A good friend designed new power supply and amplifier PCB's for the F-5 based on the large IXYS devices. A few made comments on the F-5 when using these devices, and I was not sure what to think. I would appreciate if you all could chime in on what was said, especially those that may have used the large IXYS mosfets. His amplifier runs fine and he says sounds excellent.
One guy said...
"Beautiful boards. But those MOSFETS have CIS values in the11.5 to 14.5 region, and driven by impedances in the 470 to 680 ohm region. You will have poles at 11 khz and 20 khz This means serious risk of self oscillation at 25 khz."
"And to remedy that, one will need to limit the bandwidth of the X10 gain amplifier into 2.5 khz or less.
He is sending me a set of boards so I can build one, bu5t it would be nice before hand if what this guy says in an issue.... I am really looking foreword to this project....
Thanks in advance for commenting on this... I decided to make a seperate post so it did not get burried in the main F5 post.
Mark
One guy said...
"Beautiful boards. But those MOSFETS have CIS values in the11.5 to 14.5 region, and driven by impedances in the 470 to 680 ohm region. You will have poles at 11 khz and 20 khz This means serious risk of self oscillation at 25 khz."
"And to remedy that, one will need to limit the bandwidth of the X10 gain amplifier into 2.5 khz or less.
He is sending me a set of boards so I can build one, bu5t it would be nice before hand if what this guy says in an issue.... I am really looking foreword to this project....
Thanks in advance for commenting on this... I decided to make a seperate post so it did not get burried in the main F5 post.
Mark
Attachments
Jason's sending me a set, I have been watching his build very closely too! I'll be visiting him in a few weeks when I go see my brother in Nanaimo. If anyone needs any additional info please let me know, we talk every few days.
You will pretty much still have a single pole, as the rolloffs are parallel, not in series,
and the loop stability will not be much of an issue, rather the higher capacitance
will give you a lower rolloff of open loop. If you calculate that at 11 Khz, that isn't
particularly worse than what else is out there.
Mostly I think many will find the big devices as having an interesting sound.
and the loop stability will not be much of an issue, rather the higher capacitance
will give you a lower rolloff of open loop. If you calculate that at 11 Khz, that isn't
particularly worse than what else is out there.
Mostly I think many will find the big devices as having an interesting sound.
You will pretty much still have a single pole, as the rolloffs are parallel, not in series,
and the loop stability will not be much of an issue, rather the higher capacitance
will give you a lower rolloff of open loop. If you calculate that at 11 Khz, that isn't
particularly worse than what else is out there.
Mostly I think many will find the big devices as having an interesting sound.
I've been talking to Jason and he's just loving his, pretty sure he like sit more than his Claimed aleph mini, I get to listen to his amp next week.
You don't think this will blow up Nelson ?
Is there another thread about using the IXYS MOSFETs in the F5 circuit? I found a YouTube video describing it, maybe this is one of you guys: Nelson Pass F5 with IXYS Fets !! Success ! NO SMOKE !! - YouTube
That's Jason, my friend in Nanaimo. I'll be down to his house to pick up my goodies as i'm visiting my brother there. He sent me a whole pile of screen shots and amplifier builds. Going to be building one myself.
It would be an interesting amplifier. I don’t have experience with the IXYS devices but their amp and dissipation ratings suggest you would be able to bias the amp far higher than a regular f5 with a single pair of mosfets. With enough heatsink you could probably do 3A of bias with a single pair which sounds interesting.
I know for a fact that Jason is sitting at 0.60v across the 0.47 ohm resistors and it pulls about 1.5A per side. After about 1 hour the amp sits at 49-51deg. He's got a video coming out in a week about his amp and what he has all done with it and the amount of measuring he has done.
Many people on facebook Audio Group have doubted his build, probably why Mark & myself asked Nelson about it.
Many people on facebook Audio Group have doubted his build, probably why Mark & myself asked Nelson about it.
If you are interested, I could mention to Jason you want a pair of his pcb's I know he has lots left over, let me know if interested he lives right across the water from you on the island.
Jim
James, I think you are trying to pm me. My inbox was full but it’s cleared out now.
I don’t need PCB…I have an f5 running a dual pair of toshiba mosfets at 2A bias now. I started with a single pair of mosfets…I felt the single pair had a more exciting quality but more rail voltage and bias suits my speakers better so I went with a dual pair. If the ixys mosfet allows some of that non paralleled sound to come through while still supporting increased bias levels it would be a winning combo. A lot of it is a question of how well the large die spreads heat…with multiple pairs of mosfets you can spread them out across the heatsink.
I don’t need PCB…I have an f5 running a dual pair of toshiba mosfets at 2A bias now. I started with a single pair of mosfets…I felt the single pair had a more exciting quality but more rail voltage and bias suits my speakers better so I went with a dual pair. If the ixys mosfet allows some of that non paralleled sound to come through while still supporting increased bias levels it would be a winning combo. A lot of it is a question of how well the large die spreads heat…with multiple pairs of mosfets you can spread them out across the heatsink.
Got any ideas for tests ? I believe Jason got a QA401 and has some nice test gear I gave all mine away as I changed jobs so, maybe ill have Jason build and test mine for me then ship mine up to me.
I like how the devices are really large, I saw in his video he used a little bit of thermal past and 2 m3 bolts to the heatsink, it sits pretty flat and works great.
DIY Audio - Home of HiFi Do-It-Yourself Enthusiasts This group I believe. 😉
I did think of that previously , but I confessed not once - I'm Chicken ........ to rely only on thermals to keep Gargantuas at bay
besides, solving it to use them Square Law - as Papa did in XA25, was much bigger puzzle, with greater outcome
edit: one of outcomes now is - making properly controlled F5-like , with Square-Law puck in OS, becomes trivial
besides, solving it to use them Square Law - as Papa did in XA25, was much bigger puzzle, with greater outcome
edit: one of outcomes now is - making properly controlled F5-like , with Square-Law puck in OS, becomes trivial
Last edited:
I did think of that previously , but I confessed not once - I'm Chicken ........ to rely only on thermals to keep Gargantuas at bay
besides, solving it to use them Square Law - as Papa did in XA25, was much bigger puzzle, with greater outcome
edit: one of outcomes now is - making properly controlled F5-like , with Square-Law puck in OS, becomes trivial
I agree 🙂 I don't know what we could do with out stepping on Pappa's toes 🙂 but Jason's F5 with the ixys's seem to be running fine.
Anyone know what would cause the amp to blow up ? ( not that I want to blow up 200$ in fet's but would be nice to know what to not do.😀
I think Nelson also said it should be fine with this setup, in the end its 25watts and its not being pushed to hard "BUT" there is that chance right ? of the smoke being released ?
It’s a DC coupled amplifier. We don’t worry about the mosfets, we worry about the speakers. If it was mine I’d be looking at what it look like on the scope with high capacitance input/output cables, how it behaves at unusual heatsink temps (say 70c).
Interesting that the builder got a 2nd harmonic dominant sound…possible they adjusted p3 to achieve that. Also would like to see bandwidth. The original f5 is really high bandwidth, this one might not be.
Interesting that the builder got a 2nd harmonic dominant sound…possible they adjusted p3 to achieve that. Also would like to see bandwidth. The original f5 is really high bandwidth, this one might not be.
as I said - I'm Chicken, even feeling wuss just thinking of thermals and some source resistance keeping them in little narrow safe area
they're having , even practically starved with Iq, enormous xconductane comparing to regular IRFP thingies ....... all you need is that Sparrow flying near your window farts little louder, and they gonna start welding PSU and speaker wires ....... that's at least scenario automatically unwinding in my scull
besides , I don't really like idea of their capacitances driven with eeny JFets of F5 front end, even if doubled or tripled
it works, no doubt, but that's somehow construction made purely on muscle, without elegance
not belittling anyone's efforts, just saying that I'm Chicken and that my personal taste is dictating different ways of what's pretty 🙂
they're having , even practically starved with Iq, enormous xconductane comparing to regular IRFP thingies ....... all you need is that Sparrow flying near your window farts little louder, and they gonna start welding PSU and speaker wires ....... that's at least scenario automatically unwinding in my scull
besides , I don't really like idea of their capacitances driven with eeny JFets of F5 front end, even if doubled or tripled
it works, no doubt, but that's somehow construction made purely on muscle, without elegance
not belittling anyone's efforts, just saying that I'm Chicken and that my personal taste is dictating different ways of what's pretty 🙂