The REAL John Linsley Hood Headphone Amplifier

There is a long thread for the so-called JLH Headphone Amp.
Its popularity has of course due to do with the name of Linsley Hood.
But it is also due to the availability of low-cost kits offered on ebay, etc.

In actual fact, this is not really a headphone amp designed by John.
Rather it is a scaled down version of the legendary JLH 10W Class A power amp.

John has published a few circuits specifically meant for headphone use.
A Paul Kemble web page - John Linsley Hood preamp designs.

In particular, the JLH 1985 designed specifically for headphone use is very simple to build.
http://www.angelfire.com/sd/paulkemble/jlhhdphaud.gif
Even in its original form, it consists merely of 3 pairs of bipolar transsitors and a handful of resistors and caps.
And this can be further simplified if one would do away with the auto-biasing circuit.
The topolgy is very similar to the F5 headamp, namely a CFA with negative loop feedback.

We built a modified version, just out of curiosity.
And here is a teaser first.

:)


Patrick

.
 

Attachments

  • 180330 JLH1985 Cased.JPG
    180330 JLH1985 Cased.JPG
    318.4 KB · Views: 4,920
Last edited:
As usual, we shall only publish full article after beta testing and auditioning.

But as a tribute to Mr. Linsley Hood, we shall, as an exception, also publish a public version of the Gerber files.
You can then order your own PCB and make this a low-cost project.
Essential is that you are able to handle SMD soldering down to 0805 and SOT23.
The public PCB will handle both the original and the modified versions.

Will take a while, but stay tuned.


Patrick
 
Just realised they have deleted the Schiit Magni thread from earlier, around February.
There we posted a 8 transistor 16 resistor CFA headphone amplifier circuit.
Though the JLH is even simpler, it is more load dependent.

So here is the Spice file for the CFA circuit again, just for comparison.
2SC2240/2SA970 can of course be replaced by their SMD equivalent of 2SC3324/2SA1312.


Patrick
 

Attachments

  • BJT CFA HPA V1P THD.asc
    7.3 KB · Views: 519
Perhaps a word about the auto-bias.

In circuitry terms it works very well. One can easily set the bias by changing the source resistor, e.g. to 5.6R for 120mA bias.
And if you run an AC analysis of the original JLH circuit in post #4, you will see that the -3dB bandwidth is between 5Hz and 75kHz.
The lower end is limited by the caps in the feedback network, thus masking any effects of the auto-bias.
The auto-bias will need trimming for DC (as in the Spice file or in original schematics) if the Vbe of the PNP is not 100% identical to the NPN.

We had a short discussion here some time ago :
Auto-Biasing Circuits for Complementary Followers

We remove the auto-biasing (as well as the caps in the feedback network) completely in our prototypes.
With good heatsinking (using Breeze Audio 2107 case) at 31°C working temperature after an hour, DC drift was stable to +/-10mV over time.
And bias was stable at 130mA with no thermal run-away.
So in practice there is really no need for the autobiasing, as it does not eliminate DC offset trimming.
It only helps to prevent thermal run-away, which was solved by good heatsinking and the large-ish source resistor.

This removes a further 2 transistors, 4 resistors, and 2 caps from the BoM.
So the resulting circuit contains only 4 transistors, 9 resistors, and 2 caps.
Only the F5-HA has less.


Patrick

.
 

Attachments

  • JLH 1985 HPA Original No Auto-Bias THD.asc
    7.4 KB · Views: 494
You could always replace the output transistor "BE" resistor with a diode-resistor, making it a faux current mirror. Then glue the (both) diodes to the heatsink. Or make the diodes a diode-connected bjt; same type as output bjts and bolt them to the metal. A KISS approach.
 
One of the original links in post #1 has been taken off, for whatever reasons.
So I am attaching a copy of that pdf here.
It has a very comprehensive description of the original circuit.
Worth studying in detail.

Our article later will only concentrate on the modifications we took, as well as some comparisons to e.g. the F5-HA.


Patrick

.
 

Attachments

  • JLH 1985 headphone.pdf
    57.2 KB · Views: 1,564
I am very interested to see how this turns out.

Also, I noticed the scaled down JLH 1969 amps and I got to ask, is there going to be a difference between those and this? (given same PSUs)

I for the life of me cannot calculate the power output of them all I know is their bias current is 220mA and supply voltage is +-12vdc.
Would that be 1.5W into 8 ohms and 375mW into 32ohms?
 
Independent Audition

Here is the quote from the independent auditioner :

"The JLH is an amplifier that does sound very good on itself.
It does everything right...
It has pretty much all the detail as the other XEN headphone amps (Pono as source).
No glare whatsoever and a good and stable sound stage.
But it does so in a more polite way than the F5-HA or the UTHAiM.
I especially like the neutral and natural character.
There is no hint of digital artefacts or the hardness I know from other amps.

A nice tribute to John Lindsey Hood."

But then we would not expect anything else from Mr. Linsley Hood.

As already mentioned, the F5-HA is the closest in topology to the JLH but using FETs instead.
In the article to be published (probably next week), I shall try to explain this neutral character.
Also explained in detail would be its behaviour with low and high impedance loads.
And how to get the best out of this very simple deisgn.

Stay tuned.


Patrick
 
Just want to add that we only built the modified version without C3, C4 (plus a few more changes).
So the auditioning was done using that.
Headphone used for that audition was a Focal Elear (100ohm).
But we have also tested before for days with a modified Superlux HD662 Evo 2 (32 ohm), and a Sennheiser HD250 Linear II (300 ohm).
So we know for sure it works with loads from 30R to 300R.

PCB will accomodate both versions.


Patrick