Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Headphone Systems Everything to do with Headphones

JLH Headphone Amp
JLH Headphone Amp
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 12th December 2018, 06:36 AM   #801
Hugh Jazz is offline Hugh Jazz  Taiwan
Banned
 
Join Date: Feb 2013
Quote:
Originally Posted by Hugh Jazz View Post
Thanks to all who have contributed to this thread - finished reading it at last :-)

Summary of the changes to pre-built Glozone (one channel is given but of course, both channels were changed):

reduced gain to 4x by replacing R17 470R with 1.5K
replaced ceramic capacitor C11 330p with Wima poly 220p to extend the bandwidth to about 200KHz
replaced ceramic C15 220p with Wima poly 100p for an input low pass below ~300KHz
replaced C23 with 100uF SilmicII with "reversed" polarity to correct silkscreen/schematic error
replaced C1 with 220uF SilmicII
replaced C7 with 100uF SilmicII
replaced R1 5R1 with 10R to lower current through the output transistors to 60mA from 115mA (for a 300ohm load, with a gain of 4x giving max 12V peak to peak output, 40mA is required, so ~60mA is a minimum, subject to critical listening)
I'm leaving those changes to break in before assessing what to do next. Currently, the start up offset is 110mV and falls very quickly to around 70mV. Within 10 mins it is around 20mV. I trimmed it to zero after an hour and it fluctuates within +/-2mV when fully warmed up. A 30VA 12-0-12 transformer gets warm, and the heatsinks are warm but not hot. They were at 55degC with 5.1R and I guess they are a little under 40degC now. This is fine.

Yet to do?
Replace DZ1 and C20 to create a lower noise voltage reference - green led+1n4007 diode? Adjust R21 to fine tune the voltage.
Perhaps reduce R1 to 9.1, 8.2, 7.5, 6.8, etc to see if this improves sound quality
C9 is 100nF not the 33nF in the schematic. This sounds fine and aids stability so perhaps no change needed but try 33nF ... ? I need to look up zobels for HD650 to tailor this to both amp+cans
C13 appears to be 470nF not 1uF - no need to change that given I don't plan to change the input resistance.
The rectifier diodes are fast recovery types with a large cap bank and the 7812/7912 have a large cap bank after them too, so this psu doesn't seem to be limiting the amp. If it ain't broke...?
So the issue was C1 - 220uF is too large - it dulled the treble and made the amp sound too warm for HD650. I fitted a Nichicon KZ 100uF and the amp sounded great again.

I tried different values for R1 from 5 to 10 ohms - couldn't hear a difference but I guess there might be a measurable performance boost? I even tried adding a VAR and dialing it to see if I could hear a difference. I couldn't, but with more time to listen at length, maybe it would be possible. In any event, it sounds great with 5 or 10R and the latter is less heat so I left 10 in.

I replaced C9 with a Wima 22nF and that seemed to have a very small effect on very fine details. It's such a small change that I could be imagining it, or I could have almost immediately got used to it and started to wonder if I imagined it.

I also replaced the bulk electros after the regulators with 4 fancy blue Nichicons 16V 3300uF.

I haven't changed the zener - it seems there's more noise from the source than the amp so I can't see the point, since warm up is quick enough now. I'm using a circuit protector similar to the one AMB sells, and that is set to click on at 40mV offset. That's less time than the laptop takes to boot so no problem.

So, that's it done - box it and enjoy it. It compares very well with the other amps I have such as TPA6120 and LM49610+LM49600, so I'm very pleased. JLH did a fab job
  Reply With Quote
Old 12th December 2018, 08:13 AM   #802
Hugh Jazz is offline Hugh Jazz  Taiwan
Banned
 
Join Date: Feb 2013
Well, with a little more time to experiment today I decided to try leds + diode instead of zener. I tested an smd green led with an FR207 diode thinking this would be about 2.8V, but I was surprised to find the led on its own was 2.85V. So I fitted the led alone and that worked like a charm. It starts up faster and the offset is more stable. Noise doesn't seem to have reduced significantly though, but again, that might be the source - a 16 bit nos dac. Anyway, here's a pic to show off the blue "individual tuned" Nichicons.

BTW if anyone has run SIMs or tested this amp with different bias current, please let me know if 10R is worse than 5.1R for a 300ohm load. Thanks!
Attached Images
File Type: jpg P1010218.jpg (729.0 KB, 187 views)

Last edited by Hugh Jazz; 12th December 2018 at 08:17 AM.
  Reply With Quote
Old 12th December 2018, 09:02 AM   #803
Hugh Jazz is offline Hugh Jazz  Taiwan
Banned
 
Join Date: Feb 2013
Following sgrossklass advice, I increased R21 to reduce the current through the LED - 2k2 resistor drops the voltage across the LEDs to 2.67V with about 4.5mA passing through it. From cold, it took 90 seconds for the offset to drop to 40mV with the zener - with the LED and 2k2 it takes 22 seconds from cold. :-) Noise does not seem to be significantly different - full volume on the amp, Foobar muted, there is a tiny faint hiss - quite hard to hear but it's there.

These are the LEDs if anyone is curious. ??????? 0805 LED ??? ?? ?? LED?-???
  Reply With Quote
Old 12th December 2018, 11:46 AM   #804
Hugh Jazz is offline Hugh Jazz  Taiwan
Banned
 
Join Date: Feb 2013
Quote:
Originally Posted by Hugh Jazz View Post
I tried different values for R1 from 5 to 10 ohms - couldn't hear a difference but I guess there might be a measurable performance boost? I even tried adding a VAR and dialing it to see if I could hear a difference. I couldn't, but with more time to listen at length, maybe it would be possible. In any event, it sounds great with 5 or 10R and the latter is less heat so I left 10 in.
I realised I made a schoolboy mistake wiring up R1 But happily I realised and tried again and this time I could hear a difference. Anything over about 8 ohms has what I think is a little distortion evident, but at 7 ohms I can't hear a difference between this and 5. It's a very small difference but definitely 10R is not good. 7 is fine and again, keeps the heat down so it'll be happier in a case.

I also realised I hadn't let the amp cool down enough - it still takes roughly 90 seconds to get the offset down to 40mV even with the LEDs.

And finally, the blue Nichicons are rare audio grade KT, one grade down from the premo KA. I hadn't seen this before - I'm a sucker for good looking caps so I bought them cos they were pretty. Uh huh.
  Reply With Quote
Old 13th December 2018, 06:49 PM   #805
blackdod is offline blackdod  Europe
diyAudio Member
 
Join Date: Jun 2011
Location: Spain
Well done Hugh, Is the offset low of 40mV? IMO 40mV is much.

For this reason I would like try a servo.
  Reply With Quote
Old 13th December 2018, 11:12 PM   #806
Hugh Jazz is offline Hugh Jazz  Taiwan
Banned
 
Join Date: Feb 2013
A servo will mean instant on - very convenient! I have instant on with the most recent TPA6120 I just built TPA6120 Headphone Amplifier HIFI AMP Board Kit for DIY ac/dc 12v 20v dual power supply-in Integrated Circuits from Electronic Components & Supplies on Aliexpress.com | Alibaba Group that kit has fake caps and was missing any smd ceramic caps, but I tested all the caps and they were fine - just re-skinned cheap caps it seems. I needed to add 2x 100nF and 2x 1uF X7R and 2x 10uF electro smd to build it - they fit under the heatsink.

The offset on my Glozone JLH settles to 2mV +/-2mV in about 20-30 mins. 40mV is the trigger level for the headphone protector circuit, which is very similar to AMB's e12 design - that takes 90 seconds or so. The ε12 muting / protect circuit
  Reply With Quote
Old 14th December 2018, 01:05 AM   #807
sgrossklass is offline sgrossklass  Germany
diyAudio Member
 
Join Date: Nov 2004
Location: Germany
Quote:
Originally Posted by Hugh Jazz View Post
Noise does not seem to be significantly different - full volume on the amp, Foobar muted, there is a tiny faint hiss - quite hard to hear but it's there.
That's normal and expected, especially since you say you have a 16 bit DAC. I take it you could blow your ears off when playing music at normal output levels then? With 4k7/1k5 gain setting resistors, 2 Vrms ought to give you over 8 Vrms on the output - probably the JLH would be clipping first. What's your normal volume pot setting? I imagine something between 9 and 12 o'clock.

Looks like you would need some much more sensitive headphones or IEMs to pick up on headphone amp noise at all. (Or you could cheat with another amplifier on the output...) A DAC of 96 dB SNR at 2 Vrms should still give 131 ÁV of amp output noise if cranked up all the way, which is not a small amount for a headphone amp but with some 102 dB SPL / 1 V cans would still only give a moderate 24 dB SPL of output. The JLH's circuitry itself could have thermal noise levels as low as 3 ÁV, which is quite low for a headphone amp by any standard (not quite O2 level but no more than a few dB off either). Clearly, you are not going to hear any of that if you can barely pick up on DAC noise levels 33 dB higher. Amplifier noise could be substantially higher than 3 ÁV and you still wouldn't hear it.

Last edited by sgrossklass; 14th December 2018 at 01:10 AM.
  Reply With Quote

Reply


JLH Headphone AmpHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
JLH headphone amplifier steveh49 Headphone Systems 8 14th May 2014 04:31 PM
JLH Headphone amp aethelbert Headphone Systems 4 14th July 2009 12:14 AM
JLH headphone amp with FET's and servo nelsonvandal Headphone Systems 17 30th August 2008 10:26 PM
PCB for JLH amp FaTTy Solid State 0 27th March 2007 05:40 PM


New To Site? Need Help?

All times are GMT. The time now is 03:45 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki