Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › M³ amplifier build discussions thread
New Posts  All Forums:Forum Nav:

M³ amplifier build discussions thread - Page 9  

post #121 of 828
3/12/05 at 6:06pm
That depends on the thermal compound. Some thermal compounds are electrical insulators, others, well, aren't. Honestly, I'd say that the safest solution is just to avoid using too much thermal compound. After all, temperature-wise, less is better; the goal is to fill in the tiny cracks and gaps on each surface, not to make a sandwich. If you intend on using very little but just want to be as careful as possible, you could always look for a compound that's specified as an electrical insulator.
post #122 of 828
3/12/05 at 8:49pm
I have tried several thermal compunds on computer CPU's, and can recommend Arctic Silvers Ceramique. It is easy to work with compared to most others and it is non-conductive and has very good thermal characteristics.
On top of this it is relatively cheap and easy to find on computer stores that sells CPU heatsinks.
post #123 of 828
3/12/05 at 11:29pm
Quote:
Originally Posted by amb
It's whatever the RMAA software does. If you're not familiar with the RMAA software, you simply press a button and it performs a series of tests automatically, analyzes the data and displays the results. I don't know specifically what output level it performs each tests in, and it's difficult for me to measure them because the test tones are complex and short in duration, but I'm sure it's nowhere near full output.
Guess Rightmark is still in it's infancy stage... Normally 10mW to maintain a good THD reading is required...
post #124 of 828
3/13/05 at 1:59am
Thread Starter 
Quote:
Originally Posted by jasonhanjk
Guess Rightmark is still in it's infancy stage... Normally 10mW to maintain a good THD reading is required...
10mW is a power output figure and it varies with load, given a certain output voltage: P = V^2 / R

I run my tests with 33 ohm dummy loads at the output of the amp, for 10mW into 33 ohms it works out to just under 0.6Vrms. On the other hand, with a 330 ohm load the amp will be making only 1mW at the same voltage, and have to output over 1.8Vrms to make 10mW. RMAA doesn't vary the gain based on load, so the 10mW figure is a bit nebulous.

At any rate, I think the M³'s 0.001% THD result is pretty good whatever the level is .
post #125 of 828
3/13/05 at 3:32am
True. MMM is really good...

Since adjusting Rightmark volume would create more problem, there is another way of doing this. By varying the load resistor, you can get the required testing 10mW. An osciloscope is needed to measure the -1dB voltage first though.

If it can reach 100mW, guess what. You guys should form a company and make yourself rich...
post #126 of 828
3/13/05 at 4:13am
Once you start doing it for money, it is no longer fun, plus it is hard to make much money selling headphone amps.
post #127 of 828
3/13/05 at 4:22am
Thread Starter 
Morsel is absolute right. It's fun when it's not a job.

BTW the M³ has no trouble producing 100mW of output. In fact it could deliver a few watts of power into an 8 ohm speaker if a capable power supply is used. With efficient speakers you can get a roomful of sound.
post #128 of 828
3/13/05 at 4:51am
Ummm. Yes and no.

Let's see how MMM goes a few months later.
post #129 of 828
3/13/05 at 2:23pm

Disaster... redemption?

Quote:
It's fun when it's not a job.
There are other times the fun comes to a screeching halt...

The M³ was fully assembled, and entering final check.

STEPS putting out 32.0v - check.
Resistance between power rails - check.
Resistance between pins 4 & 7 (and each to O) - check.
OPA627 inserted into G, and power on - 29.9v between 4 & 7 - check
Voltage across R5G - 25.2v... uhhhh
Voltage across R5G with BIASG maxed - 22.4v (or so) - FUBAR.

Immediate power down, and check R5G - 99.9.... K. ICK.

Simply cannot explain - no excuse.

OK, so all R5 and R8 were immediately replaced (carefully) with 100 ohm, but I lifted (burned off if more appropriate) a pad on R5R, one on R5G, and one on R8G-, even exercising extreme care.

Not surprisingly, power rails have a short. The good news is that the TP just below C5L (as well as pin 4 and Pin 7) to OL is not shorted. Same for C5R, pin 4, and pin 7 to OR. But, each is shorted on G.

Are there any additional troubleshooting tips to determine whether the problem is with R5 or R8, or both?

What is the workaround for the abominated pad?
post #130 of 828
3/13/05 at 2:43pm
Thread Starter 
Seems like you're using too hot a soldering iron (and/or heating the pads for too long) to be doing damage like that. What are you using?

Also, check carefully for solder bridges. The R5 and R8 resistors are not anywhere near signal ground so they're probably unrelated.
post #131 of 828
3/13/05 at 3:19pm
Yes, I was using the RS Solder Station on the 40w setting to uninstall...

Could this be from a toasted component? And why just the G side? Flux has been removed, but I'll look again for solder bridges.

Edit: No solder bridges - opinions about how to procede?
post #132 of 828
3/13/05 at 5:35pm

I'm done!

Got my lunchbox M³ finished and it sounds great.

Thank you to Team M³!

Pics here: http://www6.head-fi.org/forums/showp...&postcount=378
post #133 of 828
3/13/05 at 6:40pm
Thread Starter 
Quote:
Originally Posted by phobus
lunchbox M³
Very cute!

Quote:
Originally Posted by pabbi1
Could this be from a toasted component? And why just the G side? Flux has been removed, but I'll look again for solder bridges.
Edit: No solder bridges - opinions about how to procede?
When you say that there is a short from a power rail to "G", are you referring to the power rail that feeds the MOSFETs (before the capacitance multiplier) or the rail that feeds the opamps? Which rail? And where is "G" exactly?
post #134 of 828
3/13/05 at 7:25pm
G, as I called it, is the ground channel, as in RxG, QxG, CxG, QxG.

Ok, I need to start from the beginning. When I measure between V+, and V- (on a Molex connector), I get a "1" on the 2000K scale. Immediately.

When I measure the 4 & 7 pins to IG, I get resistance (takes a while to interpolate down).

What should I be measuring to test the V+ V- issue?

Please disregard my previous spurious "guesses".
post #135 of 828
3/13/05 at 9:36pm
Thread Starter 
Quote:
Originally Posted by pabbi1
Ok, I need to start from the beginning. When I measure between V+, and V- (on a Molex connector), I get a "1" on the 2000K scale. Immediately.
Try a lower resistance scale on your meter, like 200 ohms. The measurement should read something across V+ and V- due to charging up the bank of C7 capacitors, but should slowly go to "infinite". In fact, instead of measuring across V+ to V- on the connector, measure between the N-channel and P-channel MOSFET mounting tabs (this corresponds to the drain pins of the MOSFETs). This takes D1 out of the equation. If you still get a low and constant resistance reading, then your MOSFETs are probably toast on that channel. You can also verify this by measuring resistance between the drain and source pins of each MOSFET.

I am a little concerned because you had your BIAS pot maxed at one point. There should almost never be a need to go to max on that pot. You should always start from the minimum (fully counter-clockwize) position. Also, before you even fiddle with that pot, you need to get the BIASOP pot set properly to establish an 5mA current through the Vbe multiplier (start with the BIASOP pot set to middle). This is all spelled out step-by-step on my site.

Quote:
When I measure the 4 & 7 pins to IG, I get resistance (takes a while to interpolate down).
You're seeing the charging effect of the C5 capacitors. With the opamps removed from sockets, this should also go to "infinity" on a low resistance scale.
New Posts  All Forums:Forum Nav:
  Return Home
This thread is locked  
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › M³ amplifier build discussions thread