Intro and PSU opinions

[Giggity]

Hello all.  I thought I would make a proper intro.  First, I am a DIYaholic.  This applies to many things.  But within this context, my knowledge is limited.  Not a total noob but no electronics degree either.  Outside of a few amp repairs, my experience includes a few guitar pedal builds (2 kits and one from a PCB and order the parts separately), A "Spyder" PSU for said guitar pedals and a +/-15 & +48vdc PSU for my Soundcraft mixer.  Though, after building it, I found out how to repair the original SMPS.

Recording wise, it's just a hobby and only of and for myself.  The Soundcraft mixer (a Spirit E8) is only used for the mic pres and a outboard volume control for the monitors (also DIY).  Interface is a M-Audio 2496.  I like the PCI card over a USB interface.

Now, on to my current project (about time he shut up ;D).  I'm looking to do a mashup of a couple schematics I have found.  Lucky, my budget and skills match..... both pretty poor.  I want to do a preamp like the T15 from Seventh Circle combined with a 3 channel para EQ.  Should be simple enough for my feeble skills.  Also, given the room I have to record in (a basement spare bedroom), anything better would be a waste, acoustically speaking.

But first, I want to get the PSU sorted.  I took the PSU design I built for the Soundcraft and massaged it a bit. It's a similar design to others I have seen here and elsewhere.  I laid it out in LTSpice and ran the sim.  It looks pretty good.  Below are the so-so schematic that I used for the sim and the sim results.  I say "so-so schematic" because the PNP and the NPN are generic and the is no switch I like in the LTSpice library.

Please, give it a look and let me know what you think.

THANKS!

On the graph, green is pre regulator +48, yellow is post regulator +48 out, red is +15 post reg and aqua is -15 post reg

 

[joaquins]

You should add a load to the PSU for the simulation, for the phantom 10mA per channel is quite well, so 48V/10mA 4k8Ω for one channel, so 4k8/Nº of channels.

Then the load for the other part should depend on the channel strips used, count the current on it and get the resistor, or a current source dropping that amount, should be good enough for a first approach and see how much noise you are having from the PS it self, with good decoupling at each opamp you should be pretty close to that, A better approx would be load it with a noise current source, so you know how it behaves in a more real situation, but maybe too much for a PS where you over estimate the caps, for not big money in a one time build, and call it a day.

Other than that should be pretty fine, if you want to be really puristic the sources has some resistance as the wiring of the transformer but not a big deal, 15V AC is low for a 16V PS probably, I know you have the peak voltage of 21V but ripple may be an issue if the current goes high, for really low current you can get away with it, I get nice 18V from a supply like yours for testing low current single stages, 16V stands pretty fine but still... I guess your AC sources are 21V peak, but I'm not sure about that, check it.

JS

 

[Giggity]

Thanks for the pointers, joaquins.  It will just be a 2 channel pre so I added 2k4Ω for a load on the phantom power.  Of course it dropped (down to about 46vdc.  So I changed VR3 to 9K and it was good.  I intend to use trim pots on VR1, 2, 3 to fine adjust the voltage.

On the +/-15 rails, at this point, I'm not sure of the load so, I simmed loads until it started to sag.  I got to a 150mA load (100Ω) before it started to get some ripple at startup.  That will allow 75mA per channel, per rail.  I have no idea how to calc the load of a circuit.  And given that I'm not sure what the final circuit will look like, all I can do is make a semi educated guess.

I derived the 21vac peak from the 15vac RMS and multiply by 1.41 and which gives 21.15vac which lowered to 21 as a nice even number.  Better to be lower than to high.

Here are the schematics I have tentatively based this on.  The pre is the T15 from Seventh Circle Audio http://www.seventhcircleaudio.com/ and  the parametric EQ is from Sabro Tone http://www.sabrotone.com/?p=578.  The para design it intended as a guitar pedal but, I figure with some work, it could be used in a preamp.  In know the opamps can handle about double the volts than what is shown in the schematic for the para EQ.  Besides that, I would like to change them to something better than the shown TL074.  Possibly something from THAT or a Burr Brown from Ti.  The schematics can be seen much better when viewed at their respective websites linked to above (particularly the one from Seventh Circle Audio).

 

[joaquins]

At the P48 I would look at the Q1 as the source of the problem of such unregulation, getting up VR3 is not a solution, is just a solution for that particular load, not good. I don't know what they tried to do, but a collector in the direct supply path in that way doesn't look good. In any case P48 regulation isn't critical, don't let it go over 48V even unloaded, it's ok 46V loaded than 50 unloaded, well, it's the same for this values, phantom power standard is ok up to 52V IIRC, but is barely OK 38V, not ok at all 58V. Leave it as it was or take out the transistors at the output and re regulate the voltage for +48, then you can add a ~220Ω resistor going to each mic input and a large cap, ~220µF taking in the + pin the other end of the 220R and both 6k8, then the - pin of the cap directly to the pin 1 of the XLR and that directly to the case.

About the ±15V I would put bigger caps at  C1 and C2, but still should be enough as it is for more than 150mA, rule of thumb is to use 1000µF per Amp, so should be good for about 0.5A, maybe the problem is in the transformer used, as I said before. In PS isn't always good to be lower, if it's higher will dissipate more power and you may miss that in your heat calculations, but not a problem here, still you'll need some heat sink. I don't know what are the diodes doing on the output, but I can live with them and will protect the output stage for some estrange abuse, not doing nothing the rest of the time. There are some nice calculators online or a really nice XLS going around with more things in mind that you probably have about that, it heated me at some considerations the first time I saw it and it really helped to calculate everything, from the transformer being used to the heat sink, and all the details in the middle also. 150mA seems a little low to be safe at your case, and if you have spare current in a PS for a couple more of channels you can do a slave unit to use it anyway. For C5, C6 and C13 of your PS I use higher values normally, the rest of the P48 seems fine, the ones at the ±15V would harm to be twice as much.

I'm not in my best moment right now to do any more for you today, I'm certainly not in the mood for any more math today since last 12h with a couple of guys from the university with something to deliver on monday, still half way to go on sunday for that. In other occasion I help you to finish this with all the numbers around, you could simply go for any already designed PS here but has a lot of fun and learn about making the numbers for it, at least once.

See you around.

JS

 

[Giggity]

Thanks again, joaquins.  Since the PSU is basically sound. I will go ahead and draw up the whole thing.  That will probably take most of the day tomorrow to get it right and looking presentable.  I will probably it in EAGLE software as I find it easier to use for drawing a schematic.  The down side is no simulation.

I'll also look into other PSU designs.  I like the single transformer with dual secondaries approach though.

 

[PRR]

> I have no idea how to calc the load of a circuit.

Use Google to find the data-sheets of the chips. Find the Supply Current, times the number of chips.

There's usually two numbers, Typ and Max. It would be conservative to use Max.

There will also be a Dynamic Current due to audio loads (internal and external). For TL072 and similar chips driving mostly >10K, you can usually ignore this, or round-up 25%-50%.

What my SPICE will NOT tell me: when things overheat. If I drop raw 22V to reg 15V at 150mA, in a naked TO-220 package, it will get HOT. A little more, it will shut-down. I can get near 1 Amp but that takes a big hunk of metal. Fortunately, the modern regs will shut-down politely without damage. (Back in the old days we MELTED stuff.)

For a simple studio I wonder if you are over-building. I've never fooled with soft-off Phantom, and think that's a lot of parts and sag for a frill. Nor would I spend a lot of time making a Balanced Output for a basement studio (depends how nasty the basement is); much good recording has been done on totally un-balanced gear. All that servo-work is impressive but a few BFCs will get the DC down to zero, and "all" listening paths have so many coupling-caps that a few more in the studio may do no harm. The output RF filter and catch diodes may be needed in high-pressure careless-assistant commercial studios (and to protect the kit-seller from idiot buyers) but are rarely essential, especially when the user has a soldering-iron. You have enough stuff there that I don't think you want the MAX1044 voltage-trickster... feed the EQ plain +/-15V as God and Burr-Brown intended.

 

[Giggity]

Yea....that is one thing you can over do without any ill effects...heatsinking.

So long as I don't get any pops when the P48 is turned on, I have no problem doing away with the soft start.

I was only going to do a unbalanced out.  After the polarity switch on the pre, I was only going to feed one phase or the other into the para EQ.  The cable runs are short and I have never had a problem with noise in the past so..........

I probably should have erased the MAX1044 part out of the para EQ.  It's sole purpose it to power it from a 9vdc battery so it can be used as a guitar pedal.

I had already downloaded the BB OPAxx34 datasheet.  I think this is the one I need to replace the TL074.  I just checked it's current draw and it's 5mA/amp.  So that will be 20mA right there for one channel.

I'll have my work cut out for me tomorrow drawing this thing up.

 

[joaquins]

I think you have a pretty good idea of the PS now, about replacing those TL07X, is not always a good thing, replacing a TL074 with NE5534 for example will make a bad choice if unity gain is used and no proper compensation, also the bias current is much higher so unbalance on input pins will make offset worse and noise will be higher in high impedance parts, so, replacing it is not always a good thing, if it was designed for TL074, and you want to change it, you also have to change a couple of things more, scale the impedance down is one, more current from PS is needed, stability gets worse, and it keeps going...

JS

 

[Giggity]

I did some more tweeking on the PSU and I believe it should be good to go.  I dumped the 48v soft start.  Changed some caps to higher values.  Ran a sim and it looks OK.  Using trimmers on the regulators, I can set it up to +/-15 and +48 with out a load and tune it once a load is on.  A 25VA transformer should be enough for both mic preamps.  Here's what I came up with.....

Getting better on the quality of the pics as well.  No more screen shots.  :

Still working on the schematic for the mic pre.  Will post when I have something.

 

[Speedskater]

Life can sometimes be simpler if you use a separate power transformer for the phantom voltage supply. In any case use a different 'ground' symbol for the phantom supply, as it doesn't go to the Main Audio Ground (MAG) point.

 

[joaquins]

It looks good, give it a shot, if not you only need to add more caps, I'd strongly recommend to use the protection diodes... just 6 more parts and a lot less of trouble.

Life can sometimes be simpler if you use a separate power transformer for the phantom voltage supply. In any case use a different 'ground' symbol for the phantom supply, as it doesn't go to the Main Audio Ground (MAG) point.

It could go without problem, just one decoupling stage for it, with a resistor at 48V and a cap to chassis close to the input... I don't think that another transformer is really needed, not a bad idea, but not necessary for 2 channel preamp.

JS

 

[Giggity]

I added the caps and the diodes.  I actually had the diodes in, on a previous version.  They don't hurt anything, and are there if needed.

Knowing that I am not going to have a balanced output, would either of these simple preamps be worth entertaining?  Instead of modding  the Seventh Circle pre shown above.  The first is from ESP http://sound.westhost.com/project66.htm and the second is one I found on a blog named "Paul in the Lab" http://www.paulinthelab.com/2012/04/balenced-to-line-microphone-preamp.html.  I know I haven't mentioned it but, part of this DIY will also include etching a PCB for it and I have never done that before.  I would like to keep it simpler the better.  Time to get away from perf and Veroboard.

Also, phantom power can be added to either one of these preamps, as I'm sure you are aware.

Just a guess but, I would assume that any of these would be better than the mic preamps I have been using that are in the Soundcraft Spirit E8. The bottom schematic is for the mic preamps in the Soundcraft for comparison.

FWIW......I am pulling the mic signal from the tip of the Insert jack on the Soundcraft.

But, if there is nothing to be gained by going to any of these mic preamps, I can just stick to the ones in the Soundcraft and just do the parametric EQ to put in between it and the DAW.

I hope I'm not being to much of a pest.  I hate asking all kinds of questions without being able to contribute much (hopefully I can later when I learn more)

 

[joaquins]

We all started like this in some way...

I'd go for the INA or something like that, there are other options for a pre in a simple IC which are good, easy and cheap... The one on ESP is much like behringer preamp, you can get it better with better components, of course, but just a reference, I think INA is better. For discrete input there are better known preamps, like green or Amek 2500, wich for example have opamps instead of the second transistor. IC preamps goes something like this inside the IC.

JS

 

[PRR]

> better than the mic preamps I have been using that are in the Soundcraft Spirit E8.

Both are very typical of older popular-price consoles.

The "better" is that you DIY-ed it so it's sure to sound better to you.

The INA or a THAT are a simple excellent preamps with the fussy low-hiss transistors already sorted for you.

 

[ricardo]

The INA or a THAT are a simple excellent preamps with the fussy low-hiss transistors already sorted for you.

.. and hardly ANY of the more complex DIY preamps will outperform the THAT & TI datasheet & app. note circuits.  Most (all?) are worse.

Download & study all the THAT Design Notes carefully.

 

[Giggity]

I had already downloaded the design note #140.  I just took another quick look and also got #109 &#138.  It looks like most of the design notes deal with VCAs and compressors.  I was also looking at the datasheet for the BB INA217 and saw that had a layout just a bit different than the one I posted above.

Seems like this is a all around better direction then where I was headed.  I think, for my purposes, I was making it more difficult than it needed to be.  Better to take a smaller step and understand what I am doing than go over board and it be over my head.

 

[Giggity]

Ok..... I have a new one drawn up that is similar to the INA217 one above.  After reading the THAT notes, I swapped out the INA217 for a THAT1510.  It's a drop-in replacement for the INA, according to the THAT notes.  Redrew the front end to what was suggested in the THAT notes and added the phantom parts as well as a 20dB pad (all per the notes and data sheet).  The only left to do is figure out the best way to get a phase switch incorporated into it.  I'm not sure if this can be done in front of the THAT1510 or after.  I did find a schematic on the ESP website for a phase switch for unbalanced signals that could be put in after the THAT1510, but not sure where.

Also unsure if the parametric EQ posted toward the top is on par, quality wise, with this preamp.  I'll post it again below.

 

[joaquins]

For the preamp, you need some small resistors in series with C4 and C5.
Mic input pin 1 should go to chassis as soon as possible (with a short wire I mean) not the same ground that the rest of the circuit.
Also a cap to chasis for the phantom power and a resistor from +48 to it, to reference those +48V to chasis and add another stage of filtering. 220Ω 220µF may be fine for that, maybe a lower freq filter is even better.

JS

 

[Giggity]

Jumping back and forth from THAT notes 138, 140 and the 1510 datasheet, I missed the resistors just after C4 & C5.  I also changed the ground for the XLR jack.

Here are the links to the THAT notes and datasheet for reference.

http://www.thatcorp.com/datashts/dn138.pdf
http://www.thatcorp.com/datashts/dn140.pdf
http://www.thatcorp.com/datashts/THAT_1510-1512_Datasheet.pdf

I'm lost on what you are saying about the phantom power.  I drew what the THAT notes showed and explained exactly.  They did show a couple other options but, none of those used a cap and were also for specific apps that didn't pertain.

Below are the current versions of the PSU and the MicPre schematics.  Since no one has mentioned anything about the parametric EQ (posted above), I will assume that it is the equal of the MicPre and ready to do.

 

[ricardo]

Since no one has mentioned anything about the parametric EQ (posted above), I will assume that it is the equal of the MicPre and ready to do.

A 1510/12 preamp is for 'clean & accurate'.  A parametric is to twiddle the sound so is NEVER 'clean & accurate' so they can't be compared for quality or anything else.

EQs are very personal cos they have a 'sound' unlike mike preamps which should NOT have a 'sound'.  If you like the sound, its the one for you.  The Sabratone forum has comments on the sound of this fairly standard Graphic EQ type circuit.

My only suggestion would be to add a Bypass switch so you can hear what difference your twiddle settings have made to the sound.  ie take the output from IC2 instead of IC1d.  But this parametric Inverts so you need to put another inverter after IC1d to make this sensible.

I don't like Phase switches at inputs to Mike Preamps.  ALWAYS better to minimize wiring at this sensitive end of the signal chain.  Can post a link to the ESP circuit?

I'll leave joaquins to explain about filtering your P48V