XL6009 based DC/DC Converter for sensitive audio voltages like e.g. phantom power?

[trobbins]

Stick a 100uF Cap and a 10mH/10Ohm choke there, after the PCB.

To lower noise on the input side, 3,300uF/100uH/0.1Ohm/3,300uF will isolate the incoming voltage from switching noise.

Thor

The switching frequency is 400kHz, not 400Hz.

 

[thor.zmt]

The switching frequency is 400kHz, not 400Hz.

This PCB, you are quite sure? And you are sure the bodged up 34063 works nice and stable without subharmonic oscillation?

It's not what I see on the 'scope shot in the back of the pic, BWTFDIK.

Thor

 

[rock soderstrom]

Stick a 100uF Cap and a 10mH/10Ohm choke there, after the PCB.

To lower noise on the input side, 3,300uF/100uH/0.1Ohm/3,300uF will isolate the incoming voltage from switching noise.

Thanks, I have ordered a couple of suitable chokes. Yesterday, a test with my new tube microphone preamp using an ordinary RC filter yielded positive results. There is certainly still room for improvement, but overall the whole thing was promising.

The switching frequency is 400kHz, not 400Hz.

This PCB, you are quite sure? And you are sure the bodged up 34063 works nice and stable without subharmonic oscillation?

It's not what I see on the 'scope shot in the back of the pic, BWTFDIK.

Among other things, I actually "measured" something around 400Hz, which irritated me. As I said, my approach could be improved and it could also be due to the upstream power supply, possibly interference?
I am trying to understand the function of the small board in detail with this test, I am currently researching on the web. I'm beginning to understand better how it works.

The journey continues.

 

[thor.zmt]

Among other things, I actually "measured" something around 400Hz, which irritated me.

The circuit was designed to power nixies. Absolute stability and noise were not a relevant consideration.

In typical China fashion, it was then copied badly and applied badly.

As I said, my approach could be improved and it could also be due to the upstream power supply, possibly interference?View attachment 144241

Hard to say, these days SMPS typically use either 67kHz, 132kHz fixed, much higher frequencies or variable (resonant and quasi resonant) switching frequencies, QRP/RP style valley/zero switching supplies are normally targeted at 23kHz minimum frequency under maximum load and ~80khz at idle.

I am trying to understand the function of the small board in detail with this test

What we typically see is this:


Normally a linear regulator for heaters & this for the HT. Switching frequency will be below 20kHz

An alternate schematic in common use on china boards (and a bit better in theory) is this:



This should switch at ~ 30khz.

By comparison, using a suitable MOSFET as cascode on a modern Switcher IC is a much better choice:


This circuit uses a SPOT-23-6 6Pin SMT IC with 400mA maximum current and compact MOSFET with minimal additional parts and switches at 1MHz. It can be very stable and low noise (around like an adjustable 3-pin regukator).

Of course, this actually needs DESIGN WORK, not just copy/pasta, so you will not find it on china aliexpress and e-bay boards.

I have used this with the Richtek RT9297 with ~ 3.8A switch current and 1.2MHz switching frequency to generate high voltages with many Watts of power. Using a USB-C Brick supplying 20V/5A (100W) we can easily get 4A Heaters at any voltage > 18V you like and around 25W HT, using two IC's, one mosfet, one USB-C trigger chip and a few small size passive parts.

Thor

 

[rock soderstrom]

Thanks, that helps a lot.

 

[thor.zmt]

Thanks, that helps a lot.

Here two boards I got to make a HT and heater for a 5670 based circuit.

HT uses SG3525 which is a proper SMPS controller, works up to 400kHz (no idea what is actually used here, have to check) with a push-pull switching stage and bridge rectifier.

Rest is left to the user - no filtering whatsoever integrated.

LT is (fake) LM2596HV with generic elcaps that need replacing with Solid electrolyte types.

So these are a cut above the usual AliExpress junk.

Thor

 

[ccaudle]

Using a USB-C Brick supplying 20V/5A (100W)

I have seen more music equipment (e.g. controller keyboards) running from USB power these days. Any concern about the very light friction fit of most USB connectors? It helps to not break your power adapter when you accidentally pick up your phone or laptop when forgetting to unplug, but studio and stage equipment usually needs a more secure power connection.

 

[thor.zmt]

I have seen more music equipment (e.g. controller keyboards) running from USB power these days. Any concern about the very light friction fit of most USB connectors?

There are one's with a latch or other lock:

It helps to not break your power adapter when you accidentally pick up your phone or laptop when forgetting to unplug

Actually, it doesn't.

I use USB-C charging cable with a magnetic coupling precisely for that reason.

I can pick up most USB Battery banks (and mine are chunky) by picking up the phone and the usb-c cable carries the weight of the power bank.

The magnetic coupling gives way, normal USB-C cables not so easily.

but studio and stage equipment usually needs a more secure power connection.

Well, I see 2.1mm DC a lot. That is less secure than USB-C IF YOU ASK ME.

Thor

 

[ccaudle]

There are one's with a latch or other lock

Interesting, I had not seen those. Nice to see that it is standardized (for example some HDMI cables have a similar locking screw, but I don't think it is standardized in the HDMI spec, so a bit hit or miss how you get the right latching cables or whether the equipment has threaded hole needed).

I see 2.1mm DC a lot. That is less secure than USB-C

Agreed. Sometimes equipment will have a clip of some kind to hold the cable so that if the power cable gets tugged it just pulls on the clip, not the barrel connector, but still not ideal.

Using a USB-C Brick supplying 20V/5A

To get anything other than 5V you need a microcontroller to negotiate the handshake, correct? That makes it a little more difficult for a lot of DIY cases. Maybe someone makes a turnkey device you can drop in, I haven't researched it yet.

 

[thor.zmt]

Agreed. Sometimes equipment will have a clip of some kind to hold the cable so that if the power cable gets tugged it just pulls on the clip, not the barrel connector, but still not ideal.

And of course a similar clip will work the same for USB-C.

To get anything other than 5V you need a microcontroller to negotiate the handshake, correct?

There are many stand-alone IC's where you set the desired voltage with pin strapping to ground, usually jumper.

Readymade PCB's off the shelf.

That makes it a little more difficult for a lot of DIY cases.

Nope, what makes it difficult are all the people looking for reasons why it must be a bad thing.

Maybe someone makes a turnkey device you can drop in, I haven't researched it yet.

Maybe Scheena?

The catchphrase is USB PD Trigger PCB.



Thor

 

[ccaudle]

The catchphrase is USB PD Trigger

Thanks, had a short crash course in USB PD today. Looks like a great way to get decent power supplies at commodity prices.

 

[thor.zmt]

Thanks, had a short crash course in USB PD today. Looks like a great way to get decent power supplies at commodity prices.

One angle I take is that agency compliance (electrical safety, EMC, Mains harmonics) for small scale production and DIY is a huge PITA.

A 100W GaN USB power brick will pass all of it, for a fairly nominal account.

Once we have 20V @ 5A to work with, that's a decent amount of power for a wide range of project.

The 48V @ 5A (240W) variant is standardised, but not yet widely supported.

Some "bricks" on sale now offer dual "lane" 20V@5A for 200W continuous power.

With a suitable strategy (large local power storage) and if we accept that AES2 2012 applies not just to speakers but also the amplifiers driving them, 100W or 200W continuous input power allows 1600W peak Audio Power under AES2 2012.

So, USB-C has a lot of potential for audio.

Thor

 

[ReDolby]

Even at the risk of repeating myself. I would very much welcome it if this idea becomes reality. I have no idea how I can help, but I hope more people jump on this bandwagon.

Think about how a clever solution could make life easier. You set what you need and take care of the (amp) circuit instead of building a dedicated power supply for the umpteenth time.

I would jump on this bandwagon, such a great idea. Just finished another power supply for the umpteenth time and am tired of it. Count me in, not sure how I can help but am quite willing to contribute.

Thanks for a very educational thread. I have lurked for many years, learned a lot, posted infrequently but wanted to applaud this one.

 

[thor.zmt]

I would jump on this bandwagon

Let's start a project definition.

Can I propose:

Outputs:

HT : +120V ... +360V @ 200mA ... 70mA
BIAS: -40V ... -120V @ 30mA ... 10mA
P48 : +48V @ 200mA
LT :+6V ... +12V @ 8A ... 4A
AUX: +/- 12...18V @ 0.35A

Noise levels comparable with LM317 etc. 3-Pin regulators and Switching noise LC filtered into the (a few uV) noisefloor.

Inputs:

USB-C PD 100W x 2 (redundant)

Special Features:

Multiphase switchers for reduced noise and component stresses

Synchronise all switchers to a clock derived from 12MHz on board Xtal or a 256 X "Superclock" at appx 750kHz

Power on sequencing (e.g. Heater on and stable before HT turns on) and soft start

Power monitoring and interlocks in case of failure, few cute LED's so we have Blinkemlichten.

This could go two ways...

One Big Bird ... ahhhmmm Big Board with everything on it...



Or

Split into Utility / Input, Tube, P48 and AUX supply boards (4 separate boards)

Common standard for all power boards is 18...24V DC interface, single-phase ~3MHz sync clock (multiphase clocks made local), Power Good out from PCB.

Each board could also operate by itself.

One single board is much less flexible but probably covers 90% of all cases and is pretty idiot proof, but is a biggish project.

Multiple boards are very flexible. And different boards could be developed by different peeps.

In terms of flexibility, you could, for arguments sake, use multiple P48 and AUX boards to power a really big Solid State mixer at nearly 200W consumption.

Or use two Tube boards for a really big tube project. Etc. Et Al.

I just find that peeps usually overestimate the benefit of flexibility and underestimate the benefit of simple Turnkey.

Thor

PS, I will not design this and make it a ready to run project, but if someone does as open sauce project I'll help for free.

 

[rock soderstrom]

HT : +120V ... +360V @ 200mA ... 70mA
BIAS: -40V ... -120V @ 30mA ... 10mA
P48 : +48V @ 200mA
LT :+6V ... +12V @ 8A ... 4A
AUX: +/- 12...18V @ 0.35A

Noise levels comparable with LM317 etc. 3-Pin regulators and Switching noise LC filtered into the (a few uV) noisefloor.

That would be great, I think it would really cover a wide range of possible applications.

USB-C PD 100W x 2 (redundant)

USB PD now offers really amazing possibilities, the feature would certainly be nice to have, but is not the main focus for me.

Special Features:

Multiphase switchers for reduced noise and component stresses

Synchronise all switchers to a clock derived from 12MHz on board Xtal or a 256 X "Superclock" at appx 750kHz

Power on sequencing (e.g. Heater on and stable before HT turns on) and soft start

Power monitoring and interlocks in case of failure, few cute LED's so we have Blinkemlichten.

I feel like I'm in a fairy tale, these are certainly very attractive features. +1 for the "Blinkemlichten"

This could go two ways...

One Big Bird ... ahhhmmm Big Board with everything on it...

Or

Split into Utility / Input, Tube, P48 and AUX supply boards (4 separate boards)

That's a good question! I'm usually in the all inclusive camp as I really appreciate simplicity. The more I think about it, the more I favor the modular approach here. But maybe you could make two editions, for example one edition for power amps (high power, bias voltage, but no phantom) and a smaller edition for lower power preamps. Difficult decision, probably a modular approach with common sync is the best approach. You only take and pay for what you really need. Space requirements also play a role.

Each board could also operate by itself.

One single board is much less flexible but probably covers 90% of all cases and is pretty idiot proof, but is a biggish project.

true.

Multiple boards are very flexible. And different boards could be developed by different peeps.

also true

In terms of flexibility, you could, for arguments sake, use multiple P48 and AUX boards to power a really big Solid State mixer at nearly 200W consumption.

That could work, but it would bring even more complexity. In my humble opinion, I would reduce the project to the tube application, which is already complex enough.

I just find that peeps usually overestimate the benefit of flexibility and underestimate the benefit of simple Turnkey.

This is indeed often the case, but in this case I think a modular approach would be the better way. YMMV.

PS, I will not design this and make it a ready to run project, but if someone does as open sauce project I'll help for free.

It's completely understandable, we need more supporters. Volunteers please step forward.