JLH Headphone Amp

Great to see there are stil people reading this topic.

Last week I received my ebay kit and want start building it.

I'm curious, is there a "shortlist" from all the modifications needed to be made with the JHL ebay kit.
I've read almost the whole thread, but the modifications are spread over different messages and are not always the same.

Can anyone help?
 
Hi Scotsky,

Follow milescampbell thread, he made most of modifications and carried out supporting signal testing. I started from a built pcb so i had to strip off/replace components but it should be easier for you to not fit/change components.

I think the most important change is DC offset network on the input transistor. I kept the 10K adjustment pot to allow the DC output offset to be trimmed to within +/- 5mV. I thought the +/- 100mV maximum offset was too high for my 32 ohm headphones so replaced the 3V3 zener with 4 x 1N4148 diodes, 220K replaced by 68K, 100uF replaced with 180uF low ESR and the 1K (4k7 in thread) with 10K. These changes are important to remove any ripple/noise (100Hz) from the input dc bias stage otherwise you will definitely hear mains hum in your headphones. Also replace the 220pF ceramic with Wima or similar MKT. I replaced the 470nF input capacitor with a slightly higher 680nF though 1uF would probably be better. I did not replace the PSU filter capacitors (my board had 2200uF types fitted) because i used a seperate PSU with 2x1500uF/rail low ESR already fitted. The feedback capacitor change from 100uF to 220uF/470uF will probably be less noticeable but still worthwhile if you want to squeeze extra low Hz out of your bass-limited headphones. You must at least turn the original 100uF around otherwise it will be reverse-polarised.

I think its important to remember that the chinese JHL amp was quite a significant departure from the original JHL design based purely on commercial i.e. cost and not high fidelity sonic reasons. The modified headphone amplifier will sound more like the original JHL design.
 
Hi Rob,

The original JLH design used two silicon diodes (1N4148 or equivalent) to give a fixed
DC bias to the input transistor, giving a fixed DC output offset which may be too high for low impedance headphones. The modified or chinese "JHL" design incorporated a zener diode and a 10K potentiometer to allow the output DC offset to be nulled to 0V (typically +/-5mV). This was a bad design choice, i replaced the zener with 4 x 1N4148 silicon diodes and changed other components in the DC bias network to ensure that no mains hum is coupled to the headphones. Refer to the original JLH design to get the correct polarity for the diodes (opposite to the chinese JHL version). You need to set the DC offset over several hours as the output dc offset will be temperature dependant. Circuit 1 can sound very good if you make the proper modifications discussed in this thread. I dont have any experience with circuit 2 and i dont know if it is available as a kit, like circuit 1. Hope that helps.
 
That link is correct - JHL original schematic and design explanation. Only the transistors have been substituted for more modern types - the original design is from 1979 - but that should make no noticeable difference to the sound quality. Like most people, i chose the kit because it was very cheap and i could get amp built quickly. The JLH Class A design should be able to drive any headphone with ease. I also use the headphone amp as a pre-amp/buffer between my Sony Bravia TV nicam stereo output and my usual hi-fi setup - JBL Northridge E60/Harrison X-150 Mosfet amp. The sound quality was so good when watching Olympics, you could really feel and hear the whole atmosphere in the venues and it was obvious the normal TV speaker output is heavily filtered/band-limited in comparison. Good luck with your build whatever way you decide to do it and post some pics when completed.
 
It is very confusing, i know. C23 and C24 are the low frequency feedback capacitor designators on the EBAY chinese version or "JHL" headphone amp. Pretty cynical alteration of JLH to avoid possible patent infringement etc. In the original JLH design, the capacitor designator is C4 (100uF) - note the voltage polarity (+ve to GND), this is correct. To avoid any further confusion, you could post your final FidoCAD schematic here and the good people on this forum can check it over before you build?
 
MilesCampbell explains on post #133

correct polarity on the picture
 

Attachments

  • board-a.jpg
    board-a.jpg
    491.6 KB · Views: 1,201
DC-offset 150mV each channel after 30 min. open case

Hi lung,
thank you for your answer. I asked because at the time I was contemplating should the trimpots stay or should they go. I decided to keep them for the same reason as roclite, I did not want to damage small IEMs I sometimes use.

I was wondering whether two common big heat sinks for both sides of the board would lower the offset to acceptable level in which case I would remove the trimpots. If that does not turn out as expected I will desolder them, measure their resistance and replace them with a resistor of measured value.

One more thing puzzles me, although I am using over sized EI transformer (60VA) it does get very hot, what is your experience?

Price aside, this Chinese edition is not the kindest tribute to Mr. Hood and if Chiara boards were available I would have taken that route. Perhaps someone who used siliconray boards could chime in?

Kind regards
Marko
 
For what its worth, here are some voltage readings from my Headphone amp for output dc offset, no input signal or load R. Initially, i set up the trimpots on the first day of use but have not adjusted since. They were set to less than +/- 5mV initially:

CH1 CH2
Switch-on 120mV 110mV
After 1 min 60 54
After 30 min -11 -8
After 60 min -14 -11
After 120 min -12 -10

First thing to note, it might be advisable to switch amp on with headphones disconnected to reduce risk of damage, the dc offset voltage then falls rapidly.
I could probably trim again but these readings are pretty acceptable, compared to the non-trimmable circuit which has typically 100 - 150mV offset. I dont think i would spent any time trying to improve on this. The 4 series 1N4148 silicon diodes used to provide the 10K trimpot Vref of -2.25V (nominal) are mounted inside case on the pcb not on the output transistor heatsinks (two seperate heatsinks) so you wont get ideal temperature tracking anyway. Someone with more time could probably optimise the circuit and improve. If its a worry, you could always check the output dc voltage offset after being driven for several hours into headphones. My guess is that it will be still be much better than the non-trim version.
 
Hi lung,
thank you for your answer. I asked because at the time I was contemplating should the trimpots stay or should they go. I decided to keep them for the same reason as roclite, I did not want to damage small IEMs I sometimes use.

I was wondering whether two common big heat sinks for both sides of the board would lower the offset to acceptable level in which case I would remove the trimpots. If that does not turn out as expected I will desolder them, measure their resistance and replace them with a resistor of measured value.

One more thing puzzles me, although I am using over sized EI transformer (60VA) it does get very hot, what is your experience?

Price aside, this Chinese edition is not the kindest tribute to Mr. Hood and if Chiara boards were available I would have taken that route. Perhaps someone who used siliconray boards could chime in?

Kind regards
Marko

Hi muksan,

i bought a few "boards only" from this seller.

JHL Class A headphone amplifier pre-amp Preamplifier PCB | eBay

Using 50 VA EL transformer got warm, but not hot after 60 minutes
testing.

If i have time i will make a second one with bigger heatsinks and use trim pot.

I am using small cheap IEMs only for testing. To listen music i take my beyer DT 300 Ohm headphone.

Kind regards

lung
 
Hi guys, sorry for not replying sooner, been away for the weekend.

@rocklite

My measurements are consistent with yours, only had to account for 20mV offset change with the lid of the case closed (trimming was obviously done with lid off:D). Would using LEDs as a voltage reference bring any benefits in this case?

@lung

Thank you for checking transformer for me, I joined two secondaries to create center tap, perhaps small voltage difference between the two is causing additional heat.

I was referring to a diyaudio member siliconray and his dual mono design of this amp, considering the very attractive price of his boards it is surprising there is lack of builders.