Hello, everyone.
I'm wondering if this topic has been discussed before. A search for blowtorch power amp only gave me a zillion results about the line stage. Every once in a while, I revisit the idea about a F5X that utilizes super symmetry. That topic was discussed a little, but I don't think a true solution was ever found. I finally decided to put an idea down on paper to see what you all think.
Basically, a blowtorch line stage with complimentary voltage follower power stages on each end combined with the low impedance global feedback implemented in the F5.
While I wouldn't consider this truly an F5 iteration, I do believe it is in the same spirit. I also believe this satisfies the requirements for super symmetry.
Thoughts?
I'm wondering if this topic has been discussed before. A search for blowtorch power amp only gave me a zillion results about the line stage. Every once in a while, I revisit the idea about a F5X that utilizes super symmetry. That topic was discussed a little, but I don't think a true solution was ever found. I finally decided to put an idea down on paper to see what you all think.
Basically, a blowtorch line stage with complimentary voltage follower power stages on each end combined with the low impedance global feedback implemented in the F5.
While I wouldn't consider this truly an F5 iteration, I do believe it is in the same spirit. I also believe this satisfies the requirements for super symmetry.
Thoughts?
Nice circuit which probably sounds great (you be the judge) but probably doesn't quite qualify as SS as defined by the patent.
No matter, I say go for it.
😎
No matter, I say go for it.
😎
I have this on my drawing board since late last year.
All the parts have been collected and measured, including custom heatsink case.
Plan is to start building next year, after I have finished other things.
It will be called F5XL. And it will be all FET. No BJTs.
Output devices are Exicon double die laterals, matched to 3 coefficients. 😊
Apologies for hi-jacking.
Cheers,
Patrick
All the parts have been collected and measured, including custom heatsink case.
Plan is to start building next year, after I have finished other things.
It will be called F5XL. And it will be all FET. No BJTs.
Output devices are Exicon double die laterals, matched to 3 coefficients. 😊
Apologies for hi-jacking.
Cheers,
Patrick
Last edited:
Thanks for the replies, everyone. Especially the tip about the GR-25. I think that's the direction I'm headed for.
Patrick, hope you get around to the F5XL soon. I also have an F5X project that is gathering dust. It's a "normal" F5X, but I the feedback is hot switchable between the F5 style feedback and "traditional" feedback to the inputs. I wanted to be able to hear the difference on the fly. I had boards made during covid, but never got around to doing it. I'll probably add that functionality to this circuit, too.
Also, is there a particular reason you went with all FET? In the GR-25 circuit Ejam posted, my first thought was to make Q2,3,13, and 14 BJT because then you wouldn't have to overcome the Vgs and could get more gain into the folded cascode (i.e. make r13 bigger and r14 smaller). I seem to also remember reading (I think in the F5 documention) that the difference between using bjt vs fet as a cascode was negligible. But I've been away from the game for a while, so I might need a refresher.
And Mr. Pass, thank you for the endorsement. I think I'll give it a go!
Patrick, hope you get around to the F5XL soon. I also have an F5X project that is gathering dust. It's a "normal" F5X, but I the feedback is hot switchable between the F5 style feedback and "traditional" feedback to the inputs. I wanted to be able to hear the difference on the fly. I had boards made during covid, but never got around to doing it. I'll probably add that functionality to this circuit, too.
Also, is there a particular reason you went with all FET? In the GR-25 circuit Ejam posted, my first thought was to make Q2,3,13, and 14 BJT because then you wouldn't have to overcome the Vgs and could get more gain into the folded cascode (i.e. make r13 bigger and r14 smaller). I seem to also remember reading (I think in the F5 documention) that the difference between using bjt vs fet as a cascode was negligible. But I've been away from the game for a while, so I might need a refresher.
And Mr. Pass, thank you for the endorsement. I think I'll give it a go!
Also, can somebody point in the right direction for more info about Vbe multiplier bias circuits. While I was skimming through the link about the GR-25, the talked about double vs single Vbe multiplier circuits, but the bias circuit isn't in the schematic.
Thanks
Thanks
Wait. nevermind. The bias circuit is on the other page. At a glance I thought it was a voltage regulator for the front end rails. Sorry!
G25 and Blow torch are both ZGF (Zero Global Feedback).
F5XL uses negative feedback to lower Zout.
But I suggest you finish your F5X first, which is fully engineered.
It will take a year at least before F5XL will reach the same state.
Patrick
F5XL uses negative feedback to lower Zout.
But I suggest you finish your F5X first, which is fully engineered.
It will take a year at least before F5XL will reach the same state.
Patrick
The VBE multiplier circuit in post #4 of this thread, also appears in US patent number 3,995,228 , issued to Nelson Pass in November 1976. "Figure 3 is a schematic diagram of a bias circuit of the present invention" (column 3, lines 54-55)
See Fig, 1 & 2 in the same patent, which shows the "normal" Vbe multiplier, otherwise known as "amplified diode", as understood by most.
The patent describes a variant of the former.
See also :
https://en.wikipedia.org/wiki/Rubber_diode
https://sound-au.com/amp_design.htm#s31 section 3.2
www.diyaudio.com/community/threads/two-transistor-vbe-multiplier.388578/
And the attached article by Malcolm Hawksford, which is yet another variant.
Patrick
.
The patent describes a variant of the former.
See also :
https://en.wikipedia.org/wiki/Rubber_diode
https://sound-au.com/amp_design.htm#s31 section 3.2
www.diyaudio.com/community/threads/two-transistor-vbe-multiplier.388578/
And the attached article by Malcolm Hawksford, which is yet another variant.
Patrick
.
Last edited:
Kind Sirs, the CTC Blowtorch is based on a folded cascode topology, a single voltage gain stage; the F5 uses a 2 stages configuration.
Don't let the drafting of the schematics fool you, it's a BIG difference!
Don't let the drafting of the schematics fool you, it's a BIG difference!
With my poor understanding of circuitry.
The one in post #1 is still folded cascode, only that it has no load resistor to ground, but global negative feedback network.
So it is a mixture of both, if I understand correctly.
Patrick
The one in post #1 is still folded cascode, only that it has no load resistor to ground, but global negative feedback network.
So it is a mixture of both, if I understand correctly.
Patrick
Last edited:
Hi Patrick!
You're absolutely right, of course: the 1st schematic is a folded cascode (FC), single gain stage, everything else thereafter are F5's, 2 or more gain stages.
I was hoping you would use your magic to make a "real" CTC poweramp, i. e. FC front end, and a suitable output buffer.
Not easy because of the limited loop gain from a single gain stage, but doable; Charles Hansen (RIP) once shared a schematic of a similar all MOSfet prototype here.
You're absolutely right, of course: the 1st schematic is a folded cascode (FC), single gain stage, everything else thereafter are F5's, 2 or more gain stages.
I was hoping you would use your magic to make a "real" CTC poweramp, i. e. FC front end, and a suitable output buffer.
Not easy because of the limited loop gain from a single gain stage, but doable; Charles Hansen (RIP) once shared a schematic of a similar all MOSfet prototype here.
After poking around a little, I found a lot of pertinent info in the "John Curl Amp" thread. I just hadn't found it yet.
As suggested, I'm gonna throw together my shelved F5X first. But in the meantime, I'm going to finish a schematic and design some board for this project. The best part is I built five 40W UGS Power amps about 15 years ago because I wanted to have a home built surround sound system. The power supplies and chassis are already built, so I'll just have to swap out the boards. I also built an AX-100 as my first project about 20yr ago which I've dismantled but still have the heatsinks. So the plan is to build the F5 and this Curl/Pass hybrid then compare them both with the UGS Power amps I already have. The winner will get a high power version with my old monster heatsinks.
Patrick, I don't have any experience with the Exicon parts. I still have a bunch of the Fairchild FQA12P20 and FQA19N20C (like 50 of each), do you think the Exicon would be a significant upgrade?
As suggested, I'm gonna throw together my shelved F5X first. But in the meantime, I'm going to finish a schematic and design some board for this project. The best part is I built five 40W UGS Power amps about 15 years ago because I wanted to have a home built surround sound system. The power supplies and chassis are already built, so I'll just have to swap out the boards. I also built an AX-100 as my first project about 20yr ago which I've dismantled but still have the heatsinks. So the plan is to build the F5 and this Curl/Pass hybrid then compare them both with the UGS Power amps I already have. The winner will get a high power version with my old monster heatsinks.
Patrick, I don't have any experience with the Exicon parts. I still have a bunch of the Fairchild FQA12P20 and FQA19N20C (like 50 of each), do you think the Exicon would be a significant upgrade?