Crack + RS-1 = exploded driver

[wotts]

Hello from the Detroit Head-fi meet!

I recently completed a Crack build and had been using it at home. I brought it along to the meet and it seemed to be working fine - I had W1000x and T1s plugged in. Another attendee plugged in his RS-1 and then the driver just popped. We did not have music playing, and it happened 20-30 seconds after the phones were plugged in. We have since checked the amp with IEMs and XB700s to see if amp was still working, and it appears to be.

What should I check (meter) to verify the amp is working properly? I don't want to be destroying $700 cans.

Thanks,
Tim

[Paul Birkeland]

Hello Tim,

It would be a very, very good idea to check the connections of the resistors across the headphone jack on the Crack. 

These resistors keep the output capacitors in check.  To verify that they are working properly, measure the voltage at each output as you switch the crack on.  The voltage should have a very short spike of a couple of volts, then drop down immediately.

[Laudanum]

Not to contradict Paul, he's much smarter than I am.  But I want to point out that I posted a question regarding the start up voltage after Speedball and film cap installations.  The manual states a short voltage climb to about 9 volts dc in the stock Crack, then the drop, which is what I measured.  After Speedball  and Film output Caps I measured just above 20 volts dc before the drop.  I posted this and had another member measure 14 volts and another member subsequently posted that he had measured right around 20 volts as well.   This is WITHOUT headphones connected.  Voltage is far lower with the load connected.  Anyway, I dont recall if there ever was a definitive explanation offered for the higher voltage after the mods but it seems to be common.   My point is that if you measure that quick rise up to around 20 volts in the upgraded Crack, it doesnt seem to be out of the ordinary.  Mine has been in service for 9 or 10 months now I would guess, and no problems, no harm done to any of my phones.  I dont have $700 phones but surely wouldnt want to toast my $200-$300 phones either.  I dont have them plugged in when I fire up Crack, I plug them in after the spike would have occured just to play it safe.

On a related note, unplugging and plugging different phones while the amp is on never "felt" right to me.  I dont do it much because I usually just listen with one pair per session but I sometimes do.  I dont know if there is anything to really worry about from a technical standpoint.  As I mention, it just doesnt "feel" right, which very well can be without any merit at all.  But cycling the tubes on and off quickly is definitely not supposed to be good.  Maybe Paul or someone else can chime in on whether there is potentially anything to worry about with unplugging and plugging in different phones while Crack is up and running?

[Doc B.]

Actually you have answered your own question. The voltage rise only happens at start up. Once the tubes are conducting the DC voltage drops to zero. So start the amp, wait 30 seconds to plug in any cans, and just leave the amp on even if you plug in different cans. Swapping headphones is actually a wee bit more safe when the amp is running. I don't really have an explanation for what happened to the RS-1s, the description of it happening 20-30 seconds after it was plugged into a live amp with no signal seems safe. Or was the amp turned off when the RS-1s were plugged in, and then switched on? In any case I have plugged in all sorts of cans and then turned the amp on and never had a problem.

[Laudanum]

Thanks Doc, for addressing my question.  My apologies to the OP if I dragged this off track at all. 

[wotts]

I'll check the voltages again when I get back home. They all tested fine when I finished the build.


I'm pretty sure the amp was off when the cans were connected, but I wasn't paying attention. I would like to add there was a Bifrost DAC inline as well. The W1000x sounded scratchy off another amp with that DAC in the line. I am now wondering if there was some errant signal (perhaps a square wave) from the DAC that the Crack amplified and then made the RS-1 very unhappy.

Thank you for all the fast replies!

-Tim

[lextek]

I thought using low impedance headphones like the Grados would damage the Crack.

[Jim R.]

Nope, low z headphones will just not sound as good -- mostly suffering in the bass department, but will not hurt the amp.  I think Doc or PJ posted something about this somewhere, but maybe it needs to be a sticky.

Of course, there are exceptions of low z cans that do just fine with the Crack too.

HTH,

Jim

[Paul Joppa]

Technically, many phones have an impedance bump at their mechanical resonance, which is often around 100Hz. With a higher-impedance source (such as the Crack, or any source that meets the IHF/ISO standard of 120 ohms source impedance) this will emphasize that frequency region, giving lumpy and/or wooly bass. Most also have an impedance rise in the high treble, giving some emphasis (or glare or screech) to the treble. The effect is small with high-impedance units, and large with low-impedance units - that's only because the impedance is small or large with respect to 120 ohms.

Some phones manage their impedance so that the bump is small enough that it makes no difference - such phones would work well with the standard as well as with typical solid-state headphone amps which have a low output impedance (i.e. they do not follow the standard, which is in fact ignored by most).

I'm confident there is more to it than that, but this is an easily recognized and measurable effect, and must be at least part of the explanation.

Incidentally, you can in theory make an equalizing network that would flatten the impedance of any given headphone, making it immune to source impedance variations. I haven't heard of anyone doing this. My experience with loudspeakers makes me cautious about predicting whether this would work as well in practice as it does in theory  :^)  but it would be an interesting experiment.

[Laudanum]

Nope, low z headphones will just not sound as good -- mostly suffering in the bass department, but will not hurt the amp.  I think Doc or PJ posted something about this somewhere, but maybe it needs to be a sticky.
Of course, there are exceptions of low z cans that do just fine with the Crack too.
HTH,
Jim

I have a pair of M-Audio Q40 phones.  These are 64 ohm phones.  They sound OK plugged into Crack but not optimal.   I havent put my finger on it exactly.  These are actually pretty nice sounding phones and, I think, great for the money.  But they sound a bit "off" on Crack.  Very listenable but not quite right.  The bass still seems pretty well controlled (and these phones do have some bass) but they dont seem as balanced and it may be because the treble is a bit hot, as PJ describes.  I use these most often with my little AMB Mini3 when sitting at the puter.  I bet they'd be great with SEX or Smack though.  I cannot listen to my Beyers or AKG's through this little chip based amp ... Im Crack spoiled there for sure.

[wotts]

I tested all the voltages as per the manual and everything still checks out. I think this was a freak incident. I do have an o-scope I could check with as well, but no signal generator (can't find it). Is it worth connecting that? I did try other cans on the amp after checking the voltages, and I have not had further issues.

[Laudanum]

I dont know about the O-scope, Im not that tech savvy.  Does sound like it was a freak kind of thing.   The only thing I wonder about would be if the amp was turned off when you had the problem, as you mentioned 20 - 30 seconds after the phones were plugged in.  I havent timed it, but with speedball that seems about like the time period where the DC would peak on the output after turn on.  When I first built the amp, I also measured the voltage on the headphone jack at start-up with some cheaper phones plugged in.  The meter I was using doesnt have an automatic peak hold and it's near impossible to accurately note what the max voltage climbed to but it was in the millivolts.   I dont know how much DC voltage it would take and for how long to pop an average headphone driver.  But if everthing is working as normal, maybe the driver was just weak and as you suspect, just a freak kind of thing.  Plugging the phones in AFTER that DC spike seems like the safest practice.  For those that tend to leave their phones plugged in all the time, it's not that much of a chore at all to just unplug them before flipping the switch and then plugging them back in sometime after 30 seconds or so ... if that's the safest approach, why not.

I wish I had a meter with auto peak hold.  Im curious now as to the actualy max DC voltage at startup with phones plugged in.  Not worried, just curious.

[wotts]

I'll set mine to check the peak and post my findings. I have a Fluke 87V.