What DC voltage do you get from the right output to ground (metal shell of the gain switch) as the volume control goes from low to high and back to low? Leave the input jack unplugged so it is grounded and no need for headphones to be plugged in. Try the same measurement too with the left channel output.
The problem - Channel inbalance in medium volume. Either low volume or high volume works ok, but anywhere in between is off-balance.
What i've tried - Changing all the mosfets, changing the potentiometer, changing the gain switch, changing the gain resistors, changing input jacks, and I did an off mount 1/4'' jack to confirm it wasn't the output either.
When I touch my plairs from a gain resistor to the gain switch, it works. I have tried connecting the gain switch directly to ground (figuring that would solve the problem) but it did not.
What i've tried - Changing all the mosfets, changing the potentiometer, changing the gain switch, changing the gain resistors, changing input jacks, and I did an off mount 1/4'' jack to confirm it wasn't the output either.
When I touch my plairs from a gain resistor to the gain switch, it works. I have tried connecting the gain switch directly to ground (figuring that would solve the problem) but it did not.
If the DC output on the right channel doesn't change significantly from around 3mV as the pot is turned then the problem isn't output stage bias related as I was thinking it might be. However, even if the DC output voltage looks OK, you still might want to try swapping the two NJM4556 output chips if you haven't already. If the input of one has essentially shorted it could be pulling significant AC current (through C14, assuming it isn't shorted) and hence through the pot wiper. That might cause some funky effects as the pot is rotated. If the problem moves to the other channel when you swap the chips then it is a bad output chip.
The only other thing that comes to mind is the gain switch shorting issue:
http://www.diyaudio.com/forums/head...headphone-amp-diy-project-14.html#post2774393
But I don't see how that would cause the channel imbalance effect in the middle of the pot, plus you have changed the gain switch.
Last edited:
That sounds like a potentially terrible idea.
Send me a PM with your mailing address and I'll send you an envelope with a more compatible cap in it. Maybe not the same as the BOM, but at least a mono ceramic.
Sent from my DROID4 using Tapatalk 2
OK, thanks for being generous, I appreciate it. PM sent.
I looked around, and couldn't really find detailed info suggesting a tantalum would not work for C1. Actually, I can't tell exactly what C1 is for. But my beginner mind was guessing since it's in the (linear) power management section the slight difference in ESR and frequency response wouldn't matter much, AFAIK. So, what is the difference between the tantalum and the monolithic ceramic, in this particular part of this circuit?
Thanks,
Keith
I looked around, and couldn't really find detailed info suggesting a tantalum would not work for C1. Actually, I can't tell exactly what C1 is for. But my beginner mind was guessing since it's in the (linear) power management section the slight difference in ESR and frequency response wouldn't matter much, AFAIK. So, what is the difference between the tantalum and the monolithic ceramic, in this particular part of this circuit?
Thanks,
Keith
I'll take a guess that the negative voltage regulator (U6) is dead if your DC supply was tip-positive and that U5 is dead if your adapter was tip-negative.
The reason is that RocketScientist probably should have included reverse-biased rail-to-ground 1N4003 diodes on each rail. I've been meaning to post that in the modification thread for awhile. The diodes would clamp any static voltage on the opposing rail to 0.7V in the event the supply (or battery) on one rail suddenly disappears, which is what happened with your DC supply. If your DC adapter was tip-positive, then it would have forward biased D3 and lit the positive rail with the negative rail left floating by the reversed biased D4. The positive regulator would have gone into dropout, with only 12Vdc into a 12Vdc regulator, but would have passed what it could on to the positive rail, probably in the +8Vdc to +11Vdc range.
The trouble comes when rail-to-rail parts like the U2 op amp, R5/R9 voltage divider, and D7 power LED, leak positive voltage onto the (now floating) negative rail, forming a static (very low current capability) positive voltage on the negative rail that otherwise was left floating by D4 being off. RocketScientist did a very good job keeping the current draw of the parts before the mosfets low to reduce battery drain (the mosfets would be off with only one rail lit), meaning there is not much rail-to-rail bleed resistance, allowing that static voltage on the opposing rail to remain fairly high. I would guess maybe +8Vdc-or-so on the negative rail.
And that would put a reverse voltage on the negative voltage regulator. Its output pin would be at +8Vdc with respect to ground (initial conditions of the uncharged C3 and C5 left floating), which would pretty much kill the regulator instantly. They don't like being reverse-biased, even for an instant.If a reverse biased diode had been in place between ground and the negative rail it would have clamped the static reverse voltage across U6 to -0.7Vdc, which would have greatly upped the odds that it would have survived.
You can confirm the diagnosis here if you have a DMM by powering up your O2 with the correct AC adapter, then measuring the voltage from ground (metal shell of gain switch) to D1 anode for positive reg out and D5 cathode for negative reg out. One reg is probably working and the other isn't (if U6 is dead you will see a positive voltage on both D1 anode and D5 cathode, or vise versa if U5 is dead).
Whoops.
With batteries installed a tip-positive DC adapter would still power the positive rail by the action of the O2's diode logic, and charge/maintain the positive rail battery, with the U5 regulator running in dropout. But now the negative rail would be held up by the negative battery until it discharges to the point where the power management circuit trips the mosfets and shuts the whole thing down.
But the PM circuit wouldn't pick up on the fact that positive rail is on AC and the negative rail is on batteries, since it just looks for a minimum of 14Vdc across both rails combined. So the first problem here is that the O2 PM design assumes both batteries are discharging more or less equally. In this situation if the positive rail is at 11Vdc or so, the negative battery would get down to 14Vdc - 11Vdc = 3Vdc which is enough to permanently damage it. Generally 0.9Vdc per cell is considered the minimum for NiMH it seems, and this would be 3Vdc/7 = 0.42 per cell.
The second problem is that a pretty unbalanced set of rails would be presented to the op amps. Maybe starting out at +11vdc/-9.5Vdc and going down to +11Vdc/-3Vdc at the point where the PM circuit trips. For small imbalances the maximum AC swing on the op amps is just affected, clipping against the lower voltage rail first. But for that much rail imbalance (+11/-3) it could start affecting the DC bias of the op amps. I kind of doubt it would kill them although it is possible if the common mode input voltage range gets exceeded and they latch up in some fashion, but it could very well produce a high(er) DC offset on the output, which may have made out to the headphones. All of this is reversed, of course, for a tip-negative DC adapter. Then the positive rail battery might be damaged.
The good news is that the regulator chips probably did survive if there were batteries in place. The negative rail battery would have prevented the negative regulator from seeing a reverse voltage. I also don't see any path in all of this for damaging the mosfets, unless one of the output NJM4556 chips or the NJM2068 chip happened to dead short.
So the first thing to check is still those voltages out of the regulator chips on D1 and D5 to make sure the AC supplies are working OK. Then I would be inclined to just put a finger on each of the op amps to see if any are heating up excessively. Then measure the DC offset on the O2 outputs and see if it is much higher than the typical 3mV.
Finally troubleshooting.
Batteries seem ok
regulators are working properly
suggested test for U2 passes fine
DC offsets are nominal (way low)
I guess i start replacing opamps just in case?
Seems to have been the 4556's. Currently listening to Nick Cave's Dig, Lazarus, Dig!!! on Fostex T40s.
Before i case back up i want to install a panel mount power jack - i have a bamboo face and figure i will just enlarge the hole a bit and thread the jack into it, then solder little jumper wires to the pcb.
While we're on the subject, my O2 is now populated with DD marked parts in all audio opamp positions. For the life of me I can't remember what JRC/NJR means to communicate with the extra D. I'd like to think that it's for a double dose of pimpin', but maybe it's just $0.10 extra?
While I'm in there i guess i have some eensy heatsinks to put on the 4556ADD parts. Just in case they get warm.
Before i case back up i want to install a panel mount power jack - i have a bamboo face and figure i will just enlarge the hole a bit and thread the jack into it, then solder little jumper wires to the pcb.
While we're on the subject, my O2 is now populated with DD marked parts in all audio opamp positions. For the life of me I can't remember what JRC/NJR means to communicate with the extra D. I'd like to think that it's for a double dose of pimpin', but maybe it's just $0.10 extra?
While I'm in there i guess i have some eensy heatsinks to put on the 4556ADD parts. Just in case they get warm.
Last edited:
Anyone still selling pre-cut non-metallic faceplates for these?
I just sorta mangled my bamboo faceplate in an effort to fit a panel-mounted power jack, so that the power plug will stop falling out.
Aerohoff seems to have disappeared. No activity since 10/1/12 and full PM box. I sent him an email but won't be surprised if he says no.
I just sorta mangled my bamboo faceplate in an effort to fit a panel-mounted power jack, so that the power plug will stop falling out.
Aerohoff seems to have disappeared. No activity since 10/1/12 and full PM box. I sent him an email but won't be surprised if he says no.
Walter at Head-N-HiFi is selling bamboo panels and several others in the "panel color" drop-down box:
O2 Front Panel
He is in the commercial vendor forum:
http://www.diyaudio.com/forums/head-n-hifi-walter/
O2 Front Panel
He is in the commercial vendor forum:
http://www.diyaudio.com/forums/head-n-hifi-walter/
hey! awesome project! im excited to start mine. just a little issue however, pulled the BOM and found everything but one part on mouser. looking for the 270-270k-rc resistor. they seem to be sold out and have some on order, but no eta date i also checked digikey and they are all out too. also tried looking for a similar part but nothing came up. anyways just wanted to see if anyone knew where i could find these bad boys in a relatively short time... like i said im super excited to start on mine. any info would be great, thanks!
i also checked digikey and they are all out too. also tried looking for a similar part but nothing came up. anyways just wanted to see if anyone knew where i could find these bad boys in a relatively short time... like i said im super excited to start on mine. any info would be great, thanks!
It isn't that critical a part. If you need to purchase 1/4W or SMD parts to fit there I don't see a problem with that.
But Mouser shows 270k 1/8W metal film in stock, such as p/n 71-RN55D-F-270K.
you'll have more luck with that question if you go somewhere where people buy ready-made amps rather than make amps.