Vbe multiplier - bypass cap

[Donk]

Hello guys,

I'm interested in what cap value would you recommend for bypassing Vbe multiplier - AND WHY!
Some uses ~47uF elko (Slone, Self, stc...) - some uses 100nF foil/ceramic (DIY amp desingns).
Same for factory manufactured amps: 10....470nF foil cap or 1...100uF elko.

Dynamic impedance is 240R (2,4V Vbe voltage, 10mA VAS current). Sensing device is a BD139.

Thx for any response,

 

[darkfenriz]

Hi
I don't think a dynamic impedance is really 240R, probably an order lower. In my experience this cap is not mandatory at all, especially if you use decent buffering (tripple follower or bjt+mosfet).
Generally 100nF or so helps with dynamic crossover distortion and bigger one, like 47uF helps keeping the bias more constant.
Since you say it is 2,4V I guess you use ordinary darlington, that means drivers can suck a miliampere more from VAS, which is 10mA, this may result in varying LF voltage across multipier.
I suggest bigger cap here.

Regards
Adam

 

[Nordic]

It has our experience almost throughout the DX amp range that too large a value will set off ringing... my preference is no cap there.

 

[Eric Juaneda]

Hello Donk,

Quality of bypassing Vbe multiplier is very important because capacitor's DEFAULTS will be increased (multiply) by gain stage. No capacitor is very good but seems to be noisy. Hi grade capacitor like polypropylene... is very nice. Avoid electrolytic capacitors.

Eric

 

[Donk]

Wow, thanks for the replies!

Darkfenriz, the amplifier is a conventional LIN-topology amplifier with a simple resistor loaded LTP, and a simple Miller-compensated VAS stage running at 10mA current. OP stage is EF type, drivers are TIP31C/32C and the outputs are MJ15003/4 with 0.22R emmiter resistors.
I chose 30mA idle current / OP device.
Some would prefer 0.026V/0.22R=120mA but square wave response looks better at that 30mA, and crossing distortion looks lower too. And yes, the idle heat generated by the OP stage is lot lower.

Then I thinking about to avoid electrolytic capacitor (precisely constant Idle current not needed), and use a simple and cheap foil cap with 100nF value.
Would a cheaper ceramic type causing any problem? What are the opinions about using ceramic cap here?

Regards

 

[kaos]

Just my two cents, for what it’s worth, I’ve tried different values and types of capacitors in that location until I’m blue in the face and haven’t liked any of them. I’ve defaulted to not using any there. My speculation is that the dielectric absorption of the cap is amplified by the output stage resulting in dynamic bias errors. It's only a guess though.

 

[Donk]

Thx Kaos.

On the margin the amp not going to be an audiophile desing, it's for general PA purpouse.
Stability is more important then sound quality.
What would you members recommend?
1., electrolytic
2., ~100n foil
3., ~100n ceramic
4., no cap

I'd vote to 2 or 3.
No.4 (no cap) could cause instability IMHO.

 

[Bonsai]

Donk, my 2 cents worth.

I dont think the value is critical. The Vbe Zout is about 20 Ohms according to Self, and if you use a compound Vbe multiplier, its a lot lower than this.

At high frequencies, the Zout will increase, and it is also likely if you are using a balanced VAS type topology, that the VAS current will not be fixed, as is th e case with a current source loaded VAS (like self used for example) and there will be an AC component.

Under these circumstances, you will see the AC voltage across the VAS increase.

For this reason, there is a case for decoupling the Vbe multiplier with a cap. My preffered cap is a 1uF 50V stacked foil or similar. I doubt using a good quality electrolytic will harm the amps sound - I've seen a few high end designs with electrolytics in this positioin - jus t make sure they work well at high frequencies (so 10Khz upwards.)