Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › LNMP Finishing touches - wiring the battery pack and enclosing
New Posts  All Forums:Forum Nav:

LNMP Finishing touches - wiring the battery pack and enclosing - Page 4

post #46 of 66
Thread Starter 

OK, I made my measurements with CLED ON:

post #47 of 66
Thread Starter 

 

 

OH! I'm an IDIOT!

I figured it out. When I did this test:

 

Quote:

If the above tests pass, measure DC volts across R14. You should get 1.25 V

 

 was measuring like this:

 

 

Instead of measuring the difference between each side and V- like this:

 

 

 

Then getting the difference, which is 1.4V, which I figure is close enough?

 

 

YAY, Tangent! Thank you for walking with me until I realized my mistake! Uh Oh. Now on to the testing.

 

But first, I'm going to show the folks how I drilled the panels and my test point.

post #48 of 66
Thread Starter 

I didn't wait to listen to your caution about the test point, and since I'm crazy about test points, I put one in before you posted. To review:

Quote:
Originally Posted by saraengelstad View Post

If I were to drill that hole to place a test point, I was thinking the place to do it would be ~2cm to the left of C1, making sure I only hit the bottom trace, and not the top one.

Still a bad idea?

 

Quote:

Originally Posted by Tangent

Yes, I think so. After drilling the hole, you have to work out how to scrape away enough solder mask to get a solid joint. You can't just bond to the hair-thin rim of copper you expose with the drill. That's not enough for a strong bond; certainly not strong enough to be hanging test probes off of. Even if you did get a good electrical connection, you've increased the resistance of the V- trace, since solder isn't as good a conductor as copper.

 

Me, I'd just jam a probe into some likely place, like the elbow of the RLED lead.

 

If I absolutely had to create a hard test probe point on the LNMP v1.1, I'd probably surface-mount a wire loop test point to W- or B-.

 

You can also alligator-clip to the solder lugs on the DC input jack. One of those is V-.

 

Here's the result, any feedback (besides I should have waited for your answer?)

 

 

Drilled trace:

 

 

Should I put some conformal over the scratches, or will the extra solder protect the bare copper?

post #49 of 66
Thread Starter 

 

 

 

Drilling the end plate (again)

 

Here's how I drilled the rear end plate. I thought someone might appreciate seeing how I did it, since drilling holes in metal in the right places was so daunting to me.

 

First, I measured and recorded the inner and outer dimensions of the switch, the nut that holds the switch on, and the power inlet.

You can get this data from the data sheets, but I like my calipers. I got them on Amazon for $25 and they have non-titanium versions for only $15.

 

http://www.amazon.com/Neiko-01407A-Extra-Large-SAE-Metric-Conversion/dp/B000GSLKIW/ref=sr_1_1?ie=UTF8&qid=1360806207&sr=8-1&keywords=calipers

 

 

 

 

 

These are the tools I used to do the work, the calipers, a scriber to mark the metal, and a marking T rule.

http://www.amazon.com/Incra-T-RULE12-12-Inch-Precision-Marking/dp/B00004TRBX/ref=sr_1_9?s=hi&ie=UTF8&qid=1360806314&sr=1-9&keywords=t+square

http://www.amazon.com/General-Tools-88CM-Tungsten-Carbide/dp/B00004T7S1/ref=sr_1_1?s=hi&ie=UTF8&qid=1360806366&sr=1-1&keywords=scriber

 

 

 

I used the scriber to mark the midline, then I used it to mark little circles where I wanted the holes to be centered (see below).

Then I used my center punch (the red tool) to make an indentation for the drill so that it doesn't walk around when I start the holes. Amazon, again.

 

 

Then I put the old plate with the bad holes on top and compared them. I wanted the power button to be a little to the right in the new plate and made sure there was enough room for the power inlet.

 

 

 

I purposely set the switch hole above the midline because with the smaller case I started with, I needed to do so to avoid hitting the board.

Now that I don't have that problem, I should have stuck to the midline. The plastic nut that holds the switch on is awful too close to the bezel:

 

 

 

 

Both holes are on the left side of the enclosure to avoid striking the battery pack.

 

 

 

In order to drill the holes, I followed a mixture of TomB and Tangent's advise. Putting a garbage can did indeed help avoid liberally sprinkling my kitchen floor with tiny metal shavings.

For some reason, the black plates mar much more easily and noticeably than the silver/clear ones, so even though my dremel vise has soft gripping surfaces, I still used a flour sack towel to avoid scratching the paint. Here's what drilling looked like:

 

 

After driling most of the holes with my step bit, I flipped the plate over and finished from the 'pretty' side to make sure that it would turn out smooth. I think I remember something like this from shop class, but that was in the seventh grade. That was a while ago for most of us, I bet. Flipping them also gave them a nice silver outline, not that they can be seen later, but I know they look cool underneath.

 

Then I dropped the switch and inlet in to make sure the holes were the correct diameter.

 

 

YAY!

 

Then, since the black plates seem to beg picking up imperfections, after drilling, I gave it a little polishing rub with a microfiber towel. I also do this to the botttom of my boards after cleaning them so that they shine and I can detect anything that was missed during cleaning.

 

 

 

 

Finally, since I still have testing to do, I left the top of the enclosure out, but stuck the new rear plate on with the plastic nuts that Hammond includes.

 

 

Looking pretty OK, I think!

 

 

Next step... Generating a signal and checking the amp's basic functionality.

SCARY!

I hope that someone gets something out of reading this, thanks for letting me share.

-s

 

post #50 of 66
Quote:
Originally Posted by saraengelstad View Post

I always socket everything. So desolder the sockets and solder the PS ICs in directly? Could that be my problem???

 

It's fine to socket the charge controller.

 

I was thinking more about the LM317 when I said the ICs should be soldered.

 

The charge controller definitely needs to be in its socket when you test the charger.

 

Quote:
Originally Posted by saraengelstad View Post

OH! I'm an IDIOT!

I figured it out. When I did this test:

 

 was measuring like this:

 

 

Instead of measuring the difference between each side and V- like this:

 

 

Then getting the difference, which is 1.4V, which I figure is close enough?

 

No, you've still got a problem, though I'm becoming convinced the problem isn't in the LNMP.

 

You were doing the test right to begin with, with the probes across R14. The only reason I suggested measuring everything against V- in the "schematic test" is so you have a common reference, so you can compare many values directly.

 

That said, your two measurements shown above should have given exactly the same result. The fact that you get different results means the problem must be in either the equipment or technique.

 

You're measuring DC volts here, not AC?

 

This isn't one of those auto-selecting meters, that only has a "volts" setting and guesses whether you mean AC or DC?
 

I see Fluke probes. Care to disclose the meter model, too? Or better, take a pic of the meter face with the measurement showing, while the probes are across R14?

 

You can see more of this weirdness in your schematic test. There, the drop across R14 is 0.04 (3.08 - 3.04), yet somehow you manage to develop nearly 7.3 V going forward across D4?!

 

The 10.7 V measurement is plausible. The 3.0x V measurements are not just implausible, but in fact physically impossible.

 

Quote:
Originally Posted by saraengelstad View Post

I didn't wait to listen to your caution about the test point, and since I'm crazy about test points, I put one in before you posted.

 

Here's the result, any feedback (besides I should have waited for your answer?)

 

You remember how, just a few days ago, you were praising the durability of my boards, I believe in comparison to RS perfboard? In part, that's because they're double-sided. In order to tear off a double-sided board trace during desoldering, you have to also tear off the solder mask, the thru-hole plating, and the trace connected to the plating on the other side of the board. You've removed all that in making this test point.

 

It's kind of like pitching a dome tent on a slab of granite: sure, it stays up now, but what happens when the wind comes up?

 

Your test point is probably a bit stronger than one attached the same way to RS perfboard, owing to thicker copper and probably better manufacturing quality, but not nearly so much as the proper thru-holes on that board.

 

At least if you surface-mount it to W- or B- (or better, put it between these two points) you have most of the other parts of the board anchoring your test point.

 

Quote:
Originally Posted by saraengelstad View Post

For some reason, the black plates mar much more easily and noticeably than the silver/clear ones

 

I think it's just the same reason black cars get dirtier faster than white ones. The contrast makes it easier to see the same flaws.


Edited by tangent - 2/14/13 at 5:50am
post #51 of 66
Thread Starter 

OK, here's my report. I think it's fixed. I bought 700 mAh batteries and the power supply on your build list.

 

First, here's my meter:

 

Second here are my measurements:

 

1)  V- to S3B, PLED OFF: 0V

     V- to S3B, PLED ON: 10.3V (See below)

 

 

2)

PLED OFF, IC6 measurements:

Pin 1 - 3 = 5.14V

(see below)

 

 

PLED OFF, IC6 measurements: (CLED is flashing)

Pin 1 - 2 = -5.14V

(see below)

 

 

 

 

 

 

 

 

3) IC5 Tests:

PLED on, Wall supply plugged in, DC Volts from S2B to V- =  Oddly, -.32V... Capacitor thing??? Weird.

 

 

PLED Off, CLED flashing. S2B to V- = 18V (See below)

 

 

 

4) Voltage across R14: 1.25V!!!!

 

 


So.... good, right?!?!!!?

gs1000.gif

-s


Edited by saraengelstad - 2/15/13 at 5:10pm
post #52 of 66
Thread Starter 

Now. I read your article, but I'm at a loss how to generate a signal..... and do I measure the output it with the fluke, or my scope?

 

If so, how? The bnc connectors are throwing me, if I hook the output to my scope, do I need a bnc-bnc cable??

 

If I use my scope, do I just stick the probe into the bnc center, or make/buy myself a cable that terminates bnc?

post #53 of 66
Quote:
Originally Posted by saraengelstad View Post

OK, here's my report. I think it's fixed. I bought 700 mAh batteries and the power supply on your build list.

 

In that case, I'm guessing was that the problem is that your previous power supply wasn't isolated. The new one is — just checked the data sheet — and this matters in your particular build somehow. I have vague ideas why, but I'm too tired right now to think this through clearly.

 

If it wasn't that, then the only other possibilities I see are:

 

  1. bad power supply
  2. measurement error
  3. bad AAA cells

 

None of these feel especially convincing to me. I'd only bother considering them if you test the old power supply and find that it's isolated, too.

 

Quote:
Originally Posted by saraengelstad View Post

I'm at a loss how to generate a signal.....

 

 

I thought step 10 of the assembly guide was pretty clear. Test tone CD, voltage divider.

 

Quote:
 and do I measure the output it with the fluke, or my scope?

 

 

The meter.

 

You use a scope to get a picture of a fast signal, sacrificing precision. DMMs trade speed for precision. Instead of capturing thousands or millions of samples per second, as a DSO does, most DMMs only update their display every half second or so.

 

That's not to say you can't use your scope with the LNMP. But we're talking about calibration here.

 

Once you've got it calibrated to your satisfaction, plugging the LNMP's output into the scope's input can be useful. It is sometimes very helpful to get a picture of the noise your power supply is putting out, so you can see patterns. Does it look like 120 Hz hum, for example? Or are there spikes in there? Etc..

 

 

Quote:
If I use my scope, do I just stick the probe into the bnc center, or make/buy myself a cable that terminates bnc?

 

I wouldn't build such a cable. Coax isn't especially easy to terminate with anything other than coax connectors.

 

Instead, leave it to the pros: search for "Pomona BNC banana" at your favorite electronic parts distributor. Marvel. Buy. Use. Enjoy.


Edited by tangent - 2/17/13 at 12:09am
post #54 of 66
Thread Starter 

I reread your instructions on generating the signal, and this is my idea. Tell me where my reasoning starts going sideways.

 

If I burn a CD ROM, the only thing I have to put it into is my PS3, which I use for Blu-Rays because it has self-updating firmware. To get the sound out, I'd have to dig up a headphone cable adaptor and plug my cable into my reciever, which is some distance from my work area.

 

So, instead, I'm using NCH Tone generator to generate a 1 kHz sound. I'm plugging in a half-terminated headphone cable into the laptop and connecting that to the DMM. I get about 3.7mV, which I can adjust up and down using the volume control. The blue wires are the ground, and the white wires are to the tip and the sleeve of the neutrik 3.5mm connector. Here's what it looks like, all rigged up:

 

 

Then I'm going to jumper SW1 to set the LNMP 60dB calibration. Will this be adequate for the task?

 

Sir, you weren't kidding that making your own cables wouldn't be any fun. I picked up a little pack of BNC connectors, and even with the proper compression tool, the effort was a waste.

Total... Unmitigated.... Epic Fail.

 

So I bought these from Amazon. BNC - alligator for the headphone wire - LNMP connection, and same for LNMP - DMM connection:

http://www.amazon.com/gp/product/B0002JJU4G/ref=ox_ya_os_product

 

And a 75 ohm BNC-BNC to hook the LNMP to the osscilloscope for the pretty pictures:

http://www.amazon.com/gp/product/B0002J2N0Y/ref=oh_details_o00_s00_i00?ie=UTF8&psc=1

 

I tried making a connection to the scope the other night, but the guy in the room said that what I was seeing on the screen was about as meaningful as the output from an etch-a-sketch.

He uses a scope regularly in a scheme I poorly understand but has something to do with meterology, satellites, and large, fast-moving chunks of metal.

 

Thanks again!

-sara

post #55 of 66
Quote:
Originally Posted by saraengelstad View Post

I'm using NCH Tone generator

 

 

That's fine. The idea with the CD player is probably becoming fairly outdated these days. Computer sound cards and MP3 players are replacing them, but the argument in favor of using them is the same, and the basic method is the same, so I haven't bothered updating the text.

 

If you try the MP3 player idea, you don't actually want to use an MP3. You'd find a lossless format the player will accept. Apple Lossless for iPods, FLAC for many other devices, etc.

 

Quote:
I get about 3.7mV

 

Here's where you're off the rails. You've exceeded a front page spec: the maximum input signal to an LNMP running from a 10 V power supply is 10 mV on the 40 dB setting. But, you also need to test the 60 dB setting, so for that the maximum input signal level is 1 mV.

 

Also, by making your computer put out such a small signal, you're contaminating it with noise and nonlinearity. Your computer will put out the cleanest signal at its maximum output level. So, increase the amplitude as far as it will go in your sig gen program, and increase the headphone out volume level as far as it will go.

 

You should find that the signal level goes to 1-2 V. So, you then need to build a ~1000-2000x voltage divider to drop that cleanly down to 1 mV or a bit less. You want to measure the voltage level while the computer + divider lash-up is connected to the LNMP. You want the input to the combined system to be < 1 mV, so that when you measure the output, the only thing you're measuring is the effect of the amplifier. General principle at work here: change only one thing at a time.

 

And no, I'm not going to tell you how to calculate the divider, or build it, or hook it up. This is electronics Chapter 1 stuff. I will suggest that you start with a total resistance somewhere around 1K, however.

 

You will probably find that the divider doesn't give the voltage output you expect when the whole system is hooked up. The output impedance of the headphone amplifier in your laptop isn't zero, and the input impedance of the LNMP isn't zero infinite. So, you may find yourself recalculating the values to give a more useful signal level after seeing what your original calculation did.

 

The lower the divider resistance, the less noise it will add, but the more load it will place on your PC's headphone amplifier, increasing its distortion. This resistance is in parallel with the LNMP's input impedance, which is 100 ohms, which is why I suggest ~1K for the divider: so the divider doesn't contribute much to the overall resistance, yet is low enough to not add much noise.


Edited by tangent - 2/19/13 at 12:27am
post #56 of 66
Thread Starter 

I'm obviously doing something wrong, because, even with a headphone amplifier that plays that horrible, mind killing tone much louder in my headphones than a human could possibly bear, the most current I can generate is 24 mV.

 

Seriously, I can cause this noise to be so loud in headphones that I can hear the headphones in the next room. I'm obviously dearly missing something.
The only way I measure VOLTS is to get out a 1,2V battery and measure that. Which is fine.

 

Unbelievably, I can construct a voltage divider, BUT I AM TOO STUPID TO FOLLOW the simple instruction to "apply a signal"

 

Nothing I can find tells me how, beyond stripping the ends of of my sennheiser cord and JAMMING them, with icepicks and a hammer into the bananna jacks on my meter, tells me how to get the noise from an amplifier into my DMM, or anything else, for that matter. And I've no doubt that this method would also fail.

 

I feel that I am so dense that I could be used reliably to shield against neutron radiation. Because I am a girl, I feel that it is perfectly okay to admit that I am going to go cry now.

...

Having blown several hours this morning trying various things and having had a very good cry... I am considering giving up and throwing all my work in the hot tub.


Edited by saraengelstad - 2/19/13 at 6:45am
post #57 of 66
Quote:
Originally Posted by saraengelstad View Post

I'm obviously doing something wrong, because, even with a headphone amplifier that plays that horrible, mind killing tone much louder in my headphones than a human could possibly bear, the most current I can generate is 24 mV.

 

...

Having blown several hours this morning trying various things and having had a very good cry... I am considering giving up and throwing all my work in the hot tub.

 

This was originally at the end, but I think you should read it before my lunatic rant:

 

VERY very few people ever use more than 0.5v with conventional headphones - even high impedance ones! 24mV sounds totally reasonable if you are using fairly low impedance headphones. I cant normally measure how little voltage I actually use to run headphones. This is of course quite conditional on how loud you listen, and which headphones you have, and such, but still. 

 

Don't throw your toys in the tub, you deserve the relaxation more than they do. 

 

 

(brace yourself for lunatic ranting)

 

My lunatic rant was sure to get me banned so I didn't post. Trust me, it was great. 

 

In short - the guru is wrong. The mega-whoa-crazy voltage swings they assure you you need from your amp are a load of crap. 

post #58 of 66
Quote:
Originally Posted by saraengelstad View Post

the most current I can generate...

 

Voltage.

 

 

Quote:
...is 24 mV.

 

AC or DC? This is a sine wave, so you should be measuring AC voltage.

 

Your result makes sense if you're measuring DC, since that's just the DC offset of the laptop's headphone output.

 

If the sig gen's output is turned all the way up, you need to check the Windows mixer, and make sure all the sliders are all the way up there, too. You get to it with a right-click on the little speaker in the Windows system tray. Or, you can dig it out of the Control Panel.

 

Quote:
Originally Posted by nikongod View Post

the guru is wrong. The mega-whoa-crazy voltage swings they assure you you need from your amp are a load of crap. 

 

We're not talking about a headphone amplifier here. An LNMP is a low-voltage measurement device capable of 100-1000x gain in its standard configuration, with a ~1V maximum output. Therefore, in order to reduce noise and distortion, we want to put in the biggest signal we can get away with, that being 1 mV. But, it's better to scale a big, clean signal down with a voltage divider than turn down the volume on a DAC, since that just pushes the data bytes into the noise floor.

post #59 of 66
Thread Starter 

A rant that would get you banned? Wow.

 

Thank you. The problem was I was measuring DC. Habit. I hardly ever have need to measure AC. Now it's behaving the way that I expected it to. When the tone is running, I get 1.14V AC. When the tone stops, that measurement drops to near zero. That's what was scaring me, the reading didn't seem to have any relation whatsoever to the presence or lack of tone.

 

OK, building my divider so that I can get the signal down to a useable level. Pre-divider, my measurement is 1146 mV.

post #60 of 66

Good to hear it.

 

Don't be afraid to adjust the voltage a little with the Windows volume control or the sig gen amplitude control. That is, if you end up finding that the LNMP is clipping a little bit with your guess at the correct divider resistor values, don't rebuild the divider, just drop the input voltage a little.

 

Full-scale output is just an ideal.  Mainly what we're trying to avoid here is getting mislead by the noise floor and inherent distortion level of the laptop sound card. Dropping the voltage by a few dB isn't going to hurt us much here. It's when you try to drop it by 60+ dB that you run into trouble.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › LNMP Finishing touches - wiring the battery pack and enclosing