Hello,
I just assembled the Modulus 86 r1.01 board and did the final checks as recommended in the instructions.
First off, the supply voltage to the three op-amps sockets (pins 4 and 7) measure about -15.3V and +15.08V respectively which appears to be within the expected values.
After mounting the op-amps I measured the DC offset on the output plug. Initially, the offset is about -20 mV. However, it quickly increases and stabilises around +0.23 mV. However, according to the instructions it should stabilise below 20 uV, so this is somewhat higher than expected. Just wondering if I should worry about it? Also, (this might be a stupid question) should I ground the three connections in the input signal plug while measuring the output offset? I mean, if there is a tiny offset on the input, I suppose it get amplified and thus result in a DC offset on the output?
Thanks,
Soren
I just assembled the Modulus 86 r1.01 board and did the final checks as recommended in the instructions.
First off, the supply voltage to the three op-amps sockets (pins 4 and 7) measure about -15.3V and +15.08V respectively which appears to be within the expected values.
After mounting the op-amps I measured the DC offset on the output plug. Initially, the offset is about -20 mV. However, it quickly increases and stabilises around +0.23 mV. However, according to the instructions it should stabilise below 20 uV, so this is somewhat higher than expected. Just wondering if I should worry about it? Also, (this might be a stupid question) should I ground the three connections in the input signal plug while measuring the output offset? I mean, if there is a tiny offset on the input, I suppose it get amplified and thus result in a DC offset on the output?
Thanks,
Soren
I agree with you and wouldn't have even thought about it were it not for the discussion about "competence " in the power supply triggering me to go back and check the G-word article. Given the number of RF sources in my living room these days I'll take CMRR rejection.
Any knowing what the limiting factor is beats being clueless as to what is going on 🙂
It is fine to not understand. Mother history of technology advancement is like that. I would like to be proven wrong. And I could probably point out any published test involving human listening as incomplete yet still valuable. But understanding the fact that audio technology evolved mostly based on the desire to obtain personal listening experience, so the subjective factor cannot be eliminated. Rather, it should remain an important factor in audio systems development. All the measurement techniques help keep you on the shortest track to reach this goal, but if you start using the test as the final say, you lose the original motive from which the technology evolved.
I didn't find it as hard to hear as Richard makes it out if you use test tones, headphones - listen for an apparent tiny pitch change in my test tone
still no activity re my polarity test fies? - admittedly I just put the files here because I couldn't link them to the discussion on head-fi
but as the discussion developed I did install foobar2000 and its ABX plugin and it turns out that I can hear the difference, just can't articulate it clealry - foobar2000 abx, motherboard Realtek HD sound, HD600 direct from computer headphone connector
I seemed to confuse some by citing phase distortion audibility papers - which oddly seems less controversial than audibility of absolute phase
so I prepared visual for showing how polarity of a 1:1 fundamental and 2d harmonic asymmetric test tone depends on relative phase of the components
I add a phase dependent offset (magneta) to the dynamic sum (sum in blue, magenta is "0" line for blue) to show the +/-90 plots mirrored about the x axis so polarity should be clear by eyeball
I move both the fundamental and the 2nd phase to center the +/-90 sum's peaks on the y axis - "48" in the algebra pane is just the value of alpha when I did the screenshot
the varying vertical offset is just a visual aid I hoped would make the sine harmonic addition picture easier to read - less lines crossing over to keep track of - our eyes are really happier looking at a mirrored figure/seeing the inverse symmetry when the "fold" doesn't have overlap, is based off the mirroring line
green, red should be very easily seen to be mirrored about the x axis == inverses, inverted in polarity waveforms just from shifting a sine and its 2nd harmonic
direct link to animated gif in case diyAudio "pickles" the image http://img.photobucket.com/albums/v252/f5r5e5d/polarity_zps3af73b48.gif
obviously the audio doesn't want the offset - see my earlier Spice sim waveform picture - which didn't seem to convince everyone
I did get one taker for a listening test over there so far
Testing audiophile claims and myths - Page 167
back in that thread for my .wav that you can put in foobar2000 abx plugin and test for yourself
Last edited:
I buy that with a generated sinusoidal tone you can probably detect something. But is it really the case with a music signal? And if you have demonstrated audibility with a PC on board sound card that somewhat destroys the 'only the best amplifiers can show it' argument.
Fascinating stuff, thank you for bringing it to my attention. BTW cool amp in your profile pic on headfi. I was thinking about those headsinks for my mod-86 at first.
Fascinating stuff, thank you for bringing it to my attention. BTW cool amp in your profile pic on headfi. I was thinking about those headsinks for my mod-86 at first.
Perfect.
Don't worry about it. The 20 uV assumes that you use the more expensive OPA277P and I neglected some other offset sources in the quick math I did for the instructions. The lower cost OPA277PA that I specify for the MOD86 R1.0 has an intrinsic offset of 50 µV, so it's impossible to do better than that. The 200 µV you're reporting sounds very reasonable.
I should probably run a few sims and give better numbers in the instructions.
For speakers a few mV of offset is nothing to worry about.
You usually get more repeatable offset measurements by grounding the input.
~Tom
I didn't say "inaudible". It's just not the "chalk & cheese" stuff the Golden Pinnae claim.
On music, Steinbeck is the earliest I know of, 1930s in the sadly missed Bell Systems Technical Journal. There are more references in Is Linear Phase Worthwhile? Re-reading this, I'm surprised to find there isn't anything I would change.
The really useful paper is the Wharfedale Engineering Memo that my paper was based on. Alas I no longer have a copy. 🙁
You results are in line with my 1978 tests and also with Steinberg 🙂. The most critical area is 'around 100-500Hz'.
You can always dream up test signals that highlight audibility. The real question is whether its important enough to do something about it.
A DBLT is a Measurement. You need to calibrate & check your Instrument, your Blind Listening Test panel regularly. I've used DBLTs to design stuff for nearly 2 decades and am confident to use words like invariably for certain things.
- self styled Golden Pinnae (da Chalk & Cheese lot) are invariably deaf ie give random results in DBLTs
- the man in the street is more perceptive than most audio reviewers. There are (a very few) notable exceptions
- the woman in the street is even more perceptive.
- the true golden pinnae who can reliably detect eg phase distortion invariably say the effect is very small .. certainly smaller than important audible speaker distortions which stare the majority of audiophiles in the face without being seen.
IIRC, the important Absolute Polarity tests were with Bass Drum and
"Yes". These are much more difficult to pick than the type of test you describe.
If you change the phase on the fly with your test signal, this is 'easily' detectable. Caveats about checking for non-linearity bla bla in the "Is Linear Phase ..." paper.
I try to get my stuff to preserve Absolute Phase but this is cos getting it wrong gives other evils. But I've made at least 1 big boo boo in my time that I only managed to correct this Millenium. 😱
Last edited:
Yesterday we decided to try music played from an old iPod nano, and also music streamed from the Internet through an iPhone 6+. Polarity was obvious in music that was very clear and crisp, music that were on the soft and fuzzi side did not present an obvious choice.
What speakers & amp? The ones on your website?
If they are your speakers, have you a link to the THD performance vs freq. & level?
Are you using current drive?
The active speakers in my avatar. As mentioned before, distortion is not the system is not used mainly because you can get a different result depending on where you place the mic. I have not found a way to pinpoint problem causes through those measurements. But if someone has done something wrong, it is used as proof that something was done wrong when the same measurement setup is used.
This audition session was to evaluate the effect of additional current drive. From the structure, I would not have expected it to deliver the current it did. But I do intend to get to the bottom of this. Right now, the analysis shows it working that way, a quick adjustment to the test circuit demonstrates it works that way, but there is still much more to explore to figure out because there are still some questions unanswered.
If the Modulus could deliver currents through the higher impedance and still have the distortion figures under constant load, it would be a very interesting amplifier. I am sure it would then give users the impression of really powerful, which is why some people like using small power tube amps.
One of the reasons why most companies refrain from using current drive is that most multi-way speakers have high impedance at crossover points, this can cause significant sonic balance problems giving the company a bad name, and severely limiting sales.
Hi Soon and other guys,
it is quite easy to convert the modulus into a "current drive"amplifier.
I ve made that and finally decided to create my own pcb with 2 LME+LM3886 on it for an active speaker. testing is WIP and I do not want to bother the thread on this.
But if you want to convert you modulus to a current drive amp, here is the trick:
disconnect the feedback resistor R14 and replace it by 2k2 and solder a 10nf cap+100ohm resistor in paralell on it.
keep one leg up, the one supposed to go to LM3886, and instead you have to connect it to what I d call a "foot" 0.5ohm resistor used for measuring current flowing in the speaker.
This foot resistor is to be inserted in the ground going to the speaker.
it all works. The 10nf cap has to be adjusted depending on the inductance of the speaker and the 100ohm res provide better stability, Their role is to flatten the current curve as much as possible (you can measure it with rew and chose the best cap giving the most flat curve).
Stability wise this looks ok on my scope, but I ve not performed extensive tests so far and would not recomend you to plug a tweeter on it whithout securing this by yourself first.
With 0.5ohm measurement resistor, the input should be reduced to 0.5V rms to avoid too high current, but you can do the math easily.
There will be differences in the performance compared to the standard Modulus, because with that approach, the speaker becomes part of the feedback look and this brings more sensistivity to environemental emissions.
Be aware you cant use such amp without EQ as you ll have to compensate for driver/box resonance...
if you have more question and ideas around this, dont post them here 🙂 there are couple of threads on current drive on diyaudio, and suggest you start reading the book of Esa Merlainen! enjoy the journey to Current Drive, motional impedance, diaphragme acceleration control and linear phase. cheers
it is quite easy to convert the modulus into a "current drive"amplifier.
I ve made that and finally decided to create my own pcb with 2 LME+LM3886 on it for an active speaker. testing is WIP and I do not want to bother the thread on this.
But if you want to convert you modulus to a current drive amp, here is the trick:
disconnect the feedback resistor R14 and replace it by 2k2 and solder a 10nf cap+100ohm resistor in paralell on it.
keep one leg up, the one supposed to go to LM3886, and instead you have to connect it to what I d call a "foot" 0.5ohm resistor used for measuring current flowing in the speaker.
This foot resistor is to be inserted in the ground going to the speaker.
it all works. The 10nf cap has to be adjusted depending on the inductance of the speaker and the 100ohm res provide better stability, Their role is to flatten the current curve as much as possible (you can measure it with rew and chose the best cap giving the most flat curve).
Stability wise this looks ok on my scope, but I ve not performed extensive tests so far and would not recomend you to plug a tweeter on it whithout securing this by yourself first.
With 0.5ohm measurement resistor, the input should be reduced to 0.5V rms to avoid too high current, but you can do the math easily.
There will be differences in the performance compared to the standard Modulus, because with that approach, the speaker becomes part of the feedback look and this brings more sensistivity to environemental emissions.
Be aware you cant use such amp without EQ as you ll have to compensate for driver/box resonance...
if you have more question and ideas around this, dont post them here 🙂 there are couple of threads on current drive on diyaudio, and suggest you start reading the book of Esa Merlainen! enjoy the journey to Current Drive, motional impedance, diaphragme acceleration control and linear phase. cheers
Last edited:
Hi Tom,
thanks a lot for your reply. The op-amp on my board is in fact the more expensive OPA277P. I simply ordered the BOM from Mouser😕
What Power Supply Rails are you using?
What is the DC resistance of your speaker units?
What is the current you didn't expect to deliver?
+/- 24 V ~ 27V, the ones I used on the MyRef.
5.6 Ohms
delivering more current through the impedance peaks. The circuit originally used to flatten the impedance were changed to absorb some current, because there was slightly more current than necessary to keep flat current.
5.6 Ohms
delivering more current through the impedance peaks. The circuit originally used to flatten the impedance were changed to absorb some current, because there was slightly more current than necessary to keep flat current.
Well, if you have too much current going through the driver, you add a current path to bypass some so that the driver gets constant current, is there anything wrong with that?
A 8ohms speaker with an Re value of 5r6 will demand more current than an 8r0 resistor when driven with a fast starting/stopping transient signal.
Expect an 8ohms speaker driver to have about 50% more current demand than a 8r0 resistor.
Add on a 2way crossover feeding two reactive drivers and the current demand can increase beyond three times the resistor's current demand.