I want to use my B&W DM7 Mk II speakers as the L+R speakers for my Denon AV Receiver, and also as the stereo speakers for my Audiolab 8000Q/P hi-fi system.
So I thought: design an IR remote controlled amplifier switch. The current version of the schematic follows:
Now the problems:
Thanks
David
So I thought: design an IR remote controlled amplifier switch. The current version of the schematic follows:
Now the problems:
- It uses relays, so the problem of "dry-switching" will happen at some stage.
- It uses relays, so I need snubbers across the relay contacts.
They are handling up to 100Wrms into 8R so up to 30Vrms, 3.5Arms
I initially thought of Transzorbs, but they don't appear to be available right sort of voltage rating (say 80V).
If you use relays for power audio switching what do you use for snubbers? - I'd love to replace the relays with MOSFET based analogue switches, but so far haven't found any suitable parts.
That means design my own DPCO MOSFET based switches. Has anyone been there and done that? I've never done that sort of thing. - Do I need the dummy loads for solid-state class AB and class D amps, or are they only necessary for valve (tube) amps? If I don't need them I could halve the relay count!
Thanks
David
You can use back to back power mosfets switched by a optocoupler (like a TLP3906 or VOM1271) designed to drive the gates of the mosfets.
There's a long thread on this subject - https://www.diyaudio.com/community/threads/output-relays.191449/
Whatever we use (relay or MOSFET), we're still switching alternating currents (AC) in and out of inductive loads with energy storage. In general, snubbing is necessary as there would always be some energy left in the load inductance that would cause flyback across the selector switches.
With back to back MOSFETs, you could arrange for the switch-over to happen at exactly a zero-crossing of the load current, when the inductive storage is NIL. But there's still going to be a small timing error that needs to handled traditionally (RC snubber). You also need to sense the load current somehow (e.g. hall effect sensor).
Suitable parts are 2 nos of IRF540N (100V), IRF640N (200V) etc. The easiest is common gate / source, i. e. body diodes pointing away from each other. A fully floating isolated gate drive will be needed.
And you have an extra problem if the amplifier grounds are not the same. You'd need to switch them too !!
There's no need for snubbing, use mosfets that are rated for inductive load dumps - https://www.diyaudio.com/community/threads/output-relays.191449/post-2843378
Place a TVS diode across the mosfets.
Nor is it necessary for zero crossing of the load current.
Why would you need to sense the load current ? The box is switching loudspeakers, its no different to plugging the speakers in/out of the back of the amplifier.
The IRF540N, IRF640N mosfets are not suitable, use modern mosfets, characteristics to look for in datasheets - RDS on, Qg, Avalanche energy.
Place a TVS diode across the mosfets.
Nor is it necessary for zero crossing of the load current.
Why would you need to sense the load current ? The box is switching loudspeakers, its no different to plugging the speakers in/out of the back of the amplifier.
The IRF540N, IRF640N mosfets are not suitable, use modern mosfets, characteristics to look for in datasheets - RDS on, Qg, Avalanche energy.
Thank you, I just read the long thread on output relays, and I'm thinking in terms of using BUK9K29-100E Dual 100V 30A N channel parts and probably the Broadcom/Avago ASSR-V621-300E dual gate driver (though the latter is not cheap).
What's the position about the dummy loads - are they needed or not? I can halve the switch count if I don't need them.
PS Does anyone know of a 4 channel optical gate driver IC?
Thanks, David
What's the position about the dummy loads - are they needed or not? I can halve the switch count if I don't need them.
PS Does anyone know of a 4 channel optical gate driver IC?
Thanks, David
Last edited:
@Indiglo I see you used a 100V TVS across the two 100V MOSFETS in the circuit you showed. Can the voltage rating of the TVS be increased given that there are two MOSFETS in series, or is the breakdown across the SSR dependent on the Vds voltage of only one MOSFET?
I suspect it it should be rated the same as the Vds of one MOSFET, but it's always best to ask!
Thanks
David
I suspect it it should be rated the same as the Vds of one MOSFET, but it's always best to ask!
Thanks
David
Transitions during the zero crossings of the voltage and current for turn on and turn off (respectively) enable fastest possible operation (for a given MOSFET) due to the absence of Miller Effect. This is especially true when a milliampere-current gate-drive (like the ones mentioned above) is used. In in doubt, refer to the following.
https://d1.amobbs.com/bbs_upload782111/files_42/ourdev_654767J2LOLK.pdf
Please note that the avalanche energy for this device (83mJ) is the non-repetitive (single-pulse) avalanche energy, which for the IRF540N is typically 700mJ.
This is a perfect example for over engineering. In no amp I have ever see problems with switching A-B speaker pairs by a simple contact.
Putting semiconductors in line with the signal is the worst idea I can think of.
What you should take into account is the fact that there are more and more bridge type amplifiers. So switching only the positive side will give an explosive result.
I have build such a device from simple and cheap standard parts, even switched remotely. The long time problem of arcing when switching AC with a relay can be reduced if you simply double the number of relais contacts. Which means one contact will be the first to take the hit, while the second keeps it's contact surface clean and resistance low. Has worked for high end for ages.
Putting semiconductors in line with the signal is the worst idea I can think of.
What you should take into account is the fact that there are more and more bridge type amplifiers. So switching only the positive side will give an explosive result.
I have build such a device from simple and cheap standard parts, even switched remotely. The long time problem of arcing when switching AC with a relay can be reduced if you simply double the number of relais contacts. Which means one contact will be the first to take the hit, while the second keeps it's contact surface clean and resistance low. Has worked for high end for ages.
Attachments
Dear OP,
If you want MOSFETs so much, all without the snubber and kickback problems, you could always switch when there's no music running.
For an accidental switching, you could always have your TVS diode across the device(s). The P6KE series bi-directional TVS diodes (use CA suffix) are available from 6.8V to 440V ratings in with a surge power capability of 600W @ 1ms.
https://www.mouser.com/datasheet/2/345/p6ke-4848.pdf
However, if you choose to use relays, there's no need for TVS / snubber. The low leakage current advantage of relays would also be compromised.
Not at all. For protection, you'd want the TVS to avalanche before the MOSFET. This places the TVS' minimum avalanche voltage below the breakdown VDS of the MOSFET.
If you want MOSFETs so much, all without the snubber and kickback problems, you could always switch when there's no music running.
- This prevents any load current from being interrupted during turn-off, giving no arcing, no snubber, no leakage.
- This also guarantees a zero-voltage turn-on, with no Miller effect.
- Since the music begins only after the changeover is complete, a weak gate driver with slow turn on may very well be used for control.
For an accidental switching, you could always have your TVS diode across the device(s). The P6KE series bi-directional TVS diodes (use CA suffix) are available from 6.8V to 440V ratings in with a surge power capability of 600W @ 1ms.
https://www.mouser.com/datasheet/2/345/p6ke-4848.pdf
However, if you choose to use relays, there's no need for TVS / snubber. The low leakage current advantage of relays would also be compromised.
Not at all. For protection, you'd want the TVS to avalanche before the MOSFET. This places the TVS' minimum avalanche voltage below the breakdown VDS of the MOSFET.
Last edited:
That's a automotive mosfet, probably best to stick with standard single power mosfets, there's is a better chance to find one to suit the application.
Probably not, a properly designed amplifier should be able to function without a load.
There are only single and dual types suitable to drive mosfets.
Yes
See this thread for a discussion on SSR (solid state relays) -
https://www.diyaudio.com/community/threads/ssr-for-speaker-protection.318946/post-7145185
I don't think the amplifiers you mentioned in your 1st post needs a dummy load. However, most tube amp and some solid state amplifiers do require a load to operate safely, my suggestion is leave it in for a peace of mind as well as future proof.
The resistors can use higher resistance and lower wattage rating, i.e. 33Ω 25W or 16Ω 50W. You can use the Finder relays or solid state switch for audio and use €0.2 cheap power relays for dummy loads.
SONGLE Power Relay
Last edited:
The average load an audio amp for private listening is not that high, so even small relais are sufficient, if you parallel two contacts for switching reasons.
Even loud music on average is only afraction of a Watt.
The 5 Volt TTL relais modules from eBay in 1-2-4 or 8- fold are dead cheap and very conveniant and simple to combine You even can realise dummy loads for a one in a million amp.
What I read here is so complicated and, beside for maybe for some valve amps, over complicated and near to nonsense.
If you are an engineer, you calculate loads, risks and cost of a solution, but do not thinker about non existing, invented phantasy problems. That is a sign of unsufficent knowlede and in essence pure Voodoo spiced snake oiling. Just like hitting a pure mechanical switching problem with the most complicated, sensitive and expensive semiconductor. The same semiconductor you are going to abolish in another thread, because of negative influence on the signal.
If you do not know how the amp-speaker connection works and what problem you it may, have learn, read, but don't spam a whole thread.
Even loud music on average is only afraction of a Watt.
The 5 Volt TTL relais modules from eBay in 1-2-4 or 8- fold are dead cheap and very conveniant and simple to combine You even can realise dummy loads for a one in a million amp.
What I read here is so complicated and, beside for maybe for some valve amps, over complicated and near to nonsense.
If you are an engineer, you calculate loads, risks and cost of a solution, but do not thinker about non existing, invented phantasy problems. That is a sign of unsufficent knowlede and in essence pure Voodoo spiced snake oiling. Just like hitting a pure mechanical switching problem with the most complicated, sensitive and expensive semiconductor. The same semiconductor you are going to abolish in another thread, because of negative influence on the signal.
If you do not know how the amp-speaker connection works and what problem you it may, have learn, read, but don't spam a whole thread.
I've priced up the different design options (MOSFETs versus relays). I decided to go with relays, and have omitted the dummy loads. I've use two 4PDT Phoenix Contact REL-IR4/LDP-_12DC/4X21 (part number 2903676) and have used two contact sets for switching each signal:
I've made the board up as a four layer board in 2oz. copper for the outer layers and 0.5oz for the inners (+12V, GND).
Kicad files available, I'll also have a few spare PCBs once JLCPCB deliver them - they went to fab today.
Software will be done in the new year.
David
I've made the board up as a four layer board in 2oz. copper for the outer layers and 0.5oz for the inners (+12V, GND).
Kicad files available, I'll also have a few spare PCBs once JLCPCB deliver them - they went to fab today.
Software will be done in the new year.
David
Last edited:
Two things:
* Since the body diode of Q1 is already forward biased, the ON / OFF status of Q1 would not do much to the operation of the circuit. You might as well skip that MOSFET.
* Failure of the MMBF170 would short out the ATTiny port to 12V through the relay coil. Adding a 1k series resistor to the port would limit this fault current to 7mA, which should be well within the limits of the internal clamping diodes of the ATtiny.
All the best.
* Since the body diode of Q1 is already forward biased, the ON / OFF status of Q1 would not do much to the operation of the circuit. You might as well skip that MOSFET.
* Failure of the MMBF170 would short out the ATTiny port to 12V through the relay coil. Adding a 1k series resistor to the port would limit this fault current to 7mA, which should be well within the limits of the internal clamping diodes of the ATtiny.
All the best.