7 Easy Pieces - Page 6

gengcard · 24th November 2004, 05:42 AM

I decided to use MJE3055T of On Semiconductor. Is it suitable for the 100mA constant current source?

Nelson Pass · 24th November 2004, 07:52 PM

I should imagine, with a heat sink.

jh6you · 25th November 2004, 01:24 PM

For what is the CCS?

Nelson Pass · 14th February 2006, 11:34 PM

In the spirit of improving even chip amplifiers, I offer this
Eighth Easy Piece for you LM3886 users:

I have noticed that in the distortion curves, there is a little
glitch down at about 1 watt, and that this is a voltage related
glitch, not a current related one (it occurs at twice the wattage
with 4 ohms) and it is consistent.

Nelson Pass · 14th February 2006, 11:59 PM

Studying the circuit (or in this case a small section) we can
see the probable cause.

Neglecting the input buffer transistors, the amplifier has two
differential pairs taking feedback. The first, Q1 and Q2 are the
actual signal bearing transistors. Q3 and Q4 form a secondary
pair which does not process input signal (their + input is
grounded) but takes unity gain signal from the output, limited
by a pair of diodes.

When the current is run through the mute pin, Q1,2 are biased
up and Q3,4 are starved for current, so Q1,2 dominate the
performance and everything is normal. When you mute the
amplifier, the current is shifted to Q3,4 and the amplifier goes
into unity gain with a shorted input.

The transition is dependent on the amount of current going
through the mute pin. The spec sheet suggests 500 uA as the
appropriate value, and apparently the distortion curve above
reflects that. But another curve shows the significant sharing
of current between these diff pairs at half this value, and I
suggest that the glitch at 1 watt is an artifact of the incomplete
shutoff of Q3,4 and the diode nonlinearities in that unity gain
loop.

I recommend that you explore higher mute currents at greater
than 1 mA to provide more shutoff to Q3,4 during operation so
as to lower the distortion in the 1 watt region.

The LM3875 does not have this circuit, but the LM3876 does,
and is interesting to note that the mfr's curves do not show this
artifact, although it may be partially obscured by the higher
general THD level in this chip at 1 watt.

pinkmouse · 15th February 2006, 12:39 AM

Excellent work, as always Nelson. Thanks. It just shows what you can achieve by reading datasheets properly, a lesson to us all.

metalman · 15th February 2006, 03:46 AM

Thanks Nelson. I noticed the same effect by accident a while back, but I couldn't figure out why it was happening. I find tha 1mA on the mute pin isn't quite enough, and have been using about 3mA finding that it consistently provides a benefit. Now that it has been explained I feel better now that I know I wasn't just deluding myself.

Cheers, Terry

Buhl · 15th February 2006, 09:44 AM

Very interesting.....
Not beeing a shark at math, I tried to calculate the current going through the mute resistor. The mute resistor is connected to the negative rail and output, and I thus count the output as a 0 volt reference - thus, the voltage over the resistor is equal to the negative rail voltage - in my case -33.2 volts, and using a 10K mute resistor - this gives 0,00332 amps.

I notice that other layouts uses higher values like 38K - which gives current values under 1 mA.

Cheers !
Buhl

peranders · 21st November 2006, 03:09 PM

Quote:

Originally posted by Nelson Pass
In the spirit of improving even chip amplifiers, I offer this
Eighth Easy Piece for you LM3886 users:

I have noticed that in the distortion curves, there is a little
glitch down at about 1 watt, and that this is a voltage related
glitch, not a current related one (it occurs at twice the wattage
with 4 ohms) and it is consistent.

Excuse me Nelson, but isn't it a change of measurement range of the test equipment?

Nelson Pass · 21st November 2006, 06:33 PM

That's a possibility, although you note that the glitch occurs
at a specific output voltage (1 watt into 8 ohms) and my AP
doesn't do that. A quick thumbnail calculation shows that you
expect the "mute" differential pair to switch in approximately
this region if its bias has not been shut off.

Of course you can always test the proposition by measuring
an actual unit and varying the mute current...

15th February 2006, 12:39 AM	#56
pinkmouse diyAudio Moderator Join Date: Apr 2002 Location: Chatham, England	Excellent work, as always Nelson. Thanks. It just shows what you can achieve by reading datasheets properly, a lesson to us all. __________________ Al I conceive of nothing, in religion, science or philosophy, that is more than the proper thing to wear, for a while. Charles Fort

15th February 2006, 09:44 AM	#58
Buhl diyAudio Member Join Date: May 2002 Location: Aalborg, Denmark	Very interesting..... Not beeing a shark at math, I tried to calculate the current going through the mute resistor. The mute resistor is connected to the negative rail and output, and I thus count the output as a 0 volt reference - thus, the voltage over the resistor is equal to the negative rail voltage - in my case -33.2 volts, and using a 10K mute resistor - this gives 0,00332 amps. I notice that other layouts uses higher values like 38K - which gives current values under 1 mA. Cheers ! Buhl __________________ "Calculations? We don't need no stinkin calculations! We just make them really big!" Quote: Nelson Pass

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Missing links for 7 Easy Pieces	antomas	Pass Labs	10	29th January 2009 09:48 PM
WTB - 16 pieces of TDA1541A	ErikdeBest	Swap Meet	2	2nd June 2006 11:50 AM
FS: Tubes--84 pieces, NOS	gengis	Swap Meet	0	9th January 2006 01:27 AM
WTB: Three pieces of MC74VHX74	hifi	Swap Meet	1	31st August 2003 08:51 PM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	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

24th November 2004, 05:42 AM	#51
gengcard diyAudio Member Join Date: Jun 2004 Location: Bangkok	I decided to use MJE3055T of On Semiconductor. Is it suitable for the 100mA constant current source?

24th November 2004, 07:52 PM	#52
Nelson Pass The one and only Join Date: Mar 2001	I should imagine, with a heat sink.

25th November 2004, 01:24 PM	#53
jh6you Account Disabled Join Date: Aug 2002 Location: Near to the Pacific Ocean	For what is the CCS?

15th February 2006, 03:46 AM	#57
metalman diyAudio Member Join Date: Jun 2003 Location: Calgary, Alberta	Thanks Nelson. I noticed the same effect by accident a while back, but I couldn't figure out why it was happening. I find tha 1mA on the mute pin isn't quite enough, and have been using about 3mA finding that it consistently provides a benefit. Now that it has been explained I feel better now that I know I wasn't just deluding myself. Cheers, Terry

21st November 2006, 06:33 PM	#60
Nelson Pass The one and only Join Date: Mar 2001	That's a possibility, although you note that the glitch occurs at a specific output voltage (1 watt into 8 ohms) and my AP doesn't do that. A quick thumbnail calculation shows that you expect the "mute" differential pair to switch in approximately this region if its bias has not been shut off. Of course you can always test the proposition by measuring an actual unit and varying the mute current...