Lm4702

[Mike Gergen]

Anybody investigate this part yet?

 

[Gcollier]

Hmm...this one looks interesting. I ordered a couple of samples. It would give me an excuse to use up all of the thransformers I have that put out too much voltage for a gainclone. If you want to work togeteher on a board design let me know.

G.

 

[Mark St. Denis]

LM4702

Were is everyboby. LM4702 looks like a gold mine. Why are the blogs so quite? Were are the prototyping tools?

Mark
Las Vegas, Nevada

 

[dfdye]

Search = goldmine!

http://www.diyaudio.com/forums/showthread.php?postid=750854#post750854

If you are interested in getting a prototype board to help those investigations along, there is even a group buy for some boards active!

http://www.diyaudio.com/forums/showthread.php?postid=898724#post898724

 

[Mark St. Denis]

LM4702

Hello,

I am a beginner at this so I hope this question helps.
National Semiconductor has a "Build It" page at www.national.com/appinfo/power/webbench/buildit/ with a "contact us" click that is subdivided into sales, technical and others. Does anyone know how user friendly NS "Build It" and "contact technical" is? Perhaps NS is wiiling to give good technical advice to diyaudio members via postings, since they have complete development enviroments for other audio amplifiers, yet not for LM4702.

Mark St. Denis

 

[Mark St. Denis]

LM4702

Hello,

In June 2004 NS made Application Note 1192 by John De Celles and Troy Huebner AN-1192.pdf in reference to LM3886 at Analog University www.national.com/AU/design/0,4706,6_0_,00.html. In this note power dissipation is discussed on page 10 figure 11 by FFT's on Noise Floor Tests with log-scale noise floor and then concluding as quoted "as showen in figure 12 noise level is quite low and influence of the power supply is relativility small" while using note 7.2.3 typically unregulated bi-polar 385VA torodial transformer, therein showing the coupling effect to any 60Hz power line signal or other 60 Hz vibe to be minimum. Were are the FFT Noise Floor Tests for LM4702? In all the LM4702 discussions, I guess it is just assumed, that any power supply will suffice, with no interference to some peoples favorite base frequencies DC to 200Hz.

Mark

 

[Bazukaz]

Still , as i understand , the IC does not protect output transistors at all ?

 

[veteran]

Still , as i understand , the IC does not protect output transistors at all ?

Yes, You are right. There is only thermal shutdown and current protection for LM4702 output.

 

[Mark St. Denis]

LM4702

Hi Bazukas,

I think I see what you mean? (1) In 1993 John De Celles per AN-898.pdf says approximate quote "Overture Serries possesses unique protection system that should save designers costs ... these SPiKE advantages are generally only in high end and discrete amplifiers ... this type of protection usually requires costly zeners or fast recovery Schottky diodes" and then (2) In 1994 John De Celles per AN1192.pdf says approximate quote at section 4.2 "go through worst case power dissipation senarios which implies sinusoids will be used with unregulated power supplies ... and then onto section 4.3 bridged if two individual IC's are used then total power dissipation is divided by each IC" or Pdmax = Vcctot squared vs. Pdmax = Vbtl. I like bridged IC's
anyway, so the point that NS in some manner watered down their own SPike dynamic, that is O.K. by me. Sinusoids in unregulated power supplies, no problem. I hope I got your topic point, at least in partial reply.

Mark

 

[Mark St. Denis]

LM4702 Application Note 1490

Hello,

By this time, everyone can know that Application Note 1490 by National was released in May 2006 showing completed prototype of LM 4702 at www.national.com/an/AN/AN-1490.pdf#page=7 and this is good timing for their Knowledge Base.

Mark

 

[Mark St. Denis]

300 Watts per channel?

Hi,

According to http://search.national.com/iphrase/query?query=LM4702&render=1&categories=1phrase:faq the following makes sense:

100 watts per channel uses 1 output pair.

100-200 watts per channel uses 2 output pairs.

300 watts per channel uses 3 NPN/PNP output pairs.

Has anyone made close to 300 watts per channel on this chip?

National calls this a "common question".


Mark

 

[Mark St. Denis]

Hi,

Can adapt LM4702 to drive MOSFETS and VFETS and TUBES at

www.werple.net.au/~gnb/audio/lm4702.html

and can you show schematics for novice beginners for 3 stage

300 watt per channel using TUBES or hybrid transistors and

tube circuit?

Mark

 

[glennb]

My page!

Hi,
Can adapt LM4702 to drive MOSFETS and VFETS and TUBES at
www.werple.net.au/~gnb/audio/lm4702.html
and can you show schematics for novice beginners for 3 stage
300 watt per channel using TUBES or hybrid transistors and
tube circuit?
Mark

That's my page. I am currently working on a prototype using darlingtons, and will get to tubes eventually (eg. KT88 or 813 in push-pull Class A). Power to the people!

Very soon I will post show a table for easy selection of the mute resistor, and present my test results on heatsinking requirements (none, until +Vcc reaches a certain level).

I'm also working on an amp design which has no coupling capacitors on the input or in the feedback loop, which eliminates two of the impediments to excellent sound quality.

Glenn

 

[Mark St. Denis]

New Chip Driver And Old Nobel Tubes

In 1938 RCA introduced RCA 813

http://faculty.frostburg.edu/phys/latta/ee/wing813amp/813.html

and in 1957 Marconi Osram Valve Company of Hammersmith,

London introduced KT 88 www.vacuumtube.com/issue6.htm

and RCA 813 pentode requires less than 1 watt to drive it to full

power and PP class AB amp can be used for better than 200

watts of power http://homepages.ihug.co.nz/~rwillis/alex

-- so where is 2006 interface between LM4702 and these nobel

old tubes, since tubes dissipate power at 200-300 watts without

semiconductor heat sinks in output stages and do not require

protective output thermal cut off circuitry which is internal to

LM4702 semiconductor driver?

Mark

 

[glennb]

Re: My page!

...
Very soon I will post show a table for easy selection of the mute resistor, and present my test results on heatsinking requirements (none, until +Vcc reaches a certain level).
....

OK, I've described a graphical method for easy selection of the mute resistor at http://home.pacific.net.au/~gnb/audio/lm4702.html#muteresistor
and made a few other updates to my LM4702 page.

The heatsinking test results will appear in the next few days.

Glenn.

 

[glennb]

Re: New Chip Driver And Old Nobel Tubes

... RCA 813 pentode requires less than 1 watt to drive it to full
power and PP class AB amp can be used for better than 200
watts of power ... -- so where is 2006 interface between LM4702 and these nobel old tubes, since tubes dissipate power at 200-300 watts without semiconductor heat sinks in output stages and do not require protective output thermal cut off circuitry which is internal to LM4702 semiconductor driver? ...

Hold your horses! I'm still working on it! I have a neat phase-splitter design which has gain-trim (for lowest distortion) and no coupling capacitors anywhere, including on the control grids of push-pull tubes.

 

[Mark St. Denis]

Hello,

Since triple Darlingtons seem to be a topic for LM4702, here is the

orginal patent where Dr. Darington includes triple transistors at

figures 8 and figure 9:

http://andros.eecs.berkeley.edu/~hodges/DarlingtonCircuit.pdf

Just for history fans, no big rush.

Mark

 

[dshortt9]

Thanks, Mark - very interesting reading.

I think a problem with this chip is that people with enough savvy to design the Vbe, driver, output stage would rather design the front end also.
And people without the design knowledge may find building an amp with this chip intimidating.

 

[jackinnj]

Thanks, Mark - very interesting reading.

I think a problem with this chip is that people with enough savvy to design the Vbe, driver, output stage would rather design the front end also.
And people without the design knowledge may find building an amp with this chip intimidating.

Perhaps, but the active component matching you can do on one die is incredible -- that's why the claimed THD of <0.0006% is pretty stunning.

Does it matter in the listening? I plan a rigorous test with lots of listerners and lots of music (I will have to hid all that Baroque stuff) when the boards get back. I have no interest in flogging something which isn't demonstrably better.

 

[glennb]

Excuses excuses

Thanks, Mark - very interesting reading.
I think a problem with this chip is that people with enough savvy to design the Vbe, driver, output stage would rather design the front end also.
And people without the design knowledge may find building an amp with this chip intimidating.

You are entitled to this opinion, and I respect it, but the reverse situation is also quite likely. I've designed and constructed all stages of discrete audio power amplifiers over the past 20 years and done a great deal of learning and research. I relish this new possibility of having a high performance front-end in one chip, leaving me to concentrate on designing a high performance output stage to follow it.

People who have built a chip amp may actually find pleasure in the challenge of playing with a new chip which requires some out-board power components, rather than finding it intimidating. It certainly increases the output power possibilities! Hooking up a SAP-15N/P or SAP-16N/P pair of darlingtons is actually quite easy. I make no apologies for the blunders of people who don't actually know what they are doing, but learning by practical experience is a great tool.