Aksa Lender P-MOS Hybrid Aleph (ALPHA) Amplifier

... Any takers? :)
Hi X,
My understanding is class A vertical mosfet design needs very careful bias setting procedure for thermal stability, far from plug and play, yet the 20W design offers a "new" approach. A test build with passive heatsink will confirm requirement for regular build.

However, I think design with parallel mosfets sans source resistor like the 55W should be reserved only for more advance builders. When Vgs matching of the parallel p channel is not good enough, current hogging could prove to be disastrous.
 
Good comments, I like them.......

OFFSET:
On this amp it is set by the trimmer on the LTP stage supply and then maintained by the LTP, which should locate the two LTP transistors in thermal contact. This means that during thermal cycling, the offset will, like any LTP input stage, very effectively control the offset within +/-15mV, more than enough.

BIAS:
ON this circuit the bias (quiescent) current is set by the value of the upper output device CCS sense resistor and the controlling transistor. The lower output device sets voltage at the output, but NOT the current, which is set solely by the Pass CCS at the upper device. Again, thermal cycling will affect this quiescent current; as the transistor heats, its Vbe will drop and with it the quiescent, but not more than a couple of mV/C. This control transistor should NOT be mounted on the heatsink or output device, of course.

OTHER ISSUES:
These are the big issues for any PP amp, but other issues are equalising the dynamic current swings on the two output devices which ensure neither drop below about 200mA. By never turning off, their gates are easier to control, because the charge transitions are far smaller than in AB. Another issue is how to design for low impedance. I have set it to cope with down to 6R speakers, but of course at some frequency most 8R speakers will drop to 6R - or even more. Therefore the speakers should be higher rather than lower impedance.

OUTPUT QUIESCENT AND MOSFETS
Lastly, I have specified three pairs of IRFs for the 55W monster ALPHA. In reality, I would lay out the design to use 480W mosfets, and a recent revision has indicated to me that the total dissipation will be around 188W per channel for 56W into 8R. This brings the total quiescent down to only 2.675A, so using a single BIG pair of mosfets, total quiescent on each would be 96W, doable and reliable with good heatsinking and large mosfets of 480W. Larger mosfets have very high transconductance too, so replacing 3 x 7S IRFs with 1 x 25S is quite practical. IXYS make very good device for this.

HD
 

Attachments

  • FDA69N25.pdf
    720.2 KB · Views: 336
  • IXTH48P20P.pdf
    133.1 KB · Views: 263
Last edited:
Yes, Tibou, there is a direct link from fb node base T2 through R9 to output so there is DC offset control

Kean, good question!

R16 is a nested fb resistor. C7 ensures it does not really influence proceedings until around 100Hz, ensuring that damping factor, and hence loop gain, is maximised below 100Hz. After beyond 100Hz, nested feedback ensures the first and second stages are linearised; we can do this with a very linear Class A output stage.

HD
 
Additional note about sound presentation and distortion.
...I would say it sounds a bit drier than MoFo which has maybe almost 8x the H2 distortion. Both have a similar profile though. ...
...I think this amp is more neutral in presentation and soundstage and imaging should be just as good as MoFo despite MoFo being zero global feedback.
On Stereophile Interview here, Nelson Pass has this to say about 2nd harmonic:
"There was a consistent subjective observation that there was a difference not only with the level of second harmonic, but phase also. Negative-phase second harmonic tends to expand the perception of front-to-back space in the soundstage, separating instruments a bit. Positive phase does the opposite, putting things subjectively closer and "in your face." I have heard this sort of comment from people who were not in a position to have expectation bias, so I treat it seriously."

The Mofo tend to produce the positive phase 2H. Polarity of speaker connection needs to be reversed to have negative 2H, or modify the Mofo using P channel. I have not simulated nor build the Alpha but the topology suggest negative phase 2H, similar to the Alephs, hence will tend to give better soundstage depth and separation with normal speaker polarity connection. By being single ended design with negative phase 2H, I would also suspect that the ALPHA will cancel some 2H driver distortion resulting in lower overall distortion like what Eduardo de Lima found with SE tube amps.

Hopefully you can do listening comparison between the Alpha and the Mofo to include switching speaker polarity and share your findings. Some of us in Pass forums have tried and confirm the effect.
 
Founder of XSA-Labs
Joined 2012
Paid Member
When I get a stereo version going I will be glad to compare them side by side. I use banana jacks on the amps and banana plugs on the speaker cables so the switchover can be fairly quick. I need to build a speaker level AB switch box.

So maybe on the 55W monster amp JPS64 can layout spots for two optional extra MOSFETs for those wanting to try the triple output method or use simple the single devices that Hugh has specified.
 

6L6

Moderator
Joined 2010
Paid Member
This is a fantastic take on the Aleph! I love it!!

The use of jellybean parts is also worth note, that's getting more important for DIY as through-hole parts seem to be slowly (but surely) disappearing into the ether...

I'd love to build this amp. :D :D :D
 

6L6

Moderator
Joined 2010
Paid Member
Lastly, the harmonic profile has been thrown into prominence in the last ten years, but in fact has been pivotal for good sound and Jean Hiraga demonstrated this in Le Revue du Son in Paris in the mid-sixties. He said that if you can arrange the first five or six harmonics off the fundamental along a straight, descending line, with at least 10dB reduction with successive higher harmonic, the sound will be perceived as 'natural'. H2 should start no higher than 68dB down from the fundamental. This was a controversial statement at the time, and even now, because it pays little regard for THD, which is a rms summation of the harmonic percentages of the fundamental, with no information about each actual harmonic. I have been designing amps using this philosophy for twenty years, and if you give regard to the monotonic decrease of the harmonics and a moderate THD (no more than 0.03%) you can create a good sounding amplifier. There are other issues, of course, but it certainly explains why many amps with very low THD still do not sound good.


This is the most concise description of this I've ever seen.

The effect was shown to me in clear effect when I build a Nelson Pass F5 Turbo, which has a pot (known as "P3") in parallel with the input Jfet source resistors. P3 allows you to adjust the source resistance and balance it in a way that will null a good quantity of the H2, which lowers THD, but will also make the H3 dominant. Losing the smooth curve as described above makes for a fast, clean, dynamic, energetic and precise sounding amplifier. It also, in my opinion, sucks the life and soul completely out of the music.

The beauty of "P3" is, of course, the ability to adjust the source resistance in can not only decrease H2, but can increase it as well... I did that to an amp to set H2 dominant, with a spectra described as above, and long story short, that amplifier sounds beautiful.
 
Founder of XSA-Labs
Joined 2012
Paid Member
This is a fantastic take on the Aleph! I love it!!

The use of jellybean parts is also worth note, that's getting more important for DIY as through-hole parts seem to be slowly (but surely) disappearing into the ether...

I'd love to build this amp. :D :D :D

Great to hear your support 6L6! JPS64 is working on the layout and in usual JPS64 style, I am sure it will be pro and over the top good. :D

Can't wait to see the layout, usually accompanied with full 3d renders of all parts fitted.... :) He spoils us, I know...

We are keeping the board all through-hole (with optional SMT NP0 caps as I have tons of them and they are cheap and sound great). But TH pads for those wanting to use film or silver mica there.

I am also asking JPS64 to add optional pots in parallel to R5 (for DC offset independent of LTP bias), R11 & R13 to adjust Aleph CCS action independent of main bias current. If you have any other suggestions for pots etc please say so as you have extensive experience on Aleph J build.
 
Last edited:

6L6

Moderator
Joined 2010
Paid Member
I am also asking JPS64 to add optional pots in parallel to R5 (for DC offset independent of LTP bias), R11 & R13 to adjust Aleph CCS action independent of main bias current. If you have any other suggestions for pots etc please say so as you have extensive experience on Aleph J build.

My experience on the pots in the Aleph J are to have as few as possible, and have pads where a resistor can be used in lieu of the pot.

I like the idea of having bias separate from Aleph CCS current gain.

The addition of the pot in the input stage CCS is the one that I will be sure to remove if a Rev. B is ever done of the AJ boards, it's useless. Just let the CCS do it's job. (Which is set by a zener in the AJ)
 

6L6

Moderator
Joined 2010
Paid Member
I was thinking about a PCB for the smaller version. (and please please please make it as through-hole as possible...)

The Big Boy is interesting and compelling, right up to the point you realize the simplest way to cool it is to submerge the whole thing in a mountain stream... :)
 
Founder of XSA-Labs
Joined 2012
Paid Member
Or submerge it in a fish tank of mineral oil with circulating pump and heat exchangers. :)

Ok thanks for your thoughts on the pots. So no need to be UMS compatible on Big Boy 55w version then?

Two pots to specify bias current and Aleph CCS action. Can you please confirm on 20W version which resistors need the parallel pots?

Thanks,
X
 
Last edited:

6L6

Moderator
Joined 2010
Paid Member
In the 20W, R11 would be made a bit smaller and have a pot placed in parallel to adjust the Aleph CCS.

If the LTP can be balanced (and with it DC offset) with transistor matching that can be done on a cheap DMM, then just tell everybody to buy 10 or 20 and get two nice matches. The KSA992 cost ¢15 in qty. (10), so it's more than cheap enough to not matter.

What did you use for the sense resistor R18? It's 4 parallel 3W resistors in the AJ...
 
Founder of XSA-Labs
Joined 2012
Paid Member
In the 20W, R11 would be made a bit smaller and have a pot placed in parallel to adjust the Aleph CCS.

If the LTP can be balanced (and with it DC offset) with transistor matching that can be done on a cheap DMM, then just tell everybody to buy 10 or 20 and get two nice matches. The KSA992 cost ¢15 in qty. (10), so it's more than cheap enough to not matter.

What did you use for the sense resistor R18? It's 4 parallel 3W resistors in the AJ...

Good point on matching - was able to easily match a pair with transistor Hfe tester and DC offset is under 5mV.

I am using a parallel set of 0.33R 2W resistors. I only have MOx resistors at present in 0.33R and see that they generate high H3 when I had them on R17. Will replace with Panasonic metal thin film 5% 3W as soon as I order them. I have lots of Panasonic 0.22R and 0.47R 3W but no 0.33R (always seem to be out of stock when I order from Mouser). I suppose it would be good to have 2 or 3 sets of pads here since these need to dissipate a lot of AC heat.

So what about a parallel pot at R13?

So when adjusting Aleph CCS, we give it say 1kHz 2.83vrms and adjust R11 to get 2x the current going through R18 vs R17, correct or is it the other way around?
 
Last edited:

6L6

Moderator
Joined 2010
Paid Member
The Aleph CCS is covered here - PassDiy


In the AJ, the pot across the controlling transistor in the CCS (what would be a pot across R11 in this schematic) is the most direct adjustment of DC bias in the circuit. I know it also effects the AC gain of the CCS, but I'm honestly a bit fuzzy on how all that interacts.

A pot across R13 might be the way to adjust the AC gain of the Aleph...? Seems logical but you'd need to ask somebody smarter than myself if it would work that simply...
 
Last edited: