[AndrewFischer]

Has anyone tried using an LM317 as an audio amplifier?

I've seen several schematics but never an actual device.

National has a couple of schematics in the LM317 data sheet where the 317 is a current source and a LM195 or LM395 for an output transistor.

I like the idea. The resulting amp is bullet proof. LM395T is $3.95. LP395 is the same thing in a to92 case and costs about $1.50

Using the to92 case parts we could have a tiny amp. Edit -- This is a class A amplifier. Device is good for 100mA / or 1Amp in the larger package. Amp output power would be less.

It is also possible to put an audio signal into the adjust pin and use the 317 as an audio amplifier.


Brooksie has another idea here:

http://www.tubecad.com/2004/blog0005.htm


All this got me to thinking. The 317 can be had in a variety of small packages. One 317 costs about 80 cents. It is also rugged. Output shorts and overloads won't kill it.


It should be possible to build a small easy to build head amp with these parts.

But how will it sound?

 

[tangent]

Quote:

Originally Posted by AndrewFischer /img/forum/go_quote.gif One 317 costs about 50 cents. It is also bullet proof.

My experience suggests otherwise. After moving the TREAD kits from Fairchid 317s to NatSemi 317s, the number of reported instances of regulator damage have gone way down.

 

[AndrewFischer]

Quote:

Originally Posted by tangent /img/forum/go_quote.gif My experience suggests otherwise. After moving the TREAD kits from Fairchid 317s to NatSemi 317s, the number of reported instances of regulator damage have gone way down.

I'm editing the original post. They will blow up if you exceed voltage limits.


Still the 317 could be used to make a cheap, rugged class A amplifier.

LP395 / LM 317L cheap simple single end class A. Could be that by the time I add enough parts to make a practical amplifier it won't be cheap or simple.

 

[beerguy0]

I've built an amp similar to what you describe, using an LM317 as a current source, with an IRF520. I took a few pics of it.

The construction looks cheesy, but it works. I built this in a weekend with parts readily available to me at work. A good part of my job is building electrical prototypes, so I just built it the way I would any random bit of lab electronics, i.e., Vector board.

This is basically the PRR-Tori amp from Headwize. I've added the Tangent mod using JFETS to bias the opamp (AD826 at the moment) into Class A, and also added a Cmoy crossfeed. It's powered by a +/-15 VDC, 1.5A supply, with separate +5 VDC supply for the fan. The power supply was scavenged from a dead test fixture.

Being Class A, this thing runs pretty hot. I found a fan to be a necessity, but I was limited to the heatsinks at hand.

The caps at the front were coupling caps, but I bypassed them and now run it straight DC.

The opamp section. Caps are polyester film, resistors are 1% metal film. Everything is on headers and sockets.

IRF520 on the left, LM317 on the right. The resistor near the LM317 sets the current.

The JFET mod for Class A operation. Done in typical breadboard style. (I like point to point construction - and wire is optional

)

 

[fortney]

Your amp looks real good.

There was a thread on here or headwize started by a man named Raoul something from about 1994 specifically about an LM317 amp. I tried to reactivate it without success. I don't know how many such amps were built.

F

 

[ericj]

There are schematics and pictures of a class-A headphone amp (well, a current buffer, no gain) over at That Other Forum, which uses only lm317T's and passives.

Also some similar amps by korean headphone DIYers, like this:

http://www.headphoneamp.co.kr/bbs/vi...esc=asc&no=855

 

[AndrewFischer]

Cool or should I say hot? How does it sound? How much output power?

Using a MOSFET could be a lot better sounding than all those parallel bipolar transistors in the LM395.


The LM317 and LM395 have been around forever. I suspected that they were tried for audio and either forgotten or abandoned ages ago. I didn't have much of an internet connection in 1994, so I would have missed anything on Headwize from that long ago.

http://www.tubecad.com/2004/image129.gif

hmmmm.

 

[beerguy0]

Quote:

Originally Posted by AndrewFischer /img/forum/go_quote.gif Cool or should I say hot? How does it sound? How much output power? Using a MOSFET could be a lot better sounding than all those parallel bipolar transistors in the LM395. The LM317 and LM395 have been around forever. I suspected that they were tried for audio and either forgotten or abandoned ages ago. I didn't have much of an internet connection in 1994, so I would have missed anything on Headwize from that long ago. http://www.tubecad.com/2004/image129.gif hmmmm.

Personally, I like the way it sounds, but I've never had another amp to compare it with. It would be interesting to compare it to another amp, as well as to see what others thought of the sound. Who knows, maybe it sounds like fingernails on a chalkboard

Frequency response of the amplifier:

 

[2N3055]

Quote:

Originally Posted by fortney /img/forum/go_quote.gif There was a thread on here or headwize started by a man named Raoul something from about 1994 specifically about an LM317 amp. I tried to reactivate it without success. I don't know how many such amps were built. F

It was on HeadWize in 2004 -- a headphone amp using only LM317s (schematic in message 13):

http://headwize.com/ubb/showpage.php...rch=LM317;amp;

 

[amb]

Since the LM317 was designed to be a voltage regulator, not a linear amp, its distortion characteristics leaves much to be desired. Also, it is internally over-compensated to be stable while driving a large capacitive load, therefore it doesn't have much speed or bandwidth. Using it as an audio amp might be "different" and cute, just don't expect much from it with respect to actual performance.

 

[AndrewFischer]

Quote:

Originally Posted by amb /img/forum/go_quote.gif Since the LM317 was designed to be a voltage regulator, not a linear amp, its distortion characteristics leaves much to be desired.

Just as well. Building one would only distract me from better sounding projects.

 

[T-Rexx]

Andrew Fisher,
I did not want to post anything at the time I read this message until now.
I have owned an LM317K  Amplifier for years back around the 80's its a monster in size because of the heat sink size. It uses the T03 devices along with two buffer devices those devices are 2N5885 it is hooked sort of not standard and I cant explain that here.
It has two VU meters set for 100% at half power for clipping reasons. The LM317 K is a rugged device when hooked up this way each channel is at over 4 amps of current so the total amplifier power is around 8 amps. The Amplifier was built by John Bedini in the late 70's. I have had that amp to the Mac Clinics when they had them at different Hi-Fi stores.
 
At a measly 5 watts into 8 ohms the THD measured .0012% the frequency response was 5Hz to 100Khz flat as a pancake. Square wave response was flat no overshoot at all. The output has a 3 watt 9 ohm resistor in series with a .1 cap across the speaker terminals with an inductor on the cct board . I heard that John Bedini built around 20 of these class A  amps for friends and the people that have them do not give them up. Bedini worked for Acoustic Control when he was starting up his company, Bedini/Strelioff . Bedini in the early days worked for Acoustic Control building Amplifiers the drove below 2 ohms so this was a piece of cake for him. The driving IC is a LM411 Bi-Fet . I have driven all kind of speakers with this amp  nothing sounds like it not even his later amps and I have tried them all and many others. The power supply is two 27000 Uf 35 volt caps for the split supply, plus and minus 15 volts. What I might say is that it is not a lot of power but who need it when it can drive anything I care to hook it to. It even drives my friends Infinity reference system. I will try to post a picture of the AMP if I can find a chip for my camera.
 
Just thought you might want to know they are out there to be found. Oh, one other thing I found this amp in a Junk Store for 15 dollars and I did talk to Bedini to see if he did built it the answer was yes.
T

 

[wakibaki]

Elvee over on diyaudio has a few different configurations of audio amplifiers built with regulator chips:-
 
http://www.diyaudio.com/forums/chip-amps/176052-now-regulator-chip-jlh-amp.html
 
http://www.diyaudio.com/forums/chip-amps/175827-what-happens-when-you-cross-tringlotron-regulator-chip-amplifier.html
 
http://www.diyaudio.com/forums/chip-amps/176512-voltage-regulator-chip-amp-family-welcomes-new-member.html
 
http://www.diyaudio.com/forums/chip-amps/192934-se-class-regulator-chip-amp-madness.html
 
The last of these has less than 1 PPM THD, one is a JLH class A implemented with regulators, so it's a mistake to write these amps off as being incapable of high performance.
 
Elvee is a very talented designer and has at least one discrete amplifier which is completely unconventional in its topology to his credit. He is a very skilled user of LTSpice and has shown simulations of some very useful HV regulators. Not a guy you want to underestimate...
 
http://www.diyaudio.com/forums/solid-state/189599-my-little-cheap-circlophone.html
 
w

 

[JohnBedini1]

MR T
 
There was allot I did not say to you on the phone about LM 317 Amps. The Regulator first appeared in 1970, it was invented by Robert C. Dobkin and Robert J Widlar in 1970 while working for National Semiconductor. This chip was the most under rated chip ever. After my military service in 71 while attending DeVry Institute of Technology
I grounded myself in the Audio field. One of my first jobs after making tin can amplifiers for different people was Acoustic Control. At Acoustic Control I had the opportunity to meet the "Greats" in audio. Stewart Hegeman, Ed Miller, Madan Sharma, Russ Allee, I worked under Madan Sharma The inventor of the Sharma Circuit as an Engineering Tec, 1973. This is where I learned all about Amplifiers and Pre-Amps, Mixers. But I had big ideas after helping with the Model 400 bass amp. I worked with Jack Strelioff, excellent mechanical designer. We started a company together this is where we did the first high powered dual differential  amplifiers connected to a quasi output stage. The amplifiers were known for driving extreme loads like the infinity dual Watkins Woofers and were used in the QLS series speakers.
 
When I broke away from Stelioff is when I started Bedini Electronics and developed the famous 25/25 Class A Amplifier feed forward bias. Where did the LM317 Amp come into play?  Stewart Hageman and I talked about building this as a Class A  low powered amp many times with Saul Marantz  suggesting things that could be done with that regulator chip. One only has to look at it being an output transistor of unique properties to understand what can be done with it.  Yes I did build 20 of  LM317K units with a twist to the output drive by using 2N5885 power followers behind the LM317K TO3 devices. The cure to using the device was in the noble network to make sure the amplifier is loaded correctly at high frequency under any  speaker condition. The Amplifier using these devices uses about 4.2 amps of current on each channel, about 8.4 amps total. The frequency response on my test equipment at the time was 2Hz flat to100Khz the THD was very low about .012 at 1 kHz although some measured much lower then that at the McIntosh Audio Clinics. Yes you had found a rare find in that Junk Store at the time. I still have one here we use everyday nothing sounds like it, it sort of has a mysterious sound to it in that you here things you cant here on other amplifiers. The input amp was a NE5534N at the time later people switched them out for the LF411 and who knows where they all have gone now. Anyway a little history for you and others that did not know
John Bedini

 

[JohnBedini1]

Wakibaki,
Thanks for posting that information. Yes this is one chip not to underestimate at all. Again thanks for posting that information. I will suck it all up and I have seen Elvee amp over at diyaudio. Just thought I would add a little history, and if John Iverson was around today he would have more to add about this chip.
John Bedini