Keantoken's CFP cap multiplier

[RCruz]

I have got it....Very instructive.... Thank you.

Can it be used before a LM317 to create a superreg?

[keantoken]

I'm not sure why the page wouldn't load, but my friend has a rather special DIY network security system, so maybe it doesn't like you. I will have to speak with him on this. The link is valid, and loads for me.

Using the Kmultiplier before the chipreg is a great way to boost RF input rejection.

The K Multiplier

[RCruz]

Thank you.

I have been using a "normal" cap multiplier before the ship to create good 5v regs for my DAC.

Can I use the circuit as per your site schematic for a +5v reg also ?

[keantoken]

Originally Posted by jameshillj

Hello Iko,

Nice page, KMultipliers.

Not Iko, keantoken!

1.6A steady is getting into wary territory. At this point outputs are generally too slow and Hfe is too low to be unconditionally stable and still have low output impedance. In simulation this is looking like it can work.

The MJE200/210 seem like reasonable transistors in this current range, and while the MJE15033/32 are more robust they are morbidly slow in comparison. Ultimately I would need to test the circuit myself to see if it is viable. You will certainly not get all the performance at 1A that you can get from the standard Kmultipliers at 100mA.

Luckily, both Cordell and Andy_C have provided custom simulation models for the MJE13032/33 pair, so I'm reasonably sure this is the best solution. Again to be sure, I will need to test it myself or have you take some simple measurements.

For best performance the diode string needs another diode. It's not necessary but it helps.

It will help to know what is the pk-pk input ripple voltage at the point you are intending to put the Kmultipier.

EDIT: Added schematic.

[keantoken]

Originally Posted by RCruz

Thank you.

I have been using a "normal" cap multiplier before the ship to create good 5v regs for my DAC.

Can I use the circuit as per your site schematic for a +5v reg also ?

Definitely. It works at any rail voltage, C1 needs to be rated for that voltage.

[AndrewT]

Originally Posted by keantoken

...............1.6A steady is getting into wary territory. At this point outputs are generally too slow and Hfe is too low...............

would a medium power mosFET suit better?
How would the circuit be modified to accommodate a FET pass transistor?

[keantoken]

The Kmultiplier NEEDS to have BJTs for high transconductance or it cannot have low Zout, high PSRR, and low voltage drop all at once. It could be modified to use FETs, but then it would not be nearly so useful.

[jameshillj]

Ah "keantoken", sorry about that.

Those MJ200/210 are 'tuff' little things for sure - I wonder if it would be possible to use 2 in //, and reduce the strain on them without 'current hogging', etc?

Probably better to just reduce the current, I guess to about the usual 1.3A, or less.

The Power Supply is a good quality (Tortech) 600Va torroidal Txr (for both channels) 21v AC, -> byv27 diode bridges -> 4,700uF Siemens B41455 -> 0.1R -> 12,000uF 'Elna for Audio' -> IRFP240/9240 CMx -> 10,000uF SlitFoils.

Dressed this up with Schaffner line filter, CL-60s, basic dc trap, transformer tuning, R-C snubbers, good & short wiring, dc protection/delay, and so on.

[keantoken]

Sounds like you've got an advanced passive PSU section.

I thought of paralleling outputs, but the MJE200/210 are still very nonlinear and not working very well at low Vce. The MJE15032/33 on the other hand worked as well as I could hope for up to the max current, with the exception of speed. Of course this is in simulation.

I tried MJE240/250 and there was a surprising reduction in PSRR. So good Vcesat specs for the output transistor are important. And this doesn't simply mean having low Vcesat; it also has to have high Early voltage. In the OnSemi models it is very low. I don't know if they can work well even if paralleled.

[jameshillj]

Power supply the result of "long time fiddle" and chasing up "good bits" over long time.

The F5 amp is quite sensitive to power supply quality and it's relatively simple to alter the amp's sound via the supply component changes, and a little extra effort - the good quality parts make it easier, except for the size of the final product.

I'm playing around with a simple regulated version (as per Aleph style zener reg) but for all the extra heat and power loss, there isn't that much improvement to the sound, even with all the good caps throughout - a bit disappointing actually. [I may have missed something in the build, perhaps ...] and I'm quite interested to see what your much more sophisticated cct will do in same amp.

It looks like most of these adequate transistors aren't quite suitable for the higher current load - too easy!

I had a weird thought, maybe totally off target - in the F3 amp, there is a little super Jfet (Lu1014) operating with a cascoded power Fet (IRFP240 )- would these items be at all suitable for operations in the Kmx? (I've still got a few left, you see!)