| bobsinclar |
Hi guys,
I am currently building PLH and mini Aleph and lack of the final fets that my supplier will send me soon, so I jumped on something else to have my time occupied by something cool...so I jumped on Nelsons schematics for the ZV8 but had no J fets here, so I took a N channel mosfet I had lying here and a IRFP 240 and built this(picture) yes it works ...sound is not bat at all, but on sinewaves seems that the upper part of the sinewawe is cut straight, would anyone be able to have a look at this schematic ant tell me what I did wrong? i'd really like to fix it as the amp sounds really good til 5watts then it starts ti clip on the upper side of the sinewave..:bawling:
Vince |
|
|
| carpenter |
I'm no genius, but it seems to me that you're cascode fet should be idling at half your supply voltage, not 24 volts.
I'd replace the 22k fixed resistor with a pot for adjustment purposes.
Sorry if I'm wrong...
John:angel: |
|
|
| Zen Mod |
| quote: | Originally posted by carpenter
....................
Sorry if I'm wrong...
John:angel: |
for my toob brain,you aren't ;) |
|
|
| carpenter |
I was finally able to offer advise, hurray!
John:D
Thanks for the confirmation, Choky. |
|
|
| bobsinclar |
Hi Guys!
thanks for comments, I actually tried to trim the 22k (that is varianle resistor) and managed to get 9V at the gate of the IRFP 240 ...
but this adgusment does not help to solve tho whole problem...
any other idea? why i can't get the right rail value?
Thanks mates!;)
Vince |
|
|
| Zen Mod |
| quote: | Originally posted by bobsinclar
Hi Guys!
thanks for comments, I actually tried to trim the 22k (that is varianle resistor) and managed to get 9V at the gate of the IRFP 240 ...
but this adgusment does not help to solve tho whole problem...
any other idea? why i can't get the right rail value?
Thanks mates!;)
Vince |
you need to readjust biasing (consequently-Iq,too) and Rload (in this case 14E) to achieve symmetrical clipping;
when you achieve that ,with output node at ~ Ub/2 ,that's it .
in other words- tweak divider for cascode biasing,Rload and even 0R47 to ground
I can't tell you more without looking at precise graphs.......my toob brain works just in that way ;) |
|
|
| JCM |
HI, I love the looks of that thing because it's simple, and I have seen that such circuits can work pretty well.
:2c:
The bias on the lower fet : looks like a 300K resistor from drain to gate, and no resistance from gate to ground - meaning, I think, that the driving source has a big fat gate-to-source capacitance to deal with - which is not so good.
I believe that if you put a voltage divider pot on the lower fet, of a value between 10K and 25K, you can tweak it to maximum output.
You have a 14 ohm resistor dropping 13 volts, so the Idle current
would be 926 mA. For 5 watts into 8 ohms(?), we need 18 volts pk-pk, 9 volts pk, so 9/8 = 1.12 Amps Idle current minimum is needed. So there must be some modulation of the current source going on there, I think. I am not familiar with the operation of that current source you are using. I have a very simple one I like, but it can't be modulated. It uses the gate-to-source voltage to regulate itself. In my experiments, I seem to recall it provided better regulation than the Aleph current source, using dummy loads (although whether that is so important, I don't know).
:soapbox: If you use resistors in the value around or below 10K on the lower mosfet, and make it a standard voltage-divider bias, I think you will have more success. :wrench: |
|
|
| bobsinclar |
Hey thanks guys!
here is the improvment:
carpenter and zen higlighted the fact that the supply value should be divided by 2 at the drain of the 240 by trimming it's gate resistor, actually I just had to use another value of resistor and we got it:D the amp clips symetrically, but only when trimming gives me 16v at the 240 drain when I have 18v at the supply (I don't know why, but when I put half the supply at his drain clipping isn't symmetric...)
Now the amp sounds wonderful very punchy bass and sweet meds...
peak to peak I get about 12v that is to say (if i am correct 6Vrms) on a dummy 8ohm load
I'll try to work on the cascode small fet now just like JCM proposed to...
but this seem a very good start , maybe i'll had to apply some feedback? any comment is welcome!
Vince |
|
|
| bobsinclar |
Oh just forgot,
what could I do to get more power out of it without replacing the supply? I'd like to get 10wattsrms is that possible?actually I didn't choose to clip at about 4 watts:D it's just the way it ended but i'd prefer 10!;)
Vince |
|
|
| Zen Mod |
| quote: | Originally posted by bobsinclar
...................
but this seem a very good start , maybe i'll had to apply some feedback? any comment is welcome!
Vince |
agree with everything ,especially with glorifying Carpenter and me ;) ........
feedback-you already have that :devilr: ,but if you mean on regular loop job,than is time for standard resistor in line with lower transistor gate and one from ldsp node to gate........in this case you have cap induced mumbojumbo included in feedback ;) |
|
|
| Zen Mod |
| quote: | Originally posted by bobsinclar
Oh just forgot,
what could I do to get more power out of it without replacing the supply? I'd like to get 10wattsrms is that possible?actually I didn't choose to clip at about 4 watts:D it's just the way it ended but i'd prefer 10!;)
Vince |
use inductor instead R or use CCS .........whatever
in both case you need to readjust biasing divider(s)
or just make F2..........if you are lazy |
|
|
| bobsinclar |
"agree with everything ,especially with glorifying Carpenter and me ........"
hehe
:D you and Carpenter made my amp alive;) he'll never forget:cool:
thanks for the tips, I'll go on working on it following your advises on the global feedback loop...
Best regards!and many thanks Zen , and also JCM and Carpenter,
really nice place here!
Vince |
|
|
| JCM |
:magnify: ~ I just noticed I totally missed that that was a cascode design. Somehow I looked at that 14 ohm resistor as a filter, not a load resistor, but I see there's no dc cap. there; and that the output is taken from there:Ohno:
The only thing I can think to get more power is to increase the idle current, maybe increase the supply voltage. Might replace the load resistor with a current source - that might be the single biggest improvement. As it is, with the 14 ohm load resistor and an 8 ohm load, the amp is driving a 5 ohm load. |
|
|
| bobsinclar |
So JCM you suggest I just raise idle and replace the 14ohm load resistor with a current source...I'll try that out!
but what do you mean when you say that the amp is driving 5 ohm load?
Vince |
|
|
| lineup |
| quote: | Originally posted by bobsinclar
the amp clips symetrically,
but only when trimming gives me 16v at the 240 drain
when I have 18v at the supply
(I don't know why, but when I put half the supply at his drain clipping isn't symmetric...)
Now the amp sounds wonderful very punchy bass and sweet mids...
. |
To have output at half the supply is many times NOT the best solution. Sometimes is, but far from every case.
We have several good examples of this fact.
In Nelson Pass amplifiers and others.
To adjust for symmetrical clipping, if you want high and good QUALITY output
is more important, than then V/2 output.
Especially when we like in this case
deals with a non-symmetrical amplifier topology.
For so called, dual supply and totally symmetrical amplifiers
using NPN + PNP transistors in mirrored configuration,
is another case.
These uses 99.9% of the time output at half the supply.
lineup
Lineup Audio Lab
http://lineup.awardspace.com/ |
|
|
| bobsinclar |
Lineup,
thanks for the confirmation:)
Vince |
|
|
| JCM |
| The load resistor being 14 ohms, and the load being 8(?) ohms, in parallel, the load is 5 ohms, normally, anyways, I mean for less exotic kind of amps. |
|
|
| lineup |
| quote: | Originally posted by JCM
The load resistor being 14 ohms, and the load being 8(?) ohms, in parallel, the load is 5 ohms, normally, anyways, I mean for less exotic kind of amps. |
The theoretical most effective output stage and maybe also best quality higher power output,
is reached when the amplifier load resistor (DRAIN resistor)
is equal to load average impedance (LoudSpeaker).
But lower the resistor from 14 Ohm to 4/6/8 Ohm
in this case will probably increase the total power too much.
A lot of heat to dissipate :att'n:
But if you have a another good transformer with a bit lower voltage
then it might be worth a new experiment.
Some other component values would need to be adjusted, for sure,
and you may even have to use some other MOSFET.
.... it would be another amplifier .... with another performance ....
lineup |
|
|
| JCM |
| quote: | The theoretical most effective output stage is reached when the amplifier load resistor (DRAIN resistor)
is equal to load average impedance (LoudSpeaker). |
True - theoretically.
| quote: |
But lowering the resistor... will [result in] ...a lot of heat to dissipate. |
One could always take a cue from N.P. and use light bulb filaments for the load resistor. Without turning up the idle current, and maybe supply voltage, a ccs is probably the fastest, maybe only way, to crank up the power output. |
|
|
| lineup |
| quote: | Originally posted by JCM
.... a ccs is probably the fastest, maybe only way,
to crank up the power output.
|
I agree fully, JCM.
One P-Channel IRFP9240 on top as costant current source.
Not only for cranking power, but for lower circuit distortion at equal power output.
But I don't know ...
... there is something special about the sound of a Power resistor loaded Single end output!
And resistors are more linear than any semiconductor.
Even wire wound power resistors are.
It is much up to what the designer himself fancy to do. :)
There are always so many choices ... so many ways and combinations.
This is a bit of what makes audio amplifiers so interesting.
There is always a new idea to test ... in theory, simulation
.. and :clown: even in plain reality, ........ sometimes.
lineup |
|
|
| bobsinclar |
Zen suggested a coil as CCS:rolleyes: what about an impedance" transformer ?would it make sense? I have one here that comes from a broken old JBL bipolar amp from the 70'sI have to fix...
this outpot Xformer has 2 secondaries 16ohm /8ohmand but where I could get the signal :confused: maybe worth to be tested...this would look a bit like a tube amp topology if I'm correct...any idea?
Vince |
|
|
| Zen Mod |
| quote: | Originally posted by bobsinclar
Zen suggested a coil as CCS:rolleyes: what about an impedance" transformer ?would it make sense? I have one here that comes from a broken old JBL bipolar amp from the 70'sI have to fix...
this outpot Xformer has 2 secondaries 16ohm /8ohmand but where I could get the signal :confused: maybe worth to be tested...this would look a bit like a tube amp topology if I'm correct...any idea?
Vince |
just in case:
be sure that this is "proper"xformer,not autoformer as in McIntosh amps.....
regarding "where I could get the signal" it's same as any SE tube stage with OPT....click on my www button,go on "my WOT preamp " page and you'll see plain old SE stage;
drawing is almost nice as yours ;) |
|
|
| JCM |
| quote: | | ... there is something special about the sound of a Power resistor loaded Single end output! |
My experience to date is too limited to comment about the sound, but I like simplicity; the notion of having fewer parts, the most direct practical path from the source to the speakers. It's the easiest and most natural thing to put a current source in there, take out the resistor - but then it starts to not be quite as simple...| quote: |
And resistors are more linear than any semiconductor ... |
That's true, but in an amplifier, my understanding is that a current source is going to force the amp to be more linear than a resistor. Still, I liked the concept of the Zen-light so much I was willing to consider building a 5 watt version. I remember looking at it, but the power dissipation even for that output was - monstrous, by my standards.| quote: |
Zen suggested a coil as CCS what about an impedance" transformer ?would it make sense? I have one here that comes from a broken old JBL bipolar amp from the 70's I have to fix... |
It looks like NP got almost 10 watts out of a cascode design, Zen V9 with 2.2 amps idle current and 50 volt supply using light bulb filaments - the resistance was said to be 11 ohms. There might be no way around increased idle and supply. Five watts output is pretty good with that layout. Transformer ... not sure. He got almost 10% efficiency out of that, which is the same as you're getting. I'm doubtful more power can be squeezed out without turning up the juice. |
|
|
| lineup |
| quote: | Originally posted by JCM
Still, I liked the concept of the Zen-light so much I was willing to consider building a 5 watt version.
I remember looking at it,
but the power dissipation even for that output was - monstrous, by my standards.
. |
I have praised the genius ZenLite by Pass several times before in this forum.
Not many would get this idea to solve power handling issue.
And even only a brave man would be willing to try it, in reality.
It was a real Bright ;) moment in audio history, in my opinion.
( ... but I would not be much surprised if Nelson got his inspiration from some even older tryout,
by some other real pioneer.
Seems like whatever some of us come up with these days, crazy or not,
it has been done before :xeye: somebody has been there first)
I am surprised not many more of those posting here
have built one ZenLite version of their own .....
Besides, it is a spectacular amplifier to show to your friends
with that glowing bulb on top!
I already had this idea, after thinking about this very thread yesterday:
Why not try the same concept for a Single Supply single ended Class A
PREAMPLIFER
Could be called
----------------------
PreLite
----------------------
I am already thinking about what lower wattage bulb
with a suitable resistance could be fit for such an amplifier.
Maybe something used in cars ... ??
Preferably should be a a light bulb with some nice and attractive colour.
;)
lineup |
|
|
| carpenter |
Place the little preamp bulbs inside some beer bottles and call it Budlight. :clown:
John;) |
|
|
| JCM |
| quote: | | It was a real Bright moment in audio history, in my opinion. |
Definitely !
:soapbox: I had been looking at some amplifier designs, all class AB. One thing bothered me (besides xover T.H.D.), it was the issue of power supply voltage dropping in sync with the load demand. Even with a high VA tranny, it still happens. Of course, one could regulate it - and then that becomes part of the design - and more complexity. I minimized the effect in the early stages of experimental amps using not just decoupling resistors and caps, but diodes, and that, as a simple stop-gap, improved performance a lot. But still, the power stage - no getting around it, it can and does modulate the supply.
I kept thinking there ought to be a good class A design, some kind of push-pull, something, I imagined different schemes. But I found it had already been done. And the efficiency far surpassed the textbook examples of 5 or 6 % efficiency. With a class A amp, drawing the same current constantly, the power supply drooping issue is disposed of, and I can filter the supply with resistors, which I greatly prefer to do - something unthinkable with an AB.
And who doesn't love tubes ? A light bulb is ~ sort of ~ a tube - and the first triodes were created in a light bulb factory. Using bulbs for part of the circuit has its own kind of charm - I love the idea, but I was looking at 60 watt bulbs with a filament resistance of 240 ohms - way too high - even 2 is 120 ohms; four is 60 - still too high. That was why I abandoned that notion, plus I was thinking of using bipolars, which seemed to make the whole idea practically impossible. But with that Zen-light, I think N.P. has pulled a rabbit out of a hat. :wiz: I wasn't able to afford the transformer power for it, and of course I shied away from the high dissipation. But it remains a design I admire.
| quote: | | Seems like whatever some of us come up with these days, crazy or not, it has been done before somebody has been there first) |
Yes, seems to be the rule, actually.
Why not try the same concept for a Single Supply single ended Class A PREAMPLIFER ... Could be called PreLite. [/quote]
I guess one could do that. Bulbs with high enough resistance would be easy enough to find. That would be the perfect name for it, I think. :angel:
Beautiful amp there ... sweet music from a class A amp. :bulb: |
|
|
| bobsinclar |
Hehe guys, hat an interesting debate! yes I'd be tempted to try all of them bulbs/ xformer and CCS even modulated CCS(just as in the Alephs...
I am building a Zen with CCS straight from Nelson's schematic...and then I'll have a reference amp to compare to my schematic...By the first listning sessions this is much encouraging, but the amp shows great linearity...but i am curious to see what can be diffrent compared to an original Zen...
anyway I'll come soon with picts and results!and the final schematic if it turn out good...
Vince;) |
|
|
| Tyimo |
Hi Lineup!
| quote: | | Why not try the same concept for a Single Supply single ended Class A PREAMPLIFER |
It is already done! Daniel from Switzerland has built one! Look at the: the enlightened bride of zen
Greets:
Tyimo |
|
|
| lineup |
| quote: | Originally posted by Tyimo
Hi Lineup!
It is already done!
Daniel from Switzerland has built one!
Look at the: the enlightened bride of zen
Greets:
Tyimo |
thanks Tyimo
I go have a look at that link/amplifier
:) lineup
----------------------------------------------------------------
edit:
This is the correct Tyimo link
Pass Labs > the enlightened bride of zen
http://www.diyaudio.com/forums/show...&threadid=37323
|
|
|
|
