Discussion in 'Headphone Amplifiers and Combo (DAC/Amp) Units' started by purr1n, Aug 11, 2017.
Hope you set the power out to 28V and not 48V!
When I posted the question to atomicbob, I was just asking about the acopian supply differences. I was using it for a different amp, the Garage1217 Polaris which takes 48V... I figured any reply he provided might be useful to the rest of the group....
I’ve been using the Massdrop Cavalli Tube Hybrid amp for a few months now and, in my opinion, it’s a fantastic value with what it brings to the table. The synergy with the Sennheiser 650 + stock tube really surprised me and felt it was almost on par with Mr.Speakers AFO + Schiit's Mjolnir 2. We're talking about a $450 system compared to a $1500 system but in the end the AFO / MJ2 combo beats in several areas. On the MTCH I prefer the HD650s over the Aeon Flow Open.
Few things about me. I prefer a mid-forward, warm, and tonally correct sound. If a piano key hit or pluck of guitar sounds realistic then you have my attention. I live in the mids but I'll say that my guilty pleasure is the E-Mu Teaks.
PC Audio Out
MCTH + Tung-Sol 6C8G Tube / Foton Triple Mica 6N3P
ZMF Auteur / E-Mu Teaks
To get to the meat of it and the reason I'm posting. There aren't big differences when swapping the EH tube with higher end NOS tubes paired with the HD650 but there are differences when using something like the ZMF Auteurs / Clears which are revealing to changes in your chain. There are more tubes to roll with adapters like the 6C8Gs or 5670/2c51/W396As. Fair Warning: This will void your warranty if you fry the amp. I won't be responsible and this is not endorsed by Massdrop (don't kill me CEETEE). So for people who already have a LPS or are a little adventurous give it a try.
The Auteurs flavored by the 6C8G mids, wider sound stage, holographic sound, and organic tonality has me cracking smiles.
Nothing beside wiring - but if you don't have much experience doing that, I'd strongly suggest you involve someone who does. I've done bits of amateur electrical before, but I did consult an electrician acquaintance to confirm I wasn't courting disaster. For the following, insert standard disclaimers...
As @rtaylor76 said, no resistor required. Ignore the 'Remote Sense' and 'Grounding' parts of the documentation (if that's where you saw the 'resistors'). You could wire an in-line fuse holder, or use a fused, switched socket like I did (the manufacturer-specified fuse for the 28V 2A model is 1A). The output barrel plug is the standard 5.5 x 2.1 mm. Crimp connectors for the primaries on the transfomer are 0.11" fast-ons and you'll need a couple of circular sockets for the output terminals. You'll find everything necessary in the auto or household electrical section of the hardware store (or you could just solder). The jumper configurations for the primaries for local AC voltage are given in the schematics on the Galco site. Leave the 'sense' output terminals alone.
Note the output polarity when you wire up the barrel plug: + and - terminals are annotated. And as @atomicbob stressed in the measurement thread, do NOT connect the -ve output terminal of the power supply to ground. You can, though, connect a ground from the chassis to the AC input socket (I did this).
I wouldn't use a box that's too much smaller - even though it doesn't run very hot, it's good to have some air around things. I also keep it switched off when not in use (an advantage of the switched socket).
Before you plug it into the amp, I'd recommend testing with a multimeter to check that output voltage is correct and to confirm there's no ground between the output -ve and the chassis (if there is, either the unit or your wiring is faulty).
The A28MT210 Gold Box is the one I measured. The two models are nearly identical in specifications. However the Gold Box has slightly better load regulation of 1.4mV as opposed to 3mV for the Infinity. Not enough difference to worry about in this application. The AliExpress 28V LPS also works well and is probably a better choice for most, being a complete unit and available for around $125 US.
Cool! Thanks for the info @atomicbob! I'm not sure if this makes sense, but I found that if I set the voltage at 48V exactly that the amp was just a little bass shy. If I set it down to 47.8, the bass output was better but it sounded every so slightly muddy. If I set it to 47.9, it seemed to be about perfect. I would figure the voltage should be close enough not to cause any damage to the electronics. Nonetheless, I'm not sure if the MTCH would respond this way, but the Garage1217 Polaris did.
Please consider editing your posts to make it clear that your references to 48v are for a different amp... The fear is that someone reads these and blows up their mcth
We might need a new thread - benefits of LPS on audio gear.
I've had some experience wiring things up, but that was mostly robotics systems which weren't hooked up to AC.
I offered $30 on that listing and it was accepted, so now I've got 12awg wires, a fused socket, and the barrell connector on the way. What are those circular connectors called? I can't seem to find them.
EDIT: My mistake, I thought you were referring to something else. Looks like @Biodegraded has you covered
You mean the supply-end output connector, not the amp-end one @westermac refers to above? I should have said cylindrical rather than circular (rushing before dinner). I used these dual female bullets:
You could use singles, but using 18 gauge output wire folded double and crimping one end over it gave a good fit in the wire end (pull the connector out of its jacket first). The other end needed closing up a bit to give a tight fit to the terminals. They come in smaller sizes too.
If you have more questions specifically on wiring up this supply feel free to PM me - we're hijacking the thread somewhat here...
Very eager to get my butt over to @Biodegraded ‘s place to hear his LPS setup and bring mine home without blowing up my little audio nest. It’s currently a clusterfuck as residence is only temporary here, but it works.
Looking forward to finding time to do some thorough comparisons when I get some time.
I picked up a HD6XX on a whim recently, and the MCTH+LPS is a great pairing. Last time I owned a stock HD650 it just bored me... even with the Mjolnir 2 I owned at the time. I started the process of modding them but lost patience and eventually sold them.
Glad I gave them a second chance! They don't sound as rolled off as I remember, and the bass is more controlled and extended than I recall. These will complement the Focal Clear well when I need something unapologetically smooth for those craptastic masters I seem to enjoy.
I would agree. For those that don't always enjoy the laid back dreamy sound of HD650/6XX, the MCTH tends to give a more clinical sound that is quite more detailed and lively. Not that it makes them sound like HD600's, but it still has that great soundstage that is still quite involving.
I tend to still prefer the HD650 + Valhalla 2, but I could certainly live with only the MTCH, but at a slight cost of being a bit more forward and maybe slightly fatiguing (for 650's anyway). YMMV
I've been playing with my MCTH amp. Decided to explore how it sounds if pushed hard at loud levels. Plugged in my LCD-2 Classic, put on War On Drugs "Lost In A Dream" first track advanced to the middle of the song where the volume is more constant. Used a dB meter with a baffle to set the volume where peaks were at 100 dBA. Put the headphones on my head (briefly). The sound was crunchy. Obvious distortion going on. Did the same with my Liquid Glass (briefly) and Liquid Fire (briefly). The sound was clean. No crunchyness. None. I could have turned the volume up louder and still hear headroom and clean sound. It's frightening how good 100 dBA sounds with the Fire and Glass.
Is the distortion I hear with the MCTH due to the stock wall-wart power supply running out of oomph? Would an aftermarket LPS help with getting more headroom and avoiding distortion at those listening levels with the LCD-2 Classic? Could this distortion be due to a possibly noisy tube in the MCTH?
I obviously do not listen to music like War On Drugs so that it peaks at 100 dBA. That would be crazy. My usual loudish listening level is at around where the music settles around about 85 dBA or so if I want to listen loudish. But peaks obviously go higher. With classical music I find that peaks with some music can sometimes hit over 100 dBA even if the level for most of the music is at around 85 dBA. I don't want those peaks to clip or compress or get crunchy. Is the MCTH able to handle that with a headphone like the LCD-2 Classic?
A possible factor is that I'm using the MCTH with a Modi Multibit using the USB input. I'm using the Fire and Glass with an OG Gumby with an Eitr. I haven't yet tried the MCTH plugged into the Gumby and Eitr combo. I don't expect the difference between the Modi Multibit and Gumby/Eitr to make a difference with this sort of distortion.
@Hammy, I think you're just simply running the MCTH past it's peak output at those levels. Unless you're right at that turning point of distortion, I don't see how a LPS would improve max output.
First, dBA is not the correct filter to use for music at 100 dB SPL. Sound Level Meter A, B, and C scales are bandpass filters intended to behave similarly to the Human Auditory System at 40, 80 and 100 phons respectively on the ISO Contours of Equal Loudness (updated Fletcher-Munson curves.) The A-weighted scale will attenuate low frequencies by a significant amount. It is probable that LF content is being understated by 10 to 20 dB SPL. Switch to dBC and repeat your experiment. You are likely to be surprised by the results.
Next, which time integration were you using, Slow or Fast?
Slow integrates over 1 second while Fast integrates over 0.125 Second and has a faster response time, though it still can't catch transient peaks which may be 10 dB SPL higher. So depending on the Crest Factor of the stimulus, you are not observing the actual peak demands on the headamp. Use dBC, Fast and add 10 dB to the peak observed. Then you are closer to approximating the peak demands for the headamp.
Of course, an oscilloscope on the terminals of the headamp output would provide a precise indication of voltage required. Current could be approximately calculated assuming a mostly constant impedance. This would also allow observation of clipping if it is occurring which would account for the crunchy sound you describe. An LPS will not change the clipping voltage for the MCTH.
I agree with @crazychile. My LPS did not add any headroom of any kind, just more or less improved the overall SQ and had clearer/cleaner details without harshness. The low end was improved a bit in oomph, but not by a huge margin at all. Bass is just more there and more defined. It is still very clean and tight.
The LCD-2's are just hard to push and getting the MCTH to it's limits. It is a great amp, but it has it's limitations as it is not extremely loud and punchy. It has moderate power for most headphones with good overall SQ, but not the most powerful weapon.
I use fast time weighting when checking dB levels. And I usually do use C weighting for music measurements rather than the A weighting. But the meter was set at A weighting when I turned it on and I left it there for this measurement. I watched for peaks hitting 100 occasionally with fast time weighting and set the volume at that.
The specs at Massdrop say the amp has 1 W Rms at 50 ohms. The LCD-2 Classic is 70 ohms.
The distortion I heard was more of a hardness than literal crunching. I've heard real crunching when amps clip and know what that sounds like, and this wasn't doing that. Just a harder sound on the peaks than the normal smooth sound the amp should be doing. Was I clipping? I was expecting the amp to be able to play the LCD-2 Classics at this level without clipping even though the amp is just 1 W.
That hardness is evidence of an amp just beginning to clip. The levels you describe are VERY LOUD and I'm glad to know you don't actually listen at those levels. Let me demonstrate.
Comparing Sound Level Meter A and C weighting scales shows 50 Hz content will be attenuated by approximately 30 dB on A vs C while content above 1 KHz will be indicated approximately 2 to 3 dB higher:
Thus low frequency content will be under-indicated on the SLM with A-weighting accordingly.
Examining an excerpt from the test track used, Under the Pressure – The War on Drugs shows kicks with higher demands for voltage (dBFS) than the higher frequency content average level.
Under the Pressure – The War on Drugs 4.7 second excerpt view
Zoom view of 445 mS for one period between kicks with kick initial attack sound on left side of waveform viewed.
Low pass filtering for 500 Hz content that will be attenuated by A weighting:
Instantaneous kick peak demands are -3 dBFS during the attack. Sustain is lower at -12 to -15 dBFS peak. Period of 20 mS in the waveform implies 50 Hz as the fundamental frequency.
High pass filtering for 500 Hz content similar to A weighting on the SLM:
Analyzing with a 125 mS running integration the average content which will show as peaks on an SLM set for fast are approximately -28 dBFS. This is 25 dB lower than the instantaneous kick peak demand.
Between the observation under-indication by A weighting and 125 mS integration time it is reasonable to approximate the SPL as understating peak amplifier voltage demands by at least 25 to 30 dB.
Here is a dB SPL to Voltage required chart with impedance set for 70 ohms and sensitivity to 101 dB SPL / mW for LCD-2 Classic.
At 100 dB SPL the LCD-2C requires 0.69 Vpp. But the kick peak demands at 25 to 30 dB higher will require between 12.3 and 21.9 Vpp.
Using the published power specifications for MCTH, Liquid Fire and Liquid Glass the following approximate maximum voltages at 70 ohms may be calculated:
From this chart it may be determined that the aforementioned kick attack peak voltage demands likely exceed the MCTH available supply periodically but remain within the capabilities of both the Liquid Fire and Liquid Glass at these exceedingly high sound levels.
Moral of the story: 100 dBA is VERY LOUD, especially where music with high low frequency content is concerned and an SLM is a less precise method of inferring voltage demands on an amplifier.
Separate names with a comma.