JLH Headphone Amp

OOPS!

Hope
 

Attachments

  • P1020585.JPG
    P1020585.JPG
    578.1 KB · Views: 2,239
Sober rantings

Sorry about that folks managed to get some sleep.
I bought one of these on fleabay (see pic) and wasn't too impressed with the initial sound It turned out they had used horrid ceramic caps on the input,
they were first to go, this brought immediate gains in fidelity.
I then decide to pull all the ceramic bypass caps and replaced with wimas,
more improvement.
Onto the electros, feedback cap 220uF replaced with quality low ESR Sanyo cap
The four smoothing caps replaced with Nichicon low ESR
I felt there was too much gain in the amp to drive HD565 ovation cans, so the 470 ohm was raised to 2K2 .
This thing now sounds fantastic very quiet clean crisp highs and great bass.
I felt the size of the thing was overkill to drive a pair of headphones so decided
to build one on stripboard the new one is less than half size including the 317 337 power supply.
R11 is now 12 ohm R5 is 4K7 C5 R8 C8 and R12 have been omitted. C4 is now 470uF
LM336 2.5v voltage ref used for DC offset.
I cant believe how good this thing sounds for only having 5 transistors
Its dead silent no hum no hiss just great clean music.
And no it doesn't oscillate.
 

Attachments

  • P1020589.JPG
    P1020589.JPG
    790 KB · Views: 2,207
Last edited:
.... I am also interested in this circuit.

Do you have an update?

* If you mean modifications...No. I wanted to make a new board but various unrelated problems have delayed everything

What ended up as your final configuration?

* The only main difference was to use an LED instead of the two diodes D3 and D4

Does it work well as a preamp AND headphone amp, or should it be optimized for one or the other?

* Works very well either way. No modifications required.

What PS?

* I used an Lm7812 and LM7912

Other observations or comparisons?

* Sounds better than some other preamps I have like Rotel's. I don't have a megabuck preamp to compare it with. Basically with preamp in or out of the signal path to a power amp , I prefer using the preamp.

* Concluding. If adding the preamp doesn't degrade the sound ( often improves the sound in subtle ways ) and your ears like it ........it is YOUR own mega-buck preamp ! Build it. It's very musical and no loss at either extremes of the frequency range.
Note I think someone complained about offset drift in the split supply circuit. I never had that problem. Drift it might but by very little. Depending on the dissipation of the output transistors you need to use a suitable heat sink to keep the temperature rise to acceptable levels.


* cheers.

Thanks!
 
Last edited:
Just located my split supply board. It has obne preset on it for output offset correction. Will post the circuit as soon as I get some time. The past and next few days are really too busy to do that.
Gives me a chance to compare the 'preamp' function with a simple 6922 valve preamp stage !
Cheers.
 
If you really want to experiment , you can try using Jap transistor equivalents !

But with higher gain and wider bandwidth transistors you might be inviting oscillations. So IMHO a scope is a must when doing such experiments. I'm not sure if you will get accurate Spice models on all the Jap small signal transistors to run a simulation first.

I see some models every now and then. Maybe we should have a repository of working Spice models of all the devices that we generally get/use. maybe already there. I haven't looked yet. Time is in short supply.
 
Last edited:
Great Job

Ashok, Please allow me to congratulate you on a major achievement, in my humble opinion. I am an older JLH fan (I actually met and spoke to JLH) All of my home built amps and test gear is JLH based. I have been running a JLH MOSFET 100W amp for more than 20 years, still sounds good on Quad ELS.:smash:

My JLH headphone amp resembles your's somewhat and sounds good on AKG K340 Elecrtostatic headphones. The assembled board was purchased from Hong Kong for about US$16.

My previous JLH uses an IC TI TL072driver for a complimentary output class A TIP41/TIP42 pair.

Peter
 

Attachments

  • JLHphonCase0WEB.jpg
    JLHphonCase0WEB.jpg
    91.1 KB · Views: 2,108
  • JHLH'phCHINA copy.jpg
    JHLH'phCHINA copy.jpg
    34.5 KB · Views: 2,134
Sober Rantings:
I think the feedback cap was supposed to be 100 uF. What happens when you double the value?
Hotiron, Hope you replaced/didn't use the input volume pot during your listening assessment.
The four smoothing caps sit 'behind' the 78xx/317 reg. ICs so the amp does not "see" them, or the rectifiers or the transformer.
 
  • Like
Reactions: 1 user
If you double the 100uF feedback cap the LF roll off of the NFB loop will go down. F-3db = 1/( 2*pi*470*100E-6) = 3.39 Hz
That's pretty much out of the audible range. So you might not gain anything.
Your input capacitor is 1uF with an input impedance of about 220K on the amp. Hence F-3db here is = 0.72Hz ! However note that you get phase shift even if the db levels are low. So raising the F-3dB frequency will increase phase shift at low frequencies. If it is large it might affect how transients in the bass can sound.
You could get F-3db at 7.2Hz by using a 0.1uF cap instead of the 1uF cap. Chances are you can get a much better sounding input cap at 0.1uF or 0.22uF !
Cheers.
 
More sober rantings

Synchron Hi.
The NFB cap that came on my board was a very cheap and nasty 220uF and still is 220uF but its a Quality Sanyo low ESR Job.
We also have four smoothing caps after the regs which sit "between the regs and the amp",
Its quite clear in the pic which ones I've replaced.
Had I had eight of them to hand I would have replaced all eight!
Because the ones infront of the regs are just as important, any spurious crud
infront of the regs will hinder their operation and flood the amp with noise! Its Class A remember.
So this will of course happen, Once my stock is replenished.

Ashok thanks for the update But surely the NFB cap introduces phase shift also?
:rolleyes:
 
If anyone is interested in the original circuit board layouts devised by JLH, I have a number of these for Headphone amplifiers and for the symmetrical FET buffer and discrete op.amp, and the Class A headphone amps that appear on Geoff Moss' site. I also have copies of the accompanying articles.

These include "High Quality Headphone Amp" from Hi-Fi News Jan 1979 - this being a discrete component design working of +/- 12 volt rails. The design appears on Geoff Moss' site with "A Change of Tone" d.c. offset options - this was to accomodate purists who might wish to build the design without tone controls.

The original design incorporated an improved "Liniac" discrete component circuit with tone controls where the capacitors in the arrangement could be switched out of circuit to produce a flat response. The tone section is a single ended circuit producing "neglible distortion".
JLH in his original "Liniac" (Wireless World Sept 1971) produced a graph of the distortion in an open loop (no nfb) with a note that this was "almost all second harmonic". It produced 0.3% at 1 volt output and open loop gain was 2500 times falling at 6db/octave beyond several khz. Open loop gain was still 1000 times or more beyond 20khz.

I have no experience with this design although I expect it would be similar in character to the 15 watt Class A design published in Electronics World in 1996 which I built at the time and am still using.

The next design presented appeared in Wireless World issue of November 1982 as part of series of articles describing a "Modular Preamplifier". I built all the modules in the series barring the moving coil preamp. I still have this kit but I have a separate listening space and don't have a use for headphones nowadays. The headphone board uses an LF353/TL072 which might be a ''turn- off" for some. At the time of building I thought this had endearing qualities and did not feel the design was compromised in having a dual op.amp IC at it's heart.

The last design for a full complementary symmetry discrete component headphone amp appeared in ETI in June 1984. JLH described this as having "an excellent sound quality" and the circuitry was adapted by him in 1985 for moving coil cartridge stages. I have no experience of listening to either.

All of these designs operate from dual supplies. To the best of my knowledge there is only one example of a single rail headphone amp. This used a 15 volt supply and originated in an article in Hi-Fi News in 1978. I don't have this anymore but it was entitled " A Reference Standard Preamplifier" or similar. It was dedicated for vinyl sources only but it had the capacity to drive high impedance phones. I built this and found it quite enchanting. From memory the circuit was novel in that the volume control manipulated the negative feedback loop.

The accompanying articles discuss JLH's design reasons/decisions and each of these is an informative read. I was completely happy with the two designs I built and expect that builders would get satisfaction out of any JLH headphone amp project.

I am happy to scan /email material for anyone interested or to answer questions about the content of the material.
 
Symmetry

JLH also designed symmetrical amplifier circuits which he sugested might replace his previous non-symmetric designs because his listening panel thought that the fully symmetric amps sounded better.

Thanks to all who value and are willing to share the audio designs of John Linsley-Hood. He made a huge contribution to sensible audio design.

If there is a universal product that I would like to see in our discussions, it would be his power supply design for the 80-Watt MOSFET amp published in Electronics Today International, May 1989. The design has practically disappeared, I imagine, because of its apparent complexity, but it's usefulness as a component in the audio chain is invaluable.
Peter.
 
Hi Peter,
Your post made me look at the JLH page again. I decided to try out the symmetric version of the headphone amp . I have not seen anyone report about it's performance. Has ANYONE made one ?
Just got half the pcb layout done. I'm using up very precious time doing this. Hope it works out well!
 
Hi Peter,
Boards for the 80 watt Mosfet amplifier are available from APJ audio Home.
In shape the P.S.U. board looks like the one from the ETI article.

There was a later article in Electronics World + Wireless World in June 1993 that gave a simpler P.S.U. for use with a reworked version of the Mosfet amplifier. That P.S.U. like all the earlier J.L.H. examples used re-entrant short circuit protection. This worked by progressively "stealing" some of the current driving the pass transistor(or Mosfet as the case may be) and providing an absolute current limit of 3.8 amps for two channels.

This ought to be adequate for ordinary listening purposes however there is a lot of highly compressed material on CD which has little dynamic range and which has been mastered to be experienced at high sound pressure levels.

The amount of current might be increased by using seperate supply boards for each channel with should be safe enough - you would have to ensure correct earthing.

The experimentally minded might like to model some delaying RC network around the current monitor transistor so that will allow high peak currents for several microseconds without exceeding the safe operating area limits of the pass transistor.
 
I just simmed the symmetrical amp circuit. I didn't get results as good as the single supply and split supply non-symmetrical circuit. So I'm stopping work on the board. No point in breadboarding it if it performs much worse than the earlier circuits.
Cheers.
 
I will be trying out the split supply headphone amp with Jap transistors. These generally have higher gain, higher Ft and more linear in the operating range. Hopefully it will work with just a device swap and does not require modified compensation for stability. Want to see if makes any audible difference.

Has anyone already done this ?
 
Member
Joined 2004
Paid Member
I just simmed the symmetrical amp circuit. I didn't get results as good as the single supply and split supply non-symmetrical circuit. So I'm stopping work on the board. No point in breadboarding it if it performs much worse than the earlier circuits.
Cheers.
JLH appeared to me in a dream (he looks a bit like John Cleese actually) and said he's not resting very well wondering if there might be people who will trust a computer simulation more than his own work....
 
JLH appeared to me in a dream .......he's not resting very well wondering if there might be people who will trust a computer simulation more than his own work....
:D
I'll wait for JLH to appear in my dream !;)
We'll discuss the circuit in question !
...........you HAVE heard of doctors from the other side appearing to some folk and suggesting treatment for medical problems ? :)
 
Hello all

It has been on my to do list for a while to get photocopy the JLH original articles from the relevant magazines in the British Library for the following designs

Design 1
- High Quality Headphone Amp
Hi-Fi News & Record Review
Jan ‘79 p81
Source The Class-A Amplifier Site - JLH Headphone Amplifiers

Design 2
- Electronics Digest, Winter 1985/86
Source The Class-A Amplifier Site - JLH Headphone Amplifiers
OR
ETI in June 1984
http://www.diyaudio.com/forums/headphones/159202-jlh-headphone-amp-10.html#post2461964

Also JLH did some work on stabilised power supplies
- Power Supplies for Electronic Equipment Part 2 – Shunt and Series Designs
ETI
Apr ’94 p21
Source The Class-A Amplifier Site - JLH Headphone Amplifiers
and here
John Linsley Hood ripple Eater - Rock Grotto Charity Edition

By the end of the year I would like to build both the original JLH headphone amplifiers on PCB's with a solid PSU design.

- Questions -
? Would this be of interest to anyone else ?
? Does anyone have a copy of the required source material ?

Thank you
Miles