BACKGROUND:
I needed to retrofit an audio output delay to an AC-coupled headphone amplifier design. The original protracted power-on transient of 6V would be enough to destroy some lower impedance headphone drivers. There are a few schematics available online, most of them had DC-offset protection too (something I did not require). For those who need continuously active DC protection, then Ti Kan created his e12 board. There are also solid-state implementations on these forums. Then this thread came to light discussing a circuit which seemed ideal with some additional modifications. However, the space available to me was absolutely tiny. I also wanted to shoe-horn this into other line-level pre-amplifier designs without compromising space. I finally came up with this tiny design:
DETAILS:
SCHEMATIC:
BOM:
I sourced components from Rapid Electronics for a total cost of £2.83 (of which the relay is £2.22), which is a small outlay for considerable peace of mind. A more expensive Mouser cart is available: Mouser Electronics
CIRCUIT DESCRIPTION:
IC1 is a small 12V regulator to power the circuit. The output delay is controlled by an adjustable RC time constant formed by TR1 and C3 [V(rc)]. As C3 slowly charges, voltage increases at the gate of Q1 which will eventually turn on the relay [V(relay)]. D2 is a 7.5V Zener which will only conduct when the bulk of the time constant is reached [V(gate)]. This maximises the time constant and therefore minimises the value and size required of C3. R1 stops inappropriate latching of the MOSFET. D1 discharges C3 and removes issues with "warm starts" caused by switching the amplifier quickly off and on. D3 is a flyback protection diode for the relay.
OPTIONS:
BUILDING & MOUNTING:
Solder low to high components. Place diodes first with D2 in the opposite direction than D1 and D3 as per PCB. Remember not to confuse D1/3 with D2 as they appear identical. Then C1, C2, IC1, Q1. R1 is mounted vertically. Then add TR1, C3 and the relay. Power is connected to -V and +V. Audio input is connected to the side nearest the trimpot and audio output on the other side. There are two pads for audio ground. Adjust trimpot to suit your preferred delay time. NOTE: The board does not have mounting holes. I placed pad descriptions on the underside so that it can be mounted upside-down. I normally place a double-sided sticky pad on the relay and mount it this way. Alternatively you can heatshrink the entire module.
SUMMARY
For me, it was just a little way of playing with through-hole miniaturisation. The aim was to design something very cheap and very small, but effective enough to protect expensive headphones. Even though most amplifier designs are robust enough with low power on/off transients, I will probably "MOD" everything and end up sticking one of these in all my DIY headphone/pre-amps.
I have some boards for sale to cover costs available at this thread, and plan to release the Gerbers to the community around Christmas time.
I needed to retrofit an audio output delay to an AC-coupled headphone amplifier design. The original protracted power-on transient of 6V would be enough to destroy some lower impedance headphone drivers. There are a few schematics available online, most of them had DC-offset protection too (something I did not require). For those who need continuously active DC protection, then Ti Kan created his e12 board. There are also solid-state implementations on these forums. Then this thread came to light discussing a circuit which seemed ideal with some additional modifications. However, the space available to me was absolutely tiny. I also wanted to shoe-horn this into other line-level pre-amplifier designs without compromising space. I finally came up with this tiny design:
DETAILS:
- 29.2mm x 19.7mm 2-layer PCB.
- 15v to 35V voltage input with voltage regulator. Optionally bypass voltage regulator for 12V input or to use different voltage relay.
- Output delay adjustable up to 20 seconds with default build.
- Relay is rated for headphone/line level - 0.5A 125VAC or 2A 24VDC.
- Total current draw is 15mA.
SCHEMATIC:
BOM:
I sourced components from Rapid Electronics for a total cost of £2.83 (of which the relay is £2.22), which is a small outlay for considerable peace of mind. A more expensive Mouser cart is available: Mouser Electronics
CIRCUIT DESCRIPTION:
IC1 is a small 12V regulator to power the circuit. The output delay is controlled by an adjustable RC time constant formed by TR1 and C3 [V(rc)]. As C3 slowly charges, voltage increases at the gate of Q1 which will eventually turn on the relay [V(relay)]. D2 is a 7.5V Zener which will only conduct when the bulk of the time constant is reached [V(gate)]. This maximises the time constant and therefore minimises the value and size required of C3. R1 stops inappropriate latching of the MOSFET. D1 discharges C3 and removes issues with "warm starts" caused by switching the amplifier quickly off and on. D3 is a flyback protection diode for the relay.
OPTIONS:
- Trimpot and output delay time: Default 250K trimpot will result in maximum of 20 seconds. 100K provides max 6 seconds which should be suitable for most situations. 500K trimpot will give 36 seconds which may be beneficial for tube-based amps (delaying for warm ups). Either a horizontal or vertical adjustment trimpot can be used.
- 12V input voltage: bypass voltage regulator (jumper pin 1 to 3), skip C1, and use default 12V relay.
- 24V input voltage + 24V relay: bypass voltage regulator (jumper pin 1 to 3), skip C1, and use Omron G5V-2-24DC. Note, this will affect the RC time constant and you will need to use a larger trimpot resistance 500K-1M.
- LTSpice: The default build is quite universal and suits most situations. However a simulation file attached allows you to play with the values so you can find your ideal components.
BUILDING & MOUNTING:
Solder low to high components. Place diodes first with D2 in the opposite direction than D1 and D3 as per PCB. Remember not to confuse D1/3 with D2 as they appear identical. Then C1, C2, IC1, Q1. R1 is mounted vertically. Then add TR1, C3 and the relay. Power is connected to -V and +V. Audio input is connected to the side nearest the trimpot and audio output on the other side. There are two pads for audio ground. Adjust trimpot to suit your preferred delay time. NOTE: The board does not have mounting holes. I placed pad descriptions on the underside so that it can be mounted upside-down. I normally place a double-sided sticky pad on the relay and mount it this way. Alternatively you can heatshrink the entire module.
SUMMARY
For me, it was just a little way of playing with through-hole miniaturisation. The aim was to design something very cheap and very small, but effective enough to protect expensive headphones. Even though most amplifier designs are robust enough with low power on/off transients, I will probably "MOD" everything and end up sticking one of these in all my DIY headphone/pre-amps.
I have some boards for sale to cover costs available at this thread, and plan to release the Gerbers to the community around Christmas time.
Attachments
Last edited:
Looks very nice Andy - good work. Amp looks like an original Millet hybrid
Bingo. Trust one of the Millett Hybrid (revMH) PCB contributors to spot this!
JP has suggested to tie the input or output to audio-ground as the default off state which seems sensible.
Tying to ground by default (mute state) doesn't seem to be adopted in other designs I have seen. For example Ti Kan's: The ε12 muting / protect circuit
However - Muting Circuits For Audio
Thoughts?
Last edited:
I suggested to short outputs to GND (via a 100 Ohm resistor per channel if they aren't in the device) because:
- it is good design practice
- relay contacts are now in parallel to our precious audio signals and not IN the signal path
- the usual minimum current requirement for small signal relays are met
- Bad (as in choice of material or as in resistance/corrosion) relay contacts don't degrade audio signals like with series switching
- absolutely no clicks can be observed contrary to series switching when a small click may be heard because of small potential differences.
- at power off no click of discharging capacitors will be heard as the contacts short to GND again. Depends on how fast the circuit switches off.
Often people are concerned that the "surge current" may be resulting in failure but a 100 Ohm will prevent that. Any well designed device has stopper resistors in the output., certainly devices that use opamps should have those.
Also at least 2 mounting holes M3 would be nice.
- it is good design practice
- relay contacts are now in parallel to our precious audio signals and not IN the signal path
- the usual minimum current requirement for small signal relays are met
- Bad (as in choice of material or as in resistance/corrosion) relay contacts don't degrade audio signals like with series switching
- absolutely no clicks can be observed contrary to series switching when a small click may be heard because of small potential differences.
- at power off no click of discharging capacitors will be heard as the contacts short to GND again. Depends on how fast the circuit switches off.
Often people are concerned that the "surge current" may be resulting in failure but a 100 Ohm will prevent that. Any well designed device has stopper resistors in the output., certainly devices that use opamps should have those.
Also at least 2 mounting holes M3 would be nice.
Last edited:
As below? Or did you mean input to ground as per DontHertzMe, swapping ROUT/LOUT with RIN/LIN?
Attachments
Yes of course the latter. There Lin is Lout and Rin is Rout. The contacts are not in the signal path, they just short the outputs to GND. At power on the default state is outputs shorted to GND and after a while the contacts will open letting the audio signals loose on the amplifier.
Imagine a manual operated switch for series switching and imagine it too for parallel switching to GND. Now think what will happen in both situations when the contacts are bad or have a detrimental influence on the signal quality.
Imagine a manual operated switch for series switching and imagine it too for parallel switching to GND. Now think what will happen in both situations when the contacts are bad or have a detrimental influence on the signal quality.
Last edited:
Yes, I see what you mean. I suppose the only hurdle I have with this is that for headphone amplification, the ideal scenario is to have close to zero output impedance from the amplifier device. Attenuating the signal at that point seems too much of a compromise. For pre-amplifiers, this is less of a problem.
Also it would be trickier to make this universal as each amplifier design will have differences in output impendance and therefore hard for simple-beings (like me) to decide which needs a stopper resistor or not.
Then use the circuit as it is now for series switching in headphone amplifiers. Add pads to be able to use it for shorting to GND as well by connecting contacts 6 and 11 of the relay to GND by means of jumpers or 0 Ohm 0805 SMD resistors. Add instruction to those that want parallel switching to solder the aforementioned 0 Ohm resistors and to use Lin and Rin to the outputs of the device in question and ask them to verify their circuits for series resistors. In 99 % of cases these are available. A source or preamp that is not short circuit proof at the outputs is simply not well designed.
Solved
BTW I think muting circuits in headphone amplifiers should include DC protection too when these are direct coupled. You are more in the territory of power amplifiers then.
Solved
BTW I think muting circuits in headphone amplifiers should include DC protection too when these are direct coupled. You are more in the territory of power amplifiers then.
Last edited:
I think tying the NC pads (6 and 11) to GND via a resistor isn't a bad idea, (i.e. signal from the amplifier [RIN/LIN] to the MOD board would be tied to ground in the default off state). One could then very easily implement the parallel-signal-path circuit that you suggested if they knew what they were doing. Fitting even one more vertical resistor will be a challenge though! I'm very strict on the dimensions.Then use the circuit as it is now for series switching in headphone amplifiers. Add pads to be able to use it for shorting to GND as well by connecting contacts 6 and 11 of the relay to GND by means of jumpers or 0 Ohm SMD resistors. Add instruction to those that want parallel switching to use Lin and Rin to the outputs of the device in question and ask them to verify their circuits for series resistors.
Solved
Yes, I don't think it's a bad idea to be super safe. However, a lot of direct coupled designs these days have decent DC offset measurements, and it would add a lot more bulk to my original, and it has been done by Ti Kan.
p.s. Attached is the output impendance of a selection of commercial headphone amplifiers.
Attachments
Last edited:
Yes I misinterpreted the circuit to be mainly for preamps and sources (because of the small signal relay) and because I saw it on Swap Meet.
Yes they do when they are settled but not at power up.
And eh I just checked my stock for relays ... there are way smaller relays to make the PCB even smaller...
Yes they do when they are settled but not at power up.
And eh I just checked my stock for relays ... there are way smaller relays to make the PCB even smaller...
Last edited:
Yup. I can shave 5mm off the length to 24mm if I used OMRON G6S series. However it would leave me with 5mm for 4x pads
Wow! Not a bad little thing for an all-in-one onboard amplifier design. Will have to bear that in mind for a future design.
Tricky to use for retrofit including connection pads though.
Plus me and SMD don't get on.
Using that one will give way for 1 mounting hole M3 As you can see it is available as TH and as SMD type.
No, just joking as it is not as good as the larger types (which does not matter much when shorting to GND).
SMD is the way to go nowadays. It is not hard...when you use 0805 and 0603 sizes that is.
As you can see it is available as TH and as SMD type. No, just joking as it is not as good as the larger types (which does not matter much when shorting to GND).
SMD is the way to go nowadays. It is not hard...when you use 0805 and 0603 sizes that is.
Last edited:
Just seen G6K's used all over the new Topping L30 amp. It switches completely silently, so they should be good enough.
Might see how small we can go with this...
Attachments
Will start using this in my preamp when i receive the PCB's from Andy.
Could this be modified to give a delay on shutting down the relay. So instead of a few second before the relay turns on after power on, a few seconds before it turns off after power off? This would be for the HT Bypass switch that is enabled when the preamp is off...
Or would this require a totally different schematic?
Could this be modified to give a delay on shutting down the relay. So instead of a few second before the relay turns on after power on, a few seconds before it turns off after power off? This would be for the HT Bypass switch that is enabled when the preamp is off...
Or would this require a totally different schematic?
Power off will turn off the relay immediately.
It would require a completely different circuit to turn off an output before power off but relatively straightforward to modify to turn off after power off.