I just got a stack of exicon lateral mosfets both TO3 and TO247 and the level of matching is excellent.
I have found with IRFP240/9240 there is quite a large spread of Vgs and you need to buy many more to get good matching. Not so with these laterals.
I will say though it is hard to get matching between N and P for the laterals.
The N's tend to all turn on with Vgs of around 0.18V while the P's turn on around Vgs of about 0.60V. This is just using my memory here. I have the results written down at work.
I will double check to tomorrow.
To give you indication out of 16 N channel mosfets tested, I got an average of 0.18V the worst deviation was 0.16V (-0.02V). This is threshold Vgs (Id=2.5mA)
I will do further matching at several Id values eg 50mA, 100mA, 200mA, and 500mA later
Thanks a lot, for info
Last edited:
Is there actually any artwork in this thread learking about? If so could someone give me a post number? Thanks
OT excuse
OK, app accepted. 😉
Yeah AndrewT, that would be debatable but I have to remind you that these are not usual mosfets (Id) but BIGBT compound components having NPN/PNP BJT pair on output (Icnpn + Icpnp). You can see comparison of emitter resistors voltage drop in the table and these (β) were actually paired/matched for the reason of 50:50 output current spread. 🙂
What most surprised me is that IRF mosfet part of BIGBT, were not even tested nor paired but just taken from the supply bag randomly. Imagine how perfect results would be achievable if IRF's would have been matched. 🙄
At last but not least the energy transfer function of actual BIGBT is a leap larger than at power mosfet of any kind at only a fraction of input capacitance. Yeah, I can see only advantages for audio amplifier applications ... 😎
Sorry my mistake.
Only 70mV spread at 2A 5.01Vgs to 5.08Vgs.
Now run these two in parallel with the same Vgs and see how far out Id will be.
Comparing different Vgs is not the same as comparing Id at identical Vgs.
OK, app accepted. 😉
Yeah AndrewT, that would be debatable but I have to remind you that these are not usual mosfets (Id) but BIGBT compound components having NPN/PNP BJT pair on output (Icnpn + Icpnp). You can see comparison of emitter resistors voltage drop in the table and these (β) were actually paired/matched for the reason of 50:50 output current spread. 🙂
What most surprised me is that IRF mosfet part of BIGBT, were not even tested nor paired but just taken from the supply bag randomly. Imagine how perfect results would be achievable if IRF's would have been matched. 🙄
At last but not least the energy transfer function of actual BIGBT is a leap larger than at power mosfet of any kind at only a fraction of input capacitance. Yeah, I can see only advantages for audio amplifier applications ... 😎
Last edited:
And one more?
If I build this beast with only one par of fet-s per channel (N,P), will that influence on anything else except power (bias per fet, etc..) Like original schematic without changing something else.
If I build this beast with only one par of fet-s per channel (N,P), will that influence on anything else except power (bias per fet, etc..) Like original schematic without changing something else.
Josip,
It is better not to build it at all if you change so much and you do not know how to calculate the gain.
Chose some proven and tested design as DX blame.
dado
It is better not to build it at all if you change so much and you do not know how to calculate the gain.
Chose some proven and tested design as DX blame.
dado
Does someone knows the gain of this amp?
The gain should be x 31.3 I think. That's just about 30 dB.
Ah! BIGBT is your version of Insulated Gate Bipolar Transistor (IGBT).......... I have to remind you that these are not usual mosfets (Id) but BIGBT compound components having NPN/PNP BJT pair on output (Icnpn + Icpnp).
The argument still stands. The high transconductance multiplies the apparent Id mismatch for identical Vgs.
If you want to show how well matched your devices are or compare anyone else's devices, then you must hold Vgs constant and Tj constant and then show the Id differences.
A few percent difference in Vgs is NOT the same as a few percent difference in Id.
Dadod:I build practically everything that has been worth of build from Pass to symasym AAK, ugs, but I am not an expert just say beginners luck
Thank ashok
Thank ashok
Josip,
If you change a lot then amp could become unstable. You could wait for practical result of this amp( some guys are bilding it).
If one pair of output MOSFET are used for sure you should not use +-80V of the power supply. Insted 2N5565, LSK389 could be used.
dado
If you change a lot then amp could become unstable. You could wait for practical result of this amp( some guys are bilding it).
If one pair of output MOSFET are used for sure you should not use +-80V of the power supply. Insted 2N5565, LSK389 could be used.
dado
Josip,
If you change a lot then amp could become unstable. You could wait for practical result of this amp( some guys are bilding it).
If one pair of output MOSFET are used for sure you should not use +-80V of the power supply. Insted 2N5565, LSK389 could be used.
dado
I am building a Mimesis 3. And if I understood Zen Mod correctly I can build it with one pair of fets because I do not need 200W
I am building a Mimesis 3. And if I understood Zen Mod correctly I can build it with one pair of fets because I do not need 200W
yes - but you still need transconductance in output stage , so - use those mosfets from Zhoufang
catch22 isn't in overall power of mosfets , but xconductance
I am building a Mimesis 3. And if I understood Zen Mod correctly I can build it with one pair of fets because I do not need 200W
You will have to also lower the rail voltages somewhat as not to exceed the power dissipation ratings of the 1 pair of output devices.
yes - but you still need transconductance in output stage , so - use those mosfets from Zhoufang
catch22 isn't in overall power of mosfets , but xconductance
I thought the most important thing is XConduntance/Ciss
Ok Zen Mod I took your advice and go for 2SK1530 and 2SJ201. What do I need pot? to set the quiescent current? What value pot and to what value should I put it on min or max resistance
R22 in that schmtc must be made of resistor and pot in series
Ju wrote to me that in earlier post,I do not understand know if I put it in series with fixed it can not be splited 50/50 percent (just parallel)
Ju wrote to me that in earlier post,I do not understand know if I put it in series with fixed it can not be splited 50/50 percent (just parallel)
instead of R22 of 330R , use fixed resistor of 180R and trimpot of 200R , both connected in series
nomenclature from http://i189.photobucket.com/albums/z105/invictus005/GoldmundMimesis3Schematic-2.jpg
post : http://www.diyaudio.com/forums/soli...-clone-worlds-best-amplifier.html#post2371725
nomenclature from http://i189.photobucket.com/albums/z105/invictus005/GoldmundMimesis3Schematic-2.jpg
post : http://www.diyaudio.com/forums/soli...-clone-worlds-best-amplifier.html#post2371725
- Home
- Amplifiers
- Solid State
- The Very Best Amplifier I Have Ever Heard!!!!