[nikongod]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Um ... with a 25V power supply, you'll see 12.5VDC to Ground at the Left and Right channels, until the output caps charge up. Typically this takes about 5 seconds. That amount of voltage is enough to kill headphones, I think. The admonition has always been: Don't plug in your headphones until the thing has warmed up for at least 10 seconds. Same for turn off - get the headphones unplugged before you turn it off or they'll see killer offset.

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif Has anyone ever actually had headphones damaged this way? I've had continuous exposure to DC offset nuke a set before, but I've used my Millett for a good long time and never had it hurt a pair of phones when I left them plugged in.

The amp is cap coupled. there is no normal way to put DC CURRENT into the headphones, so we are safe from the 10-seconds of DC current that some direct coupled amps with SLOW dc servos can give. That settles that.

If you watch the output with no load, I agree it will go HIGH and then settle. Load it down a bit more, it settles down really quickly.

Also, its *only* a 12v spike. I have put 12v (p-p) into hd-580 for a few minutes... I was curious if the Supra would clip audibly at that level. It didnt and actually sounded nice although loud enough to be clearly heard down the hall - the headphones were on my bed, not on my head.

posts elsewhere have commented on Joe (or is it John) Grado breaking in heapdhones with a speaker amp. I guess they can get new heapdhones pretty quick if they need to, but if this technique resulted in destruction too frequently they would stop dont you think?

I am personally more concerned with what the turn-on pops could do to your ears than the headphones. the HD-580 were FAR from safe listening levels in my above example...

 

[tomb]

If Nate says he's used the older Millett Hybrid (without a delay relay) without unplugging his headphones and suffered no damage, then I can't argue with that - he's used Milletts a lot longer than the rest of us.

Still, he was concerned about a short spike of a lot less value than that just recently (4VDC for a 1/4sec - see the e12 thread on Headwize). The whole idea of exposing one's headphones to a +12V DC spike at a measurable number of seconds - either 3 to 5 sec. or 1 to 2 sec. - is contraindicated.

If you are saying p-p in your example of 12V into an HD580 and it sounded nice, then it doesn't sound like you're describing DC. Same thing with the Grado example. Plus - I'm not sure - but 32ohms is rather high for a speaker amp. Maybe the combination works out reasonably well for Grados. In any event, it doesn't sound like long-term, repeated exposure to a few seconds of pure DC. I've had it happen a few times by accident when testing a new build (something wrong, obviously), and the resulting sound is a lot worse than a pop.

Whether all of that means a hill of beans for actually destroying a headphone driver, I don't know. However, we may all have quite a few dollars invested in our headphones. Erring on the side of caution and trying to prevent any regular, repeated exposure to DC spikes seems to be a correct strategy.

 

[n_maher]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Whether all of that means a hill of beans for actually destroying a headphone driver, I don't know. However, we may all have quite a few dollars invested in our headphones. Erring on the side of caution and trying to prevent any regular, repeated exposure to DC spikes seems to be a correct strategy.

Agreed Tom, I just didn't want to be reading a bunch of panic stricken posts from folks who'd done this in the past and were now obsessing about whether or not they'd ruined their headphones.

 

[MASantos]

Ok, I installed the diamond buffers. I'm not sure everything is ok with them. I had a transistor reversed(dumb me soldering stuff at 2 AM) but corrected that quickly.

THe problem is with biasing. I followed the instructions on the buffers site but even with the trimpots fully turned to the right, I get around 25-30 mA on each channel and it drifts up to 40mA. I can't get the bias lower than 22-23mA. I have inspected every solder spot and part orientation and believe that everything is correctly assembled. COuld there be something wrong with the trimpots? Is it possible to damage them by "over turning" the screw ?

I'll try to post some pics but need to get my camera from the office tomorrow

If anyone has any ideas...

I plugged the phones and they sounded good, but with the bias at 35mA I removed the DB board to avoid damaging them.

Quote:

Originally Posted by fran /img/forum/go_quote.gif I added a e12 to my board. Works fine and gives about 10-12 seconds of mute. There are simple instructions on ambs site about how to wire it so that it cuts instantaneously on power off - and that works too! The parts for my diamond buffer just arrived today so I hope to get soldering in the next few days. Does anyone have extra diamond buffer boards? Fran

If you want the original diamond buffer boards, I have two as Nate sent me extra ones. I can ship you one on monday if you want to.

 

[nikongod]

Quote:

THe problem is with biasing. I followed the instructions on the buffers site but even with the trimpots fully turned to the right, I get around 25-30 mA on each channel and it drifts up to 40mA. I can't get the bias lower than 22-23mA. I have inspected every solder spot and part orientation and believe that everything is correctly assembled. COuld there be something wrong with the trimpots? Is it possible to damage them by "over turning" the screw ?

did you reuse the transistor you originally put in backwards? it may be dead.

is the problem similar in both channels? or just one?
Quote:

Originally Posted by tomb /img/forum/go_quote.gif If you are saying p-p in your example of 12V into an HD580 and it sounded nice, then it doesn't sound like you're describing DC. Same thing with the Grado example. Plus - I'm not sure - but 32ohms is rather high for a speaker amp. Maybe the combination works out reasonably well for Grados. In any event, it doesn't sound like long-term, repeated exposure to a few seconds of pure DC. I've had it happen a few times by accident when testing a new build (something wrong, obviously), and the resulting sound is a lot worse than a pop.

im not describing DC, rather full AC. i have never gotten DC through a capacitor in any appreciable quantity, it is not an issue as long as the cap is not defective.

my point with the examples is that DC kills heapdhones in a matter of moments, where AC is safe at absurd levels.

 

[tomb]

Quote:

Originally Posted by MASantos /img/forum/go_quote.gif Ok, I installed the diamond buffers. I'm not sure everything is ok with them. I had a transistor reversed(dumb me soldering stuff at 2 AM) but corrected that quickly. THe problem is with biasing. I followed the instructions on the buffers site but even with the trimpots fully turned to the right, I get around 25-30 mA on each channel and it drifts up to 40mA. I can't get the bias lower than 22-23mA. I have inspected every solder spot and part orientation and believe that everything is correctly assembled. COuld there be something wrong with the trimpots? Is it possible to damage them by "over turning" the screw ? I'll try to post some pics but need to get my camera from the office tomorrow If anyone has any ideas...

A couple of things:
1. Yes, it is possible to over-turn some trimmers. Some operate with a click - turn it past and it may break. I believe the Vishay-BC and the Murata's (those for sure) have a clutch that prevents damage from over-turning - they'll reach minimum or maximum and then continue to turn with no change. I think the Bournes will break, but I'm not certain about that - but there are people who've broken theirs - whatever brand they might've been.

2. Check your output transistors and your output resistors. If you used the 2.2ohm output resistors, there are some combinations that are not stable at that resistance. I haven't seem many people report that, but Steinchen does mention it as a possibility. I used the 3.3ohm and never had any trouble, but that's not on the higher edge of performance.

3. Other than that, it's possible you have another part mis-oriented or misplaced.

Those are my guesses until we see some pics.

EDIT:
Quote:

Originally Posted by nikongod /img/forum/go_quote.gif did you reuse the transistor you originally put in backwards? it may be dead.

Yes, this is a good point. It only takes a split second to fry one. If it was actually soldered in, the instant you applied power to it, it's probably gone.


Quote:

im not describing DC, rather full AC. i have never gotten DC through a capacitor in any appreciable quantity, it is not an issue as long as the cap is not defective. my point with the examples is that DC kills heapdhones in a matter of moments, where AC is safe at absurd levels.

YES - my point as well was that the coupling caps in the rev Millett Hybrid take a couple of seconds before they charge enough to block a large DC spike. This is variable depending on how often you turn the amp on and off and the timespan between uses. It sounds like we are in agreement, though.

 

[n_maher]

I've experience similar results with the DB's where there's a threshold below which they can't be adjusted. My opinion is that as long as they're able to be adjusted and remain stable at an acceptable bias value it's probably safe to leave well enough alone.

 

[MASantos]

I am using the 2.2Ohm ones. Nate got a kit together for me quite a while a go and I believe it's prettymuch stock. The transistors are the default ones and AFAIK every other part as well.

I believe that the trasistor that was reversed is still ok. I was getting 0mA on that trasistor, but after I corrected it I got current and it seems ok.

I'll try to leave them on until they get stable and them do some temp readings on the transistors. If the output transistors die on me, I'll just get new parts and build new ones, If they work, I'll be a happy new millett owner.

 

[tomb]

Quote:

Originally Posted by MASantos /img/forum/go_quote.gif I am using the 2.2Ohm ones. Nate got a kit together for me quite a while a go and I believe it's prettymuch stock. The transistors are the default ones and AFAIK every other part as well. I believe that the trasistor that was reversed is still ok. I was getting 0mA on that trasistor, but after I corrected it I got current and it seems ok. I'll try to leave them on until they get stable and them do some temp readings on the transistors. If the output transistors die on me, I'll just get new parts and build new ones, If they work, I'll be a happy new millett owner.

It's possible to turn the trimmers too quickly so that it appears that they are unstable - dialing in a setting on the trimmers before the output transistors reach some sort of thermal equilibrium. Then it will appear that they always spiral up to a higher current.

You might review Steinchen's procedure on his website - it's a good one that uses the utmost caution in giving the transistors ample time to equalize. Other than that, sounds like you have a good strategy - good luck!

 

[MASantos]

OK, I just won the DIY stipudity of the year award. After biasing the tubes with the multimeter at dcV I was using dcV instead of dcmA to bias the buffers.

I guess the lack of sleep when I build stuff during the week takes my concentration away. I was reading through everything calmly no and realised I needed to measure the current.

So nw I measured the current across the output of the buffers and the following happens: the longer I leave the probes connected to the test points, the higher the current goes and it never stops rising. I removed the probes at 30mA. Is this normal and should I leave the probes until the drifting stops?


Thanks for you patience guys!!


*** gos away and starts casework***

 

[nikongod]

it is normal for the current to rise until the output transistors reach operating temp.

it should not hit 30mA unless that was a target current.

are you measuring Ma across the test points or voltage? Just to triple check, you SHOULD measure the voltage across the test points, and look up the current in the table.

 

[MASantos]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif it is normal for the current to rise until the output transistors reach operating temp. it should not hit 30mA unless that was a target current. are you measuring Ma across the test points or voltage? Just to triple check, you SHOULD measure the voltage across the test points, and look up the current in the table.

Yeah I figured that out about the voltage/current table. I left the buffers warming up or a while and they're not drifting anymore. I'm doing the last adjustments now.

Thanks guys!

I've plugged my AKG701's and they sound really nice. I had only heard the millett at headfest07, but with all the noise in the room, I couldn't give it much attention. Next week I'll do some serious listening.

 

[fran]

With many thanks to MASantos and the fellas here, I got my diamond buffers put together tonight. I made just one boo - boo - I soldered in the wrong transistor for Q3 - but copped it the minute I did it. Other than that it all went smoothly enough, but christ, it is fiddly. Not made for people with big fingers!

Anyway, I used 3.3R for the output resistors and have it biased about 60mV or so. I think from the table the upper reaches of bias is about 85mV, so 60 should be fairly safe. The components get warm, but only about body heat or so.

But tell you what: It was well worth it! Quite an improvement in sonics and a good bit smoother sounding. Heres one pretty happy camper.

And manuel: I'm listening right now with some k701s (thanks Santa!) and it sounds bloody good. This is a cracking amp.


Fran

 

[fran]

A question for nikongod/n_maher/tomB/steinchen etc:

Once the diamond buffers are in and settled and you have biased them (leave 'em settle for say 45 mins), how stable are they? Should I be going back in there after a few hours and checking the biasing again? Or will they stay as they were?

Fran

 

[tomb]

Quote:

Originally Posted by fran /img/forum/go_quote.gif A question for nikongod/n_maher/tomB/steinchen etc: Once the diamond buffers are in and settled and you have biased them (leave 'em settle for say 45 mins), how stable are they? Should I be going back in there after a few hours and checking the biasing again? Or will they stay as they were? Fran

Once biased and settled - and 45 minutes definitely qualifies - they won't change unless a big change takes place - like putting the board into a case and closing it up. Then things may change a lot - current may increase when they get much hotter closed up in a case. That's why we use ventilation holes (or heat sinks, but not possible in this case)

Also, if you move the amp around so that it picks up some knocks, it may jar the trimmers enough to move them, but that's highly unlikely.

Otherwise, you're done.