Build update
Here's a bit of the background into how I got started on the ALPHA build. In my search for information on building a capacitance multiplier to construct a new power supply for my Pass F4 amp I discovered the SLB. On the thread I learned about the ALPHA amp, which I found to be very intriguing. I ended up purchasing a PCB along the components for a 4R 2.4 amp bias build.
Ultimately I plan to use the ALPHA in a bi-amp configuration where they will be powering the upper and mid drivers of a three-way speaker build. Albeit the ribbon tweeter in the proposed 3 way system is a nominal 6 ohm I don't believe there is a need for a 2.4 amp bias. However, because I had the parts for the build I constructed one card with the 2.4 amp bias. I just finished up the building the second card with a 1.9 amp bias. The IXTK90P20P FET for V132 broke but I was able to get the channel working with a IRFP9240 I had in my supplies.
Will the IRFP P-Channel continue to work with 6-8 ohm speakers or should I change it for a higher powered device?
I did a quick comparison listen between the two cards and the biggest difference I could hear was caused by the speakers. I will have to re-compare with some factory built speakers.
Here's a bit of the background into how I got started on the ALPHA build. In my search for information on building a capacitance multiplier to construct a new power supply for my Pass F4 amp I discovered the SLB. On the thread I learned about the ALPHA amp, which I found to be very intriguing. I ended up purchasing a PCB along the components for a 4R 2.4 amp bias build.
Ultimately I plan to use the ALPHA in a bi-amp configuration where they will be powering the upper and mid drivers of a three-way speaker build. Albeit the ribbon tweeter in the proposed 3 way system is a nominal 6 ohm I don't believe there is a need for a 2.4 amp bias. However, because I had the parts for the build I constructed one card with the 2.4 amp bias. I just finished up the building the second card with a 1.9 amp bias. The IXTK90P20P FET for V132 broke but I was able to get the channel working with a IRFP9240 I had in my supplies.
Will the IRFP P-Channel continue to work with 6-8 ohm speakers or should I change it for a higher powered device?
I did a quick comparison listen between the two cards and the biggest difference I could hear was caused by the speakers. I will have to re-compare with some factory built speakers.
Attachments
Nice work, Shawnstium. The IRFP9240 is a nominally lower power dissipation device than the IXYS. It should work and sound fine. Long lifetime might be less robust vs the IXYS. However, if you plan on using a heat pipe CPU coolers and fan cooling and it doesn’t get too hot, then it doesn’t matter.
How did the IXYS get zapped? Usually they are very robust. Unless it was static discharge then all MOSFETs are susceptible.
Or was it physically broken as in leg snapped off as you have shown? One solution would be to repair that by using a Dremel to excavate some of the epoxy casing to expose more leg to solder on a wire or large lead to salvage it.
One other suggestion is to use these helper boards to mount your MOSFETs on and use flying leads to the main PCB. Let me know if this part can be useful for you. It’s from the ABBB amp build.
Good luck!
How did the IXYS get zapped? Usually they are very robust. Unless it was static discharge then all MOSFETs are susceptible.
Or was it physically broken as in leg snapped off as you have shown? One solution would be to repair that by using a Dremel to excavate some of the epoxy casing to expose more leg to solder on a wire or large lead to salvage it.
One other suggestion is to use these helper boards to mount your MOSFETs on and use flying leads to the main PCB. Let me know if this part can be useful for you. It’s from the ABBB amp build.
Good luck!
Last edited:
The IXYS was only physically damaged and was my fault. I thought about using some solder paste to reattach the broken leg but I felt the success rate would be small. Using a Dremel to enlarge the contact area is great idea; I will have to give that a try.
The CPU heat pipes do a great job at keeping things cool. When the fan is running there is no build up of heat. Remote mounting the active devices also seems like a very practical idea.
I am excited to do some more listening comparisons this weekend to see if I can hear the difference between the different current biases and the two power supplies (XFMR with SLB vs SMPS with MOSFET Cap Mx). The biggest difference I can point out now, is that the card with the 2.4 amp has a louder turn on/off sound than the lower biased one. However this could be due to the different power supply architectures. C103 is a 1000uF on each board.
Either way, the sound is amazing and can't wait to show it off! Thank you and everyone else for their support.
The CPU heat pipes do a great job at keeping things cool. When the fan is running there is no build up of heat. Remote mounting the active devices also seems like a very practical idea.
I am excited to do some more listening comparisons this weekend to see if I can hear the difference between the different current biases and the two power supplies (XFMR with SLB vs SMPS with MOSFET Cap Mx). The biggest difference I can point out now, is that the card with the 2.4 amp has a louder turn on/off sound than the lower biased one. However this could be due to the different power supply architectures. C103 is a 1000uF on each board.
Either way, the sound is amazing and can't wait to show it off! Thank you and everyone else for their support.
Hello,
I traced some Ixys mosfets.
IXFH110N10P: blue
IXTH64N10L2: magenta
IXFH74N20P Mouser: dark green
IXFH74N20P TME: light green
IXTH52P10P: purple
IXTH90P10P Mouser: orange
IXTH90P10P TME: red
Some of the same mosfets were sourced from TME and Mouser,
strangely the trace curves of those same mosfets look a bit different.
Probably they were made during anther production run or revision.
A lot of colors to choose from for the Alpha
I traced some Ixys mosfets.
IXFH110N10P: blue
IXTH64N10L2: magenta
IXFH74N20P Mouser: dark green
IXFH74N20P TME: light green
IXTH52P10P: purple
IXTH90P10P Mouser: orange
IXTH90P10P TME: red
Some of the same mosfets were sourced from TME and Mouser,
strangely the trace curves of those same mosfets look a bit different.
Probably they were made during anther production run or revision.
A lot of colors to choose from for the Alpha
Attachments
Nice curves.
Don't forget that those IXFH74 devices tend to blow up instantly - not recommended!
Don't forget that those IXFH74 devices tend to blow up instantly - not recommended!
Hello,
I traced some Ixys mosfets.
IXFH110N10P: blue
IXTH64N10L2: magenta
IXFH74N20P Mouser: dark green
IXFH74N20P TME: light green
IXTH52P10P: purple
IXTH90P10P Mouser: orange
IXTH90P10P TME: red
Some of the same mosfets were sourced from TME and Mouser,
strangely the trace curves of those same mosfets look a bit different.
Probably they were made during anther production run or revision.
A lot of colors to choose from for the Alpha
It's the IXTH76N25T that is a Trench mosfet that Raptor blow up and should not be used.
I avoided the Trench fets, the curve traces are Trench-free
The IXFH74N20P is a PolarHTTM HiPerFETT.
Can those traces help to select candidates for the Alpha ?
Since the Alpha is SE and not traditional PP, do they need to be complementary ?
Regards,
Danny
I avoided the Trench fets, the curve traces are Trench-free
The IXFH74N20P is a PolarHTTM HiPerFETT.
Can those traces help to select candidates for the Alpha ?
Since the Alpha is SE and not traditional PP, do they need to be complementary ?
Regards,
Danny
Hi Danny,
My thoughts are that all of your selections are OK. Since this amp runs at high quiescent, much of the transconductance is linear, or at least more linear than operating in Class AB at <200mA, and because both devices are always on right through the signal cycle the gate requirement is not varying like a Class AB.
This means that even with high gate capacitances, since bootstraps by source follower operation, any of them are suitable. Even variations from gm from 6 to 50 are not that important; this is a very tolerant circuit and with around 20dB of global feedback at midrange and beyond any output signal compression is accommodated.
On this circuit both outputs are used for totally different jobs; the upper nmos is used to set the output stage current, right through the entire cycle. The lower pmos however is used solely as a source follower; it sets the voltage on the load only. These are such different jobs that the concept of complementary as in a push pull circuit is not relevant. Hard to believe, but essentially this is a single ended push pull circuit, a particularly version of a SE amplifier. Broskie has a lot to say about this type of amplifier, it's a very good topology.
Again, thanks Danny, and congratulations to Shawnstium and Raptor!
HD
My thoughts are that all of your selections are OK. Since this amp runs at high quiescent, much of the transconductance is linear, or at least more linear than operating in Class AB at <200mA, and because both devices are always on right through the signal cycle the gate requirement is not varying like a Class AB.
This means that even with high gate capacitances, since bootstraps by source follower operation, any of them are suitable. Even variations from gm from 6 to 50 are not that important; this is a very tolerant circuit and with around 20dB of global feedback at midrange and beyond any output signal compression is accommodated.
On this circuit both outputs are used for totally different jobs; the upper nmos is used to set the output stage current, right through the entire cycle. The lower pmos however is used solely as a source follower; it sets the voltage on the load only. These are such different jobs that the concept of complementary as in a push pull circuit is not relevant. Hard to believe, but essentially this is a single ended push pull circuit, a particularly version of a SE amplifier. Broskie has a lot to say about this type of amplifier, it's a very good topology.
Again, thanks Danny, and congratulations to Shawnstium and Raptor!
HD
Last edited:
Hi Hugh,
Thanks for the explanation.
It’s very nice that the Alpha can handle all those mosfets and don’t need to be matched !
I traced the mosfets at 2A and +-20v, those Ixys are definitely linear in that region.
For the P-channel I’ll take the IXTH90P10P,
for the N-channel maybe the IXFH110N10P or the IXTH64N10L2, that one is designed for linear operation as specified in the datasheet.
Regards,
Danny
Thanks for the explanation.
It’s very nice that the Alpha can handle all those mosfets and don’t need to be matched !
I traced the mosfets at 2A and +-20v, those Ixys are definitely linear in that region.
For the P-channel I’ll take the IXTH90P10P,
for the N-channel maybe the IXFH110N10P or the IXTH64N10L2, that one is designed for linear operation as specified in the datasheet.
Regards,
Danny
Good looking out and sorry for misremembering the number. There's so many of these things with similar numbers it's hard to keep track!
I'm happy to report that with the standard IRFP devices the amp is still soldiering on just fine!
I'm happy to report that with the standard IRFP devices the amp is still soldiering on just fine!
It's the IXTH76N25T that is a Trench mosfet that Raptor blow up and should not be used.
I avoided the Trench fets, the curve traces are Trench-free
The IXFH74N20P is a PolarHTTM HiPerFETT.
Can those traces help to select candidates for the Alpha ?
Since the Alpha is SE and not traditional PP, do they need to be complementary ?
Regards,
Danny
Thanks!
Yes, it's a dual mono.
The right channel has:
- 500VA 2x18v transformer
- two LT4320 ideal bridges, so it's a full bridge
- two F&T 47000uF, for pos and neg rails
- two 10mH chokes, one for each rail
- four Sikorels 22000uF, that's 44000 for each rail, pos and neg
The same for the left channel
Yes, it's a dual mono.
The right channel has:
- 500VA 2x18v transformer
- two LT4320 ideal bridges, so it's a full bridge
- two F&T 47000uF, for pos and neg rails
- two 10mH chokes, one for each rail
- four Sikorels 22000uF, that's 44000 for each rail, pos and neg
The same for the left channel
That looks great Danny - very impressive. You may be able to save those massive caps and make 2-3 more amp projects with them if you use an SLB for the PSU. SLB has ideal LT4320 bridge built in too. You will get 1mV rms ripple with only 4x 15,000uF bulk caps and no CLC needed. But would require 22v trafo and you have 18v already. Just think about it for future builds. It will make for a quieter and lighter weight amp.