Simple Symetrical Amplifier

[Samuel Jayaraj]

If BF471/472 transistors are replaced with faster ones from KSC/KSA series, would the amp need external compensation? The output device Gate stoppers could be redimensioned; it should enhance the bass characteristics of the amp. I have all the parts in my bin. I will first design a PCB. Thanks, Lazy Cat for such an elegant and yet simple design.

[Samuel Jayaraj]

Just another question - Can B550/560 be replaced with 2SC1845/2S992?

[stoc005]

I see about a gain of a bit over 19db, Idle current: P chan: 104 ma N chan:124 ma, DC out (No Signal) -167mv. This is after the bias adjust is set for around 100 ma. 8 ohm load.

Circuit uses BC557C, BC547C inputs. MPSA42/92 "VAS" and EXC16P/N20. Yes, the big 16A output lateral fet models. No output zobel or inductor.

Clipping behavior is benign. No oscillation. Overall, a very stable amp thus far with no miller caps. Wish the gain was a bit more.
-3db at 8 MEGAHERTZ. WOW.
Phase response....the best I've ever seen but that tracks the freq. response.

[Lazy Cat]

@Samuel Jayaraj
BF471/472 pair is just my suggestion for recommended BJT type, of course you can use faster ones like 2SA1478/2SC3788 from Sanyo or similar TO-126 low Wattage case BJT. As long as you'll be locked to proposed sch there is no need for extra frequency compensation for the cascodes cause common base is fixed to well filtered DC potential. Possible oscillation would only come from the input pair but certainly not in this current feedback configuration.

The value of the gate stopper resistors could be defined latter experimentally according to the best sounding result.

BC550C/BC560C are very good signal transistors and can be replaced to similar types but for the 2SC1845/2S992 I am not sure if noise figure is comparable, please check.

I agree design like this combined from only six transistors is nice DIY stuff.

[Lazy Cat]

@stoc005
Very promising LT spice results, I knew the speed of the amp would be enourmous, phase shift also the same.

Output DC offset voltage can be trimmed to zero as I already explained in post #12.

The closed feedback loop gain can be altered simply by lowering the value of 100 ohm resistors, of course both values equaly.

Proposed schematics has very low impedance values to the GND around the feedback loop also because of the strong feedback current symetrically spreaded to the input pair, this leave no chance to parasitic oscillations caused by the same capacitances. From there good stability of this design comes from, thus certainly deserving further investigations to be done.

[Lazy Cat]

Originally Posted by Bigun

A nice project - DIY at it's best !

Thanks Bigun

DIY if it is simple & good, makes a lot of fun. Concretely at this amp current feedback loop (really a kind of a loop) combined with bias DC potentials +/- 0,6 V is what makes it something special, very educational for the begginers.

PCB for this amp could be also very simple and small especially if alex mm would take the opportunity ...

[Lazy Cat]

Originally Posted by MarianB

Very interesting indeed, but most of all very simple and quick to built, good job.

I was looking for a simple low power and not very expensive schematic to built an small amp and i think i have just found it, many thanks @Lazy Cat. I presume the fets can be many tipes, nut just the ones in youre desighn, as long as they present similar specifications, all should be good wright?

So despite so many amp suggestions in another thread you decided to build this one?

[Bigun]

Be careful of PSRR, I believe the Le Monstre amp has quite poor PSRR but this was not a problem as the original used car batteries !

[Lazy Cat]

Only resistor devider for the bases of the cascodes in Le Monstre really can not provide any isolation from power supply ripple or DC variations, so poor PSSR is quite obvious. In my sch +/- 15 V zener filtered DC potential is substantially more adequate solution. If there would be CSS implemented instead of 1 k input pair current supply resistors, PSSR would be comparable to any CF amp.
As I already mention for such low power amps there is not a big problem to invest a little more in the power supply, which can be for this kind of current consumption even regulated to improve PSSR.

[Samuel Jayaraj]

Lazy Cat, thank you for your response. I think that the use of 2SC1845/2SA992 would make the amp more 'transparent' compared to BC550C/560C due to lower Cob of the former. They also exhibit higher gain. Noise should be comparable or better.

I have started with a PCB design and put the BF transistors on the ouput pair heatsink. Would it help in thermally coupling the input pair?

I take it that there is no need for a Zobel network or Output inductor, even with reactive loads. Thanks again. Wonderful design. Should soon be able to report on its sonic qualities.