Go Back   Home > Forums > >

Headphone Systems Everything to do with Headphones

Designing my headphone amp
Designing my 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
Thread Tools
Old 21st March 2021, 10:45 PM   #101
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
So given the 24v to 230v, the maths is showing a duty cycle of 89:6%, a current ripple of 1mA and the losses increasing with switching frequency.

So I also miscalculated the inductor, 47uH? No 220mH yes.. well I can use a smaller inductor and solve the additional ripple after.

If that doesn’t work the I can try half bridge and full bridge topologies. Theyre used with more power at the expense id more components.. although with less demands on them.
  Reply With Quote
Old 28th March 2021, 04:31 PM   #102
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
Amp startup and idle with no signal.

Screenshot 2021-03-28 at 17.16.46.png

Idling ~70W of 300W 24V PSU.

Need to add +30W of heaters to that but the SMPS approach appears to have paid off.

No big mF caps. So I'll be doing some more testing.

So I save on caps.. well except the WIMA MKPs are 28 for the two and the two coil craft inductors are 36.. so I could possibly find some cheaper.

Output B+ noise?

Screenshot 2021-03-28 at 17.26.30.png

Sample period 4-6 seconds from start up, so I'd expect a little more LF noise drop but that's looking very good, well at least from an ideal simulation perspective!

The output tubes have a 750uF LPF/reserve cap, the input tubes have a smaller LPF but there's some additional changes I want todo which should reduce noise further in the amp.
  Reply With Quote
Old 28th March 2021, 09:35 PM   #103
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
And full 'power'..

Screenshot 2021-03-28 at 22.24.38.png

Screenshot 2021-03-28 at 22.25.45.png

Screenshot 2021-03-28 at 22.26.36.png

So virtually no current change or power consumption change which is good.

Also the noise fft - this is with a 10Khz 3.16V input and a 30Hz 3.16V input on the other channel (there's a little cross talk).

I think I can live with that.

I've already started researching/designing the PCB layout, just trying to keep the di/dt loops small but I think this will get redesigned to make the out cap loop even smaller but at the same time I don't want to cross the HV switching and LV sense/FB paths.


Annoyingly the MOSFET has a centre drain and a gate on one side - when I want to mount it on the edge of the PCB along with the diode.
  Reply With Quote
Old 31st March 2021, 11:34 AM   #104
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
New v2 redesign on the SMPS booster and it all looks good. I've also gone through the components, located real values and updated the models for caps, inductors, etc. The model's switching is now currently running at 57KHz.

The max I've got out of the boost so far at 14A input - is 1.2A at 270V.. but that had a 37 mosfet running on it. The limit is the current through the inductors - so additional inductions would be needed however at that current we'd end up needing to look at the PCB and busbars.

So I'm happy with the performance, I'll limit it to 13A max, which works perfectly for the 230-250V with 600mA.

So my attention is starting to turn back to the amp itself, in terms of the model and finalising with the components I'd previously found or replacing with better/more suited ones.

I've also switched the 2200uF cap to Nichicon Muse ES which is bipolar (although only 50V - if the cathode is 50V then something is really wrong and the fuse should blow). Keeping the ESR and therefore the impedance as low as possible, instead of using a 2200uF I'm using 6x330uF in parallel. The result is that the ESR drops from 0.48ohm to 0.08ohm.

Next up is some adjustments on the B+ rails. They're current configured to use Nichicon 450V 300uF and 750uF but I can't find the matching configuration (I didn't note the model number). So I've started looking for some lower ESR (my thinking here is ESR is a bit like the accelerator vs the ripple which is the top speed), dynamics is a goal.

I know that "audio" caps are meant to have different internal builds but for now - I'm happy getting the thing running. I can then replace caps as I go. I know that Elina are considered very dynamic vs Panasonic/nichicon.. but I think there will be bigger issues to solve first.

The SMPS itself has a 20uF and 80uF PP 6mOhm caps that are required and seem a better investment at this time.

Current SMPS Boost model:

Screenshot 2021-03-31 at 12.34.43.png

The boot time with the full amp is very quick and the Rsense allows the current inrush to be limited too as a bonus.

It's also a good point to get to, given work is likely to slow a little over summer.

Last edited by NickKUK; 31st March 2021 at 11:39 AM.
  Reply With Quote
Old 8th April 2021, 07:22 AM   #105
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
So I have my SMPS booster modelled with the large signal test, ignoring the fact that the amp needs some tuning to reduce the harmonic content, the SMPS+3080-based regulator givens sub 1mV ripple. I've tweaked the booster with a little snubbing, increased to 75KHz, added two more low ESR caps at the source as they were starting to get maxed in ripple current, added 4x20uF caps after the diode to take the current faster then added a slightly tweaked maida-christen LT3080 based regulator - currently limited to 500mA. Also the model now has 150mVpp ripple added into the initial source 24V.

This is 10Khz + 30Hz at 3.16V input. The model FFT shows 5.0 seconds to 12 seconds of simulation time (so far it's taken two nights and a day).

Screenshot 2021-04-08 at 08.08.14.png

Screenshot 2021-04-08 at 08.12.10.png

Note the lack of background noise

The model looks like this:
Screenshot 2021-04-08 at 08.29.39.png

Last edited by NickKUK; 8th April 2021 at 07:33 AM.
  Reply With Quote
Old 12th April 2021, 05:32 PM   #106
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
Now the boost power supply is in fine tuning.. it's about time to revisit the amp and start fine tuning that too..

This is a channel, and it's looking at tad more complex than the previous post..

Screenshot 2021-04-12 at 18.10.25.png

Each tube transconductance will differ so making each output tube adjustable becomes necessary. I've also retuned a little for 200V B+ - this may mean I need a regulator per channel to act as noise barrier, but for now this will do.

So each tube has it's own output cap bank, feedback and CCS. As each tube needs some adjustment this helps prevent it becomes a web of parallel adjustments only for temperature and ageing to compound things further.

I've explored using a CCS per channel, the this work nicely, but I've also tried making this 'stiffer' (essentially making it a more stringent regulator) however this reduces impedance and so the valve pulls the current from the CCS in preference to pulling from the output cap bank, reducing the effectiveness.

I'm now starting to look at the THD. 1V input at 1Khz:

Fourier components of V(output-left)
DC component:2.23821e-07

Harmonic	Frequency	 Fourier 	Normalized	 Phase  	Normalized
 Number 	  [Hz]   	Component	 Component	[degree]	Phase [deg]
    1   	 1.000e+3	 1.094e+0	 1.000e+0	    0.04	    0.00
    2   	 2.000e+3	 2.073e-2	 1.895e-2	  -89.99	  -90.02
    3   	 3.000e+3	 4.223e-4	 3.860e-4	    0.11	    0.07
    4   	 4.000e+3	 2.381e-5	 2.176e-5	   90.25	   90.21
    5   	 5.000e+3	 2.270e-6	 2.075e-6	 -179.63	 -179.66
    6   	 6.000e+3	 2.801e-7	 2.561e-7	  -90.03	  -90.07
    7   	 7.000e+3	 3.936e-8	 3.598e-8	    0.38	    0.34
    8   	 8.000e+3	 9.387e-9	 8.581e-9	  103.14	  103.10
    9   	 9.000e+3	 8.857e-10	 8.096e-10	  120.05	  120.01
Total Harmonic Distortion: 1.895205%(1.895206%)
Hmm 1.9% THD.. can do better. I suspect this is close enough for a simulation but I may adjust the operating point a little to get to 1%. I did leave the 1mV PSU ripple in but I doubt that would cause a difference..

Last edited by NickKUK; 12th April 2021 at 05:36 PM.
  Reply With Quote
Old 12th April 2021, 08:26 PM   #107
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
Front end output I see this
Direct Newton iteration for .op point succeeded.
Fourier components of V(l-out-grid-start)
DC component:6.3806

Harmonic	Frequency	 Fourier 	Normalized	 Phase  	Normalized
 Number 	  [Hz]   	Component	 Component	[degree]	Phase [deg]
    1   	 1.000e+3	 2.339e+0	 1.000e+0	   -0.01	    0.00
    2   	 2.000e+3	 1.893e-2	 8.093e-3	  -90.06	  -90.05
    3   	 3.000e+3	 8.430e-5	 3.604e-5	  179.67	  179.68
    4   	 4.000e+3	 3.402e-5	 1.454e-5	  -90.25	  -90.24
    5   	 5.000e+3	 4.635e-6	 1.982e-6	   -0.38	   -0.37
    6   	 6.000e+3	 6.208e-7	 2.654e-7	   89.26	   89.27
    7   	 7.000e+3	 1.067e-7	 4.561e-8	 -179.66	 -179.65
    8   	 8.000e+3	 6.734e-9	 2.879e-9	 -151.86	 -151.85
    9   	 9.000e+3	 2.148e-8	 9.184e-9	  -22.21	  -22.20
Total Harmonic Distortion: 0.809335%(0.809335%)
  Reply With Quote
Old 12th April 2021, 10:10 PM   #108
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
Fourier components of V(output-left)
DC component:-8.41261e-06

Harmonic	Frequency	 Fourier 	Normalized	 Phase  	Normalized
 Number 	  [Hz]   	Component	 Component	[degree]	Phase [deg]
    1   	 1.000e+3	 4.735e-1	 1.000e+0	    0.09	    0.00
    2   	 2.000e+3	 1.238e-2	 2.615e-2	  -89.95	  -90.03
    3   	 3.000e+3	 1.576e-5	 3.329e-5	  175.58	  175.49
    4   	 4.000e+3	 1.984e-5	 4.190e-5	   90.27	   90.18
    5   	 5.000e+3	 7.441e-7	 1.571e-6	 -178.81	 -178.89
    6   	 6.000e+3	 1.570e-7	 3.315e-7	  -91.52	  -91.61
    7   	 7.000e+3	 1.796e-8	 3.792e-8	   -2.03	   -2.12
    8   	 8.000e+3	 8.152e-9	 1.722e-8	  149.38	  149.29
    9   	 9.000e+3	 6.971e-10	 1.472e-9	  152.51	  152.42
Total Harmonic Distortion: 2.615248%(2.615248%)
So this has a larger second harmonic that the original setup, but the the 3rd is lower.

Screenshot 2021-04-12 at 23.06.32.png

Only thing is this produces less output, however shows you can adjust the harmonics. For now the original settings are probably the best starting point.

Last edited by NickKUK; 12th April 2021 at 10:13 PM.
  Reply With Quote
Old 13th April 2021, 01:05 PM   #109
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
The ECC99 load lines as they are currently showing the near vertical low impedance line - anode/cathode it doesn't matter they're both steep:

Screenshot 2021-04-13 at 14.02.11.png

If you look at the left, that bend in the 150V line is what's giving us some of that loving THD..

So first up - I'm looking to tune the output section without NFB running.. lowing the voltage to 135V but keeping the same current operating point.

Last edited by NickKUK; 13th April 2021 at 01:09 PM.
  Reply With Quote
Old 13th April 2021, 08:12 PM   #110
NickKUK is offline NickKUK
diyAudio Member
Join Date: Dec 2019
At 55ohms that's Total Harmonic Distortion: 1.828005%(1.828005%) external 180deg feedback from the output itself.
2 2.000e+3 2.033e-2 1.827e-2 -89.98 -90.02
3 3.000e+3 5.210e-4 4.684e-4 0.12 0.08
4 4.000e+3 3.907e-5 3.512e-5 90.28 90.24

This rises using the internal original feedback, 1.912218%(1.912218%), however I'm also aware that ltspice has some variation in terms each run.
2 2.000e+3 2.182e-2 1.912e-2 -89.96 -90.02
3 3.000e+3 5.702e-4 4.996e-4 0.18 0.13
4 4.000e+3 4.789e-5 4.196e-5 90.36 90.31

I think there's obviously some tuning that can be done of the feedback too and I could bang my head against a brick wall resulting in paralysis by analysis.

So for my first build I think this will do. I can tune it once the hardware is running in front of me.

I am having a quick shifty (look) at the front end 12BH7A tuning. It's occurred to me that putting a heap of voltage gain into a ecc99 output wouldn't help that much, the THD and in general the distortions, would rise too high given the output we're after.

As a basic design there's not much you can do (other than play with feedback). I can see why people for OTL use either masses of tubes to plumb for PP configuration/fets for low impedance. I think if I need more oophm I would put a couple of fet followers.
  Reply With Quote


Designing my headphone ampHide this!Advertise here!
Thread Tools

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
I'm designing a TPA6120 based headphone amp, is there anyone can give me some advice shadownee Headphone Systems 13 31st August 2012 04:50 AM
C&C request: designing a headphone amp Thalyn Chip Amps 0 14th April 2012 12:12 PM
Designing a super low distortion headphone amp. Rainwulf Chip Amps 41 1st April 2011 12:24 AM
designing horn for compression driver & tweeters jayam000 Multi-Way 0 15th April 2009 09:26 AM
Designing for guys who like designing Sch3mat1c Tubes / Valves 14 31st August 2003 11:46 AM

New To Site? Need Help?

All times are GMT. The time now is 09:55 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Copyright ©1999-2021 diyAudio