Crossover distortion in push-pull

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Vgs values of 10Vgs to 20Vgs is near saturation level and near destruction level.

Note also the junction peak dissipation 1280Wpk for the 140N and 2880Wpk for the 240.
How fast will the junction heat up?
How long can that power pulse last before you damage the device?

I don't think you are looking at the correct data for your application.
 
Member
Joined 2010
Paid Member
Love Virus
There are no precise complementary matches for power Mosfets - just close as feasible for manufacture. The closest match for IRF/P 9140 is IRF/P 240 and not what might look more obvious.

Crossover distortion in Mosfets is not minimal at some particular bias level, as with transistors. Rather. it starts at worst with zero and trends slowly down until full class A bias is applied. However, we don't all want low power burning boxes so we find that for most builders, where adequate NFB is applied, a good compromise begins with around 150mA bias current per pair of these devices.

Check the many threads here discussing designs using these most popular VFET pairs and check the facts in Bob Cordell's book covering your concerns and interest well.
 
Last edited:
Qgs is not very important in this application. The datasheets reflect the intention of these devices as being designed for switching purposes. Here is a related post.

Exactly, they're usually not intended for lineair operation, but can be used as such with (much) lower currents, the most important is their power derating stat to figure out tolerable dissipation with lineair operation. If the higher capacitance isn't an issue, one with a higher transconductance (and those HEXfets have a lot of it) is always better. These conditions makes them excellent as cascode devices since the parasitic capacitance has no effect in common gate configuration and the high transconductance reduces Vgs modulation making the cascode more robust.

I personally never liked the idea of using these as output devices because of the non-lineair behaviour at higher currents that are common to output stages. Laterals are so much better in that regard :)
 
Last edited:
Love Virus
There are no precise complementary matches for power Mosfets - just close as feasible for manufacture. The closest match for IRF/P 9140 is IRF/P 240 and not what might look more obvious.

Also note these are not the N-suffix parts. The 'improved' N suffix parts are improved for switching use, charge storage and transconductance is improved but the die is smaller, resulting in much lower peak power dissipation handling ability, exactly what you do not want in a linear application.
 
Also note these are not the N-suffix parts. The 'improved' N suffix parts are improved for switching use, charge storage and transconductance is improved but the die is smaller, resulting in much lower peak power dissipation handling ability, exactly what you do not want in a linear application.

Your mean is none N-suffix parts better than N-suffix parts in audio?
 
completley wrong - the big gate, parasitic C of power MOSFET make them worse performing at "subthreshold" currents than properly sized MOSFETs running at a good fraction of their Idmax, properly chosen BJT Q are often better

this is textbook stuff - please don't "help" others with incorrect info

for better info try http://www.diyaudio.com/forums/solid-state/101745-bob-cordell-interview-bjt-vs-mosfet.html

Jcx, you gotta stay realistic. With devices that have an Id leakage current of say 50uA, you're not going to run these at 100uA quiescent. The maximum rated continuous current as specified in the datasheet for these devices is practically meaningless if you want to exploit lineair operation as that is the current they can handle when fully saturated. For example, that IRF620 is still fine dealing with several 10s of mA which is still above the sub treshold for them. But, for lineaur small (current) signal operation the Vgs/Ids curve at that small scale is still linear. FYI I've experimented a lot with HEXfets in control loops and have quite a good understanding of how they behave for real.
 
Member
Joined 2010
Paid Member
Quasi Mos?

I'm not clear on what your comment about tubes is there, LV, but some members build quasi-complementary Mosfet amplifiers for this very reason. It's not as common as approximate complementary pairs but there are some advantages to either topology.

You may need specific advice for the complements you wish to use as the degree of mismatch in the pair will affect the choice of best topology, as will the circuit you plan to use/adapt. Even if they are not best suited to Audio, Power Mosfets have some big advantages: they are cheap, tough and plentiful.

Practically, the choice is yours but you must read the theory behind these choices and learn how to make the best compromise yourself. Try to use the most common parts so that others will be familiar with the issues if you need assistance when it comes to designing or building things on your own. You can't learn much by starting a lot of Q & A threads other than a few random facts, recollections and references to documents, IMHO. The better choice for any newb, young or old, is to follow a construction thread (where the design is decided, posted and the guy(s) are buying parts and doing rather than just talking about more possibilities). Then research the topics as you follow the thread from the beginning with the search button, Wikipaedia or Google etc, if you have no access to a library or useful books on audio electronics.

For me, the best money I ever spent on electronics was on good books - not parts nor PCBs nor cameras to show things off even, though I've used a lot of stuff. :D
 
I'm not clear on what your comment about tubes is there, LV, but some members build quasi-complementary Mosfet amplifiers for this very reason. It's not as common as approximate complementary pairs but there are some advantages to either topology.

You may need specific advice for the complements you wish to use as the degree of mismatch in the pair will affect the choice of best topology, as will the circuit you plan to use/adapt. Even if they are not best suited to Audio, Power Mosfets have some big advantages: they are cheap, tough and plentiful.

Practically, the choice is yours but you must read the theory behind these choices and learn how to make the best compromise yourself. Try to use the most common parts so that others will be familiar with the issues if you need assistance when it comes to designing or building things on your own. You can't learn much by starting a lot of Q & A threads other than a few random facts, recollections and references to documents, IMHO. The better choice for any newb, young or old, is to follow a construction thread (where the design is decided, posted and the guy(s) are buying parts and doing rather than just talking about more possibilities). Then research the topics as you follow the thread from the beginning with the search button, Wikipaedia or Google etc, if you have no access to a library or useful books on audio electronics.

For me, the best money I ever spent on electronics was on good books - not parts nor PCBs nor cameras to show things off even, though I've used a lot of stuff. :D

Thank you for your suggestion, I think I should buy an oscilloscope, because often only our subjective hearing is not perfect.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.