So I'm just wondering about these MOSFETs. Could I do a straight swap for them in say an amp that uses 2SK1058 and complementary so I could have 2.5 times the dissipation and effectively have nearly three pairs of 2SK1058/2SJ162? But the three pairs would require more current from the VAS to drive them is this the same for the Exicons as they are double die?
ECW20N20-Z
Thanks
Boscoe
ECW20N20-Z
Thanks
Boscoe
The ECW20N20-Z has a Ciss of 900pF compared to the 2SK1058's 600pF, therefore, one pair of the Exicon's would still be easier to drive than two pairs of the Renesas parts (assuming my reasoning is correct).
The Exicon parts are stocked in the US as a ALF16N20 (Semelab) with a K or W suffix for to-3 or to-264 packages. The Exicon datasheets say these are 20A max. parts. The Semelab datasheet says these are 16A max. parts. Pretty sure these are the same parts with double dies.
As to replacements of the K1058/J162, which are 7A parts.....I'd use (1) ECW20N20 (ALF16N20) to replace (2) 2SK1058. My conservative engineering is showing....
As to driving them......My experiments with Spice models of the double die parts shows that any reasonably designed amp that uses 2SK/2SJ parts has no trouble driving them. Might have to tune the series R in the gate though. Input capacitance is about the same for either part.
As to replacements of the K1058/J162, which are 7A parts.....I'd use (1) ECW20N20 (ALF16N20) to replace (2) 2SK1058. My conservative engineering is showing....
As to driving them......My experiments with Spice models of the double die parts shows that any reasonably designed amp that uses 2SK/2SJ parts has no trouble driving them. Might have to tune the series R in the gate though. Input capacitance is about the same for either part.
If you check Profusion.com (Exicon) and Semelab (Alfet) websites, the specs are identical for these brandings and the range of available lateral Mosfets is much the same.
The EXC10N20 or ALF08N16 are said to be drop-ins for the 2SK1058/J162 and still slightly higher rated. They work for any design I've seen them substituted in anyway.
As far as getting the most out of double-die mosfets, wouldn't you have to go to a TO3 can to get the thermal dissipation to match 2 x TO247 or TO 264?
The EXC10N20 or ALF08N16 are said to be drop-ins for the 2SK1058/J162 and still slightly higher rated. They work for any design I've seen them substituted in anyway.
As far as getting the most out of double-die mosfets, wouldn't you have to go to a TO3 can to get the thermal dissipation to match 2 x TO247 or TO 264?
Hi,
What are the advantages of using these devices ( laterals ) in comparison to say 2SJ201 &2SK1530, particularly Dual die devices besides power handling capabilities..
What are the advantages of using these devices ( laterals ) in comparison to say 2SJ201 &2SK1530, particularly Dual die devices besides power handling capabilities..
Laterals have very good thermal stability. If you look to datasheets you could find that laterals are decreasing output current at higher temperatures. That is opposite to the vertical MOSFETS.
That's right - they're about the only type of power transistor where you don't need to worry to some degree, about thermal servo controls. They self-limit and share current superbly. I have Hitachi Mosfet amps from the 1980s which are still fully functional, unbreakable even into dead shorts with no protection circuits, relays or gizmos at all.
That's over 30 years of regular hard work and abuse at up to 150W/8R. Distortion figures are not as low as BJTs but these are in the order of 0.OO5% 1kHz so there is little to argue about anyway.
The downsides are the range, cost and availability which prevent their use in monster PA amps - where the audio profits are. Neither do you find them in industrial application for which most VMosfet audio parts were intended.
That's over 30 years of regular hard work and abuse at up to 150W/8R. Distortion figures are not as low as BJTs but these are in the order of 0.OO5% 1kHz so there is little to argue about anyway.
The downsides are the range, cost and availability which prevent their use in monster PA amps - where the audio profits are. Neither do you find them in industrial application for which most VMosfet audio parts were intended.
Ian Finch, Igabco
Thanks very much for taking time to answer my query 🙂. Just wondering are these devices suitable for Hi-Fi / Ultra Hi-fi use like class A?
Haven't seen them used in Class A amp design with high /moderate bias current.
Regards.
Thanks very much for taking time to answer my query 🙂. Just wondering are these devices suitable for Hi-Fi / Ultra Hi-fi use like class A?
Haven't seen them used in Class A amp design with high /moderate bias current.
Regards.
class A Mosfets
Lateral Mosfets were solely intended for Hi-Fi Audio application, AFAIK.
Class A was not a significant force when they were developed but the point about devices that don't switch off is that linearity is less a problem,for sound quality so often, really crap devices sound ok and have nice distortion profiles in class A. The JLH 10W amp is a good example with it's horror-story 2N3055s eek:
You get the worst efficiency with Mosfet class A, particularly single ended, where you don't even think about it. It's just a furnace for a few watts unless you only want a headphone amplifier. The distortion with some of these SE amps is often over 3% so linearity can't be a concern but big transistors should be just as capable of heating your room whilst getting a little more power out as well.
Linear class A, push-pull, is another matter. The Silicon Chip 20W transistor model, for example, has unbelieveably low THD+Noise. Here you would need at least lateral Mosfets if that was your choice of outputs. However this isn't what many audiophiles like - some prefer a range of distortion to sweeten the audio up. It should be said that latFETS are biased at different voltage, don't need thermal compensation and so aren't drop-in's for VFETS but there are plenty of schematics around.
Nelson Pass's legendary SE approach probably relies on specific VMOS distortion for the particular sound quality his fans like 😎
Lateral Mosfets were solely intended for Hi-Fi Audio application, AFAIK.
Class A was not a significant force when they were developed but the point about devices that don't switch off is that linearity is less a problem,for sound quality so often, really crap devices sound ok and have nice distortion profiles in class A. The JLH 10W amp is a good example with it's horror-story 2N3055s eek:
You get the worst efficiency with Mosfet class A, particularly single ended, where you don't even think about it. It's just a furnace for a few watts unless you only want a headphone amplifier. The distortion with some of these SE amps is often over 3% so linearity can't be a concern but big transistors should be just as capable of heating your room whilst getting a little more power out as well.
Linear class A, push-pull, is another matter. The Silicon Chip 20W transistor model, for example, has unbelieveably low THD+Noise. Here you would need at least lateral Mosfets if that was your choice of outputs. However this isn't what many audiophiles like - some prefer a range of distortion to sweeten the audio up. It should be said that latFETS are biased at different voltage, don't need thermal compensation and so aren't drop-in's for VFETS but there are plenty of schematics around.
Nelson Pass's legendary SE approach probably relies on specific VMOS distortion for the particular sound quality his fans like 😎
Last edited:
Thanks for the info.
I don't understand class a people says they have more fidelity yet a good ab amp will beat the class a amp pretty much every single time on distortion. Distortion is the inverse of fidelity, fact!
I don't understand class a people says they have more fidelity yet a good ab amp will beat the class a amp pretty much every single time on distortion. Distortion is the inverse of fidelity, fact!
no reason you cant run them in class A with enough bias. i'll be trying out dropping the rails down to +/-25-30v (40-45v on the LME) and upping the bias over 1A on one of my 'the wire' Semelab + lme49830 amps, for about 30wpc class A. It should be interesting to compare versus the same amp with +/-55v (70v on the lme) and 500-600ma bias. seeing how the trade-off of better linearity in Class A vs better linearity from higher voltage works out.
Class A IS higher performance, there is no question there, the same amp with Class A bias will have lower distortion (already vanishingly low in this case) and better crossover distortion, but whether its worth it is personal taste I guess. I dont disagree that a well designed AB, particularly with todays parts, doesnt really lose out enough to make it worth worrying about; Class A doesnt have to be a cliche
Class A IS higher performance, there is no question there, the same amp with Class A bias will have lower distortion (already vanishingly low in this case) and better crossover distortion, but whether its worth it is personal taste I guess. I dont disagree that a well designed AB, particularly with todays parts, doesnt really lose out enough to make it worth worrying about; Class A doesnt have to be a cliche
Unless massive, class A just doesn't have enough power to drive speakers properly especially when driving passives. It's not just the amp you have to consider.
I am going to be provocative again and suggest that most amplifiers are ClassA.
It's just the output stage of some, particularly power amps and opamps, that are not ClassA.
All those ClassA stages and ClassA amplifiers are very capable of low levels of distortion. Even the single ended ones (and they dominate the ClassA stages at the front end and middle of almost all linear amplifiers) have low levels of distortion. Look at some of the modern opamps, look at the "blameless" series of power amplifiers.
It's just the output stage of some, particularly power amps and opamps, that are not ClassA.
All those ClassA stages and ClassA amplifiers are very capable of low levels of distortion. Even the single ended ones (and they dominate the ClassA stages at the front end and middle of almost all linear amplifiers) have low levels of distortion. Look at some of the modern opamps, look at the "blameless" series of power amplifiers.
Is it really?
High Fidelity is reproduction as close to the original as possible. Any deviance from the original is distortion.
Hi Boscoe
Perhaps the word "fidelity" is being stretched in the direction of "niceness" when people talk subjectively about preferred amplification?
What the specifications of AB amplifiers don't tell you is what is happeneing to distortion at below the first watt....in the microdetails
which are shot to ribbons by crossover distortion at low level in standard class AB IMO. Class A keeps the integrity down to inaudible
so for general cases this is what we find. It is the reason why hi-end manufacturers whomp up the bias to cover a few watts of output
level before the transition to AB. The distortion figures for these designs are usually worse but SQ now could well be "hi-end" 😀
Perhaps the word "fidelity" is being stretched in the direction of "niceness" when people talk subjectively about preferred amplification?
What the specifications of AB amplifiers don't tell you is what is happeneing to distortion at below the first watt....in the microdetails
which are shot to ribbons by crossover distortion at low level in standard class AB IMO. Class A keeps the integrity down to inaudible
so for general cases this is what we find. It is the reason why hi-end manufacturers whomp up the bias to cover a few watts of output
level before the transition to AB. The distortion figures for these designs are usually worse but SQ now could well be "hi-end" 😀
- Status
- Not open for further replies.