Hmmm. Not sure I get your point. My idea was to get the output stage (buffer) just a few mm from the load circuit. So the inductance of the load carrying wire is negligleable. The remote sense still has higher inductance but since that circuit is higher impedance the impact of the L is relatively lower. What do you think?
Edit: in my idea, the sense would still go back sveral inches to the control circuit (error amp, ref etc).
jd
I said expertS. And you don't know which I meant. I meant those people who have for the past several years not only given their position but backed it up with experiments and measurements. You and your 'references' have limited themselves to explaining when they had lunch together and what car they drive. Some references.
jd
Quote without comment:'4.2.5 Interelectrode Parasitic Capacitances'
"Capacitance is a dynamic characteristic, as it is measured at a high frequency, generally 1MHz. For any FET, whether it is a SIT or a power MOSFET, we collect the various parasitic capacitances and identify them as Ciss, Coss, and Crss.
Each of these capacitances are complex, consisting of numerous parasitic elements. Furthermore, these capacitances are voltage-dependent, raising to high values at low voltages and frequently decaying to quite low values at high voltages. ... etc., etc."
Ref: Ed Oxner, one of my mentors.
"Capacitance is a dynamic characteristic, as it is measured at a high frequency, generally 1MHz. For any FET, whether it is a SIT or a power MOSFET, we collect the various parasitic capacitances and identify them as Ciss, Coss, and Crss.
Each of these capacitances are complex, consisting of numerous parasitic elements. Furthermore, these capacitances are voltage-dependent, raising to high values at low voltages and frequently decaying to quite low values at high voltages. ... etc., etc."
Ref: Ed Oxner, one of my mentors.
Last edited:
In my experiments the inductance and resistance of the remote sense leads had almost no significant effect (tested with about 15cm length), whereas a few cm of directly connected wire did have. This I tested in a real shunt reg with a sine wave load of about 56kHz. Before testing it for real I simulated the parasitic properties of the wire, somewhat exaggerated too, and the results were very good. Please don't be offended, I mention this for other beginners like me, who will benefit: it is essential to measure the voltage differentially at the load, V_positive-V_negative, and not V_positive - GND. I used the diff amp channel A-B on my ancient oscilloscope, at the 100uV/div, with probe A connected to V_positive and probe B connected to V_negative, and their grounds to the regulator ground.
I was reacting to your post implying that feedback causes TIM, PIM, whatever as a blanket statement. Does Ed Oxner write somewhere that this is the case? Show us, if not, please stop trying to mislead us that the above has anything to do with that issue.
jd
I have no interest in being 'helpfull' when you perpetrate fables that have been shown over and over again to be just that, fables. Anything else you can count on me, as everyone here.
jd
But I still don't quite see why the linearity should not follow the curve I linked to earlier, which is quite close to that of a triode?
Magura
Yes, fully agree; that's why I want the driving stage of the reg close to the load.
Ahhh! I was talking about a series reg. Could that be the confusion? Hmmm.
No experience here, but I would guess that also with shunt I would want the load-carrying wire minimal in length. Are we on the same sheet?
Are you kidding?? I'm excited to have at last an engineering discussion here
jd
Sigh. I'm calling you nothing. I'm just saying that some of the 'info' you give is not true, incorrect, has been shown to be false, doesn't jive. I've no idea why you do that, so I can't judge you. And I'm not interested in judgeing you or anyone anyway.
jd
- Status
- Not open for further replies.