DC-DC vs DC-AC

[jdbakker]

[quote author="lagoausente"]I´m quite surprised with a DC-.Dc converter from Bellnix[/quote]
URL ?

[quote author="lagoausente"]that claims only 3mVpp of Ripple an noise, while all other brands have around 75mpp or more.[/quote]
Over what bandwidth ? Ripple/noise figures are useless without a bandwidth specification.

JDB.

 

[JohnRoberts]

I suspect a "P-P" ripple + noise spec assumes a wideband measurement.

JR

 

[jdbakker]

[quote author="JohnRoberts"]I suspect a "P-P" ripple + noise spec assumes a wideband measurement.[/quote]
Sure, but how wideband ? Would be useful to get an idea of the effect of PSU noise.

JDB
[not that I'm saying they are, but it wouldn't be the first time a manufacturer were to pick a 'convenient' bandwidth for a noise spec]

 

[mikep]

[quote author="jdbakker"]
Sure, but how wideband ? Would be useful to get an idea of the effect of PSU noise.
[/quote]

just this week at work I had to measure the noise spectrums of about a half a dozen commercial switchers. some interesting results. the different size supplies within a product family (from a manufacturer that will go unnamed) were all quite different from each other. like, as different as the ones from other manufacturers I measured. the bigger supplies were a little nosier than the smaller ones for the same load. none of the supplies had any measurable power over 100MHz. most had very little above 10MHz but it varies alot. all the PSUs i tested the fundamental of the switching frequency was missing from the spectra, or very low. it appears to be "all harmonics".

mike

 

[mwkeene]

Yeah, I mean noise is definitely a function of topology first and foremost. A forward converter is gonna have less noise than a flyback, all other things being equal. After that you can add snubbers and clamps to help attenuate the leakage spikes. And of course dont forget PCB layout. Parasitic inductances on a pcb will cause all kinds of crazy spikes.

Go with the linear reg and the beads. Just remember to give yourself a few volts of headroom between DC/DC output and the preamp supply voltage, so your linear regulator can regulate. If you use an LDO you will need less headroom. Whats your battery voltage by the way? I seem to remember you saying you needed 20V for the preamp. 24V out from the switcher regulated to 20V should be fine.

The only better solution would be to design the switcher yourself. That would be an education in itself.

80mV pk-pk sounds like a non isolated boost or buck. I got something like that with a 10 watt boost from 12 to 48 volts, with no filtering other than a few ceramic and electrolytic caps. The forward converter uses an LC filter so ripple and noise should be lower. The parasitic winding capacitances in the filter inductor and the ESR of caps tends to let that hi frequency stuff pass however.

-Mike

 

[jdbakker]

[quote author="mwkeene"]If you use an LDO you will need less headroom.[/quote]
All LDOs that I have seen have far worse ripple rejection when operated near dropout than a garden variety 317 with 3V between Vin and Vout. If you can afford turning some more battery power into heat, you're better off without an LDO.

JDB.
[who still believes the OP is better off with a big battery pack and no switchers]

 

[lagoausente]

[quote author="jdbakker"][quote author="lagoausente"]I´m quite surprised with a DC-.Dc converter from Bellnix[/quote]
URL ?

[quote author="lagoausente"]that claims only 3mVpp of Ripple an noise, while all other brands have around 75mpp or more.[/quote]
Over what bandwidth ? Ripple/noise figures are useless without a bandwidth specification.

JDB.[/quote]


URL: http://www.bellnix.com/products/dcdc/br-lb/br-lb.htm

most ones don´t specifies the bandwith, but all have similar noise unless
Bellnix ones.

 

[lagoausente]

Go with the linear reg and the beads. Just remember to give yourself a few volts of headroom between DC/DC output and the preamp supply voltage, so your linear regulator can regulate. If you use an LDO you will need less headroom. Whats your battery voltage by the way? I seem to remember you saying you needed 20V for the preamp. 24V out from the switcher regulated to 20V should be fine.

Yes, the preamp needs 20V, and the 24V converter is isolated, Mean Well: http://www.switching-powersupplies.com/shop/dc-dc-converter/module-type/dke15.html
but all ones I have checked was around those Vpp of noise, unless the Bellnix.
http://www.bellnix.com/products/dcdc/br-lb/br-lb.htm But this one has +/-15V, so I´m not sure if would be suitable for the Supergreen.

The only better solution would be to design the switcher yourself. That would be an education in itself.

Yes, but having in account my knowledge status,and the dificult of switchers I will need much time to that, so better for now, go for a practical solution to can get the preamp running for making some music. I will need to read some big books, and get skills, what I´m planning to do not also for DIY, but also for studing an Engeneering, not sure if I will can do it.

All LDOs that I have seen have far worse ripple rejection when operated near dropout than a garden variety 317 with 3V between Vin and Vout. If you can afford turning some more battery power into heat, you're better off without an LDO.

What´s LDO?

 

[jdbakker]

[quote author="lagoausente"][quote author="jdbakker"][quote author="lagoausente"]I´m quite surprised with a DC-.Dc converter from Bellnix[/quote]
URL ?[/quote]
http://www.bellnix.com/products/dcdc/br-lb/br-lb.htm[/quote]
Its noise performance sure looks good on paper. From the scope plot it seems like mostly switching transients (low in energy, hard to filter out entirely).

These models are 5V in only. Worse, their efficiency is 64%, so you need to take 54% extra weight in batteries with you.

Sorry to nag, but why again are you doing this ? If you want to learn, by all means buy/build a DC-DC converter. If you want a system that's actually optimally useable in the field, just chain together a lot of AA batteries as was suggested some time ago. This has the advantage of being 100% efficient.

[quote author="lagoausente"]so better for now, go for a practical solution to can get the preamp running for making some music.[/quote]

So build a simple op-amp mic pre. Try it. It'll be cheap, easy to power from batteries and you might like the sound enough to not need the Green with its cumbersome bipolar 15..18V supply. 'Good enough in practice' beats 'optimal in theory' every time.

JDB.
[switching efficiency is proportional to switch turn on/off-speed, which is proportional to output noise (for a given topology). So you fundamentally cannot get a silent switcher that's very efficient]

 

[mwkeene]

[quote author="jdbakker"][quote author="mwkeene"]If you use an LDO you will need less headroom.[/quote]
All LDOs that I have seen have far worse ripple rejection when operated near dropout than a garden variety 317 with 3V between Vin and Vout. If you can afford turning some more battery power into heat, you're better off without an LDO.

JDB.
[who still believes the OP is better off with a big battery pack and no switchers][/quote]


LDOs can have some good rejection between 100kHz and 1MHz, up to 60dB. The problem I see is there arent as many options in terms of LDOs at higher current levels in the 20 - 24 volt range.

What's LDO?

LDO stands for Low Dropout Regulator, which means that you dont need as much voltage between Vin and Vout to guarantee your pass device does not saturate. Its basically a PNP instead of an NPN darlington pair, which requires more voltage to stay linear.

 

[jdbakker]

[quote author="I"][quote author="lagoausente"]so better for now, go for a practical solution to can get the preamp running for making some music.[/quote]
So build a simple op-amp mic pre.[/quote]
Your profile doesn't state where you live. If it's hard to get parts, you may wish to get a kit like the ThatMic by forum member mcs. I have one, and I like it. It uses the THAT1510, which effectively is a Green pre on a chip. Depending on your output headroom requirements you could power one from eight AA batteries for about two hundred hours, or use an existing 12V supply with a virtual ground circuit.

JDB.

 

[lagoausente]

These models are 5V in only. Worse, their efficiency is 64%, so you need to take 54% extra weight in batteries with you.

I think maybe efficiency can be not a problem, since the preamp itself has not much consumption, so even must be double, is not very high. Have suggested +/- 50 ma.

Sorry to nag, but why again are you doing this ? If you want to learn, by all means buy/build a DC-DC converter. If you want a system that's actually optimally useable in the field, just chain together a lot of AA batteries as was suggested some time ago. This has the advantage of being 100% efficient.

I also consider the AA battery option, and also have though about testing the three posibiliies, linear, AA, and converter.
Yes, I want to learn, but in the range of time/knowledge I have now. So would be considered as a medium solution. You don´t nag at all. All info have been posted here I´ll came back in the future, with more calm to can learn more in depth. I consider as a first step, starting with a comparison, linear psu, AA baterries, DC-DC converter + linear regulator, and see what happen. If the converter has a considerable higher noise, could be the start point to get in the DC-DC design or filtering.

So build a simple op-amp mic pre. Try it. It'll be cheap, easy to power from batteries and you might like the sound enough to not need the Green with its cumbersome bipolar 15..18V supply. 'Good enough in practice' beats 'optimal in theory' every time.

Any suggestion? would be nice also to build one, and compare with the Supergreen (3 options), and with the Mp20.

 

[jdbakker]

[quote author="mwkeene"]LDOs can have some good rejection between 100kHz and 1MHz, up to 60dB.[/quote]
When operated near dropout ? Which ones ? I know several that have good rejection with a few volts across them, but then they're not very LDO any more.

JD 'genuinely interested' B.

 

[jdbakker]

[quote author="lagoausente"][quote author="I"]So build a simple op-amp mic pre.[/quote]
Any suggestion? would be nice also to build one, and compare with the Supergreen (3 options), and with the Mp20.[/quote]
I mentioned the NYD one a week ago, and Mikkel's ThatMic I just posted about is another option. There's CJ's if you can get a transformer, or this one which is being discussed in the Lab right now. The article mentions +/-15V supplies, I have it working on +/-12V (2x8AA), and I suspect it'll work fine on +/-9V.

JDB.
[and if all else fails, see the Mic Pre Meta]

 

[lagoausente]

I live in Spain. I think I have much information now, and things to check, like those Mic pre kits, AA options, filtering with beads etc.
Thanks everybody.

 

[mwkeene]

[quote author="jdbakker"][quote author="mwkeene"]LDOs can have some good rejection between 100kHz and 1MHz, up to 60dB.[/quote]
When operated near dropout ? Which ones ? I know several that have good rejection with a few volts across them, but then they're not very LDO any more.

JD 'genuinely interested' B.[/quote]

Well, the TPS715xx regs from TI have 40 to 50dB rejection at 100kHz and 1MHz, with higher rejection between (full load). Thats with 1 volt across them. They wont work for this application because they only output 15V max.

I'm no LDO expert, and I'm generally using them in the 3.3V to 5V range. You might have trouble finding one at higher voltages.

-Mike

 

[lagoausente]

I think Portico uses Motien DC-Dc converters, V7 that have 100 mvpp of noise. I guess if they take care about a proper filtering, or not. Or maybe just sold because is "Neve" ?

 

[lagoausente]

A just got reply from Bellnix. The minimun order is 10 units, each unit is US$ 45,6. I guess if there would be someone interested on it, maybe for the low noise specs, 3mvpp.

DC in: 5V, DC out +/- 15V.

the URL: http://www.bellnix.com/products/dcdc/br-lb/br-lb.htm

I´ll put also on the Black Market.

 

[mwkeene]

With 64% efficiency those things might have a linear reg onboard. That, to me, seems to be why they're not getting 80-90%. Buy one or two, try and learn. It's the best way, as long as you're willing to spend some money in the name of experience.

 

[lagoausente]

[quote author="mwkeene"]With 64% efficiency those things might have a linear reg onboard. That, to me, seems to be why they're not getting 80-90%. Buy one or two, try and learn. It's the best way, as long as you're willing to spend some money in the name of experience.[/quote]

Min order is 10 units.
lower performance could be aswell for diferent swithing curve I think, as claimed on AN70 on linear technologies. The linear Lt1533 IC also has 5V input, so maybe the Bellnix can be using it. Due to the way the 1533 switchs gets less efficiency but can get 100uV of noise on the circuit showed on AN70.

http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1040,C1130,P1535,D4159