Is there any speaker protection against damaging DC offset?
Having eviscerated a few expensive bass unit voice coils over the years due to this kind of mishap I get a bit nervous about nor having it now, especially if using hard to replace drive units.
Can this chip supply enough current to drive several pairs of MOSFETS?
Thanks.
Having eviscerated a few expensive bass unit voice coils over the years due to this kind of mishap I get a bit nervous about nor having it now, especially if using hard to replace drive units.
Can this chip supply enough current to drive several pairs of MOSFETS?
Thanks.
Hi Ian,
All very good questions!
There is no DC protection, and no intention of providing it. If you use this amp within any sort of reasonable power envelope, I don't see a reason to worry about DC protection. Most good quality power amplifiers don't have this, at least none of the amplifiers I've ever owned have. My entire life, I've never once blown a driver due to a failed output device or amplifier. I did once blow a ribbon tweeter due to my own ignorance.
You could try fast blow fuses set to just above your maximum power output, but you risk nuisance tripping. You could also try one of the many relay based circuits that can be found around this site, but my personal opinion is that usually by the time those things have reacted, it's too late anyhow.
This chip could indeed drive several mosfets in parallel without any problem at all. You would need to be cautious with wiring them up so as to not have a rats nest of wires hanging off the end of a very high BW amplifier. Each fet should have its own gate resistor, but that would be easy to fly wire. in place. I would put the practical limit at about 4 parallel devices per rail.
The answer to this is NO!
Maybe I should repeat that for anyone who wasn't paying attention:
YOU CANNOT USE NON-LATERAL MOSFETS WITH THIS DESIGN!!!!!!!!
The number one reason is temperature co-efficient. Lateral mosfets are wonderful in that the hotter they get, the lower Ids gets for a given Vgs. This is called a negative temperature coefficient and it means that if you set the bias at a certain level, the hotter the output transistors get, the lower the bias goes.
Normal mosfets and BJT's have a positive temperature coefficient which means that as they get warmer, Ids increases for a given Vgs. In other words, as the amp warms up, it will draw a little more bias current, then it'll get a little warmer and draw a little more bias current and next thing you know your amplifier is on fire.
This design was specifically done for lateral mosfets, and that's why there's no bias compensation network and no emitter resistors. Lateral fets are also less likely to oscillate and they don't have secondary breakdown like a BJT would.
There are a variety of lateral fets you can use in this amp, but you cannot use anything that's not a lateral fet.
Cheers,
Owen
All very good questions!
There is no DC protection, and no intention of providing it. If you use this amp within any sort of reasonable power envelope, I don't see a reason to worry about DC protection. Most good quality power amplifiers don't have this, at least none of the amplifiers I've ever owned have. My entire life, I've never once blown a driver due to a failed output device or amplifier. I did once blow a ribbon tweeter due to my own ignorance.
You could try fast blow fuses set to just above your maximum power output, but you risk nuisance tripping. You could also try one of the many relay based circuits that can be found around this site, but my personal opinion is that usually by the time those things have reacted, it's too late anyhow.
This chip could indeed drive several mosfets in parallel without any problem at all. You would need to be cautious with wiring them up so as to not have a rats nest of wires hanging off the end of a very high BW amplifier. Each fet should have its own gate resistor, but that would be easy to fly wire. in place. I would put the practical limit at about 4 parallel devices per rail.
The answer to this is NO!
Maybe I should repeat that for anyone who wasn't paying attention:
YOU CANNOT USE NON-LATERAL MOSFETS WITH THIS DESIGN!!!!!!!!
The number one reason is temperature co-efficient. Lateral mosfets are wonderful in that the hotter they get, the lower Ids gets for a given Vgs. This is called a negative temperature coefficient and it means that if you set the bias at a certain level, the hotter the output transistors get, the lower the bias goes.
Normal mosfets and BJT's have a positive temperature coefficient which means that as they get warmer, Ids increases for a given Vgs. In other words, as the amp warms up, it will draw a little more bias current, then it'll get a little warmer and draw a little more bias current and next thing you know your amplifier is on fire.
This design was specifically done for lateral mosfets, and that's why there's no bias compensation network and no emitter resistors. Lateral fets are also less likely to oscillate and they don't have secondary breakdown like a BJT would.
There are a variety of lateral fets you can use in this amp, but you cannot use anything that's not a lateral fet.
Cheers,
Owen
Thanks for your answers 
It was DC from source components that did the damage.
A wire came adrift from a CD player PSU and the output stage applied the 12V rail to the amplifier. This fried two bass units and deleted two MOSFETS in a de-speaker protection fused Hafler DH200. 'Cause fuses sound bad, wherever they are. I don't use them, not even in the mains supply. For the Hi-Fi spur I have just one 30amp at the meter board in it's own unswitched holder as switches sound bad too. I pull some plugs to turn off the amps. I bought a mercury switch recently as I wondered if a liquid switch would do it. It didn't; very thin and coarse sound. I made up a 16 pole relay for one mono block. No, sound went a bit thin, a bit coarse and bit dull. One or two or three 0.1uF X caps across the contact helped the mid and top clarity but not the thinness.
Another time I was using a scope and probing around in a player and slipped with the probe tip. The output stage op-amps latched to the rail and did the same as before.
From then on I used protection. The DC coupled Meridian 605's, which I've extensively upgraded, have a sensitive DC detection and switch off the mains supply with a relay. The relay does damage the sound quality but speaker relays are worse. Seems like any protection anywhere has a noticeably bad effect I've been glad of that protection a few times since.
I wondered in the mute feature could be employed to shut down this 49830 if DC were detected on the output, or even on the input?
I think my MOSFETS are complimentary pairs of 2SJ162 and 2SK1058. The pin out is G.S.D. And some www digging just now tells me they are Lateral.
What attracted me to your 49830 was the mention of a balanced going straight in rather than via some more components.
So if I were to do this it would be:
- What load do you plan to drive? The small speakers drop to only about 2 ohms at 110Hz. Rated at 8Ohms I think. The big speakers at about 6 Ohms.
- What power output do you want? Maybe 300W with the big ones. They are about 87dB and eat power. Years ago I bought some custom 1000VA Toroid Tx's with extra secondaries for the regulated early stage. They are still waiting for an amp. With these a MOSFET amp might make about 290Watts it seems, IIRC the post bridge late stage voltage is about 90VDC and early about 110VDC. If I did it again I'd have had them bit lower so I could safely use the Nichicon KZ 100V caps for the early stage as they sounded a bit better than most others I tired at the time.
- Do you plan to run differential or SE input? SE First then Diff once I make a Balanced Source.
- Will you build a higher voltage regulated supply for the front end? Yes
It was DC from source components that did the damage.
A wire came adrift from a CD player PSU and the output stage applied the 12V rail to the amplifier. This fried two bass units and deleted two MOSFETS in a de-speaker protection fused Hafler DH200. 'Cause fuses sound bad, wherever they are. I don't use them, not even in the mains supply. For the Hi-Fi spur I have just one 30amp at the meter board in it's own unswitched holder as switches sound bad too. I pull some plugs to turn off the amps. I bought a mercury switch recently as I wondered if a liquid switch would do it. It didn't; very thin and coarse sound. I made up a 16 pole relay for one mono block. No, sound went a bit thin, a bit coarse and bit dull. One or two or three 0.1uF X caps across the contact helped the mid and top clarity but not the thinness.
Another time I was using a scope and probing around in a player and slipped with the probe tip. The output stage op-amps latched to the rail and did the same as before.
From then on I used protection. The DC coupled Meridian 605's, which I've extensively upgraded, have a sensitive DC detection and switch off the mains supply with a relay. The relay does damage the sound quality but speaker relays are worse. Seems like any protection anywhere has a noticeably bad effect I've been glad of that protection a few times since.
I wondered in the mute feature could be employed to shut down this 49830 if DC were detected on the output, or even on the input?
I think my MOSFETS are complimentary pairs of 2SJ162 and 2SK1058. The pin out is G.S.D. And some www digging just now tells me they are Lateral.
What attracted me to your 49830 was the mention of a balanced going straight in rather than via some more components.
So if I were to do this it would be:
- What load do you plan to drive? The small speakers drop to only about 2 ohms at 110Hz. Rated at 8Ohms I think. The big speakers at about 6 Ohms.
- What power output do you want? Maybe 300W with the big ones. They are about 87dB and eat power. Years ago I bought some custom 1000VA Toroid Tx's with extra secondaries for the regulated early stage. They are still waiting for an amp. With these a MOSFET amp might make about 290Watts it seems, IIRC the post bridge late stage voltage is about 90VDC and early about 110VDC. If I did it again I'd have had them bit lower so I could safely use the Nichicon KZ 100V caps for the early stage as they sounded a bit better than most others I tired at the time.
- Do you plan to run differential or SE input? SE First then Diff once I make a Balanced Source.
- Will you build a higher voltage regulated supply for the front end? Yes
Last edited:
Perhaps giving up on DC coupling would be the solution to your problem? You could drop a few thousand on some Duelund CAST-cu caps to couple the inputs, but unfortunately they won't fit on the PCB.
In all honesty though, I'm glad you described your issues and related them to user errors in your front end circuitry. I've been there and done that, and sometimes having a safety net just means you'll learn the lesson the hard way a little later than you otherwise would have. Simple things likes "If you're working on it, unplug it" and "check twice connect once" will go a long ways towards ensuring the safety of your downstream gear, and more importantly, your own personal safety.
The mosfets you mentioned definitely won't work with this amplifier as they're regular enhancement mode fets. You'll have to splurge on a few laterals if you want to build this amp.
Cheers,
Owen
In all honesty though, I'm glad you described your issues and related them to user errors in your front end circuitry. I've been there and done that, and sometimes having a safety net just means you'll learn the lesson the hard way a little later than you otherwise would have. Simple things likes "If you're working on it, unplug it" and "check twice connect once" will go a long ways towards ensuring the safety of your downstream gear, and more importantly, your own personal safety.
The mosfets you mentioned definitely won't work with this amplifier as they're regular enhancement mode fets. You'll have to splurge on a few laterals if you want to build this amp.
Cheers,
Owen
I do not agree, I will never use amplifier without DC Speaker protection.
Triac Crowbar is cheap abour 3 EUR and protect all your speaker
some exepnsive esoteric amplifiers dont have VI Limiter and DC protection, this is crazy
and this folks tell you in Fairs beautiful story about improving sound without this proetctions
What a Fuuck
Last edited:
ahh, i didnt realize, never used them before and took it opc had, so just reiterated that most laterals will work
cant say ive owned an amp with protection before either. only one of the dacs i'm using in my setup is DC coupled, actually its the one i was going to use with this amp, but i'm not running such high power i'm not that fussed. i'll look into it and may try to see what i think, but we'll see. was thinking of trying the low RDSON fet switch method
Last edited:
Until recently most of the amps I've had had protection until I took it all off to improve sound quality. But I've been trying a bit, in a clunky unelegant way, to achieve both.
Taking a risk is OK is you have drive units that can be purchased or plenty of spare cash. Many of my pals are using older but very good sounding Focal kevlar units that can't and they mostly can't afford to just bin the speakers and get new ones of equivalent quality. Some are using even older Kef B139's in large Transmission Lines speakers that make a great clean subsonic bass. No longer available, I had about or actually the last units from the UK's Kef repair centre based near me as they were the 4 units I smoked.
Taking a risk is OK is you have drive units that can be purchased or plenty of spare cash. Many of my pals are using older but very good sounding Focal kevlar units that can't and they mostly can't afford to just bin the speakers and get new ones of equivalent quality. Some are using even older Kef B139's in large Transmission Lines speakers that make a great clean subsonic bass. No longer available, I had about or actually the last units from the UK's Kef repair centre based near me as they were the 4 units I smoked.
Last edited:
A pair of boards please.
Thanks
In which case, no either way, Owen can you please put me down for a pair of boards should more become available.
Thanks
Last edited:
Would that be like fig 26.3b on page 529?
Designing Audio Power Amplifiers - Bob Cordell - Google Books
I presume it's OK to show this here as I found it in the public domain on Google Books and so not copyright?
The real grist would seem to be on the following couple of pages which are withheld.
Last edited:
@opc
Would it be possible to add a small 2.54mm pin header between the MUTE pin and GND (in parallel with C75)? The reason is that im thinking about having an external mute button and it would be nice to be able to connect an optocoupler (to avoid a long cable/antenna) there to short the mute pin to gnd.
The datasheet recommends to ground trough R38 but that would cause trouble with this design so i hope it works just as well to ground the mute pin directly.
While on that signal, why do you have the extra vias between R38 and C75? Wouldn't it be better to draw the entire trace on the red layer (top?)?
@ Everyone
How does it look with a GB for power supply?
The connex website is up and running again and the SMPS500R (77.99$) or SMPS800R (99.99$) looks like a good match for this project. Cristi here at the forum is the guy building these supplies (if im not mistaken) and he seems very helpful to make modifications to adapt the supplies to the user preferences. This post on how to get higher AUX voltage should be interesting for the people here. So we should be able to get a tailor made supply for a given rail voltage at reasonable price.
If this works we could probably get an efficient supply with high power output and a regulated higher voltage supply as requested. This should be in the 100$ ballpark + shipment.
The question is what would the cost for a normal toroid/rectifier design end up on for say a 60 50 + -50 -60 VDC supply? And most importantly would you be able to hear/measure any significant difference between the two?
Would it be possible to add a small 2.54mm pin header between the MUTE pin and GND (in parallel with C75)? The reason is that im thinking about having an external mute button and it would be nice to be able to connect an optocoupler (to avoid a long cable/antenna) there to short the mute pin to gnd.
The datasheet recommends to ground trough R38 but that would cause trouble with this design so i hope it works just as well to ground the mute pin directly.
While on that signal, why do you have the extra vias between R38 and C75? Wouldn't it be better to draw the entire trace on the red layer (top?)?
@ Everyone
How does it look with a GB for power supply?
The connex website is up and running again and the SMPS500R (77.99$) or SMPS800R (99.99$) looks like a good match for this project. Cristi here at the forum is the guy building these supplies (if im not mistaken) and he seems very helpful to make modifications to adapt the supplies to the user preferences. This post on how to get higher AUX voltage should be interesting for the people here. So we should be able to get a tailor made supply for a given rail voltage at reasonable price.
If this works we could probably get an efficient supply with high power output and a regulated higher voltage supply as requested. This should be in the 100$ ballpark + shipment.
The question is what would the cost for a normal toroid/rectifier design end up on for say a 60 50 + -50 -60 VDC supply? And most importantly would you be able to hear/measure any significant difference between the two?
I can only add that with the various power amps I've had over the last 18 years and those of friends, everything you do make a difference to the sound, often a huge change.
Apart from each different size and make of toroid sounding different, I worked my way up to 1000VA ones. All custom made. You don't need to be pulling big power and current for a big Tx to confer a better sound. I had a 1000VA on my CD player DAC / analogue stage eventually which blew the small ones totally away even though there was only 100mA draw, discrete regs, 50,000uF of Black Gate and Nichicon KZ goodness between Br and regs and 1000uf of BG's and Oscons after them. Each different big Tx was totally different sounding, some really sweet and nice, other full and bassy. Different makes and type and quantities of main res caps, fuses and switches and relays at all anywhere, which way around you connect the mains cable to the toroid, every different type of IEC connector and mains cable. I've even soldered an amplifier Tx direct to the 25mm2 mains supply coming from the meter board with only the one main 20amp fuse there, was so much better My pal couldn't turn the amp off ever, but it was a compromise that he was happy with, and that one was DC coupled with no speaker protection. Each different small caps across the main res, even if they are a bank of Nichicon grade 4 at 80,000uF, makes a difference. I now take my loud speaker earth return directly to the main res caps earth star, that was well worth doing.
I'd be very keen to know if a switch mode set up was free from all this.
Apart from each different size and make of toroid sounding different, I worked my way up to 1000VA ones. All custom made. You don't need to be pulling big power and current for a big Tx to confer a better sound. I had a 1000VA on my CD player DAC / analogue stage eventually which blew the small ones totally away even though there was only 100mA draw, discrete regs, 50,000uF of Black Gate and Nichicon KZ goodness between Br and regs and 1000uf of BG's and Oscons after them. Each different big Tx was totally different sounding, some really sweet and nice, other full and bassy. Different makes and type and quantities of main res caps, fuses and switches and relays at all anywhere, which way around you connect the mains cable to the toroid, every different type of IEC connector and mains cable. I've even soldered an amplifier Tx direct to the 25mm2 mains supply coming from the meter board with only the one main 20amp fuse there, was so much better
My pal couldn't turn the amp off ever, but it was a compromise that he was happy with, and that one was DC coupled with no speaker protection. Each different small caps across the main res, even if they are a bank of Nichicon grade 4 at 80,000uF, makes a difference. I now take my loud speaker earth return directly to the main res caps earth star, that was well worth doing.I'd be very keen to know if a switch mode set up was free from all this.
Last edited: