you, I'd run 60 volt rails on the output fets, and run the LME49830 on a separate supply at about 65-70 volts. This lets you swing the fets right to the rail, which means you need less rail voltage to attain a desired output power, and you can bias higher without dissipating as much power. Keep in mind that the designs I've built with this chip really like a lot of bias, I've taken THD+N measurements at 10mA increments from 20mA to 2A, and the performance just keeps getting better the higher you go.

Best of luck!
Owen
 
My requirement is to develop a 150 watt mono block amplifier to drive 8 Ohms speakers.
this can easily be done with a 40+40Vac transformer and 63V capacitors using BJTs. If adopting FETs requires the next voltage step up in transformer and capacitors then it is not good value for your money. eg. a 3pair Leach clone on 40Vac delivers >170W into 8r0 and can safely drive 6ohm speakers. Some even use it to drive 4ohm speaker, but I won't and don't because the SOAR of 3pair does not meet my criteria for a 4ohm amplifier.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.