LM3886 PCB vs Point-to-Point (with data) - diyAudio
Go Back   Home > Forums > Amplifiers > Chip Amps

Chip Amps Amplifiers based on integrated circuits

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 7th March 2014, 07:38 AM   #1
tomchr is offline tomchr  United States
diyAudio Member
 
tomchr's Avatar
 
Join Date: Feb 2009
Location: Greater Seattle Area
Default LM3886 PCB vs Point-to-Point (with data)

First off, let me say that I do not intend to cause a shouting match between the PCB lovers and the Point-to-Point (P2P) enthusiasts. I just noticed something interesting while prototyping so I decided to investigate and document it...

The story is this: I'm designing an amp based on the LM3886. My intent is to improve dramatically on the performance of the original LM3886. The final amp will reside on a PCB, but I figured I would build a prototype first. That way I could get the bugs ironed out before committing big bucks to a PCB fab run. I built the prototype using P2P wiring as that's the quickest and I expected it to provide as good performance as a PCB. Indeed the performance at 1 kHz was reasonably close to the data sheet figures.

However, I was rather surprised when I measured the THD versus frequency. At 20 kHz the THD was an order of magnitude worse than the data sheet would indicate. This was rather surprising to me as the P2P circuit is pretty tight. The feedback network is right on the IC pins. As is the decoupling caps. The distance from the 1000 uF supply caps to the IC is not more than 40~50 mm. The circuit was fed from a quality lab supply.

I decided to spin a quick prototype PCB. On the PCB is the exact same circuit as I used for the P2P setup. All the components came from the same batches of parts. Yet, on the PCB, I am able to reproduce the data sheet performance. The THD is much improved. The ground plane on the PCB is connected to the power ground. Signal ground runs as a separate trace and joins the power ground at the center between the two 1000 uF supply caps.

As seen from the transient responses, neither circuit exhibits tendencies towards instability. The transient response is nice and well controlled for both the PCB and the P2P circuit.

The schematic attached shows the main parts of the circuit. In addition, there are two 1000 uF electrolytic caps bypassed by two 47 nF ceramic caps where the power enters the board.

I suspect the return paths for the various electric fields in the circuit are better controlled in the PCB-based circuit. I am guessing that is the explanation for the difference.

The abrupt change in THD around 1.5 W is caused by range switching in the HP 8903A distortion analyzer. The uptick in THD above 6 W for the 20 Hz trace is caused by brownout of the power supply.

The supply voltage was +/-28 V.

Equipment: HP 8903A distortion analyzer. Tektronix 2465B 400 MHz oscilloscope (using 20 MHz bandwidth setting to avoid interference from a nearby FM transmitter). HP 6228B power supply.

Just thought I'd share...

~Tom
Attached Images
File Type: png LM3886_Schematic.png (17.3 KB, 1609 views)
File Type: jpg LM3886_PCB_Transient10kHz.jpg (271.1 KB, 1578 views)
File Type: png LM3886_PCB_THD_vs_Freq_15W.png (80.2 KB, 1486 views)
File Type: png LM3886_PCB_THD_PowerFreq.png (135.6 KB, 1369 views)
File Type: jpg LM3886_P2P_15R22nF.jpg (241.5 KB, 1330 views)
File Type: png LM3886_P2P_THD_vs_Freq_15W.png (82.8 KB, 339 views)
File Type: png LM3886_P2P_THD_vs_PowerFreq_15R22nF.png (119.9 KB, 362 views)
File Type: jpg LM3886_PCB.jpg (348.7 KB, 722 views)
File Type: jpg LM3886_P2P_Full.jpg (419.5 KB, 743 views)
File Type: jpg LM3886_P2P_Detail.jpg (425.5 KB, 662 views)
__________________
21st Century Maida Regulator, Universal Filament Regulator, Damn Good 300B, Novar Spud, and more...
Neurochrome : : Audio - http://www.neurochrome.com/audio - Engineering : : Done : : Right

Last edited by tomchr; 7th March 2014 at 08:01 AM. Reason: Managed attachments.
  Reply With Quote
Old 7th March 2014, 07:49 AM   #2
diyAudio Member
 
Join Date: Aug 2013
Nice findings indeed...one question though, is that PCB for a Myref circuit?

Best regards.

Edit: And now it's gone

Last edited by sghr220; 7th March 2014 at 07:54 AM.
  Reply With Quote
Old 7th March 2014, 07:58 AM   #3
tomchr is offline tomchr  United States
diyAudio Member
 
tomchr's Avatar
 
Join Date: Feb 2009
Location: Greater Seattle Area
I think we may have crossed edits... I realized I uploaded the wrong file for one of the pictures so I went back and changed it.

Nope. Not a Myref. Something better...

~Tom
__________________
21st Century Maida Regulator, Universal Filament Regulator, Damn Good 300B, Novar Spud, and more...
Neurochrome : : Audio - http://www.neurochrome.com/audio - Engineering : : Done : : Right
  Reply With Quote
Old 7th March 2014, 08:09 AM   #4
diyAudio Member
 
Join Date: Aug 2013
Quote:
Originally Posted by tomchr View Post
I think we may have crossed edits... I realized I uploaded the wrong file for one of the pictures so I went back and changed it.

Nope. Not a Myref. Something better...

~Tom
That's intriguing, what is it then?
  Reply With Quote
Old 7th March 2014, 09:11 AM   #5
laplace is offline laplace  Australia
diyAudio Member
 
Join Date: Jan 2014
About those big bucks... you know you can get 10 of 50x50mm PCBs for about $10-$12, right? Probably even cheaper with a bit of googling, but this is a vendor I've used before.
__________________
Open-Source F/Stop Timer
on flickr
  Reply With Quote
Old 7th March 2014, 09:22 AM   #6
AndrewT is online now AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Have you used the black speaker terminal as your Main Audio Ground?

Where is the Power Ground?
Are the two sets of decoupling capacitors and the Output Zobel connected to the Power Ground?

Does every circuit follow strictly the two wire close coupled pairs technique?
__________________
regards Andrew T.
  Reply With Quote
Old 7th March 2014, 01:22 PM   #7
diyAudio Member
 
availlyrics's Avatar
 
Join Date: May 2011
Can you post PCB layout?
__________________
Thanks,
availlyrics
  Reply With Quote
Old 7th March 2014, 02:51 PM   #8
catalin is offline catalin  Romania
diyAudio Member
 
Join Date: Jun 2007
Interesting facts ,we need more pictures and the better design schematic over the My_ref
__________________
"please try to listen to some music through the amplifier instead. Life is so short ..."
  Reply With Quote
Old 7th March 2014, 03:00 PM   #9
diyAudio Member
 
Join Date: Aug 2013
Quote:
Originally Posted by catalin View Post
Interesting facts ,we need more pictures and the better design schematic over the My_ref
That we do need.
  Reply With Quote
Old 7th March 2014, 05:08 PM   #10
tomchr is offline tomchr  United States
diyAudio Member
 
tomchr's Avatar
 
Join Date: Feb 2009
Location: Greater Seattle Area
Quote:
Originally Posted by AndrewT View Post
Have you used the black speaker terminal as your Main Audio Ground?
There are two black terminals. The one with the banana test lead plugged in is the power ground. This goes to the ground star (where all the passives come together). From the ground star runs a wire to the other black terminal (speaker negative) and to the shield of the input RCA connector.

During the test, the differential input on the HP8903A was used. I.e. its ground was lifted. This eliminated the possibility of a ground loop (and did improve the THD measurements).

Quote:
Originally Posted by AndrewT View Post
Are the two sets of decoupling capacitors and the Output Zobel connected to the Power Ground?
The output Zobel is on the speaker terminals. That should be pretty obvious in the picture. I changed the Zobel to 22 nF polypropylene and 15 Ω (just as I have on the PCB) before I took the measurements as the 2.0 Ω + 150 nF shown in the picture caused stability issues.

Quote:
Originally Posted by AndrewT View Post
Does every circuit follow strictly the two wire close coupled pairs technique?
In the P2P setup, the input connection and output connections are twisted pairs. They're short, so they don't have many twists in them. But they are as closely coupled as I could make them. On the PCB, I run the traces pseudo-differentially. I.e. signal forward and return are tightly coupled.

Quote:
Originally Posted by availlyrics View Post
Can you post PCB layout?
The relevant bits attached. I've grayed out the op-amp area (both for clarity and to avoid further confusion/speculation). Keep in mind that this is a prototype circuit slapped together quickly. There are some obvious optimizations to be made on the PCB. I'll leave those as an exercise for the reader...

BTW, I did not use the L||R Zobel network that you see in the PCB. R3 was a wire link (black wire on the board in the picture). Same is true in the P2P circuit.

Quote:
Originally Posted by sghr220 View Post
That we do need.
I think so anyway... I'm afraid I've led the discussion astray by indicating that I'm working on something cool. Sorry about that. I would like to refocus this thread on LM3886 P2P vs PCB rather than letting it spin into a rumor mill. Once I'm ready with my circuit, I'll post a new thread and get a page started on my website. That's probably a month or two out. (Those are Engineering Months, so figure 3-4 calendar months... )

~Tom
Attached Images
File Type: png BPA200C_TestCkt_PCB_LM3886area.png (63.4 KB, 709 views)
__________________
21st Century Maida Regulator, Universal Filament Regulator, Damn Good 300B, Novar Spud, and more...
Neurochrome : : Audio - http://www.neurochrome.com/audio - Engineering : : Done : : Right

Last edited by tomchr; 7th March 2014 at 05:13 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Point 2 Point (no PCB) for TDA7293, TDA7294, TDA7295, TDA7296. danielwritesbac Chip Amps 149 10th July 2013 08:18 PM
Point to point or pcb? Edwar65 Tubes / Valves 111 16th December 2012 09:54 AM
LM3386: point to point or PCB? tbj Chip Amps 21 6th November 2012 01:21 PM
Point-to-Point LM3886 6L6 Chip Amps 34 13th June 2011 04:26 PM
Point to point or use a PCB jarthel Tubes / Valves 13 10th June 2003 09:15 PM


New To Site? Need Help?

All times are GMT. The time now is 06:48 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2