And you need a lot laterals, as their transconductance is a bit weak compared to verticals.
Personally I'm not very keen on laterals in power amps as verticals perform better at high bias and for low bias one gets better linearity from (high-beta) bipolars.
But in the end I would see no reason not build an amp because of its use of laterals in the output stage (and one can mod it to verticals anyway).
Have fun, Hannes
What? no pots? Just kidding, pots can fail, it sounds like a novel idea, especially to experiment and compare sound differences wrt bias current.
IMHO, vertical fets are better to my likening. Laterals cost too much. I can crank out 100Wrms from 2 pair of TO-220 hexfets and they cost 1 dollar each. If you overdrive or screw up your circuit and the transistor blows up, you’re not out that much. As stated by Hannes their transconductance is much higher. Verticals are faster too. As for linearity, with a good error correction scheme (EEK!
complexity ) Verticals can match up with BJT's, and even better them on heavy reactive loads since they have no secondary breakdown.
IMHO, vertical fets are better to my likening. Laterals cost too much. I can crank out 100Wrms from 2 pair of TO-220 hexfets and they cost 1 dollar each. If you overdrive or screw up your circuit and the transistor blows up, you’re not out that much. As stated by Hannes their transconductance is much higher. Verticals are faster too. As for linearity, with a good error correction scheme (EEK!
complexity ) Verticals can match up with BJT's, and even better them on heavy reactive loads since they have no secondary breakdown.
I'm listening I think they both have advantages and disadvantages.
If you want to know what I have been doing I have been seeing how my Lateral Mosfet design "takes" to vertical types. Very few changes seem to be needed, mainly the adition of a Vgs multiplier with good thermal tracking of the outputs. Slight change to some resistor values of the drivers but that's about it.
From a single pair of IRF9140/140 the performance on test was impressive. Particularly surprising was how well they coped with low impedance loads and most surprising of all -- stability seeing as it was a bit of a lash up.
thanks guys, keep the comments coming.
I think they both have advantages and disadvantages. If you want to know what I have been doing I have been seeing how my Lateral Mosfet design "takes" to vertical types. Very few changes seem to be needed, mainly the adition of a Vgs multiplier with good thermal tracking of the outputs. Slight change to some resistor values of the drivers but that's about it.
From a single pair of IRF9140/140 the performance on test was impressive. Particularly surprising was how well they coped with low impedance loads and most surprising of all -- stability seeing as it was a bit of a lash up.
thanks guys, keep the comments coming.
Mooly said:I'm listening
If you want to know what I have been doing I have been seeing how my Lateral Mosfet design "takes" to vertical types. Very few changes seem to be needed, mainly the adition of a Vgs multiplier with good thermal tracking of the outputs. Slight change to some resistor values of the drivers but that's about it.
From a single pair of IRF9140/140 the performance on test was impressive. Particularly surprising was how well they coped with low impedance loads and most surprising of all -- stability seeing as it was a bit of a lash up.
thanks guys, keep the comments coming.
I Mooly,
I use laterals almost exclusively in my designs nowadays. Granted they cost more. but I have no negative things to report about them (2SK1058/2SJ162). I never use a Vbe multiplier and I have never ran into thermal problems.
Hi Nico,
"Nothing negative to report" only the tempco. And we like that . And they are tough, I once years ago accidently shorted the speaker output to one of the rails on an earlier amp. Flash/bang, but no harm done.
Just interested really in what I can do with the verticals, can I match the sonics of,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=119151&perpage=20&highlight=&pagenumber=1
The verticals are just so cheap now, not that that's a reason in itself of course.
"Nothing negative to report" only the tempco. And we like that
. And they are tough, I once years ago accidently shorted the speaker output to one of the rails on an earlier amp. Flash/bang, but no harm done. Just interested really in what I can do with the verticals, can I match the sonics of,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=119151&perpage=20&highlight=&pagenumber=1
The verticals are just so cheap now, not that that's a reason in itself of course.
Mooly said:Hi Nico,
"Nothing negative to report" only the tempco. And we like that
Just interested really in what I can do with the verticals, can I match the sonics of,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=119151&perpage=20&highlight=&pagenumber=1
The verticals are just so cheap now, not that that's a reason in itself of course.
I prefer verticals, but one can build a very good amplifier out of either type.
I think it is fair to say that one can get into more trouble with verticals than with laterals, so they need a bit more care and feeding to be applied safely and optimally. Since they are much faster than laterals, the verticals are more prone to oscillation unless good measures are taken to avoid it. To get optimum stability with verticals while not throwing away all of their speed, sometimes a gate zobel network is needed.
While verticals can produce amazingly large currents (usually a desirable good thing), this can sometimes be their downfall if there is not some kind of short circuit protection. Laterals just seem to be naturally more robust to screwups.
Cheers,
Bob
Mooly said:Hi Nico,
"Nothing negative to report" only the tempco. And we like that
Just interested really in what I can do with the verticals, can I match the sonics of,
http://www.diyaudio.com/forums/showthread.php?s=&threadid=119151&perpage=20&highlight=&pagenumber=1
The verticals are just so cheap now, not that that's a reason in itself of course.
I find the lateral fets very interesting devices, I'd use them but too bad they are expensive. I'm not that surprised the transistors survived the rail-to-output short though.
I deliberately tried to blow a lateral FET up last year:
http://www.diyaudio.com/forums/showthread.php?threadid=111116
50V 10A power supply over D-S and hit it with 9V on the gate on a too small heatsink... And it survived! Did it ten times and it still works!
Expect for the transconductance decreasing a lot as it heats up there seems to be another pretty non-obvious mechanism protecting it. Seems like the internal protection zeners start conducting like a thyristor when the temperature gets high enough. This clamps gate voltage to ~1V provided gate current drive is limited and at this gate voltage and temperature the transistors don't let much current through. It was 0.2A with the transistor I tested on the heatsink I had.
This might explain why they are so hard to blow up as many have experienced.
To take advantage of this one has to make sure:
* External gate zeners limit the gate voltage so the current can't go high enough to blow the internal drain wire or damage the transistor in other ways. Probably more of a problem with the metal cans as the plastic ones have the wire embedded in epoxy, increasing it's overload capacity.
* Gate current is limited. The gate resistors and/or VAS should limit the gate current to a not-to-high value if the transistor starts drawing current into the gate. With a standard topology current limited VAS is needed anyway as the external zeners should be protected.
If the VAS is not current limited the VAS transistor is probably the first component to blow with a shorted output. This will then take out resistors and stuff, hopefully the output stage gate resistors will fuse without outputs blowing but the amp is broken nevertheless.
well.. i did mean there is a bias pot, but the jumpers would allow quick change from a 1.0-2.0V range to 2-4V, 3-8V ranges in order to quickly get "in the ballpark" for different output device combinations. using a simple 8 pin (2x4) header and a jumper block. this would be for prototyping only. i wouldn't want this feature available to consumers. too much temptation for the curious mind.....
i've seen it before.... "hmm...what does this do????....... SNAP!!!! SSSSZZZZTTBOOOOMMMM....."
i've seen it before.... "hmm...what does this do????....... SNAP!!!! SSSSZZZZTTBOOOOMMMM....."
Mooly, good to see you. as long as the amp sounded good i would have no problem using L-MOSFET in the OPS. i (very slightly) prefer the sound of bipolars to mosfets actually, but thats just my purely subjective opinion.
ROFL! Yeah i've seen this a few times too and been guilty of it myself as well from time to time i must admit. ahhh, cats and curiosity.
unclejed613 said:
ROFL! Yeah i've seen this a few times too and been guilty of it myself as well from time to time i must admit. ahhh, cats and curiosity.
Bob wrote,
"A Zobel network on the gate" interesting thought, I hadn't even considered that.
I haven't run into any stability problems so far with the verticals in a CFP arrangement, I use a 100mhz 'scope and all seems well.
180 ohms on the gates, I am amazed how docile they seem.
Are they going to bite when I least expect it ?
Quiescent current was another worry. Folks seem to say "more is better" when it comes to VFets. Again I found the non linearity using the CFP stage dissappeared at around 100ma, but a Vgs multiplier in contact with the outputs seems a must.
Zobel on the gate eh !!
Thanks
"A Zobel network on the gate" interesting thought, I hadn't even considered that.
I haven't run into any stability problems so far with the verticals in a CFP arrangement, I use a 100mhz 'scope and all seems well.
180 ohms on the gates, I am amazed how docile they seem.
Are they going to bite when I least expect it ?
Quiescent current was another worry. Folks seem to say "more is better" when it comes to VFets. Again I found the non linearity using the CFP stage dissappeared at around 100ma, but a Vgs multiplier in contact with the outputs seems a must.
Zobel on the gate eh !!
Thanks
unlike tubes, gates should never draw any current. unless they're shorted to the channel or they have internal zener protection. with RF there's also the possibility of RF current into the gate capacitance.
would the zobels go to the source or ground? probably makes sense to go to the source since that would shunt any RF to the source.
i have a book here with a test circuit for testing various "drop in" output stages. its basically a generic class B amp with (socketed?) "DUT" spots where you connect BJT's, darlingtons, Sziklai pairs, or MOSFETs. that's why the idea of the jumper configurable bias.
VFETs actually have a positive tempco only up to a point (up to about 100mA or so, sometimes as high as 1 amp) after that it turns negative like L-MOSFETs. that explains why they behave better with higher bias current.
would the zobels go to the source or ground? probably makes sense to go to the source since that would shunt any RF to the source.
i have a book here with a test circuit for testing various "drop in" output stages. its basically a generic class B amp with (socketed?) "DUT" spots where you connect BJT's, darlingtons, Sziklai pairs, or MOSFETs. that's why the idea of the jumper configurable bias.
VFETs actually have a positive tempco only up to a point (up to about 100mA or so, sometimes as high as 1 amp) after that it turns negative like L-MOSFETs. that explains why they behave better with higher bias current.
"unlike tubes, gates should never draw any current. unless they're shorted to the channel or they have internal zener protection. with RF there's also the possibility of RF current into the gate capacitance."
The laterals have internal zener protection. They seem to latch up at lower voltage at high temperature protecting the transistor.
The laterals have internal zener protection. They seem to latch up at lower voltage at high temperature protecting the transistor.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.