Lme49830

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

Segran

Member
Joined 2008
Hi Spittin Llama and other pro's, I have a LME49830/MOSFET PCB layout (sorry nickarsow, it's not SMD...) to which I would like to get some feed-back. This is supposed to be the final version, I have made improvements to the signal ground and feed-back loop compared to the previous prototypes.

Thanks in advance!
 

Attachments

  • lme49830.pdf
    95.6 KB · Views: 865
S

Segran

Member
Joined 2008
Here's the schematic to go with the layout. The layout is used both for LME49810 and 49830 with MOSFET/BJT's. That's why R12, J3, D6, D7 are present on the PCB, but not used with LME4930/MOSFET's.

I have found C14 and C15 in the Leach type feedback RC-net to cause more distortion (THD, THD+N and IMD) so they are not to be mounted. Anyone else that have made the same experience, or is it due to an "inferior" PCB layout?

I have made some measurements using Arta (great stuff) and my Asus Xonar Essence STX sound card (also great stuff - dynamic range >124dB, 0,0003%THD+N on Line out, 118dB and 0,0002%THD+N on Line in, measured using AP!).

In the MOSFET version I get 0,00067%THD, and 0,00055%THD on the BJT, measured at 110W@7R0 (that's due to my 2x48VDC suplly as opposed to the 60VDC in the schematic. IMD also slightly better on the BJT version.

However, the odd harmonics is higher on the BJT, which probably explains why I (and several others) deem the MOSFET to sound more pleasant to the ear in spite of the measured data.
 

Attachments

  • lme49830 mosfet 150w.pdf
    38 KB · Views: 891
N

nickarsow

Member
Joined 2009
interlux-bg.com
Dear Sergan,
Thanks for your reply. Now I'm working on a LME49830 module - a separate module which could be suited for any amplifier design.
It contains 1 lme49830 with a heat sink , a preamp lme49710 , surrounding SMD components and some capacitors at power pins.
The module will have 2 rows of pins - 10pins and 15pins.
The problem I'm thinking on is how to realise the groundon the module , because it has just an analogue groung.
Could I use top and bottom poligon copper for the analogue ground?
Best regards
Nick
 
S

Segran

Member
Joined 2008
The problem I'm thinking on is how to realise the groundon the module , because it has just an analogue groung.
Click to expand...

I'm sure there are more experienced guys around thsi forum to give you better answers than me!

If you look at my schematics you will notice that I use different ground symbols for analog Signal GND and PWR GND, including decoupling cap's.

The two "grounds" are joined by R31, it's stated 82R but could be anthing in the range from 10 to 100R. The important thing is to keep the two grounds separate all the way to your Star ground point (teh ONE and ONLY). I use separate ground wires for Signal GND and PWR GND.

BTW, since you will have a lower voltage supply for the LME49710, you have to give some thougt to how you connect the GND leads from the two DC supplies - again only connected to the Star GND point. Just a piece of advice, this had me going in circles for weeks until I found my wrong doings!
 
S

Segran

Member
Joined 2008
Please note the module has no POWER FETS on it , so it has only one ground - analogue
Click to expand...

Ok, but I guess you have a few decoupling caps around the chips. My point is that the ground connection for all decoupling/bypass cap's should have it's own ground wire to the Star GND point, not mixed with the analog signal GND. I was amazed myself how much noise you pick up from those...:)
 
A

AndrewT

R.I.P.
Joined 2004
Segran said:
[BI guess you have a few decoupling caps around the chips. My point is that the ground connection for all decoupling/bypass cap's should have it's own ground wire to the Star GND point, not mixed with the analog signal GND. I was amazed myself how much noise you pick up from those...:) [/B]
Click to expand...
The Signal Ground must not be contaminated with any dirty (spiky) currents. All dirty currents must return to the Audio Ground (star grnd) on a separate connection from the clean currents returning through the Signal Ground.
 
N

nickarsow

Member
Joined 2009
interlux-bg.com
Dear Segran,
Please excuse me I can't send you the sch file now , because till monday I'm out of my work , but monday morning I'll send you the file.
Nick. Preliminary - the schematics is based on an original National's file and a 1/2 LM833 , 2 bypass electrolytic for the LME49830 and 2 for LM833. 2 outputs through resistors to the FETs and 3 connections to the external bias NPN. Nothing else - very simple , but the idea is that module could be used by anyone anywhere where needed.
Regards
Nick
 
S

Segran

Member
Joined 2008
Hi Nick,

my advice (and AndrewT's) is still to use separate ground connections for power (decoupling caps and pin 3 onLME49830) and the analog signals and connect the both using separate wires all the way to the Star GND point.

This will minimize the noise level to a minimum. As it is done on the PCB now you will have a lot of 100Hz noise.

Connect the GND leads of R4, R6, R7, C7 and C10 to the Signal Ground, and all the remaining capacitors to Power Ground.

Good luck!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom