I've tried a couple of these DC/DC supply boards and so far my testing hasn't been quite up to par. I've added a decent amount of filtering (47uF 1k resistor into another 22uF cap) and it reduced the noise by a lot but I can still hear it especially if I use compressor after the preamp stages.
Is the idea of DC/DC for a high voltage preamp supply (200VDC), just not practical?
Is the idea of DC/DC for a high voltage preamp supply (200VDC), just not practical?
It is entirely practical, but the noise avoidance is a bit different than you're used to with mains-based supplies.
One of the very important things is the grounding scheme and how you route/ground input signals.
Something that can't be transferred in a single post, I would suggest reading manu app notes on the subject.
Recom, XT power, Traco have those.
Jan
One of the very important things is the grounding scheme and how you route/ground input signals.
Something that can't be transferred in a single post, I would suggest reading manu app notes on the subject.
Recom, XT power, Traco have those.
Jan
it depends on the type of dcdc. most likely a flyback version, that generates ground current noise. the source is the primary to secondary capacitance of the trasnformer where the high dvdt signals are occuring. a sine based llc converter with multiple lower voltage rectified outputs, stacked at the dc side may solve it.
Firstly for the majority of DC-DC converter you will have a minimum current load both in terms of Power Factor Correction (mains side switching) and for the boost/buck. The modern designs will vary the switching. Most are around 50-100KHz but some can drop further but they tend to stick outside of audio range (that doesn't stop power sucking from the HF amplification).
Now how quiet do you want it?
You will find noise below 1MHz can be targeted using things like a Maida-style regulator. I've done that at HT voltage but needs a minimum current draw for the LDOs to keep working which leads me to point 2. The next stage you can use a shunt to provide a minimum constant load - if you use a current based shunt that will be faster and less noise than a voltage comparator based shunt. As the frequencies go up the PSRR of these drops hence the limit of 1MHz as a rough cut off.
Above 1MHz it's all inductors. A good inductor filter design at this point will nail noise nicely. Getting high frequency high voltage inductors can be a pain as all the SMT tend to be focused on low voltage/current ratings.
For my ADC power (low power) I use SMPS input, LDO regulation to 15V (you'd use Maida style here) and then finally inductors. I have a 24MHz clock source on the power line and there's no noise as I also keep the current loops as small as possible. On my tube amp - a cap filter and then a maida works really well.
For SMPS 50-100KHz switching just use a Maida and a shunt if you need to more minimum current draw to keep the LDO working.
Now how quiet do you want it?
You will find noise below 1MHz can be targeted using things like a Maida-style regulator. I've done that at HT voltage but needs a minimum current draw for the LDOs to keep working which leads me to point 2. The next stage you can use a shunt to provide a minimum constant load - if you use a current based shunt that will be faster and less noise than a voltage comparator based shunt. As the frequencies go up the PSRR of these drops hence the limit of 1MHz as a rough cut off.
Above 1MHz it's all inductors. A good inductor filter design at this point will nail noise nicely. Getting high frequency high voltage inductors can be a pain as all the SMT tend to be focused on low voltage/current ratings.
For my ADC power (low power) I use SMPS input, LDO regulation to 15V (you'd use Maida style here) and then finally inductors. I have a 24MHz clock source on the power line and there's no noise as I also keep the current loops as small as possible. On my tube amp - a cap filter and then a maida works really well.
For SMPS 50-100KHz switching just use a Maida and a shunt if you need to more minimum current draw to keep the LDO working.
I would definitively go for a typical power transformer for a preamp, if space permits. The required power is very low, small size, weight and simplicity.
If you use an SMPS, try ones that are electrically isolated, something like the one below (Aliexpress). I have not used it with a preamp, just with a SE, it was dead silent on 90db speakers.
If you use an SMPS, try ones that are electrically isolated, something like the one below (Aliexpress). I have not used it with a preamp, just with a SE, it was dead silent on 90db speakers.
I normally do go the usual 12VAC wall wart to 12vac transformer used in reverse route for 12ax7s. But I have a case where I have 12VDC in the chassis I'm adding a tube stage to. So switching seems to make sense.
The issue with wall warts is that they do not usually include an earth wire, which is required if you use a metal chassis that is grounded with the power supply.
I think that's a feature rather than an issue. To the best of my knowledge, protective earth is not required when the power supply is double insulated, and not having it helps to avoid ground loops.
Last edited:
It is a requirement for me, at least for anything I build that has a metal case in any way connected to a power supply.
Yes, but wall warts usually have no earth ground. That was what I warned about.
I always use standard UK power cables with earth connection.
I always use standard UK power cables with earth connection.
Wouldn't 200VDC made using AC from a tiny transformer reverse fed by 12VAC on the secondary have pretty minimal current. I'd not think it'd to be dangerous. More like a shock from an electric fence, no?
I guess it does have a few capacitors smoothing out the DC. But still wouldn't it still be fairly weak?
We're talking a tiny amount of current through a 12ax7 tube.
I guess it does have a few capacitors smoothing out the DC. But still wouldn't it still be fairly weak?
We're talking a tiny amount of current through a 12ax7 tube.
I use the 150W dc/ac inverter shown in post #6, with its isolated secondaries. I use a 12Vdc wall wart with a current rating to suit the application, and many have no PE connection to 12Vdc, and not having a PE connection can alter ground loop noise, but either way there will be ground loop noise of some sort to manage.
I'm not sure I appreciate the concern of not having a PE connection to a metal chassis that is powered via double-insulated supply - that is how the world goes around - same as if your equipment was battery powered and you touched the metal chassis.
Like mains ripple, there will always be some residual level of high frequency ripple when using a dc/ac or dc/dc. But ripple is managed whether it is mains frequency or 50kHz - either way it can be measured and exists. But if someone is 'hearing' noise from high frequency ripple then I'd suggest that is from poor filtering and ripple management - just as if mains frequency hum was being heard. Some find that ripple management is a black art but it is just electronics awareness.
I'm not sure I appreciate the concern of not having a PE connection to a metal chassis that is powered via double-insulated supply - that is how the world goes around - same as if your equipment was battery powered and you touched the metal chassis.
Like mains ripple, there will always be some residual level of high frequency ripple when using a dc/ac or dc/dc. But ripple is managed whether it is mains frequency or 50kHz - either way it can be measured and exists. But if someone is 'hearing' noise from high frequency ripple then I'd suggest that is from poor filtering and ripple management - just as if mains frequency hum was being heard. Some find that ripple management is a black art but it is just electronics awareness.
I discovered that not all DC-DC chips are the same - and that some use fake chips.
Mine was supposed to switch at 150kHz, but instead switched at 50kHz with a lot of noise, and i researched the problem and discovered that fake chips were widespread.
I'd previously bought the identical looking regulators and measured them, but they had the genuine chips in them.
So I think you have to go to a 'known good' source and measure the ripple, and see if it ties wup with what is claimed and what the chip maker claims.
Fundamentally they should work well: but the fakes are very noisy, and sometimes a cheap SMPS can be a real screamer, radiating dirty switching noise everywhere which is impossible to suppress as much becomes RF.
Insulation can fail. A battery powered device is no in any way connected to the mains, unless you plug it to a charger.
Let's not forget this is DIY, not design houses operating with processes and procedures, certifications etc.
I do not say everybody has to follow the advice, I say that is what I recommend doing.
The premise is to use a wall wart that is 'country authorised' and compliant to standards, and hence has insulation that is deemed safe for millions of supplies being touched by millions of users. I'm not suggesting a diy device that is powered directly from the mains.
I understand your point, yes, in an ideal world that would be completely safe. But with all crap that is being sold in Amazon and the likes, and even some reputable suppliers, I prefer not to take risks. But that's me.
I designed and starting build ing a SMPS for boosting from 24V to 300V puts considerable stress on components plus generates considerable noise both back via the input supply (affects other components) and the output supply. Also finding a transformer that suits the requirements is an absolute ball ache.
It gets more difficult as the boost switching needs to vary to maintain the voltage as the load changes (otherwise you'll get overvoltage) that makes filtering a pain to optimise. The SMPS should have low ESR caps on the input side but it wouldn't surprise me if they weren't and added to the pending destruction of the SMPS..
It gets more difficult as the boost switching needs to vary to maintain the voltage as the load changes (otherwise you'll get overvoltage) that makes filtering a pain to optimise. The SMPS should have low ESR caps on the input side but it wouldn't surprise me if they weren't and added to the pending destruction of the SMPS..
The 150W dc-ac module in post #6 is unregulated, which imho is an advantage as the output voltage (B+ and neg bias if configured for that as well) is fairly stiffly aligned to the nominal 12Vdc input rail. The output voltage can then easily be 'trimmed' down by inserting 1 or more series diodes in to the 12Vdc input feed (sort of like a mains transformer primary with taps).