The MONGREL (supersym II)

[Art M]

I thought that the series Base resistors on the Driver and Output devices were supposed to take care of any Cranky emitter follower oscillation due to inadvertent negative impedances ?

Just for backup I was considering optional RC's C to B in just the output devices as discussed by Cordell ?

[dadod]

Originally Posted by ostripper

Cherry spoke of the BD139/140 pair. He said they oscillate at 100mhz if driven with over 10cm trace length. He recommended shorter traces and the stopper caps placed within 1cm of the driver C-B. I adhered to this with the above boards for both resistors and caps(and trace length -40-80mm). My modular design forces me to have 4-8cm between VAS OP and Vbe/driver. The Nikko and my JVC BOTH have the 22-47pf caps from B-C on the drivers and long VAS to driver/Vbe traces.

In the post above the addition of the caps prompted the suggestion of "HMC" "hybrid miller comp" . All I know is that it DOES work , even my ears tell me that.

OS

To put a capacitor between drivers base and collector(+ or - power supply) is electrically the same as put it between base and a ground(shunt compensation?). It is not Miller compensation.
dado

[ostripper]

Originally Posted by dadod

To put a capacitor between drivers base and collector(+ or - power supply) is electrically the same as put it between base and a ground(shunt compensation?). It is not Miller compensation.
dado

I didn't say it (that it was miller comp.) This was said on the Cordell book thread.

OS

[artu]

Originally Posted by dadod

To put a capacitor between drivers base and collector(+ or - power supply) is electrically the same as put it between base and a ground(shunt compensation?). It is not Miller compensation.
dado

Hi Dado
Semantics ... even if the collector if fixed there is also a parametric "Miller cap." between B-C (with less "capacitance swing" than a loaded collector, but it swings), so the shunt cap. helps linearising this capacitance, and acts also as a LPF (R-C) R=Vas output impedance, it can be a "Miller-Shunt" if you want.

Cheers
Arturo

[ostripper]

After drooling over $30,000 luxman "space amps" , I decided to redo my already tested LX voltage board (below 1) and do the "correct" setup and pair it with a triple. I need a sub amp badly.

Not any triple ... but B. cordell's "DBT" (diamond buffer triple - below 2). He has it as just a concept in his book , but after rereading my "trouble with triples" thread , it decided to go ahead with the simulation/CAD work. This OP stage just has to compensate one Vbe , will be driven by a hefty output pair
( the drivers = njw0281/0302) , and basically be a scalable "mini arc welder".

I almost have the "DBT" layout done , once I do ... it will exist in the real world. The LX/DBT combo simulates very nicely (attached) ... in fact , the best non-tmc THD-transient performance I have seen. The PCB will have the cap multipliers , PMA's CFP bias spreader , single or dual voltage (without jumpering) capability.

All these subcircuits have been thoroughly tested both for being "idiotproof" and maximum fidelity. The whole package will be a choice of the AX,EX, or LX. The power boards will be the PB120F (EF2 - simple) and the PB150 (this DBT triple). with the triple scalability , i can just make the board longer , clone OP pairs and make a 10 device 80V rail monster that will not "droop" (or drive 2R loads).


WAIT till you see these !!! No more BS- the performance will be superior.
PS - I won't compromise and use angle mounted PCB's ... WHY ??? The pb150 driving a 4R sub with 3-5 pair NJW21193/4's needs excellent thermal conduction (direct mounting) to have the best MTBF. One could use 2 angles to adapt my board , but what is so hard about drilling 7 holes and using self tapping screws ??? (example board- below 3 ... "DBT" / 3-pair OP)

If this works well for my sub , I WILL buy a new $160 par-metal amp case , put the nikko out to the barn and have yet another amp.
OS

[dadod]

OS, did you check OLG, it does not look good.
dado

[ostripper]

No , but I did now (below) . What is wrong ?? Gain Xover at 375khz w/ 87 degree margin. I did cap. load with squarewave , minimal ringing. Triple is very stable.

I already built the LX and ran it for 4 months , so I know it has no "bugs", besides having a high impedance VAS (needs a triple to really perform).

OS

[dadod]

Hi OS, you are correct. Actually I took -20dB as 0 dB so in this respect it looks OK, but 25dB of OLG at 20kHz is it not that to low?
dado

[Macleod]

Originally Posted by dadod

...but 25dB of OLG at 20kHz is it not that to low?

It seems that the PA driver arrangement leads to a reduced VAS load and then a reduced distorsion. Thus a lower OLG may be needed to have similar DHT.

DHT is 0.002849% for 20V peak 20kHz. Quite good for no TMC.

Sure, we are waiting for the TMC version with increased OLG from OS...

[Macleod]

Clipping may need some work I think. Sticking and traces of oscillations.