BUZ-amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,
I once pulled a pair of complementary BUZ900/905 lateral MOSFETs from a scrapped subwoofer amplifier and still haven’t found a good circuit to put them into. I have looked at some of the Jean-Marc Plantefève designs but personaly would prefer to use BJT input to JFETs. So as I am not good at all to make a design of my own, I did some copy and paste from circuits of others. Here is what I ended up. So to all circuit wizards -please help me out - is that feasible design or should better I forget it?

Regards,
Argo
 

Attachments

  • mosfet amp.jpg
    mosfet amp.jpg
    60 KB · Views: 1,884
Hi Argo

Sorry, I'm not qualified to comment on the circuit you posted but how about using the BUZ in a JLH MOSFET design. There have been several published over the years that might be suitable. email me if you are interested and I will send you the schematics.

ps Did you complete the JLH Class-AB? If so I would be interested in the results as I have another constructor who is considering building this circuit.

Geoff
 
I never tried a similar design, but for me it looks promising.
Could result in a fast but at the same time stable design.
The output stage seems somehow close to a ZEN....
If you find the right bias... promising....

Hm, bias might be critical in reality.
In fact I am not sure if your DC bias of the output stage
will show a satisfying temperature behaviour.

If you observe issues with oscillation/ringing, then my proposals would be:
1. Make the slowest pole more dominant and more linear.
==> caps from gate to drain of the Mosfets.
2. Shift phase of the feedback signal towards earlier.
==> cap parallel to the feedback resistor.

Bye and looking forward to your results (if you go for this amp)
Markus
 
Geoff,

how good to see you here again.
I believe you are not talking about the JHL MOSFET designs published here by some forum members not long time ago. Tried these but with no great results :flame: (No offence to the authors of these experimental circuits, just couldn’t get mine working properly). If you are talking about the designs made by John Linsley Hood, of course I am more than interested to look at them.

No :sorry:, unfortunately I didn’t complete the JLH Class-AB. Restless soul as I am, got carried away with various other building plans just before firing it up. Now I have cannibalized the amp boards for components and used the heatsinks meant to them to build a voltage regulators by the schematic you once sent me. I don’t know if I will return to build it again some time soon. I’d like to finish an amp with these BUZes before. I need to learn finishing things :headbash:.

Markus,

Thanks for encouraging words and hints :cool: .

Does anybody see some :bomb: odd component values there. Cheers,

Cheers,

Argo
 

fab

Member
Joined 2004
Paid Member
Argo,

I will try to describe your design: Complementary symetric differential cascode input stage with passive current sources and no "standard" VAS (or VAS formed by the output stage).
I believe that this simple circuit topology may be better for a preamp (class A) than a power amp. I will explain why. It is true that power mosfet does not require a high drive current compared to power BJT but you would need a high diff input current (thus medium power transistors) which in turns will load too much the input stage by the mosfets. Also, unless they are matched very closely, the bias of the mosfets will be different and imply a different current for positive and negative diff input stage which is unusual.

Also, R21 and R20 should be much higher (R21 is usually closed to R1). The overall open loop gain is only about 16 to 30 depending on the transconductance of the mosfet (usually less than 2, thus mosfet stage gain 2 * 8 =16) and you have a "desired" closed loop gain of 11. In practice, because of the low open loop gain, the closed loop gain would be about 6.5 to 8. The amp output resistance would be too high which in turns creates a low damping factor, and so on... Unless your intent is to perform a "non overal feedback" design or extremely low feedback design, you need more open loop gain. Again for a preamp since the output load would be something like at least 10K ohms then the open loop gain would be about
Sorry if I may sound a little pessimistic about your design but I just want to help you and no more. Good luck!

Thanks

Fab
 
Fab,

Thank you for fare warning. Better safe :angel: than sorry :redhot:. I have built too many :clock: :$: non-working :wchair: circuits by now. But I didn’t quite got it - do you suggest to forget about this topology or is there a chance to modify it for better performance? Or do you perhaps have a suggestion for another nice input stage ;) ?

Argo
 
Hi Argo!
With the second schematic you will only decouple the differential from the MosFets.
But you will still have the low gain and a high comparably high output stage.
If you go for such much number of stages, you should probably
go for a more standard design.
I.e.
You could use Q9 and Q10 for a traditional VAS.
By this you will not need anymore gain from the output stage
and you could connect Q11 and Q12 as a normal voltage follower.
...entire circuit would become pretty much standard by this....
But if you are not experienced in own designs and all the related calculation and background for frequency compensation you will probably be more lucky if you pick tested and proved design and only use your BUZ for the output.


In fact your first design was somehow charming for me, because it had the following one straight line.
A 2-stage-amp instead of the typicall 3-stage-design.
Low feedback.
Very fast cascode differential input . Dominant pole given by the
MosFets reverse capacitance and the driving resistance....
Also the low resistances in the feedback matched perfectly to this
approach, as this would shift another parasitic pole to very high frequencies.....
Alltogether charming and worth a trial, even if we consider
all the obvious drawbacks which FAB listed.
The ZEN amps have proven that such a low gain+high output impedance design can sound pretty fine.
Your design would combine a ZEN with some more gain and
a DC-coupled signal path... again charming... at least for me
I am not sure if all your values would be really fortunate....
...cannot answer this just by a short look.....


Most new designs require a lot of work, fine tuning, fundamental know how and PATIENCE !!!
I must admit that when I was trying my first own amplifier designs (about 20 years back ? :scratch:) it took me two years to learn some basics until my amps started to work acceptable.... not always fun and really catastrophic for my performance in school... but a passion was born... Well, when I started to work I "forgot" this passion... Now, last winter, I found this forum (accidentially :cheerful: ).... and the fire comes back.....
I am finding a lot of inspiration here, and sometimes ideas
which are new or charming for me.
What did I learn? Nothing!!! I know that it is silly to go
for a own design, if you want a fast success.
But my passion does not wake up, if somebody gives me a prooved design to build....
Even my NAD Receiver did not "survive" more than one day without modification......simply cannot keep my fingers off...

Do you feel the same? What is your intension?
Curiosity? Then blow off your sleep for the next years. :D
Setting up your DIY amp successful? Then pick a proven design.

Cheers
Markus
 
Hi Marcus,
Thanks again for encouragement. While I really appreciate Fabs effort to talk sense to me - what the heck - I have built quite few circuits, which after much hassle and time spent, have given me nothing but a disappointment because of the sound I heard from them after fire up, that one more failure cannot kill my interest to this hobby. So I will build it according to the first schematic, thanks to your enthusiasm, you have also stirred up in me by now.
As I told it’s not my schematic. I copied a most part of it from here http://member.newsguy.com/~stigerik/html/mpa_30_power_amp.htm and the rest I added from the schematic from here http://home.tiscalinet.ch/cooltune/Projects/Megalith01.htm . Yes I am curious what it comes out. If it turns out bad, I will always have a safe wonderful sounding circuit - JLH class A with Geoff’s updates. I could be happy with JLH amp, enjoy it for the rest of my life and have other interests and not build a thing ever but this is not the case. The thing, what drives me, seems not to be a listening of the good equipment but rather building and experimenting with it.

Thanks Torben for the link. I have seen that cicuit before and I like it very much too.
 
Argo,
one more point. You never said which if Plantefeves circuits
you were inspired by, but note that most of his circuits are
current feedback, not voltage feedback like yours. Nothing
wrong with that, it just raises the question if there is much
left that resembles his cirucuits?
 

fab

Member
Joined 2004
Paid Member
Argo,

To respond to your question, your input stage is fine. Your second schematic, as pointed by Chocoholic, has still low gain and high output impedance so I would not go for it. However, my comments are only on electronic characteristics so you know. I do not pretend that low open loop gain sounds worse or better than higher open loop gain. The THD will be higher (but is it an important element?) and the damping factor will be low. For the latter, I am not convinced how it can sound better, but "sound" is subjective to the listeners.

I must say that I did like also your design (maybe it is because it is fully symetric, uses mosfet, cascode diff input, etc). The MPA30 and Megalith have higher open loop gain thus lower output impedance and output transistors are isolated from input stage.
the MPA30 is a low overall feedback and the Megalith is a local feedback design on gain stages.

this is really your decision to continue with your design. Maybe, it is still possible to use the input of MPA30 and output of Megalith to meet your original intent ..but it is tool ate and I must go to sleep...

Fab
 
Christer,

One of circuits of Mr Plantefeves I was inspired of was named "Mr Lanne's amp" (may be its not even his circuit but was published in his web site among his other designs). It is now deleted of from his page and replaced by circuit called Ampereton. I have built both of them but was not really satisfied by their performance. One thing I noticed that they are damn hard to adjust for desired quiescent current and zero dc offset at the same time and need very well matched input jfets.
Also the amount of distortion (both even and odd harmonics) of the Ampereton by simulations (I didn’t measure the actual amplifier) is quite high and was clearly audible.
I am aware that Plantefeves designs incorporate current feedback instead voltage feedback. Does it makes difference, I lack basic knowledge and experience to tell. And you are right - the circuits I have shown do not resemble his circuits any more but it was not my goal to copy them.
I haven’t still tried his Zenquito or Buzquito amp. They seem to be built by many enthusiasts and are well regarded. So at the end of my experiences I can always use these circuits to use the BUZ transistors I have (If they are not blown up by then).

Fab,

Thanks again four your mindful notes. Using mosfets in circuit would not be my first preference otherwise because so far I haven’t heard a really satisfying mosfet amp. But this time I just have them at hand and I am too stubborn and tight-fisted to throw or give them away before I am absolutely sure that they are not good for anything. I really hope to prove myself wrong. I still refuse to consider that they are inherently flawed for best audio and believe that there must be a trick(s) to make them sound good. Hence my determination for these experiments and all sugestions and corrections are more than welcome.



Argo
 
Hi Argo!
Thanks for your links to the Megalith desings...
Well, on the other hand I am not really lucky.
The Megalith is close to some ideas I was thinking of during the
last weeks...
...to close...
OK, I have to accept that there was a guy who had the same ideas already 25 years back!!!!!!! :xeye: :bawling:

Bye Markus
 

fab

Member
Joined 2004
Paid Member
Argo,

Trying to figure out your intent based on your comments on this thread, here is what I began to have in mind before I got to sleep yesterday.

You will recognize your symetric diff input stage (MPA30) and mosfet in voltage gain configuration ( Megalith output stage) with local feedback. Also, it has a VAS to increase open loop gain and the Megalith output stage. It has bias current adjustment to go either class AB or up to class A. Since you can adjust resistors values to your needs ,you are able to either have a very low feedback or high feedback design. It is just an adaptation from your original concept and I do not pretend that it can give outstanding results (but it could too!). What do you think?

Fab
 

Attachments

  • fab_concept_based_on_argo.gif
    fab_concept_based_on_argo.gif
    18.1 KB · Views: 969
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.