Hello Folks,
for an active Speaker project I am (re)using VFET Power Amp Modules and Transformer from old Speakers (let's leave the reasoning for that off topic here 😉). The Power Modules need +/-50V for the Driving Stages and +/-42V for the Power Stages.
I want to use LT4320 based active rectification (have the madules already), the toroid transformer's windings are shown in the principle schematic on the left. The My question is: Which of the LT4320 based detail circuits A or B on the right will work (if any)? If both should work, which one is better for a Power Amp?
(Circuits A and B show positive half of supply only, the negative half would be designed like "mirror image")
Thanks a lot for suggestions and explanations!
Winfried
for an active Speaker project I am (re)using VFET Power Amp Modules and Transformer from old Speakers (let's leave the reasoning for that off topic here 😉). The Power Modules need +/-50V for the Driving Stages and +/-42V for the Power Stages.
I want to use LT4320 based active rectification (have the madules already), the toroid transformer's windings are shown in the principle schematic on the left. The My question is: Which of the LT4320 based detail circuits A or B on the right will work (if any)? If both should work, which one is better for a Power Amp?
(Circuits A and B show positive half of supply only, the negative half would be designed like "mirror image")
Thanks a lot for suggestions and explanations!
Winfried
Last edited:
Both will work but the bottom one looks like an unnecessary complication, requiring higher voltage rating for the upper rectifier fets.
Thanks for the quick response and the good news! So I can go on from here and consider other issues:
The original circuitry employed two DiodeBridge Rectifiers, one for the +/42V and the other for the +/-50V. This seems to me is more closely resembled in version B rather than in A.
Thinking about that more, my concern with version A rose that if the 42V-rail current "increases or decreases" the actusal 42V-rail voltage (due to varying load currents), this will directly "modulate" the 50V rail as well, due to their series connection. Version B would probably be more "stable" in this respect. What do you think?
Thanks for raising the Udsmax concern: I do have SQJQ980EL DualNMOSFETs which should suffice for the 50V rails, the 42V-rails get TPH1R306PLs which I have as well.
Regards,
Winfried
The original circuitry employed two DiodeBridge Rectifiers, one for the +/42V and the other for the +/-50V. This seems to me is more closely resembled in version B rather than in A.
Thinking about that more, my concern with version A rose that if the 42V-rail current "increases or decreases" the actusal 42V-rail voltage (due to varying load currents), this will directly "modulate" the 50V rail as well, due to their series connection. Version B would probably be more "stable" in this respect. What do you think?
Thanks for raising the Udsmax concern: I do have SQJQ980EL DualNMOSFETs which should suffice for the 50V rails, the 42V-rails get TPH1R306PLs which I have as well.
Regards,
Winfried
Last edited:
Not sure i see any meaningful difference in the degree of modulation the 50v rail sees. In one case the AC sides are in series and in the other the DC. The rectifiers do not present a barrier against current draw.
The only way to avoid this is with a separate dedicated winding for the full 50v or even better, a separate transformer. This would definitely be my preference.
The only way to avoid this is with a separate dedicated winding for the full 50v or even better, a separate transformer. This would definitely be my preference.
There is another solution to rail voltage modulation. Active rectification provides 1.5 V more compared to the diode bridge, so there will be + 3 V headroom on 50 V rail. It is enough to use linear regulator and have clean supply for driving stages. I’ve gone this route (linear regulated PS) for my pending amplifier build.
I don't know why I'm not sure about B (for active rectifier). The A is really safe.
But I would use regular diodes for +-50 VDC stage. Of cause there is some sense to use Ideal Bridge Rectifier in a low-voltage high power stage, but high-voltage low-current driver stage may be powered even by single-phase diode rectifiers (with high enough filter capacitance).
So I vote for option C: two Ideal Bridge Rectifiers for +-42 VDC (each for 42 VDC), and one classic diode (bridge) rectifier for +-50 VDC (with 63-80 VDC rated caps)
But I would use regular diodes for +-50 VDC stage. Of cause there is some sense to use Ideal Bridge Rectifier in a low-voltage high power stage, but high-voltage low-current driver stage may be powered even by single-phase diode rectifiers (with high enough filter capacitance).
So I vote for option C: two Ideal Bridge Rectifiers for +-42 VDC (each for 42 VDC), and one classic diode (bridge) rectifier for +-50 VDC (with 63-80 VDC rated caps)
Last edited:
Applications for the LT4320 go up to 70V, so there's sufficient headroom left and from my experience the ideal bridge rectifier has audible advantages in audio circuitry over diode rectification, so I plan to stay with that. Granted: The supply is immensly more complex and costly this way.
What a great idea! It makes sense to me to stabilise the 50V supplies for the Amps' Voltage Amplification stages (like I do for the OPA circuits on the +/-15V anyway). I'll clearly consider that option.
Thanks for all suggestions!
Winfried
Hi Folks,
I‘ve tried to find a simple 50V Regulator Solution and came across the TL783KC, but could not find in the datasheet if ~3V Input to Output difference will work...
I‘d be grateful for advice or other recommendations for the 50V Regulator. The current draw is <400mA.
Thanks,
Winfried
I‘ve tried to find a simple 50V Regulator Solution and came across the TL783KC, but could not find in the datasheet if ~3V Input to Output difference will work...
I‘d be grateful for advice or other recommendations for the 50V Regulator. The current draw is <400mA.
Thanks,
Winfried
Last edited:
Figure 10. from datasheet specifies the dropout voltage. At 50 V and 400 mA, dropout is 10 V. Considering that dropout voltage is absolute minimum voltage difference at which regulator still maintains regulation, it is really not suitable for your case.
Thank you "tombo",
well, 10V min. drop just looked unbelievable to me, so your advice is well taken.
In the mean time I have "discovered" the LM3x7HV which look to me like being well usable up to 60V Input and need min. 3V voltage drop. With the low constant current draw, the standard application (with protecting diodes) seems sufficient to me.
So what about this schematic?
As four PwrAmp boards need +/-50V supply, I could potentially place dedicated regulator circuits on each Pwr-Module... 😎
Regards,
Winfried
well, 10V min. drop just looked unbelievable to me, so your advice is well taken.
In the mean time I have "discovered" the LM3x7HV which look to me like being well usable up to 60V Input and need min. 3V voltage drop. With the low constant current draw, the standard application (with protecting diodes) seems sufficient to me.
So what about this schematic?
As four PwrAmp boards need +/-50V supply, I could potentially place dedicated regulator circuits on each Pwr-Module... 😎
Regards,
Winfried
That will work OK. Although no output capacitors are required at the output of LM3xx regulators for stability reasons, I would put there low ESR 220 uF capacitor (like Panasonic FC series) because of transient response and noise. LM3xx regulators have really poor response on load transient (datasheet figure 12.) without any output capacitor.
I'm not sure if B will be possible with the LT3420, but I agree it looks complicated.
However, 28 volts AC makes 42 dc yes but 10 volts AC makes 14 DC, adding them (figure A) makes 56 volts, not 50. That might be oké for your application, but just wanted to point that out ;-)
The LT can remain active with an input until around 6.x volts AC/9.x volts DC. Below it just uses the fets passive: the built in diodes, so if 56 volts DC is a problem, there's room in both transformers to solve the issue, if needed.
However, 28 volts AC makes 42 dc yes but 10 volts AC makes 14 DC, adding them (figure A) makes 56 volts, not 50. That might be oké for your application, but just wanted to point that out ;-)
The LT can remain active with an input until around 6.x volts AC/9.x volts DC. Below it just uses the fets passive: the built in diodes, so if 56 volts DC is a problem, there's room in both transformers to solve the issue, if needed.
Hello all,
the output caps are actually planned and additional buffering of the rails on each module as shown in the updated, more elaborate schematic:
Greetings,
Winfried
the output caps are actually planned and additional buffering of the rails on each module as shown in the updated, more elaborate schematic:
Greetings,
Winfried
What's the idea behind connecting the secondary transformer windings, when you reference them again later after rectifying them?
Edit: now your secondary DC voltage is depending on the relative phase of the secondary windings.
Edit: now your secondary DC voltage is depending on the relative phase of the secondary windings.
Last edited:
@ mterbekke
Thanks for your feedback! Unfortunately, I don's get from it what I should do...
So, let's take one step back: The original supply schematic is below and the goal is to use LT4320 based active bridge rectification.
Please share your proposal how to do it best with ABRs.
Thanks a lot and kind Regards,
Winfried
Thanks for your feedback! Unfortunately, I don's get from it what I should do...
So, let's take one step back: The original supply schematic is below and the goal is to use LT4320 based active bridge rectification.
Please share your proposal how to do it best with ABRs.
Thanks a lot and kind Regards,
Winfried
The only way to generate a symmetrical supply is to use 2 windings per ideal bridge, use 2 of them in series and use that connection to derive the GND.
Center tabs for mid supply in the classical sense, destroy the bridge.
Center tabs for mid supply in the classical sense, destroy the bridge.
Can you kindly provide a simple drawing according to your above instruction " 2 windings per ideal bridge, use 2 of them in series and use that connection to derive the GND".
Thanks
Imagine to build this:
As per the first picture of this thread, Circuit A.
1 pair of secondary windings go to 1 bridge.
You build 2 of those and connect them in series at the output, as of circuit A in the first picture of this thread, but:
The connection made between them at the output (as drawn in 1) you now refer to as ground, instead of the bottom output.
See the symmetrical supply come into existence before your eyes?
That also means the last bottom capacitor at the output in fig. A needs it's lower connection disconnected to gnd (already has gnd at his top connection remember?) , you can place them next to "c1", lose the bottom GND connection to it and be done with it.
Followed that?
Good luck!
As per the first picture of this thread, Circuit A.
1 pair of secondary windings go to 1 bridge.
You build 2 of those and connect them in series at the output, as of circuit A in the first picture of this thread, but:
The connection made between them at the output (as drawn in 1) you now refer to as ground, instead of the bottom output.
See the symmetrical supply come into existence before your eyes?
That also means the last bottom capacitor at the output in fig. A needs it's lower connection disconnected to gnd (already has gnd at his top connection remember?) , you can place them next to "c1", lose the bottom GND connection to it and be done with it.
Followed that?
Good luck!