There have been multiple conversations on many threads about the potential benefits of removing source resistors from the AC current flow;
Nelson write in the F5Turbo article about the application of diodes for the purpose but more recently in several threads there are discussions regarding the potential use of supercapacitors for bypassing the source resistors:
F5 - `Update on Pa's new F5?', some comments in a tread on the F5m.
And of course most recently in the ACA Mini - the article itself
After some searching and reading I am curious why they are not used more often and I am hoping I could gain some enlightenment on the subject from the great and wise?
Nelson write in the F5Turbo article about the application of diodes for the purpose but more recently in several threads there are discussions regarding the potential use of supercapacitors for bypassing the source resistors:
F5 - `Update on Pa's new F5?', some comments in a tread on the F5m.
And of course most recently in the ACA Mini - the article itself
After some searching and reading I am curious why they are not used more often and I am hoping I could gain some enlightenment on the subject from the great and wise?
it's Ok for ACA Mini, but I'm not sure their current rating is anything enough for bigger amps and possible current through each cap in that arrangement
though, there are years when I last time checked in supercap datasheets ....... was hoping to use them for low voltage PSU but nada
so, check ripple current figures in datasheet, calc what can be asked in amp of your interest and you'll know
ZM's Square Law OS Fetish is solved in different way
though, there are years when I last time checked in supercap datasheets ....... was hoping to use them for low voltage PSU but nada
so, check ripple current figures in datasheet, calc what can be asked in amp of your interest and you'll know
ZM's Square Law OS Fetish is solved in different way
Omniscient ZM!
cheers
Here is the one Nelson suggested for the Mini ACA SCCR20B335PRBLE
Peak current appears a pretty sensible 3A:
https://www.mouser.com/datasheet/2/40/AVX_SCC_LE-3084833.pdf
cheers
Here is the one Nelson suggested for the Mini ACA SCCR20B335PRBLE
Peak current appears a pretty sensible 3A:
https://www.mouser.com/datasheet/2/40/AVX_SCC_LE-3084833.pdf
Ummmm not to flog a fatally wounded equine quaduped....
A typical FW power supply of +/-24V and a Vgs of ~4V this implies a max. voltage swing of +/-20V which is 2.5A peak into 8 Ohms. So am just surprised this idea doesn't feature much beyond the occasional question.
A typical FW power supply of +/-24V and a Vgs of ~4V this implies a max. voltage swing of +/-20V which is 2.5A peak into 8 Ohms. So am just surprised this idea doesn't feature much beyond the occasional question.
Supercapacitors for regenerative braking of electric vehicles can handle dozens of amperes.
Please keep in mind that the capacitors need to get charged before you blow up the MOSFET.
That is, I assume the reason why there is a source resistor in the first place, if it is not desirable for the signal path, is to stabilize MOSFET biasing or as part of some sort of protection. Shunt such a resistor with a really big capacitor and you will blow up your MOSFET again, just like you would without the resistor.
Please keep in mind that the capacitors need to get charged before you blow up the MOSFET.
That is, I assume the reason why there is a source resistor in the first place, if it is not desirable for the signal path, is to stabilize MOSFET biasing or as part of some sort of protection. Shunt such a resistor with a really big capacitor and you will blow up your MOSFET again, just like you would without the resistor.
Disclaimer: I may well be wrong here but...
My understanding is that source resistors are a source of local feedback (degeneration), and as you surmise, stabilise the DC bias: They reduce the impact of positive thermal drift of Vgs, by sucking up an increasing amount of voltage as bias current increases. They also reduce the impact of mismatched n & p devices (and hence influence the amount and phase of resultant second harmonic distortion). However, they also reduce the gain and limit the maximum available current on large signal excursions towards the rail voltage
DC bias current is limited by the source resistors and inclusion of a supercapacitor in parallel allows the AC component (signal) to bypass as outlined by Nelson in the ACA Mini article. He notes this has no impact on DC bias and minimal impact on thermal loading of devices.
It appears a great idea but I may be missing something. I don't see the technique used a lot. Perhaps in this era of disappearing through hole components and rare Toshiba jfets, supercapacitors are one of the few new toys for us to employ(?) but the idea has been around for a while and where I have seen reference to it, the conversation has been brief and the idea tends to be discarded -perhaps it just doesn't sound great(?).
My understanding is that source resistors are a source of local feedback (degeneration), and as you surmise, stabilise the DC bias: They reduce the impact of positive thermal drift of Vgs, by sucking up an increasing amount of voltage as bias current increases. They also reduce the impact of mismatched n & p devices (and hence influence the amount and phase of resultant second harmonic distortion). However, they also reduce the gain and limit the maximum available current on large signal excursions towards the rail voltage
DC bias current is limited by the source resistors and inclusion of a supercapacitor in parallel allows the AC component (signal) to bypass as outlined by Nelson in the ACA Mini article. He notes this has no impact on DC bias and minimal impact on thermal loading of devices.
It appears a great idea but I may be missing something. I don't see the technique used a lot. Perhaps in this era of disappearing through hole components and rare Toshiba jfets, supercapacitors are one of the few new toys for us to employ(?) but the idea has been around for a while and where I have seen reference to it, the conversation has been brief and the idea tends to be discarded -perhaps it just doesn't sound great(?).
I have worked with 300 F supercapacitors for a non-audio project. In order to measure the leakage current of one of them over time, I charged it to 2.5 V with a simple LM317 voltage regulator with an auxiliary circuit that limited the current to 1 A.
The current actually dropped to around 700 mA before the capacitor was fully charged, because the heatsink I used was too small to support 1 A of output current during the full 12 minutes and 30 seconds needed to pump a charge of 300 F • 2.5 V = 750 C into the capacitor at a 1 A = 1 C/s rate. The LM317 went into thermal protection and further limited the current.
The current actually dropped to around 700 mA before the capacitor was fully charged, because the heatsink I used was too small to support 1 A of output current during the full 12 minutes and 30 seconds needed to pump a charge of 300 F • 2.5 V = 750 C into the capacitor at a 1 A = 1 C/s rate. The LM317 went into thermal protection and further limited the current.
