Hey, with tongue-in-cheek, Ostripper, I'm going to be contentious and suggest that with your rate of progress you'll soon be able to listen to an amp just by looking at the schematic like a number of inmates here.
Yes , I noted on the real amp (FA3) they didn't make much
difference as far as sound with the holman topology.
Also , using the triple , the vas sees just uA's of loading.
I wish.. 🙂
One can see the relationship betweendesign and sound , but you have to build it to really see.
Andy C.'s ".step param" and timestep - frequency directives
really helped to fine tune the minor errata.
OS
Hi OSTRIPPER,
For the F2T amp of yours what is the actual size of the board (pcb)
and can I use 2sc5200/2sa1943 for the output?
Regards,
Stcboy
For the F2T amp of yours what is the actual size of the board (pcb)
and can I use 2sc5200/2sa1943 for the output?
Regards,
Stcboy
homemodder said:
Something similar is used in an Onkyo amp IIRC.
Nice work OS, I simmed the Onkyo years ago but the Apt method is better.
Pete B.
Yes, 2sc5200/a1943 , mjl21193/4 (close fit on board with .2"
spacing between devices) FJL5200/1943. Any to-3p or
to-264 device will fit perfectly and perform the same.
Any device with Hfe 35 - 80/ Ft 4-30mhz/ 200v+ Vce/150+W
dissipation shound work fine.
Board is 6" X 6" single sided. diode board is 1 X1 , output
board is 1 X 5 and has inductor /zoble for both channels.
The 1200DPI artworks will show as 6 x6 image size.
You could reduce them to .JPG without loss of quality
since they are just 2bit images (black/white).
OS
PB2 said:
I beg to differ, I simulated a varient of that armstrong, it has lower THD, the bandbwith is an order of magnitude higher and has higher slewrate and this prickled my interest but simulator is one thing and what a amp sounds like is another.
And so the previous night and yesterday I quickly built the varient front end on veroboard to compare to a standard Apt frontend for a listen. The outputstage is a triple but employing toshiba vertical fets as final outputs. I was afraid the bass would deminish which is the fort of apt. Well I was wrong, it stumps the apt s sound by far, sorry. The bass is still there, ableit its faster cleaner but my, it takes the mids and treble into another performance bracket. Then comparison to a apt varient which I tried and found better than the standard, the vas cascode is referenced to the output, this alone is better for mids and treble that standard apt, but nope still not upto it. The standard apts sound is very little effected by the cascode. It benefits hugely from using a cascoded jfet input instead of the bjts and in this guise it comes closer to the armstrong topology or whatever topology you wish to call it with the extra ltp, but not quite.
One should first build and investigate, before making comments on a design. It was designed like that for a porpose, not because the guy decided heck lets add to the parts count for no reason and fool people for it does nothing for the sound but adds 4 or 5 dollars cost.
So if onkyo like the pioneer also mentioned, have also gone and added a extra ltp, they were quite clever and new how to make that topology sound darn good. Even Im surprised, I dont in particular like the standard apt, compaired to modern high end amps, it has no feet to stand on. All listening tests were made in comparison with electrocompaniet Nemo monoblocks using BW 803s speakers and three different cdplayers. With some fine tuning of compensation and it will most probably be the 4 th amp on my list of best sounding DIYs.
homemodder said:
Jump to conclusions often? Did I say the standard Apt? I also did not say that the Onkyo was exactly the same.
You don't have to lecture me about building and listening.
No point in continuing here ... LOL!
Pete B.
I don't think you can build the "standard APT" (they don't make the devices anymore) , also the original APT, with it's
low- tech current mirror ,was not very linear and barely
beat a bootstrap compensated "schoolbook" amp.
You add the wilson CM and all this changes , once you get
the "CM balance" right (100R trimmer), the ltp "locks" into
balance - even mismatching the pair has little effect. 😎
I did this mod on my FA3 , and the sound is very "symasym"
like, but with even smoother mids and highs.
I decided to add the triple OP stage later because I might
use a FA2T for HT sub duty.
On my peerless sub the APT varient has all the "slam" it
needs (scary with drums/movie soundtracks).
Slew wise, the only reason for the 68pF millers is that I already
have them, at only 25v/us , I admit they are less than ideal.
With 33/39 pF , this changes everything... 50v/us and a very sharp 20k SW.
OS
Its not a lecture, its my findings, I dont lecture or preach to anyone. I built it with modern devices, even higher specked ones than are shown in OS s schematics. I did use a wilson employing matched 2sc3067 and went even further to get currents as balanced as possible, they are minuscle and I only used 27pf which was enough to stabilize the circuit as I used mosfets in the triple which helped.
Os there nothing wrong with your circuit and it works, these other designs just used a different design approach and sometimes this comes off and sometimes not, in this case it does. There performance can be enhanced even more by also getting better current balance, something I didnt even attempt but can be seen on the pioneer. The bass performance is one area no one wil be able to say anything wrong about, this armstrong shares the same bass impact, techno and house music fans heaven.
PB2, I dont jump to conclusions often, when I think it has merit Ill try it, but I did make that mistake with the buzquito because it was not much talked around here, finally I got a reason to build one, not on my expense and got paid to do it, what a revelation it is, half the complexity of most amps here. If I remeber correctly you clearly stated that the apt method was better, jumping to a conclusion without any sort of technical investigation or listening test. 😀 Well your conclusion is wrong. My apt build is not standard in any way and I did mention what I changed, if it were to just be standard as designed I would not even have bothered with it. You did mention that something similar is used in the onkyo, similar topology or just similar current source or similar ouputstage, as we were talking of topology I took it as being similar topology.
Os there nothing wrong with your circuit and it works, these other designs just used a different design approach and sometimes this comes off and sometimes not, in this case it does. There performance can be enhanced even more by also getting better current balance, something I didnt even attempt but can be seen on the pioneer. The bass performance is one area no one wil be able to say anything wrong about, this armstrong shares the same bass impact, techno and house music fans heaven.
PB2, I dont jump to conclusions often, when I think it has merit Ill try it, but I did make that mistake with the buzquito because it was not much talked around here, finally I got a reason to build one, not on my expense and got paid to do it, what a revelation it is, half the complexity of most amps here. If I remeber correctly you clearly stated that the apt method was better, jumping to a conclusion without any sort of technical investigation or listening test. 😀 Well your conclusion is wrong. My apt build is not standard in any way and I did mention what I changed, if it were to just be standard as designed I would not even have bothered with it. You did mention that something similar is used in the onkyo, similar topology or just similar current source or similar ouputstage, as we were talking of topology I took it as being similar topology.
I am glad to know nothing is wrong 😉. I do know it works
(listening now). I did try the dual differential (pioneer)
and only accessed it's performance relying on simulations.
I concluded that it's higher device count did not justify
a couple 1000th's percent distortion improvement.
Both were "neck to neck" both in balance and THD20,
better devices gave me a few 1000th's more,but
the MPSA92/42 are cheap and spec out at 3pF-Cob/
50mhz Ft... not too bad.
Would not the OPS be a bigger factor in bass performance.?
Both VGS's (APT and the pioneer)
were quite flat down to 5 hz and not loaded at
all by the OPS (triple).
I noticed the difference between a stndard EF and a triple
immediately. You most likely have far better loudspeakers
than me (peerless 12" sub/sony 3-way's for full range) , so
shortcomings might be more apparent to you..
OS
Hi ,mr. Chua. What you are saying is that the "can" has a different sound ?? Why? , thermal coefficient(metal). die
size? The flatpacks and the cans have similar specs.
Modifying the 2T board will be simple , is that why you use
to-3's exclusively on your amps?
Really , I want my first 2 for driving my
satellites (60hz-20khz), but I will make a third for my
12" peerless.
I already have 2 pair 15024/5's on a modded C200,but want
more power
.
OS
size? The flatpacks and the cans have similar specs.
Modifying the 2T board will be simple , is that why you use
to-3's exclusively on your amps?
Really , I want my first 2 for driving my
satellites (60hz-20khz), but I will make a third for my
12" peerless.
I already have 2 pair 15024/5's on a modded C200,but want
more power
.OS
Hi OS
Mike will do.
I prefer metal cans for bass. Don't know why they sound different. May well be due to their construction. Basically, the MJ15024/5 sound more muscular with bass. This is one of the reasons for using them in my kits. The problem with these older TO3s are the mids and highs. It takes quite a bit of work to get them to sound smooth like the newer 30MHz power transistors.
For subwoofer use, I believe the TO3 outputs in your circuit will perform better. Your amp is quite "fast". I expect the bass to have a higher resolution - tighter and more dynamic with the MJ15024/5 or MJ21193/4.
Nothing scientific about this. A purely subjective opinion.
Mike
Mike will do.I prefer metal cans for bass. Don't know why they sound different. May well be due to their construction. Basically, the MJ15024/5 sound more muscular with bass. This is one of the reasons for using them in my kits. The problem with these older TO3s are the mids and highs. It takes quite a bit of work to get them to sound smooth like the newer 30MHz power transistors.
For subwoofer use, I believe the TO3 outputs in your circuit will perform better. Your amp is quite "fast". I expect the bass to have a higher resolution - tighter and more dynamic with the MJ15024/5 or MJ21193/4.
Nothing scientific about this. A purely subjective opinion.
Mike
I downloaded the models for the MJ's , and they do perform
(simulate) differently.
I stepped the bias and the THD bottoms out (.007%) at a much
lower bias (50ma), overall THD was higher at 20K but
the same as the flatties at 1k. Overall ,the design performed
flawlessly with the new devices (triple EF + "burly drivers").
So it seems... the higher Cob/lower Ft means more minority
carriers (larger junction) . As far as how well they will work,
open loop gain stayed exactly the same, but I took a "hit"
with my slew.
For a full range amp I figure I would just have to reduce Cdom,
recompensate, and maybe reduce or eliminate the basestoppers.
I will make a to-3 PCB tommorrow.
OS
For the difference between bass of triple and ef I cannot comment, havent used ef for more than ten years now. The cfp ef is very similar to triple, I cant hear any diffrence, maybe theres something there but at subsonic frequencies. I do believe the push pull has something to do with the bass, my self based circuit is dc coupled, I use jfets, no cap on ltp feedback node so it will probably amplify dc and comparing the self circuit with apt, the apt has more profound bass, with the same outputstage and power supply. One would have to examine the output characterestics of the push pull stage to find what is diffrent, most probably different impedance characteristics. Im done with triples with BJTs, now only fets as final stage, I dont have the patience to deal with small oscilation hickups anymore although I find the fets to have less bass impact. Im seriously considering apt type circuit for subwoofer use in car amplifiers.
Mike has now had the guts to say that metal cans help, I would not just come out and say it, some members will come tell you youre crazy but I also find this to be tha case. 😀
Maybe its because the the higher frequencies are not so profound and this is what we percieve, I cannot explain it either and the above reason is also maybe the reason behind the push pull topology amp but I doubt this as the armstrong type cicuit has excellent mids and highs yet in the bass department it has more oomph.
Which manufaturer makes mpsa with such low cob, last time I checked figures were around 6 pf.
I have BW803s speakers, I find them to sound very good, perfect to partner electrocompaniet amp, they are not 12 inch sub types that can rattle windows at low volumes though. In my spare car its a different story, bass heaven, there I have 4 12 inchers rockford punch series subs, each dedicated with my self based monoblock amp that can push 400w each. I guess the other 4 6 x 9 fullrange speakers also contribute some to the bass. This is for competition, not recommendable for listening while sitting inside the car. I had to glue the back numberplate onto its plastic frame for it would regurlarly pop out and I keep losing some bolts at the back of the car, not to mention rear lights failures, My girlfriend calls it the speaker box on wheels. 😀 😀 😀
Mike has now had the guts to say that metal cans help, I would not just come out and say it, some members will come tell you youre crazy but I also find this to be tha case. 😀
Maybe its because the the higher frequencies are not so profound and this is what we percieve, I cannot explain it either and the above reason is also maybe the reason behind the push pull topology amp but I doubt this as the armstrong type cicuit has excellent mids and highs yet in the bass department it has more oomph.
Which manufaturer makes mpsa with such low cob, last time I checked figures were around 6 pf.
I have BW803s speakers, I find them to sound very good, perfect to partner electrocompaniet amp, they are not 12 inch sub types that can rattle windows at low volumes though. In my spare car its a different story, bass heaven, there I have 4 12 inchers rockford punch series subs, each dedicated with my self based monoblock amp that can push 400w each. I guess the other 4 6 x 9 fullrange speakers also contribute some to the bass. This is for competition, not recommendable for listening while sitting inside the car. I had to glue the back numberplate onto its plastic frame for it would regurlarly pop out and I keep losing some bolts at the back of the car, not to mention rear lights failures, My girlfriend calls it the speaker box on wheels. 😀 😀 😀
Yes , movie sountracks highlight another aspect of amp
performance , the subsonics are there. Car systems often
highlight the 45 Hz band , still serious bass , but not
HT content, also.. the car's acoustics are a world away
from a living room.(all that BZZZZZZT..)
On the same VGS , I have noticed how different devices "react".
My mj15024/5's are absolutely more convincing than the
small NJL's I use for the frugalamps when powering the
HT sub.
I suspect it has to do with the dissipation of the devices.
(250 W vs. 150W) as far as the metal , I don't know.
It is a lot of work to mount TO-3's , so I might try
the flat TO-264 MJL21193/4 (they have a higher SOA than
all the TO-3's) so I won't have to do all that drilling!!
That would be true psychoacoustically with fullrange applications,
but with the peerless sub we are dealing with 15 -60hz.
The new KSA92/42's are 3-4 pf. (epitaxial)
OS
- Status
- Not open for further replies.