I’ve been wrapping my novice audio brain around speaker loads and sensitivity recently. It seems like sensitivity gets talked about a lot, but (as in the above ML speakers) the load itself seems a more important consideration when thinking about amp/speaker pairing. Is that a fair assumption?
Thank you Nelson Pass and everyone for your comments. Confirmed my suspicions that the F6 should be happy as the impedance / spectral energy demands shouldn't be too overwhelming! 
Just wanted to make sure there wasn't any concern as it's my first FirstWatt amp (coming from an LC Audio no feedback Millennium XP amp which drives the MLs with ease) and therefore a bit of an investment with new case, PSU, transformer etc
Will order up the bits on the DIYAudio store and enjoy a summer project here
Just wanted to make sure there wasn't any concern as it's my first FirstWatt amp (coming from an LC Audio no feedback Millennium XP amp which drives the MLs with ease) and therefore a bit of an investment with new case, PSU, transformer etc
Will order up the bits on the DIYAudio store and enjoy a summer project here
I’m using FQH44N10 instead of IRFP150. Finally fired up my F6 today, and am impressed with how great it sounds. I’ve pushed the power rails to 26.4V, with 1.42 Amps of quiescent current for now. The power supplies and heat sinks are designed to support 1.8 Amps of current if that proves beneficial.
Hi TungstenAudio,
Is that a direct swap?
Anything needs to be done on the front end?
I am on +/- 25.4Vdc, recently bias up to 1.35A, sound pretty good than before. Please share with us your mods.
Thanks
Is that a direct swap?
Anything needs to be done on the front end?
I am on +/- 25.4Vdc, recently bias up to 1.35A, sound pretty good than before. Please share with us your mods.
Thanks
Yeah these are great devices in this application, the 150V FQA28N15 parts are also great.
Last edited:
Nothing will need to be done. at most maybe swapping out biasing Zener references but unlikely.
Nothing special needed to support the FQH parts. I am using LM329 voltage references in place of the zener diodes, but I would have used those anyway.
The higher voltage called for a different input buffer, so I implemented a Diamond buffer very similar to the Austin input stage of the M2x.
The higher voltage called for a different input buffer, so I implemented a Diamond buffer very similar to the Austin input stage of the M2x.
TA,
Congratulations on your successful tweaks and thank you for sharing.
However, these are a bunch of mods you've implemented at the same time:
1. Brand new input buffer
2. Higher voltage power rails
3. Higher bias
4. Different MOSFETs (to handle higher bias)
Questions:
1. Would ~26V be too high for the default 2SK170/SJ74? I recall reading somewhere that 26-27V would be ok (I might be wrong)?
2. Even if we keep ~24-25V voltage rails, the sound is expected to "improve" at higher bias for MOSFETs (IRF or others)?
3. Any measurements to show what changed in the THD level or harmonic profile?
My build of the F6 does have a number of modifications from the standard circuit and PSU which are documented in the F6 Illustrated Build Guide. Partly because of that, this build has taken me longer to complete as I worked through a couple issues related to initial prototypes.
This amp has been inspired by several fairly diverse sources, including the M2x, a couple Bablefish designs and some of the original F6 discussions in this thread. To a lesser extent I also tried to use parts that I have acquired over the last year or so for other ideas. In large part, the main goal of this amp was to use a set of FQH44N10 parts that I discovered while looking for parts similar to the IRFP140 and IRFP150. The FQH parts looked very good on paper, and I'm happy to say I'm very pleased with how they sound as well. In order to best support the parametric performance of the FQH Mosfets, I made several other changes to the main board design of the F6:
1. Power supply voltages were increased to a target range of 26 to 27 Volts.
2. Quiescent currents were increased to a target range of 1.4 to 1.8 Amps.
3. Source resistances at the output stage were lowered to 0.50 Ohms for R1 and 0.40 Ohms for R2.
4. The low side of C1 was connected directly to the ouput node, rather than the Source of Q1. C2 is connected as specified by the F6 Mosfet DIY schematic.
5. Z1 and Z2 were replaced by LM329, 6.9V voltage sources. R7 and R8 were lowered to 4.3 kΩ.
6. The simple input buffer was replaced by a Diamond buffer inspired by the Austin input stage of the M2x.
7. The SLB power supplies available from one of the group buys are used with Antek AS-3222 power transformers.
So, I really have no idea what a stock F6 sounds like, in either power JFet or Mosfet configuration. I do really like what I'm hearing out of my build so far. Following an overnight thermal soak, another short listening session this morning showed that the amp has changed very little from the couple hours after initial power up in my system. I still have a lot more listening to do and some measurements to make. As it stands, this F6 is giving my Aleph J a run for its money, which says a lot. The two amps have different characters, and it would be difficult to choose a favorite.
I can say that I intend to follow up on some ideas I have regarding the power supply. It has been my experience that Cap Mx types of supplies often benefit from large bulk capacitors in the 10 mF to 20 mF range (10,000 uF to 20,000 uF) to better support low frequency performance in the bottom octave or two. I will try this over the next day or two after I've hooked the amp up to my HP 339A to take some baseline measurements. I will also be increasing the quiescent current, probably before adding the big power supply caps.
This amp has been inspired by several fairly diverse sources, including the M2x, a couple Bablefish designs and some of the original F6 discussions in this thread. To a lesser extent I also tried to use parts that I have acquired over the last year or so for other ideas. In large part, the main goal of this amp was to use a set of FQH44N10 parts that I discovered while looking for parts similar to the IRFP140 and IRFP150. The FQH parts looked very good on paper, and I'm happy to say I'm very pleased with how they sound as well. In order to best support the parametric performance of the FQH Mosfets, I made several other changes to the main board design of the F6:
1. Power supply voltages were increased to a target range of 26 to 27 Volts.
2. Quiescent currents were increased to a target range of 1.4 to 1.8 Amps.
3. Source resistances at the output stage were lowered to 0.50 Ohms for R1 and 0.40 Ohms for R2.
4. The low side of C1 was connected directly to the ouput node, rather than the Source of Q1. C2 is connected as specified by the F6 Mosfet DIY schematic.
5. Z1 and Z2 were replaced by LM329, 6.9V voltage sources. R7 and R8 were lowered to 4.3 kΩ.
6. The simple input buffer was replaced by a Diamond buffer inspired by the Austin input stage of the M2x.
7. The SLB power supplies available from one of the group buys are used with Antek AS-3222 power transformers.
So, I really have no idea what a stock F6 sounds like, in either power JFet or Mosfet configuration. I do really like what I'm hearing out of my build so far. Following an overnight thermal soak, another short listening session this morning showed that the amp has changed very little from the couple hours after initial power up in my system. I still have a lot more listening to do and some measurements to make. As it stands, this F6 is giving my Aleph J a run for its money, which says a lot. The two amps have different characters, and it would be difficult to choose a favorite.
I can say that I intend to follow up on some ideas I have regarding the power supply. It has been my experience that Cap Mx types of supplies often benefit from large bulk capacitors in the 10 mF to 20 mF range (10,000 uF to 20,000 uF) to better support low frequency performance in the bottom octave or two. I will try this over the next day or two after I've hooked the amp up to my HP 339A to take some baseline measurements. I will also be increasing the quiescent current, probably before adding the big power supply caps.
Ordering a transformer for F6 build. Which would be the best choice?
VA CORE CODE CORE SIZE OD mm HT mm WT kg REG %
300 H 60/100/40 110 58 2.6 6
300 X7 60/110/40 120 58 3.0 5
400 TLP 80/180/20 190 37 5.0 6.5
400 T 60/110/40 120 57 3.0 6
500 I 60/125/45 135 63 4.5 5
Sorry about the formatting - looks great in preview....
VA CORE CODE CORE SIZE OD mm HT mm WT kg REG %
300 H 60/100/40 110 58 2.6 6
300 X7 60/110/40 120 58 3.0 5
400 TLP 80/180/20 190 37 5.0 6.5
400 T 60/110/40 120 57 3.0 6
500 I 60/125/45 135 63 4.5 5
Sorry about the formatting - looks great in preview....