In the discussion with Merlin el mago it was additionally talked about the effect of changing R1 to increae the current to 300-400mA for better performance.
I think that the discussion was referring to V1.0 - is that correct?
Is something like that diserable for V1.2, too? Or is setting 200mA the best choice?
In terms of regulated output to 5V/15V what is the maximum power without compromising performance?
I think that the discussion was referring to V1.0 - is that correct?
Is something like that diserable for V1.2, too? Or is setting 200mA the best choice?
In terms of regulated output to 5V/15V what is the maximum power without compromising performance?
Only the MKP Vref capacitors. Tolerance no problem. Try some subjectively. Contentious area. 47uF choice lytic is another option to audition for Vref cap.
300mA will further lower Zo in 1.2 as well. Only its performance is more complete on get go, so the benefit somewhat less. Judge against heat and sinking area.
Maximum power is relevant to consumption.
I'm really fascinated by your enthusiasm to answer all my "silly" questions, Salas. Thanks a lot.
You are speaking of VRef Capacitors in plural:
C2 is obvious with 4,7uF - Probably Mundorf is not the worst choice and they are easily available in Germany.
There are two additional capacitors in the design:
C1 is 33pF which is not available as MKP, is it?
C3 is part of the Zobel-network that you adviced to replace with a single medium ESR 22-47uF electrolytic cap. Are you talking of that one, too?
For 5V and 15V respectively: should I lean more towards 22uF or 47uF?
Does using Vishay Dale resistors benefit the quality? The price is not too extreme and I have often read about those resistors being superior in this forum.
Talking about the power supply for the digital system (Tent-clock, receiver and CLPD) - is using Mundorf capacitors overkill?
Further following the discussions with Merlin el mango, the symmetric supply I need for +/-15V will need some tweaking to match both channels.
I have seen some V1.0 boards that had places to solder multiple resistors in parallel - I guess that it has been done for exactly that reason.
Which resistors would need that kind of tweaking?
At least R1, but are there further positions?
Talking about the adjustments - isn't the noise introduced by a trimmer far worse than from good resistors?
Should I design the pcb so that I can replace the trimmer with some resistors in parallel/series to solder after the exact resistance has been determined?
Or is the noise florr from the other components far too high for that effect to matter?
As you might guess, I am just a software engineer and although I had electronics, the time is far too long gone and not that much of interst to me anyhow, so I'm really sorry if my questions sound abolutely silly...
You are speaking of VRef Capacitors in plural:
C2 is obvious with 4,7uF - Probably Mundorf is not the worst choice and they are easily available in Germany.
There are two additional capacitors in the design:
C1 is 33pF which is not available as MKP, is it?
C3 is part of the Zobel-network that you adviced to replace with a single medium ESR 22-47uF electrolytic cap. Are you talking of that one, too?
For 5V and 15V respectively: should I lean more towards 22uF or 47uF?
Does using Vishay Dale resistors benefit the quality? The price is not too extreme and I have often read about those resistors being superior in this forum.
Talking about the power supply for the digital system (Tent-clock, receiver and CLPD) - is using Mundorf capacitors overkill?
Is it set via R1 only still? Anything else that needs tweaking (beside heatsinking for the o IRFP9240s)?
Further following the discussions with Merlin el mango, the symmetric supply I need for +/-15V will need some tweaking to match both channels.
I have seen some V1.0 boards that had places to solder multiple resistors in parallel - I guess that it has been done for exactly that reason.
Which resistors would need that kind of tweaking?
At least R1, but are there further positions?
Talking about the adjustments - isn't the noise introduced by a trimmer far worse than from good resistors?
Should I design the pcb so that I can replace the trimmer with some resistors in parallel/series to solder after the exact resistance has been determined?
Or is the noise florr from the other components far too high for that effect to matter?
As you might guess, I am just a software engineer and although I had electronics, the time is far too long gone and not that much of interst to me anyhow, so I'm really sorry if my questions sound abolutely silly...
Yes, C2 I am talking as MKP Vref cap, your choice. For the 33pF use C0G/NP0 or Silver Mica.
In between 22 to 47uF there is always 33uF. Normal electrolytic. Not those super low ESR, not MKT/MKP.
Use a Dale for 27k Rref. The others can be generic.
R1 is setting the CCS current 0.6V/R1=CCS current. Use 1W single resistor.
Don't know what boards you have seen. Maybe Quanghao's SSHV? That one I remember it has 2W multiples but there the dissipation is high due to high voltage.
The noise floor isn't high at all, its good to have place for a fixed resistor to replace the trimmer after fixing permanently, a wiper is a moving contact, not ideal guarantee. That could a Dale too.
In between 22 to 47uF there is always 33uF. Normal electrolytic. Not those super low ESR, not MKT/MKP.
Use a Dale for 27k Rref. The others can be generic.
R1 is setting the CCS current 0.6V/R1=CCS current. Use 1W single resistor.
Don't know what boards you have seen. Maybe Quanghao's SSHV? That one I remember it has 2W multiples but there the dissipation is high due to high voltage.
The noise floor isn't high at all, its good to have place for a fixed resistor to replace the trimmer after fixing permanently, a wiper is a moving contact, not ideal guarantee. That could a Dale too.
Not a typo, the PSU board of Dac End2 has a rectifier before the shunt reg.
I have Vgs 3.9V for IRFP9240 and 3.57V for IRFP240. I think its OK.
I used 2x100nF MKP2 as I read in "Building a symmetrical psu B1 buffer" thread that others found this the best. I will try a 47uF Oscon see how it sounds. As I understood, the function of this cap is noise filtering of leds. Why a smaller value can give better subjective performance than a bigger?
I tried bigger cap over Vref, not helping,... then
I found the problem: my test board's mains switch was probably arcing badly If I simply unplug the trafo from the mains everything is perfect. I will change it and see.May I suggest Wima MKP10, easily available in Germany too, for C2 & C3?
Moreover (post 1894 & 2134), Salas also stated "up to 10uF MKP or MKT" for C2 that seems even better IMO.
If you can find a vendor offering Wima MKP10 10uF/100V pls let me know. I'm hunting hi and lo for them.....
I used 2x100nF MKP2 as I read in "Building a symmetrical psu B1 buffer" thread that others found this the best. I will try a 47uF Oscon see how it sounds. As I understood, the function of this cap is noise filtering of leds. Why a smaller value can give better subjective performance than a bigger?
I tried bigger cap over Vref, not helping,... then
Glad you found the odd problem. The DCB1 does not have a resistive element too, as yours got. If there is one, they form a high pass together with the noise filter cap. The experiments by enough members on the Leds only ref DCB1 thread have shown that the superiority in construction and characteristics of a small MKP cap outweighed subjectively lytics. But your mileage may vary.
In the past I experienced instability when placing oscon caps after regulators... particularly the negative ones.
OsCon sound very good decoupling digital circuits but need to be near the chip and away from the regulator.
Because the previous cap's ESR was intrinsic and calculated to the loop gain phase margin of the design. The OsCon dropped that. A reg is a DC amp. All complexities of amp design apply. And it must be better than your audio gain cct in DC so to get out of the way. That is why a reg has such an impact when initially seems like performing a secondary task.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.