Good work on your troubleshooting! That LED must have either been bad from the start or possibly static damage during assembly. LEDs are pretty tough though. I can't remember the last time I killed one. Probably just defective.
RocketScientist is the best one to assist with the blown chip issue. I can't think of any link between the bad LED and the blown chip. All the bad LED would do is jam the power management circuit "on" all the time.
robodoc - I just went back and read your original posting again. It sounds like the LED was originally working OK, but then died at the same instant the NJM4556 exploded.
http://www.diyaudio.com/forums/head...headphone-amp-diy-project-22.html#post2803953
Interesting! Definitely will take RocketScientist to figure that one out. Even a pretty healthly voltage spike on the rails shouldn't take out the LED since it has the series current limiting resistor.
http://www.diyaudio.com/forums/head...headphone-amp-diy-project-22.html#post2803953
Interesting! Definitely will take RocketScientist to figure that one out. Even a pretty healthly voltage spike on the rails shouldn't take out the LED since it has the series current limiting resistor.
Last edited:
i'm thinking of doing a long battery-life desktop version of the O2 and wondering how limiting that would be. I know it sounds crazy but providing for a niche is always highly satisfying, and it's not truly mutually exclusive. Because the op-amps are pluggable there's no real risk any way. On that note: I was wondering how much the long-battery-life resistor recommendations might add on top of the 20-30 hours? Is it worth the bother? As it happens I wouldn't know what resistors to get anyway, but I was thinking of exploring the possibility although it would seriously limit the device, of course. But the exclusivity would be marvellous.
I have both contact and IR thermometers--both by Fluke and very accurate.
I don't know what "draws 35ma, and the regs have 15ma at the outputs" means. They should be the same number. And if your amp is drawing 35 mA idling with no batteries, something is wrong or your mods have really messed up the amp.
At 35 mA and 24 volt unregulated, you get 24-12 = 12 volts drop and .035*12 = 0.42 watts not 0.615 watts.
The junction temp rise would be, worst case, 0.42 * 60 = 25C. So even with a warmish ambient of 30C that gives 55C junction temp. By definition the case temp has to be lower than the junction temp. Your 77C is way out of line. So either something is wrong with your measurements, your regs are damaged and dumping a bunch of current through the ground terminal (very unlikely), or you need a better thermometer.
As they have posted, many others are using the O2 with the HD650 and similar headphones, including me, and are very happy with the sound and nothing gets uncomfortably hot or overheats.
I suspect one or more of your mods has messed up your amp. It may well be oscillating at ultrasonic frequencies which would explain your bad sound complaints and high current draw. I would suggest restoring it to the default "stock" build and then you should get the same great performance as the rest of us. If not, there's still an error with your build.
Yes thats the way the things were going.
But ... now with new LED and both NJM4556 exchanged to new ones the o2 works like hell .... MUSIC!!!
Still have no idea what happened with those parts, if one of them was defect or damaged by installation or something else? Now its working, so obviouslyI did no soldering errors etc. Maybe RocketScientist will find an explanation. I'm not into that electronics so much to interpret this special case.
Big big thanks to agdr and RocketScientist for the helpful hints and THANKS to all of you for assisting my troubleshooting.
The gain stage resistor change will only help battery life a small amount but it may lower distortion slightly under some circumstances. The downside is slightly higher noise. The easiest change is to approximately triple the values of R16, R17, R19, R21, R22, R23. That provides an easier load for U1 to drive. You have to change all six of those resistors.
The biggest issue is still what headphones you want to drive. If they're 50 ohms or above the low power version should work fairly well. And under those conditions, increasing R10, R11, R15, and R18 to 10 ohms will improve battery life still further as it will reduce quiescent DC offset current between the op amp pairs. The output impedance would be 5 ohms which is still fine for any (non balanced armature) headphones 40 ohms or higher.
You're welcome. Glad it's working! There are lots of possibilities as to what may have gone wrong. It might have been something as simple as a lead clipping that was laying under the PC board when you first powered the amp up. Such a clipping can create a nasty short and easily damage components with just a slight move of the board while testing or using the amp.
how does one even do that? with 2 x double soic8 adapters? without proper compensation i can only imagine what that looks like on the scopeThey don't get hot enough to "hurt." More like hot shower hot as you said lol. What would be the expected component life span of the regulators through every day use for a few hours(maybe a worst case scenario)?
I'm using the wau16-400 so those temperatures seem too be right. I have a temperature probe in my lab I could use but I'm not sure how accurate it will be. What I'm also unsure of is why did I measure a few volts over the adapter rating on it's output. Is that the "unloaded" voltage?
@dewesiuk, you are correct about the unloaded voltage. And the regulators may outlive you running at a 60C junction temp for a few hours a day. A lot of commercial gear runs parts much hotter than that. You don't even want to know what some of the junction temps are in a typical 55 inch LCD TV. Pioneer had an entire series of stereo receivers that ran the regulators so hot the PC board (over several years of the receiver being stacked with other gear and stuffed in a hot closed equipment rack or similar) degraded, turned black, and eventually the pads on the PC board failed before the regulators did.
For those wondering, I took a class on DFR--design for reliability. It was taught by the head of engineering for one the biggest DFR consulting firms in the country. He had lots of interesting stories to share.
For those wondering, I took a class on DFR--design for reliability. It was taught by the head of engineering for one the biggest DFR consulting firms in the country. He had lots of interesting stories to share.
Amazon.com: Hot Air Rises and Heat Sinks: Everything You Know About Cooling Electronics Is Wrong (9780791800744): Tony Kordyban: Books
Personally, I'd just run a 12V transformer and avoid the issue altogether. Don't see the point of overvoltage on the regulator inputs . . .
Those things are rugged bastards haha. Another thing I noticed is that the electrolytic caps beside the regulators also warm up too(I'm guessing that's either because they're quite close to the regulators or their voltage is more than half the rated 35 volts?
Last edited:
@dewardh, that's why the WAU12-200 is the primary part number in the BOM. But, for those with low line voltage, or who want to do sine wave testing of the O2, the WAU16-400 does have some slight benefits. Because there's a lot of ripple from the half wave supply under load, if you heavily load the amp or have low line voltage, the regulators fall out of regulation on the "troughs" (low points) of the ripple. This is true even when a DMM still reads a high enough unregulated DC voltage.
The good news is the O2 has such high PSRR even when the regulators drop out of regulation under adverse conditions with a 12 VAC transformer, it barely changes the measured performance of the amp and hence is very likely inaudible. But it's still not ideal.
The caps get warm just from the regulators. You can run a 79xx into thermal limiting and boil water with it and it will survive. Really.
The good news is the O2 has such high PSRR even when the regulators drop out of regulation under adverse conditions with a 12 VAC transformer, it barely changes the measured performance of the amp and hence is very likely inaudible. But it's still not ideal.
The caps get warm just from the regulators. You can run a 79xx into thermal limiting and boil water with it and it will survive. Really.
Quick BOM question - I placed my mouser order based on the BOM effective 11/16. I got an update from mouser a day or two ago that one the parts is not expected to be available until early January. (Specifically, the 270-2.74M-RC 2.74Mohms 1% Xicon Metal Film Resistor)
I noticed you have a new BOM up as of 12/2. Should I go ahead and cancel my pending order and reorder per the new BOM? I poked around a little, but didn't find anything explaining the BOM update. Was this just based on part availability? Thanks!
I noticed you have a new BOM up as of 12/2. Should I go ahead and cancel my pending order and reorder per the new BOM? I poked around a little, but didn't find anything explaining the BOM update. Was this just based on part availability? Thanks!
I'd address low line voltage at the source, since it would negatively impact everything else (except incandescent bulb life) as well. And I confine sine wave testing to the bench, where I know I have adequate power. The "troughs" are most easily addressed with bigger caps (the physical size of which should not be an issue for an AC powered application).
RocketScientist reduced R25 from 2.74M to 1.5M and R9 from 40.2k to 33k. He wrote that the old values work OK but the new values provide slightly better performance of the power management circuit in low battery conditions. So yes, nuking your pending order and re-ordering from the new BOM would be a good way to go.
AC vs. DC there.
20VAC coming in from the transformer will be somewhere around [20Vac * sqrt(2)] - 0.7Vdc[diode drop] = 27.5Vdc for a half wave rectifier, with no load, since the caps charge up to more-or-less the peak of the AC waveform. That is the DC voltage coming into the rectifier then, of course. Add a load to that and you will get somewhere around (27.5Vdc * 0.9) = 24.7Vdc for an average value, depending upon load current.
Last edited:
@DubiousMike. You don't need to cancel the entire order but you can cancel just the 2.7M resistor if you want and have a source for any resistor in the 1.2M - 1.8M range--even a 1/4 watt (stand it on end). It's not part of the audio circuit so it's not a critical part. You can also leave R9 at 40K if you want. I explained the 11/30 posts here that same day in this thread. The 12/2 changes were also explained here on 12/2 and involved increasing C6 and C7 to 1.0 uF just as a precautionary measure. I don't have any evidence there's any benefit to that change. You can use the same cap as for C1 in the C6 and C7 locations by re-bending the leads or the part in the 12/2 BOM with the correct lead spacing.
Do you not know how your own amplifier works?
Not trying to be an *** but I put some numbers up that were basically inline with other numbers. So I found your response a tad smarmy.
The reason I ask about the case temps is that if you aren't calculating the specific heat capacity for the volume of air in the case, along with the thermal conductivity of the case then you're just assuming what the temperature of the case will be.
Last edited:
Somewhere it has been previously written that the O2 is "a datasheet circuit with a Wikipedia output." So detractors have therefore chosen to focus on the relatively unconventional split supply to attack the design and designer.
CSA, it really does lean toward clueless. The reference is to the case of the regulator, the plastic part with leads and tab sticking out, not the B2 or whatever is used to enclose the amp.
I think these detractors would like nothing better than to make the O2 a banned topic here or anywhere else.
CSA, it really does lean toward clueless. The reference is to the case of the regulator, the plastic part with leads and tab sticking out, not the B2 or whatever is used to enclose the amp.
I think these detractors would like nothing better than to make the O2 a banned topic here or anywhere else.
Last edited: