I learned about power on/off ordering the hard way

[aroide]

Hi all.  Just sharing my stupidity so others won't repeat what I did.  I just hooked my BeePre up in my soon-to-be listening room.  The signal flow is DAC->BeePre->Bryston 3BST.  My Magnepan 1.7s are still being built, so to check out my system hook-up, I dragged my Epos M5s from the bedroom to the listening room.  Everything worked, though I have a HUGE gain problem with the BeePre/Bryston/Epos where I have tons of noise, and can only turn up the volume to about 5% with LOUD volume. 

Anyways, I was messing around and assumed my amp was off, it was not.  I turned off the BeePre and immediately saw the Epos speaker cones thrust out for a second or so and then wisps of smoke came out of the speaker.  My Bryston was signaling overload.  I FRIED my speakers with some kind of full power DC current load.    When powering up and down, the BeePre puts out some wacky signals.  Crud....

A few thoughts and questions based on my experience:
- is the Bryston input AC coupled?  It seems like it isn't as the cone was clearly getting a strong DC current before it fried.  If the BeePre output such a low frequency signal, the output cap wouldn't transmit it.  If DC coupled, the amp will transmit all kinds of crud and could easily overload speakers if the rules are violated.
- I guess I'll find out, but I hope that the Maggie 1.7s will change the gain problem I have.  The Epos M5s are fairly sensitive 4ohm speakers.  Does anyone have experiences with BeePre feeding a Bryston 3B or 4B? 

If I have a gain issue, I assume I can either 1) mod the gain of the BeePre or 2) insert an attenuator.  Anyone have insight into different speakers and changing gain?

Now to see if Epos can repair my speakers...

The lesson to be learned.... NEVER violate the amp on last, off first rule.

[Paul Birkeland]

- is the Bryston input AC coupled?  It seems like it isn't as the cone was clearly getting a strong DC current before it fried.  If the BeePre output such a low frequency signal, the output cap wouldn't transmit it.  If DC coupled, the amp will transmit all kinds of crud and could easily overload speakers if the rules are violated.

Yeah, it's really too bad that there isn't a cap at the input of the Bryston.  This has been common practice for decades, but having "DC coupled" in one's product literature seems more important these days. 

- I guess I'll find out, but I hope that the Maggie 1.7s will change the gain problem I have.  The Epos M5s are fairly sensitive 4ohm speakers.  Does anyone have experiences with BeePre feeding a Bryston 3B or 4B? 

Yes, the Maggies should reduce the sensitivity.  It's also advisable to pad down the input of your amplifier if it is incredibly sensitive.  This may also have prevented the turn-off transient from mattering if it had been padded down a bit.

Now to see if Epos can repair my speakers...

Did the woofer coils burn?  You could probably get a recone or a new pair of drivers pretty easily. 

The Bryston has about 30dB of gain, which is very high.  IMO, 20dB would be on the high side. With so much gain, the turn-off transient is relatively large in comparison to your signal voltage, and thus you have your issue. 

[aroide]

Thanks for the help!   I assume that if I need to reduce gain further I will add an L-pad or similar to the specific output on my BeePre.  Its easler to add it there than at my amp.  I have read that some other attenuators are sonically better... bridged-T or Pi.  Any experience with different attenuator designs?  I am assuming Holco resistors would be best, but any experience with other brands?

Given my very high gain on the Bryston, what gain would be 'reasonable'?  I have a small room and want to get rid of the high noise I have today.  In my initial test rig after build, the BeePre was between my DAC and a Pioneer receiver.  The BeePre was drop dead quiet in that config.  The Maggies are 86db/500Hz/2.83V.  Is there a desired sweet spot for gain given my config?

[johnsonad]

I used a MQ design with a P&G pot and 10k:10k transformers as an L pad when I used a Quad 303 instead of my current amps.  It did the trick though you need to get your impedances right when you can.  I happened to have this passive preamp sitting around and it worked well.  BH has a new passive pre that would do the trick though it's a lot more than a couple of 50k alpha pots.

[Paul Birkeland]

Thanks for the help!   I assume that if I need to reduce gain further I will add an L-pad or similar to the specific output on my BeePre.  Its easler to add it there than at my amp.

If you add it at the output of the BeePre, you'll very likely increase the output impedance of the preamp, and degrade the performance significantly.  If you don't want to build it into the amp, try the Harrison Labs attenuators that Parts Express has.

Given my very high gain on the Bryston, what gain would be 'reasonable'?  I have a small room and want to get rid of the high noise I have today.

Start with -12dB and see how you like that.  Even with less sensitive speakers, this is still likely a desirable setting. 

[aroide]

I have Harrison -12db attenuators on the way.  If they impact sound quality at all, I'll open them up and change resistors to implement a perfect impedance match l-pad with really nice resistors.    I looked at my Bryston schematic and hardware mod options.  I can't see voiding my 20 yr warranty to mod the amp.  But I can use the Harrison hardware as a vehicle for my perfect L-pad.

I'll update the forum when I have some before/after results.

By the way, major kudos to Epos.  They responded quickly and offered to sell me new drivers for me to install in my m5s.  The price with shipping and pound/dollar exchange rate + the risk that I have more damage to the crossover and/or tweeters means I will be shopping for new bedroom speakers instead (Wharfdale Diamond 10.1s will suffice for my bedroom system needs).  But it was really pleasant to see Epos' response and options to help me out.

[Paul Birkeland]

The BeePre does not pass DC through its output - this is just a shutdown transient and an incredibly sensitive system. 

A cap at the input may actually pass this as well, as there is dV/dT in the shutdown transient that would pass through an input cap.

[aroide]

I received the Harrison labs -12dB attenuators.  Measuring resistance (and visually checking by opening the attenuator case), the series resistor is 6.8k and parallels is 2.8k.  When I calculate impedance I get 9.08k.

I don't have my speakers yet to hear attenuators with my Bryston amp, but adding to my PPA headphone amp makes a HUGE improvement in dynamics, reduction of noise, and overall a great improvement to the sound quality.  And the PPA volume knob is closer to where it should be when used with line out from a receiver.

To be as accurate as possible with impedance (50k imput impedance on the Bryston 3BST), I will use the attenuator housing, but replace the resistors with 36k serial and 16k parallel.  The attenuators unscrew and there is a nice PCB inside with surface mounted resistors.  This change will give me 48k impedance.  I have some Shinkoh tantalum film resistors on order and will document the conversion process and results with pics also.  Stay tuned...

[Paul Birkeland]

There isn't necessarily a huge need to replace the resistors in the attenuator.  The BeePre will drive 9K just fine, but you may prefer the sound of the higher impedance configuration.

[aroide]

Yes Paul, I'm in that "I've just gotta know" state where I have to try stuff.  In the worst case, I'm out $22 or so for new attenuators.  I really want to hear how the tantalum resistors sound also.  Based on google searches, people really like those resistors.  But one could argue that almost any decent resistor will sound good.

[aroide]

My Harrison Labs attenuator mod was a success.  I can hear a difference in detail with my headphone amp.  Still can't try out with the Bryston as my speakers are still not here (shipping next week though!).

My mod was to change out the resistors in the Harrison Labs -12db attenuator so that impedance matched my Bryston amp exactly.  The attenuators had a 9k ohm impedance and my amp is 50K.

1st pic is the unopened attenuator.  Unscrew the male rca end and you get the 2nd picture.  I have to say that the guts of this attenuator are quite nice with good parts and layout.  Note the ground plane behind the resistor board in pic 3.

I tried to remove the chip resistors and attach the 36k series and 16k parallel Shinkoh 1/2W resistors to the board.  The board traces and pads are too delicate and they broke. I then attached the resistors as in pic 4.  You may not be able to see well, but the center connection from the female socket is attached via a wire that BARELY reaches the brass housing.

In the end, it worked well.  And these resistors do sound great.  Maybe its the impedance change, but there is more detail now.  Its subtle, but I can hear it right away with my headphones.

[dave-tx]

Thanks for posting your mod, I've got a similar setup to yours (BeePre->Bryston->Maggies), and would love to hear what you think after you get your 1.7s in place.

[aroide]

The Maggie 1.7s arrived.   Even with little breakin on all components, the system sounds fan-freakin-tastic.   The -12db attenuation is just the right amount for the Bryston.  I hear zero background noise, no hum, and the beequiet has some headroom.  If I go to 0db it is LOUD in my small room... more headroom than I will ever need and no clipping on the Bryston.  For normal listening the beequiet is mid-range values.   The modded attenuators sound great.  I don't have non-modded to compare to.  I'm a very happy camper with a system at a level I've never had before.

For reference, here is my whole system signal flow:  flac files on macbook pro running audirvana with usb out->audioquest carbon usb->valab x10 dac (ESS9018 based)->Nordost Heimdall rca->beepre with beequiet, output cap mod (Claritycap MR), bottlehead power cord->audioquest columbia RCA->modded Harrison labs -12db attenuator->Bryston 3BST amp->audioquest gibraltar speaker cable->magnepan 1.7s.