SMPS + Class A microphone preamp discoveries

[thor.zmt]

I saw the other post but I’m still not quite sure I understand the topology you’re describing. Would it be possible for you to provide a schematic please?

Principle schematic.

Thor

 

[MicUlli]

Hi Thor,
fine concept, clear and straightforward Another member using TINA, I like it.

 

[Potato Cakes]

View attachment 102704

My suggestion. No input capacitor shown. A large value input capacitor should not be needed with SMPS.

This circuit drops out less DC (~0.5V), has lower output impedance and around 6dB more noise suppression than the same circuit as the Darlington configuration.

Thor

So I wired up this circuit and while I was able to make it work without any load, but the output voltage drops to 1V when connected to the audio circuit. The thing I am trying to power draws about 450mA so maybe the BD140 needs to be swapped out with something that can handle more current.

Thanks!

Paul

 

[thor.zmt]

So I wired up this circuit and while I was able to make it work without any load, but the output voltage drops to 1V when connected to the audio circuit.

I suspect you mixed up transistor leadouts somewhere.

I routinely use this circuit (with extremely similar SMT Parts) in commercial equipment. Like so:



FWIW that is a 2012 design for a Headphone Amplifier which was sold across a number a minor revisions and reached >> 10kU.

The thing I am trying to power draws about 450mA so maybe the BD140 needs to be swapped out with something that can handle more current.

I doubt it. Check that you got transistor pinning right.

Thor

 

[emrr]

Couple threads exploring off the shelf switcher noise.

https://www.proaudiodesignforum.com/forum/php/viewtopic.php?t=943
https://www.proaudiodesignforum.com/forum/php/viewtopic.php?t=945

 

[Bo Deadly]

FWIW that is a 2012 design for a Headphone Amplifier which was sold across a number a minor revisions and reached >> 10kU.

Ahh, TPA6120 with no power limit.



So did people have to sign a waiver to purchase this device?

 

[thor.zmt]

Ahh, TPA6120 with no power limit.

View attachment 103150

So did people have to sign a waiver to purchase this device?

???

You are referring to the shorted out 10 Ohm resistor?

Actually, on the PCB the "short" forms a planar (air core) inductor.



The Amplifier was reliable with a dead short indefinite. No waiver needed.

One thing to watch is to not overdrive the input common mode range, the TPA6120 will go into phase inversion in that case.

Thor

 

[Bo Deadly]

???

You are referring to the shorted out 10 Ohm resistor?

No, I mean people are going to destroy their eardrums. I don't see much between the +-9V to the phones. With a 32 ohm load that's multiple watts of power. At least add some series resistance and give people a chance to rip the headphones off of their head.

 

[Potato Cakes]

I suspect you mixed up transistor leadouts somewhere.

I routinely use this circuit (with extremely similar SMT Parts) in commercial equipment. Like so:

View attachment 103132

FWIW that is a 2012 design for a Headphone Amplifier which was sold across a number a minor revisions and reached >> 10kU.



I doubt it. Check that you got transistor pinning right.

Thor

I'll look at it again. Just to clarify, the 40 Ohm resistor that is between V+ and V-, is that the to simulate circuit load or is that actually being used before the audio circuit. I saw that and it seemed strange to me that a resistor value that low would be strapped across the two voltages.


Thanks!

Paul

 

[thor.zmt]

Just to clarify, the 40 Ohm resistor that is between V+ and V-, is that the to simulate circuit load

Of course. 24V / 40 Ohm = 0.6A

Thor

 

[thor.zmt]

No, I mean people are going to destroy their eardrums. I don't see much between the +-9V to the phones. With a 32 ohm load that's multiple watts of power.

Correct. Try Magnetic planar headphones, ~95dB/1V.

At least add some series resistance

Bad idea for the same reason as it is with speakers.

Minimum gain is unity and there is a device called "Volume Control" designed in. It is used to adjust the volume to below maximum across a wide range.

Thor

 

[Potato Cakes]

Of course. 24V / 40 Ohm = 0.6A

Thor

That's what I thought. Just wanted clarify that I understood what that resistor was doing in the schematic.

Thanks!

Paul

 

[Potato Cakes]

Turns out that I was missing a connection on the input side. It seems to be working but the voltage drop is 4V on the output. Also, much more noise then before but there may be an issue with this particular build. I don't have the other capacitance multiplier circuit I built to make a comparison as that is already in use with another build for a client. I'll retest after I rebuild the audio circuit.

Thanks!

Paul

 

[Potato Cakes]

Just for fun, I added a 2200uF cap after the CM and it tapered the high frequency noise some. Still not what it needs to be.

Thanks!

Paul

 

[Potato Cakes]

Well, it seems that the regulators I am using to adjust the voltages are what is causing the noise. The previous pic was CM then IC regulators then circuit. This is that it looks like with regulators THEN CM then audio circuit. I still need to rebuild the audio circuit and do more testing but this is what I am expecting when this audio circuit is working properly.

Thanks!

Paul

 

[Potato Cakes]

Noise after burn in.

 

[thor.zmt]

Well, it seems that the regulators I am using to adjust the voltages are what is causing the noise. The previous pic was CM then IC regulators then circuit. This is that it looks like with regulators THEN CM then audio circuit.

That's possible. The Inverted Darlington Capacitor Multiplier (IDCM) has very low output impedance, compared to the classic Darlington CM (DCM) so it is a good choice after a noisy regulator. Noise is really down to the control transistor.

Drop out is one Vbe plus saturation voltage of control transistor at the current it is used at, I now commonly use 0.1mA nomial and BC550 or 2SC2240 equivalents in SMD.

A high Beta and sustained Beta transistor is recommended for the pass device. Modern TO-220F devices aimed at Amplifier Output stage drivers are a good choice, as are "old faithful" BD139/140.

Some more recent "logic level" Mosfets can have as little as 0.5V threshold voltage and a very high transconductance. These make good pass transistors but commonly only come in really difficult to use (for DIY, not mass production) leadless SMD packages.

Capacitor microphonics can impact the circuit, so the capacitor on the base should have low microphonics. I often use CRCR chains for the reference voltage, putting a 1N4148 across each resistor can speed up turn on time a lot.

Finally, the circuit can be extended to become a highly precise "open loop(ish)" LDO if you use a TL431 programmable reference and a dual transitor with one as control transistor and the second as bias transistor in series with the 431.

So the reference chain is RC and 431 Bias Circuit RC to base of control transistor.

Thor

 

[Yash]

Good to see you here, Thor.

I loved my iFi iDSD Nano DAC, especially when I got the firmware for DSD256 for free and then improved on the USB Isolation and gave it clean power from an external Linear Regulated PSU. Read on another site that I don't post on that you parted ways with them.

Looking forward to learning a lot from you. Was reading you on DIYHiFi until you-know-what...

Best of 2023 and onwards for all your endeavours.

 

[Cqwet Dbdfte]

Hi Thor,
fine concept, clear and straightforward Another member using TINA, I like it.

TINA, ugh, did not like it. LTspice is intuitive, fast, you can learn it in a few minutes. Massive size user-group.

 

[ct.waveform]

My suggestion. No input capacitor shown.

No input capacitor? Yuk! Maybe as a field dressing. That offends me in so many different places. LOL.

I've done almost exactly this after an SMPS but with a voltage regulator where you have your cap multiplier. It works well. I think your R1 is far too high. Mine was for a whole power amp and I used 27R there. Nor was it an especially robust NPN. ZTX753 into TIP36C, iirc. I also used a small stopper resistor as the ztx is a bit lively. The, rather airy, thinking was to save money on the power transistor and control it with something decent that can make up for its deficiencies.

PS The ztx753 will take 2A (or so they say) but dissipating the heat is the problem with those E-lines.