News:

Forum may be experiencing issues.

Main Menu

Cupcake smoke

Started by Ang3lus, March 09, 2015, 09:51:49 PM

Previous topic - Next topic

mgwhit

#15
Quote from: Ang3lus on March 15, 2015, 09:20:14 PM
22UJ100

but the package i got it in says 100n

I actually decided to look at your photos.  Duh.  It really says µ22J100, and, yes, the position of the µ is important -- it's like a decimal point.  That means it's 0.22µF, or, as you guessed, 220nF.  J = 5% tolerance, and the 100 means it's rated for up to 100V.

Rob, thanks for the op amp information!

RobA

First thing to note is that mine is the older 2012 version so it doesn't have the power supply filtering resistor or extra cap and the polarity protection is the crowbar type. None of that should really matter and I stuck a 10Ω resistor in front of the power supply input to the PCB so I could check the current draw, so that's more the same now.

Here's my voltages:
Power Supply 9.05V
JRC4558
1  4.97V
2  4.97V
3  4.87V
4  Gnd
5  Gnd
6  8.48V
7  8.48V
8  9.02V

So, so much for my saying I hadn't seen the JRC4558 latch up this way. The one on my board is actually latched up. But given that, it is stable and not oscillating and the total current draw of the board is still only sitting at about 2.7mA. So, it isn't causing any issues. That could be different with other op amps though. For example, I've seen an NE5532 draw a ton of current on the breadboard when it latched up.
Affiliations: Music Unfolding (musicunfolding.com), software based effects and Rock•it Frog (rock.it-frog.com), DIY effects (coming soon).

RobA

The JRC4558 latching up in the circuit got me intrigued, so I set up my breadboard with one op-amp of the dual with the input at half supply using two 10k resistors and the other half with the input connected to ground. I grabbed a handful of op amps out of the bag popped in there and took the measurements of the output voltages and the total current draw. I then moved the grounded input to half supply using two 100k resistors and took the current measurement again. The table below shows the results.

The NE5532 really doesn't like its input at ground in a single-supply setting. It oscillates and draws a bunch of extra current. The OPA2134 and OPA2604 are either oscillating or on the verge of it. The TLE2072, TL072, LM833, TL062, NJM4558, MC33078, NJM2082, and NJM4580 all latch up, but they don't seem to draw any extra current because of it. The LM1458 is weird sitting at about 2.0V and maybe oscillating but it still draws on the low end of the current. The LM358 handles the voltage really well and has a low current draw, but it might sound like crap. The TLV2372 is similar but it's optimized for really low voltage power rails and not really aimed at audio at this power supply voltages. It could sound good here though. The TS912 and TS922 look good as they always do and might be a good choice. The TLC272 and TLC2272 both did really well and would be good to look at if you were set on using a FET input op amp. The one that was kinda surprising to me was the NJM2904. It didn't latch up at all and used very little current. It'd probably sound pretty good in there.

If I were going to build a new one, I'd try the TS912. They've sounded really good everywhere I've tried them and their performance in the latch up test here was perfect. Too bad they've EOLed the PTH versions.


OPA       GND(V)     Half(V) Current(mA) Current with both at half supply (mA)
LM358   0      4.5      0.6      0.5
LM358   0      4.5      0.6      0.6
LM1458   1.97      4.5      0.87      0.8
TLV2372   0      4.5      0.9      1.2
JRC2082   8.9      4.5      2.9      4.1
OPA2134   0.137   4.5      8.7      8.1
OPA2134   ~0.220   4.5      8.6      8.2
TL072   8.46      4.5      2.8      3.3
TL072   8.46      4.5      2.7      3.2
TS912   0.035   4.5      0.5      0.6
TS922   0.021   4.5      2.4      2.2
LM833   8.48      4.5      2.5      3.9
LM833   8.48      4.5      2.7      3.9
NE5532   ~ 1.135   4.5      ~ 18.0   6.3
NE5532   ~ 1.135   4.5      ~ 21.0   6.2
TLE2072   8.33      4.5      2.7      3.1
TLC272   0      4.5      0.8      1.4
TLC2272   0      4.5      1.6      2.2
MC33078   8.41      4.5      2.6      3.3
JRC4580   8.48      4.5      3.1      4.7
JRC4558D   8.5      4.5      2.4      3.5
JRC4558D   8.5      4.5      2.4      3.5
TL062   8.47      4.5      0.2      0.3
JRC2904   0      4.5      0.7      0.7
OPA2604   ~1.0      4.5      9.4      10.0
Affiliations: Music Unfolding (musicunfolding.com), software based effects and Rock•it Frog (rock.it-frog.com), DIY effects (coming soon).

mgwhit

#18
Congratulations, Rob!  When the Madbean wiki is created, you will have already written the entry for "Unused Op Amp".  ;)  Seriously, I'm filing this stuff away.

Quote from: Ang3lus on March 14, 2015, 08:07:29 PM
Took some readings with the multimeter

from pins 1-8:

3.34v
3.3v
3.315v
0v
5.93v
5.78v
5.73v
0v

that don't seem right ?

Ang3lus, looking at what Rob and Jon have added here, it occurs to me that your voltages make total sense (a.) if you're using a depleted battery, and (b.) if you counted the pins incorrectly.

You have voltage on pin 5 where you should have 0V (ground) and you have 0V on pin 8, where you should have your highest voltage (supply).  This is exactly what I would expect if you were counting pins from top to bottom on both sides.  You should count pins starting at the top left pin (1, just left of the notch) going down and continuing counter clockwise from there.  In other words, you count down the left and then up the right sides of the IC.

(Jon obviously figured this out, but didn't mention the pin counting error.  I was too stupid to catch that the first time around.)

Ang3lus

You are corret.

I just came to post that i stupidly mis-counted the pins (i SWORE it woulden't happen to me, huge embarrasment)

the correct voltages in the correct pinout is

pins 1-8:

3.34v
3.3v
3.315v
0v
0v
5.73v
5.78v
5.93v

I'm using a power supply, not a depleted battery, could the 47uf cap or 1n5817 be defective and not rectifying correctly ?

cause i do get 9v at the source.

i'm guessing the 1n5817 is for polarity protection only ?

RobA

I missed the pin counting too. Those readings do make more sense now.

The 1n5817 is for polarity protection. Get a measurement of the voltage drop across the 1n5817 and across R12.
Affiliations: Music Unfolding (musicunfolding.com), software based effects and Rock•it Frog (rock.it-frog.com), DIY effects (coming soon).

mgwhit

I agree with Rob.  Check across those two components.  The drop across D2 should be less than 450mV.  The drop across R12 should really only be about 50mV (assuming 2.5mA/op amp and no R13+LED).  If the drop across D2 is high, then, yeah, it's a bad diode.  If the drop across R12 is high, you've either got an incorrect (high) value in R12 or you have a short that is causing excessive current draw.

Ang3lus

checked the entire circuit with a good brand new multimeter today.

coulden't find anything.

until i checked the grounds and found out i had one ground on the switch loose, it had an intermittent connection, soldered it up well now and is all good. voltages are as expected.

serves me right for being dumb and lazy and not wrapping the wires around the switch lugs.

thanks for all your help guys, i think i would have given up on this if it weren't for you.

mgwhit

Cool -- glad you found it.  Have fun!