I hope my latest design is acceptable to you 😛
Edit: still gotta think of a way to get rid of the initial voltage spike though...
Edit: still gotta think of a way to get rid of the initial voltage spike though...
Last edited:
Now you have the noise and the spike to solve and then the heat 😉 I think the choice for class A hampers (I don't mention SMPS, no I don't!) things a little bit, many an opamp based class AB design can be fitted in smaller casings. The small PCB in my post is a +/- 15V very low noise linear PSU for a class AB HPA.
A class A design demands a somewhat larger ventilated casing with enough space around the respective sections and of course it needs a more powerful PSU. Those are design parameters that keep standing in the way of small size and ... waste heat disposal so eventually longevity. So it seems a class AB HPA fed by rechargeable batteries and a small casing are a good marriage.
A class A design demands a somewhat larger ventilated casing with enough space around the respective sections and of course it needs a more powerful PSU. Those are design parameters that keep standing in the way of small size and ... waste heat disposal so eventually longevity. So it seems a class AB HPA fed by rechargeable batteries and a small casing are a good marriage.
Last edited:
If the SMPS has a soft start there might not be any voltage overshoot.
BTW, here are some cute flat transformers... Same thickness as your SMPS...
BTW, here are some cute flat transformers... Same thickness as your SMPS...
Now try a diode instead of your .317 resistor and look at the switching transient. 1A Schottky
the coils cannot push current back into the switcher, but pour it into the caps. To get rid of the ripple put the .317 resistors in line with each cap.
If your coils push back your smps will go into hiccup mode, not so good.
the coils cannot push current back into the switcher, but pour it into the caps. To get rid of the ripple put the .317 resistors in line with each cap.
If your coils push back your smps will go into hiccup mode, not so good.
Start up time is 30ms so I think I'm actually good there 😀
Those transformers really do look cute! I'm tempted to try making a PSU using those now...
Those transformers really do look cute! I'm tempted to try making a PSU using those now...
Ultra-Low Noise Preamp Power Supply: +/-12 V, 650 mA
There was some discussion about the RAC20-xxDK and filtering in that thread as well.
There was some discussion about the RAC20-xxDK and filtering in that thread as well.
There are also pretty flat toroids for low power.
JP where did you get this tiny R-core transformer?
JP where did you get this tiny R-core transformer?
You can get them from Talema but there is also a Chinese? version from YHDC. Just search AliExpress for PTC15 or similar. I came across the brand when looking at the internals of Soncoz SGD1 DAC.
Everything in the AC band is attenuated by over 40dB now, interesting.
Still a spike on the output which is solved by using the resistor and schottky.
So essentially the same idea as independently suggested here and executed quite nicely with extra features that come by choice for the SMPS (always powered on in this case) but at a price that really can not compete with a very good linear PSU. If switchers are more expensive than LPS the world is upside down... Nothing new here, a recent LDO is necessary to have clean PSU lines. So the beloved and suitable for audio SMPS only works optimal for audio when followed either by hefty filters or/and a linear LDO regulator 😀 The cat running after its own tail...
An excellent solution for the HPA by BucketInABucket though!
Bought one from a member here and found another one. They come from used equipment. Like we know from Kitamura Kiden they are excellent. Very small size with 2 x 12V 6VA. Strayfield almost none. Best I can find for the purpose. No EMI creation, no postfiltering, no pulse/hiccup mode when loaded capacitively but dangerous 230V mains work with old fashioned strange devices called fuse holders and "mains switches" 🙂 Standby power usage even better than highest EU regulations.
Last edited:
I can highly recommend XP Power ECL series, digikey have stock. Their big advantage, beside good quality, is that they support unlimited capacitive load....
I believe big linear power supplies are obsolete....
CLCLC is overkill, just CLC is fine, remember to parallel some ceramics.
I use a XP POWER ECL30UT02 myself, with 10000 uF on each analog rail....
Last edited:
+/- 18V is not a common supply voltage, it is just somewhat of a standard for audio stuff. Either you are satisfied with +/-15V or you usually have to regulate +/-24V down to +/-18V and dissipate the heat somehow.
You are 100% right. Big lumpy linear and heavy PSU's for power amplifiers are indeed obsolete in recent designed power amplifiers. Any small low power source device (under let's say 25VA) like a DAC or a streamer can use a small relatively lightweight linear PSU that outperforms SMPS by a margin. Commercial interests aside that is, then SMPS are of course the best!
Last edited:
Yeah, weirdly enough RAC do have 18V modules but two of them are actually the size of this so I'm either gonna have to deal with it or go LPS.
Is the +/- 18V a hard requirement or a matter of preference? The Altra can work OK with +/- 15V I see. It can produce from 13 Vp-p to 21 V p-p which both are instant death for your ears. I think in the real world a fraction of this already is enough to damage ears. The sum of all design choices maybe better is evaluated before actual construction of the device.
Last edited:
Just wanting a challenge, using the XP Power module and a CLCLC (overkill, I know) filter means it's instantly solved and is realistically the solution I'm gonna go for 😀
I just got carried away with theorizing about other stuff...
I just got carried away with theorizing about other stuff...
Würth Elektronik has some nice instructional PDFs on how and how not to implement such filters in terms of layout and also design kits with all the components needed to develop power filters. Sadly the kits are a bit out of my budget. What I also like about Würth Elektronik is that they specify their parts well and test components meant to be used in such filters up to the actual frequencies of interest. I have seen too many datasheets where the impedance curve is cut off where it actually starts to get interesting (at least that was the impression I got when I was searching for some ferrites recently).
I should mention that I'll also be using this same circuit for the Elliott Sound Products Project 113...which also has similar power supply requirements.