A muting delay & DC offset protection circuit from amb

[guzzler]

thanks amb, I'll update the layout sometime

g

 

[AtomBoy]

I hope y'all don't mind me following allong here. I've updated my copy of the schematic here:

http://www.swcp.com/~atomboy/amps/e12.Jul15.sch

 

[amb]

Quote:

Originally Posted by AtomBoy I hope y'all don't mind me following allong here. I've updated my copy of the schematic here: http://www.swcp.com/~atomboy/amps/e12.Jul15.sch

Looks fine.

 

[guzzler]

Final version (replaces one above, which was actually wrong). Actually quite like it now, except the long wire to the darlington

Eagle .brd file

Changed: updated to latest schematic (except 24V regulation stage), made ground continuous between channel tracks to reduce crosstalk.

Beware of the Darlington transistor, I don't think the pinouts are standard. On the silkscreen, the corresponding pins are:

1. collector
2. base
3. emitter

I'd be interested to see some pics if someone etched one of these, I'm out of chemicals just now

 

[amb]

guzzler and Atomboy, before you go too far with this, I did more testing with the TL072 + dual transistor ground splitter and I have to say that it really doesn't work well. The TL072 has a very high output impedance (see its datasheet, it's got a big resistor in series with its output) and as a result the ground potential shift under load is still significant, and unstable (drifts all over the place). I tried substituting with a NE5532 (which has much better output current capability than the TL072) and it was better but the shift was still too big. To "fix" this would require making the two transistors operate in class AB rather than class B. This means adding diodes and resistors to bias them, and therefore also increasing quiescent current draw.

I am convinced at this point that this is simply not a good way to go. The two transistors and biasing resistors/diodes will eat up more board space than a single DIP-8, so I really recommend staying with the OPA551 opamp like in my rev. 1.1 schematic. It works great and steady as a rock.

 

[guzzler]

Pain, that.... I'll have a further dig about for a more suitable opamp. If not, it'll be back to the amb rev 1.1

fitting in some biasing diodes wouldn't eat much space, but is starting to lose the cost advantage (not that diodes are expensive)

 

[guzzler]

Back to Rev 1.1, after some discussion it seemed the easiest way out of the impedance problems. Anyway, the Eagle .brd files are available here and here is a picture of the finished boards (on a rather shockingly red carpet):

hope you like them

 

[Pars]

Nice Guzzler! You gonna have any extras left over or make more up? If so, I'd be interested.

Chris

 

[Pars]

Amb, all,

I began my own layout/schematic of this with the following enhancements, directed towards use of my Gilmore (or whatever) as a preamp:

1) Manual mute switch bypassing the delay... a simple SPST switch in the relay drive leg

2) Replacement of the R11/C7 and Q5/Q6 darlington pair with a 555 timer circuit driving the relay directly. This should allow for a wider range of delay with smaller caps (i.e., 1M resistor, 10uf tantalum for 11 second delay, etc.). The 555 will reliably drive one relay (has in my Counterpoint pre for >15 years). The 555 output goes low after the time delay, so is opposite the scheme used here, but is a drop in replacement. If a darlington pair to drive the relay is desired, a third 2N3904 functioning as an inverter is required.

3) Reversing the relay, and substituting an RTE24012F DPDT with both NC and NO contacts accessible. By including jumper pads in the proper sequence(s), it should allow for both

a) the original audio thru the relay, or optionally,
b) the use of the relay to short the amp output to ground thru the NC contacts and a pair of Shottkys as Dr. Gilmore suggested, thereby not routing the audio thru the relay itself.

Additionally, when used in the normal fashion of the audio going thru the relay, a pair of jumper pads would be included to connect the NC contacts to ground, thereby shorting the outputs to ground (but not the amp itself). I would think this would be a good idea, but comments welcome.

4) Inclusion of LED drivers (2N 3904 darlington pairs) for both delay/fault state and on state.

Comments/criticism welcome. Also, I assume when Dr. GIlmre states: Quote:

Originally Posted by kevin gilmore 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.

that he means the Shottkys are in parallel but reversed from one another and connected to ground? And I would assume that you choose the Shottkys based upon the lowest forward drop? Any suggested Shottkys to use here?

Thanks,

Chris

 

[amb]

Quote:

Originally Posted by Pars 1) Manual mute switch bypassing the delay... a simple SPST switch in the relay drive leg

That's fine, or you could put the SPST switch in parallel with C7. When the switch is closed, it mutes immediately, but when the switch is opened, it goes through the delay interval before un-muting.

Quote:

2) Replacement of the R11/C7 and Q5/Q6 darlington pair with a 555 timer circuit driving the relay directly.

That's fine too. There is no magic in the darlington transistor drive in my circuit, it's simple and does the job. Make whatever mod you desire in your version.

Quote:

3) Reversing the relay, and substituting an RTE24012F DPDT with both NC and NO contacts accessible. By including jumper pads in the proper sequence(s), it should allow for both

The RTE24012F will work, and the smaller-footprint Omron G6A and G5V2 series relays should work well too. The Omron relays will fit the PCB layout from guzzler (but the Tyco RT series won't). While I am on the subject about guzzler's layout, he combined the Q5 and Q6 into a single darlington transistor. I recommend using a MPSA14 for that application.

Quote:

a) the original audio thru the relay, or optionally, b) the use of the relay to short the amp output to ground thru the NC contacts and a pair of Shottkys as Dr. Gilmore suggested, thereby not routing the audio thru the relay itself.

The schottky scheme will still allow 0.3V or so of output, so it's not really a short. Since this circuit will protect against DC offset deviations as little as 70mV, that leaves a range of voltage from 70mV to 0.3V where the circuit has activated but still allows offset through to the headphones. Also, that 0.3V gap is still going to pass a thump of turn-on transient.

Quote:

Additionally, when used in the normal fashion of the audio going thru the relay, a pair of jumper pads would be included to connect the NC contacts to ground, thereby shorting the outputs to ground (but not the amp itself). I would think this would be a good idea, but comments welcome.

This won't hurt anything, but isn't necessary either.

Quote:

4) Inclusion of LED drivers (2N 3904 darlington pairs) for both delay/fault state and on state.

You can simply hook up an LED + resistor in parallel with D2 to indicate 'on' state, and an LED + resistor across the C-E junction of Q6 (or your 555 timer IC output equivalent) to indicate 'mute'.

 

[Pars]

Quote:

Originally Posted by amb That's fine, or you could put the SPST switch in parallel with C7. When the switch is closed, it mutes immediately, but when the switch is opened, it goes through the delay interval before un-muting.

Yeah, but I wanted the switch to be immediate (unless there is a fault).
Quote:

Originally Posted by amb The RTE24012F will work, and the smaller-footprint Omron G6A and G5V2 series relays should work well too. The Omron relays will fit the PCB layout from guzzler (but the Tyco RT series won't). While I am on the subject about guzzler's layout, he combined the Q5 and Q6 into a single darlington transistor. I recommend using a MPSA14 for that application.

I'll take a look at those. Thanks!
Quote:

Originally Posted by amb The schottky scheme will still allow 0.3V or so of output, so it's not really a short. Since this circuit will protect against DC offset deviations as little as 70mV, that leaves a range of voltage from 70mV to 0.3V where the circuit has activated but still allows offset through to the headphones. Also, that 0.3V gap is still going to pass a thump of turn-on transient.

Yeah, I realize that. Some people have commented that they didn't want a relay in the audio chain (one person who contacted me about doing boards backed out after learning that the audio was going thru a relay). I personally would rather have the protection in case I was running Grados or something at the time, as I would think 300mV would toast them at least in time. Having the fault LED might help save you though. Without just dead shorting the output, I don't see how this can be avoided. 0.3V is the lowest quoted value I have seen so far. Just drawing it up on paper (the configurable relay section) results in 3 jumpers per channel to do all of that. And, of course, you can use whatever wire you want to for the jumpers

Quote:

Originally Posted by amb You can simply hook up an LED + resistor in parallel with D2 to indicate 'on' state, and an LED + resistor across the C-E junction of Q6 (or your 555 timer IC output equivalent) to indicate 'mute'.

Where's the fun in that?

My goal was/is to make a more or less universal board that can be configured to suit whatever thoughts and/or biases the user might have.

 

[Garbz]

hey amb this thread just reminded me, this circuit couldn't be modified to work from a balanced source could it? my guess is the input of the comparitor would screw with something.

 

[amb]

Quote:

Originally Posted by Garbz hey amb this thread just reminded me, this circuit couldn't be modified to work from a balanced source could it? my guess is the input of the comparitor would screw with something.

If by "balanced source" you mean an amp with balanced outputs (like dynamid or dynamight), then the answer yes, the ε12 supports them without need for modification. See my web page for this circuit (or the beginning of this thread).

 

[Pars]

Quote:

Originally Posted by guzzler Back to Rev 1.1, after some discussion it seemed the easiest way out of the impedance problems. Anyway, the Eagle .brd files are available here and here is a picture of the finished boards (on a rather shockingly red carpet):

Guzzler,

I may be daft, but looking at your .brd file while doing my layout, for Q1 and Q2, are the transistors reversed? If the layout file is from the top (this is how I think that Eagle works... could be wrong), it would appear that the V+ and V- board inputs are connected to the emitters of the NPN and PNP when they should be connected to the collectors? Comparing the board file to your pics of the finished boards, that would at least appear to be the case with std. 2N3904s/3906s (std. E-B-C pinout)?

Thanks,

Chris

 

[guzzler]

Looks alright to me Chris. Which way up is your standard pinout? The pinout I have here from the Rapid Electronics catalogue is with the D facing down the way (like a bowl), with E-B-C left to right along the top. With the layout, the D is facing up so pins should be the other way round. I've attached the relevant portion of the board; check the other transistors as well