IMO the drop in THD is significant.... due to reduced fet C. But then I am for obtaining the lowest distortion possible from a circuit.
If you do not use a notch filter with your test setup, your 24bit converter isnt good enough to measure anything below -100dB. .... the numbers will be meaningless. Else were in DIY audio I have compared the sound-card type converters and QA400 to my AP-2722 and my ShibaSoku 725D to know this.
THx-RNMarsh
If you do not use a notch filter with your test setup, your 24bit converter isnt good enough to measure anything below -100dB. .... the numbers will be meaningless. Else were in DIY audio I have compared the sound-card type converters and QA400 to my AP-2722 and my ShibaSoku 725D to know this.
THx-RNMarsh
Last edited:
While this may easily be true, there's surely something to be seen below -100dB. The numbers might not be accurate, but if I had i.e. a k2 of -100dB before and -120dB after, this should be a meaningful difference to show that the performance got better. Please have a look at the measurement results here which were done with the same soundcard. They look somewhat better, although this time the output level was 6.5Vrms instead of 1Vrms.
Listening now to the Marsh Headphone Amp from Jack's kit.
First got it fired up last night. Running on a +/-12V MeanWell switching DC-DC converter and a 12V 2A wallwart. On first listen there was a *slight* hiss in the right channel -like tape hiss. Hardly noticable but was there compared to the left channel. Also *slightly* more gain in that channel and a little (very little) distorted HF. Swapped input and output cable channels, and isolated it to the headphone amp. Shorted out that channel's input and it was quiet. Reflowed all the solder joints, but it is still there. Thermally, after an hour listening, the outputs were 127-132F without heat sinks. I reset the output DC offset to 0mV, and they stayed put.
Today, I powered up and I checked offset at 9 & 7 mV, and after an hour was 0mV for both channels. Nice.
Right now it is toooooo involving listening to music with my LCD-3s to stop and debug.
I will post later on my impressions, but so far it is VERY favorable!
Thanks to Dr. Marsh for his elegant design, and to Jack for his kit.
First got it fired up last night. Running on a +/-12V MeanWell switching DC-DC converter and a 12V 2A wallwart. On first listen there was a *slight* hiss in the right channel -like tape hiss. Hardly noticable but was there compared to the left channel. Also *slightly* more gain in that channel and a little (very little) distorted HF. Swapped input and output cable channels, and isolated it to the headphone amp. Shorted out that channel's input and it was quiet. Reflowed all the solder joints, but it is still there. Thermally, after an hour listening, the outputs were 127-132F without heat sinks. I reset the output DC offset to 0mV, and they stayed put.
Today, I powered up and I checked offset at 9 & 7 mV, and after an hour was 0mV for both channels. Nice.
Right now it is toooooo involving listening to music with my LCD-3s to stop and debug.
I will post later on my impressions, but so far it is VERY favorable!
Thanks to Dr. Marsh for his elegant design, and to Jack for his kit.
Last edited:
Listening now to the Marsh Headphone Amp from Jack's kit.
First got it fired up last night. Running on a +/-12V MeanWell switching DC-DC converter and a 12V 2A wallwart. On first listen there was a *slight* hiss in the right channel -like tape hiss. Hardly noticable but was there compared to the left channel. Also *slightly* more gain in that channel and a little (very little) distorted HF. Swapped input and output cable channels, and isolated it to the headphone amp. Shorted out that channel's input and it was quiet. Reflowed all the solder joints, but it is still there. Thermally, after an hour listening, the outputs were 127-132F without heat sinks. I reset the output DC offset to 0mV, and they stayed put.
Today, I powered up and I checked offset at 9 & 7 mV, and fter an hour twas 0mV for both channels. Nice.
Right now it is toooooo involving listening to music with my LCD-3s to stop and debug.
I will post later on my impressions, but so far it is VERY favorable!
Thanks to Dr. Marsh for his elegant design, and to Jack for his kit.
First got it fired up last night. Running on a +/-12V MeanWell switching DC-DC converter and a 12V 2A wallwart. On first listen there was a *slight* hiss in the right channel -like tape hiss. Hardly noticable but was there compared to the left channel. Also *slightly* more gain in that channel and a little (very little) distorted HF. Swapped input and output cable channels, and isolated it to the headphone amp. Shorted out that channel's input and it was quiet. Reflowed all the solder joints, but it is still there. Thermally, after an hour listening, the outputs were 127-132F without heat sinks. I reset the output DC offset to 0mV, and they stayed put.
Today, I powered up and I checked offset at 9 & 7 mV, and fter an hour twas 0mV for both channels. Nice.
Right now it is toooooo involving listening to music with my LCD-3s to stop and debug.
I will post later on my impressions, but so far it is VERY favorable!
Thanks to Dr. Marsh for his elegant design, and to Jack for his kit.
Preamp -- Yes, you can say it is lower.... and that is a useful indicator to know.... but the dac is so non-linear below -100db as to be a random number generator.
Steve --- It is good to hear (sic) your results with the LCD3.... I just heard that headphone for my self this week at the Bangkok Hi-End show and a Weiss dac. Really nice! Wondered what it would be like with a better headphone circuit though
Do use higher PS voltages when you can. THx-RNMarsh
Last edited:
Here we go:
Vpos, THD%, k2
10V 0.0047 -113.97
11V 0.0028 -118.79
12V 0.0021 -122.71
13V 0.0020 -123.07
14V 0.0020 -123.78
15V 0.0017 -128.83
16V 0.0016 -129.68
17V 0.0015 -131.83
18V 0.0015 -131.26
This was achieved with 1kHz at 0.5V RMS across a resistive load of 1k.
I noticed that it only needs about 6V peak into 1k at 18V rails for the sinewave to visibly flatten at the top. That's only 1/3 of the rail voltage, which seems to be rather unsatisfying...
Vpos, THD%, k2
10V 0.0047 -113.97
11V 0.0028 -118.79
12V 0.0021 -122.71
13V 0.0020 -123.07
14V 0.0020 -123.78
15V 0.0017 -128.83
16V 0.0016 -129.68
17V 0.0015 -131.83
18V 0.0015 -131.26
This was achieved with 1kHz at 0.5V RMS across a resistive load of 1k.
I noticed that it only needs about 6V peak into 1k at 18V rails for the sinewave to visibly flatten at the top. That's only 1/3 of the rail voltage, which seems to be rather unsatisfying...
This shows well that the THD goes down ... significantly IMO.... with PS voltage...... -114dB to -131dB. Either way they are good numbers. My goal was <-100DB at 1 volt rms output.
Regarding the lower than expected max output levels.... it is a byproduct of using a lot of local stage nfb. But, this will drive headphones well where 1 v rms is super loud, I think it is fine for generating high SPL. However, you can also get even higher (than needed) output levels when going to 24 vdc PS levels. I also tested at +/-24 vdc using the substitute OPS transistors and all was stable and reliable. Try it. THD is even lower at lower Z loads (!) and gives even higher output voltage levels..... [which you do NOT need nor can use.]
I did not listen to it at the higher ps voltages.... I use 15 v supplies and can never listen to 1 volt levels as it is painfully loud. But I might try running it off 24v when I return from Bangkok just to know if i can detect any audible difference.
THx-RNMarsh
Regarding the lower than expected max output levels.... it is a byproduct of using a lot of local stage nfb. But, this will drive headphones well where 1 v rms is super loud, I think it is fine for generating high SPL. However, you can also get even higher (than needed) output levels when going to 24 vdc PS levels. I also tested at +/-24 vdc using the substitute OPS transistors and all was stable and reliable. Try it. THD is even lower at lower Z loads (!) and gives even higher output voltage levels..... [which you do NOT need nor can use.]
I did not listen to it at the higher ps voltages.... I use 15 v supplies and can never listen to 1 volt levels as it is painfully loud. But I might try running it off 24v when I return from Bangkok just to know if i can detect any audible difference.
THx-RNMarsh
Last edited:
That is quite a drop in measured THD between 12V and 18V! I must try this by swapping out my Mean Well +/-12 DC-DC switching PS for a +/-18VDC version someday soon Should still work with my 12VDC 2A wall wart.
I rerouted my grounding and am really enjoying the amp with my LCD-3F headphones. It is the best sounding amp of the limited number I have used. I used to own a Schiit Mjolnir, used it for a while and sold it after I hooked up a little PP 6P1P (similar to 6AQ5) Chinese tube amp. I also home auditioned a Violectric HPA V200 and Auralic Taurus MkII -meh...both solid state amps. I tend to prefer tube amps so I DIYed a PP 6V6 amp based on Poindexter's Musical Machine to do dual duty driving my Altec 19s and headphones. Went all out on building it with RTX coupling caps and Electroprint partial silver stranded secondaries OPTs. That was easily my reference with my LCDs for several years.
That was until I build Richard's headphone amp. Now that is my go to amp when listening to my orthodynamics. Simply everything is better using it, and has caused me to reevaluate solid state amps after prefering and using tubes for 30+ years.
Thank you Dr Marsh for publishing such an ear opening design -especially for us DIYers.
I rerouted my grounding and am really enjoying the amp with my LCD-3F headphones. It is the best sounding amp of the limited number I have used. I used to own a Schiit Mjolnir, used it for a while and sold it after I hooked up a little PP 6P1P (similar to 6AQ5) Chinese tube amp. I also home auditioned a Violectric HPA V200 and Auralic Taurus MkII -meh...both solid state amps. I tend to prefer tube amps so I DIYed a PP 6V6 amp based on Poindexter's Musical Machine to do dual duty driving my Altec 19s and headphones. Went all out on building it with RTX coupling caps and Electroprint partial silver stranded secondaries OPTs. That was easily my reference with my LCDs for several years.
That was until I build Richard's headphone amp. Now that is my go to amp when listening to my orthodynamics. Simply everything is better using it, and has caused me to reevaluate solid state amps after prefering and using tubes for 30+ years.
Thank you Dr Marsh for publishing such an ear opening design -especially for us DIYers.
Last edited:
Thank you, Steve.
BTW -- I found a site which will help anyone who is worried about how much output power/voltage is needed with their headphones. It includes a spread sheet to automate the calculations. I use 100dB SPL as max instead of the 110dB default.
go to apexhifi.com/specs
THx-RNMarsh
BTW -- I found a site which will help anyone who is worried about how much output power/voltage is needed with their headphones. It includes a spread sheet to automate the calculations. I use 100dB SPL as max instead of the 110dB default.
go to apexhifi.com/specs
THx-RNMarsh
Undistorted output is surely enough for driving headphones up to earsplitting volumes, but it won't easily double as a preamp. Seems I've been expecting some more .
Unfortunately (or fortunately?) I've not yet been able to hear any difference between the RNMarch, Ti's MMM, my ulti.preamp.org or everything else I've used round the house to drive my two Beyers. The only thing I notice every time I use the cans is that I don't like to listen to them anymore after an hour or so. That's why I'm currently working on a biophonic equalizer, which will hopefully be the cure. Some preliminary tests using digital eq'ing on the PC were rather promising already.
.Unfortunately (or fortunately?) I've not yet been able to hear any difference between the RNMarch, Ti's MMM, my ulti.preamp.org or everything else I've used round the house to drive my two Beyers
. The only thing I notice every time I use the cans is that I don't like to listen to them anymore after an hour or so. That's why I'm currently working on a biophonic equalizer, which will hopefully be the cure. Some preliminary tests using digital eq'ing on the PC were rather promising already.
Undistorted output is surely enough for driving headphones up to earsplitting volumes, but it won't easily double as a preamp. Seems I've been expecting some more
Run it with 24 volt rails and you'll be able to drive most any power amp to full output.
THx-RNMarsh
I liked the Beyers best back then when I had the opportunity to test several makes and models, but didn't test each one for an hour of course.
Can't afford some new cans at the moment, my wife would kill me. Just got myself a new camera (you can sink serious money in photography, too) and am planning on upgrading my 8+ years old computer, not to speak of a nice wide-gamut screen...
The EQ is currently working in the simulator. I've used a tested Elektor design where I have a pcb lying around for and adjusted the values according to a graph found around the interwebs. Added a sharp notch filter at the "Linkwitz Frequency" of around 7.5kHz and some crossfeed a la Jan Meier. Did not build it yet, but it should work out.
Can't afford some new cans at the moment, my wife would kill me
. Just got myself a new camera (you can sink serious money in photography, too) and am planning on upgrading my 8+ years old computer, not to speak of a nice wide-gamut screen... The EQ is currently working in the simulator. I've used a tested Elektor design where I have a pcb lying around for and adjusted the values according to a graph found around the interwebs. Added a sharp notch filter at the "Linkwitz Frequency" of around 7.5kHz and some crossfeed a la Jan Meier. Did not build it yet, but it should work out.
Attachments
Now that it's finally finished, I'll use my Ulti-Preamp instead. When running for half an hour your amp doubles as a hottie though, which is rather convenient in the evening now that the nights are getting colder
.It will be interesting to know if you like how it sounds with it EQ'ed with that curve. I have a 1/2 octave EQ to play with and can set its controls to hear the affect of EQ. Also some software can be used to change EQ with. Didnt work for me.
What DID work was to EQ the headphone and ear canal to flat freq response at the ear drum. Its all described in the same headphone article in Linear Audio.
THx-RNMarsh
Last edited:
I did actually try something along those lines. Used REW (Room Equalisation Wizard) as a software to pick out the peaks and dips to then resemble them using the Electri-Q plugin for Winamp. The effect was not as pleasing to me as dialling in a roughly biophonic curve using the same plugin. I also did notice that the dips and peaks used to vary by an octave or more depending on how and where I put the mic probe under the earpad (i.e. from the front or bottom, a little lower or higher, put it in some more or not as far, etc.), which made the whole affair a little awkward
.