Quote:



I don't have any working experience with the HA-5002 serving as a rail splitter. In theory it should work, but I don't know whether it would be happy with large-ish 22uF caps from its output to the rails. The BUF634 has been known to work well in this scenario.

Quote:



The idea is surely good but complex. Why not use precision parts instead?

My QRV-04 headphone amp with OPA627 or AD8610 is totally click/thump free.

Quote:




This will be useful in Tube amps and high power amps (Dynahi for example). Most tube amps that I have used had the on/off thump. Some had it loud and some had it long.

Has anyone built this already? My parts are coming in...but I am tempted to run to Radioshack and use generic parts for the time being just to play with the circuit.

These thumbs can be reduced in the design stage if you have those on the agenda but I admit that sometimes it's hard to realize if you will have such problems.

I'll guess that you'll have more problems with amp which are AC coupling in one way or other and also single ended amps.

Quote:



Sure, but even normally thump-free amps might malfunction and take out your expensive headphones with them, and the fact that some amps out there are not thump-free means that an inexpensive circuit like this is probably good insurance. I am not saying that a protection circuit like this is a must, each DIY builder will have to make his/her own decision.

Quote:



Since the max supply voltage for the ε12 as designed is +/-30V and must come from the amp's rails, it is not a good fit for most tube amps. However, if a tube amp is in the traditional 2-channel configuration (which should cover 99+% of them) you could build the ε12 without the BUF634 rail splitter portion, use a separate +/-15V PSU to power it, and tie the ε12's "VG" to the amp's real ground and it will work well.

Quote:



LOL, I had just started this thread yesterday! Yes, you can use Ratshack parts but they don't carry the BUF634. If you happen to have one in your drawer then it's cool. Another thing to be aware of is that if you use another relay than the Tyco I specified, be sure that its coil current is no more than around 30mA so that the BUF634 won't get too hot.

well... i hope, one of the first builders will provide us with some pics... could be really helpful for us noobs. gsferrari... please...

I do a very similar thing except i don't like anything in series with
the audio signal if at all possible. So i use the normally closed contacts
on the relay and then tie the contacts between the audio outputs
thru the relay to a pair of back to back shottky diodes. That way the dc level
limits to +/-.3 volt until the relay pulls in.

Nice job AMB.

Quote:



Neither do I, but there are lots of contacts within the audio chain, from both ends of interconnect cable, to your headphone jack, to possibly yet another set within the headphones (a la Sennheiser). If a high enough quality relay is used the degradation should be minimal.

Quote:



Yeah, that would do it but it's still possibly up to +/- 0.3V, which is quite a lot. Also, some amps don't take kindly to being shorted out for a few seconds, or in the case of an actual malfunction, possibly a lot longer than that (until the user turns the power off, after the smoke and smell of burning parts... ).

Quote:



Thanks.

Quote:



Can a regular splitter be used instead of the 634, which would supply the opamp? The relay driver does not need to be referenced to virtual ground, it can have a regular 7812 and tied to the original supply directly, right?

I have some spare 7812 and tle around, that why...

Quote:



If by "regular splitter" you mean a TLE2426, the answer is no, because the TLE2426 does not have enough current capacity.

Quote:



While it's possible to use a 7824 regulator instead of the 2N3904/3906 plus zener, I do not recommend that because the 78xx regulator has higher drop-out voltage than the discrete transistors, making the supply rail voltage requirement higher (at least +/-15V or 30V total). You cannot use a 7812 regulator because that would only make a +/-6V supply after the BUF634, and that is insufficient to pull the relay.

EDIT: After I re-read your question I think I understand what you're asking better. You are wondering whether it's possible to use a 7812 and a TLE2426 to make +/-6V just for the opamp, but change Q3, Q4, Q5 and Q6 to be referenced to -6V instead of the VG. The answer is no, because Q5 and Q6 is triggered by the opamp which normally has an output voltage of less than +/- tens of mV, and if Q3/Q4 is to be referenced to -6V they will turn on and cause C7 never to be charged, and thus the relay will not turn on.

Amb,

If one were to want to add a fault LED or other indication that it is either a) in delay, or b) a DC offset fault has occurred, I assume that this could be accomplished by monitoring the point between R11 and C7? In normal operation (relay on), this should be at about 1.4V (the base-emitter junctions of Q5 and Q6)? During fault, it should be close to 0V WRT VGnd due to Q3 or Q4 bypassing C7?

-Chris

Quote:



Don't put an LED there because it will drain down the C7 capacitor. If you want a fault LED, wire a LED + 3.3K series resistor from the collector of Q6 to the VG.

Radio Shack used to sell a 12V-powered red blinking LED (no series resistor necessary). If you could find something like similar, it would be cool to use that here.

Quote:



Would this be true even if the LED were driven by some TTL such as an OR gate or a pair of inverters? I would think these would have a high enough input inpedence to have no effect on C7?

Agreed, the blinker would be cool.

-Chris

Quote:



A CMOS logic gate should be ok, although the 1.4V B-E junction voltage there is too low to be a good "true" level.