F6 Amplifier

I will not insist. Wired as sent. I will vary Rs between .47 and .22 and see if it is noticable. I will eventually measure, but want to listen and see if i can say what it will show before it is done. I had to use 470R on gates to handle small oscillation with just output hooked up and no FE. I had 47R originally. Perhaps they could be reinstated with GFB use. Will try.
 
Rs is important from two reasons - bias stability , and dynamic influence .

you can delete entire modulation mechanismus ( xformer) and Rs is still in place where needs to be , regarding bias and it's stability

dynamic influence - even if not accounted in xconductance of part itself , it ads to output impedance of amp , contributing to THD form



regarding bias and it's stability:

Pic A there is dc bias stabilisation.
Pic B i see none. (Or it's mechanism is unknown to me.)
 

Attachments

  • P1060092.jpg
    P1060092.jpg
    58.7 KB · Views: 304
Official Court Jester
Joined 2003
Paid Member
The moterboating problem went away when I increased the size of the bias caps from around 22uf up to 470uf. I am not 100% convinced that there isn't the possibility for oscillation, but the frequency now will be below about .05 Hz. Also, the motorboating was seen in a somewhat different circuit shown in post #941 http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-95.html#post3133892. In principle, that circuit should perform similarly to the current one. Looking at the Spice AC analysis, there were very low frequencies where the phase margin was questionable.


Sorry for making confusion. I am so thick.

ZM, i see you keep a big cap across biasing diodes. I remember Papa said to take them off when Ihquam was troubleshooting his motorboating.

Ihquam, where this motorboating came from at the end, you did'nt tell really, i think?
 
"there will not be breathing , when fed with CCS"

So you have experienced a similar low frequency oscillation. How much modulation of the bias voltage (or current) have you seen? Do you fully understand the feedback mechanism?

I have a drawer full of 470uF/35V/125C industrial Elna's

for my ears , they're same as BG well broken in , both bypassed with 10n-1uF MKC

so , stubborn or what , I'm seeding them and green LED's wherever I can ....... :rofl:

there will not be breathing , when fed with CCS
 
The transformer secondaries are set up to drive the Fets in a matched
fashion, if that is what you're asking.

The Fets can be a wide variety of devices. In this case, R100's.

I no longer apply for patents.

I haven't seen this feedback mechanism as such, but you can be
sure it's been done one way or another, probably a long time ago.

:cool:

I've played with a very similar idea about 7 years ago and exchanged some emails about it with Ben Dunkan. Here is the basic circuit. At the time I couldn't find a similar feedback circuit - at least not in LF applications, however I also thought that something like this has been done a long time ago.

Cheers

Alex
 

Attachments

  • Transfollower.gif
    Transfollower.gif
    4.4 KB · Views: 299
Preliminary results make me think less feedback, not more. I currently have .47Rs, bias of 1.3A per fet, 8dB of gain. I have some 330R mills, so Iam gong to try that. BTW, Bypass cap is necessary. I have 1K Nichicon Muse and they add their electro sound to the amp. Don't stone me please. I have some 470uF Silmic II's to try. I will be byassing with fancy Film more than likely. Although i cant grasp why you would want to do so considering the balanced nature of the circuit, I have considered putting two resistors in series to form the Rs Total. Put pot across top one and pull modulation point from their. Wouldnt lose all you DF and would allow for Zen adj.