Anyone tried DIY'ing cheap, clean mic-preamps?

S

SourIce

New member
I'm curious to see how feasible it is. It'd be alot cheaper (obviously) if I could spring for a better interface with more inputs by using simple, clean DIY pre's - I also don't mean many of the expensive DIY kits to build boutique or vintage clones.

I remember reading an article somewhere that said for alot of the preamps on the market the parts are quite cheap, most of the research is making up the R&D costs, but it was light on actual information and useful schematics. My soddering skils are fine and I can follow a good schematic, but I'm no electrical engineer. I hoped that someone might have tried this already and be able to point me to useful information.
 
haven't tried this one, but it doesn't get much simpler or cheaper:

K1803: VELLEMAN INC.: Education & Hobby Kits

40dB isn't a lot of gain though, so you'd have to make up some of that gain on your interface. For US$15 and low part count, it can't hurt. I have a project 'in the works' but it isn't ready for prime time yet, and is unproven, so I'll post it when I'm done. Isn't really simple though, and involves some kludging.

You could probably get a decent older board (Tascam M30?) and use those pre's and direct outs though, for less than the "boutiquey" preamps. Not sure where you are located, but your market should support.

A lot depends on your skill level, and desire to really get into the weeds. There are plenty of projects at GroupDIY, which vary in "boutiqeuness" and the need to source exotic parts especially transformers. "Simple" generally doesn't enter into the equation. "Cheap" is relative, depending on your interface, this works out to about US$30 per preamp give or take.

Behringer: ULTRAGAIN PRO-8 DIGITAL ADA8000
 
I actually can't believe I never thought about investing in an old board. Thats a great idea and probably alot more cost effective :cool:

Are there any other similar older models I should be on the look out for?
 
I'm only really familiar with Teac/Tascam, as a brand of choice, and most of them have fantastic routing options, so will have direct outs. Depending on the vintage you may also get transformer input pre's. The direct outs is the important part. Do a search in the analog forum here. You'll find a lot of recommendations, as there are quite a few people that bypass the busses and go straight to the tape deck.
 
White Strat recently did just that. I think it's the same model you mentioned; the M30. Something like 32 channels for $500!! :eek:
 
That Vellman kits Looks like it doesnt get much simpler. Thanks Jinn. Banked in the archive if ever i need it.
 
A few caveats and tips?

Excellent tho' many of the Vellmann kits are, that pre amp circuit is no good for mics.
It uses shunt feedback and thus a 100k series resistor. For ultra low noise we want ZERO resistance in the mic input circuit (and the lowest possible feedback resistors).

That Jim's Audio PCB looks interesting but I would like to see a schematic? Cannot see any discrete ULNoise transistors there and a hybrid circuit is needed for the lowest possible noise, even the otherwise excellent NE5534 cannot quite compete with selected discretes for noise.

The NE CAN be used however if you allow a transformer input. Folks like OEP/Walter here will furnish application notes and their transformers are very good value. Jensen too will I am sure have suitable circuits and their transformers are world class (if more expensive!)
Traff input circuits are very simple and virtually RF proof, a not inconsiderable problem with ironfree designs!

Also treat yourself to "Small Signal Audio Design" by Douglas Self..VERY little math and a wealth of good sense and circuits.

Power supplies: An oft neglected factor in home built electronics. Use AA carriers to produce + and - supplies. 12 + 12 volts is plenty to test circuits for noise and will just give a slight headroom limitation. When you come to a mains supply do not exceed 2x 17volts, Duggy S tells you why and how!

Dave.
 
Excellent tho' many of the Vellmann kits are, that pre amp circuit is no good for mics.
It uses shunt feedback and thus a 100k series resistor. For ultra low noise we want ZERO resistance in the mic input circuit (and the lowest possible feedback resistors).
Click to expand...

You sir, are a steely eyed missile man!

I'm not up on those things, and always appreciate the "tips and caveats."

ecc83 said:
A few caveats and tips?

That Jim's Audio PCB looks interesting but I would like to see a schematic?

The NE CAN be used however if you allow a transformer input. Folks like OEP/Walter here will furnish application notes and their transformers are very good value. Jensen too will I am sure have suitable circuits and their transformers are world class (if more expensive!)
Traff input circuits are very simple and virtually RF proof, a not inconsiderable problem with ironfree designs!

Also treat yourself to "Small Signal Audio Design" by Douglas Self..VERY little math and a wealth of good sense and circuits.

Power supplies: An oft neglected factor in home built electronics. Use AA carriers to produce + and - supplies. 12 + 12 volts is plenty to test circuits for noise and will just give a slight headroom limitation. When you come to a mains supply do not exceed 2x 17volts, Duggy S tells you why and how!

Dave.
Click to expand...



The schemo for the Jim's is in the ad, just scroll down a little bit.

As for trafos how critical is the mu metal shield, and can the lack be compensated? Edcor has some nice inexpensive options, although I think the highest turns ratio is 1:8.


I have a discrete thing I'm working on (basically a kludge of two other circuits) but I want to try it out first and then post for criticism.

I haven't yet taken your advice and upgraded my library... :-(
 
The Jim's Audio boards use a more or less bog standard reference design from the INA217 datasheet. The layout is sufficiently generic that you can substitute other amplifiers with few or no other component changes. (The basic schematic is shown on the ebay page, and the seller emails a board schematic and a BOM after purchase.)

For me, at least, one attraction of these boards is that the output stage is left to the constructor, with plenty of options available.

Examples:

Impedance balanced using 2 resistors (as per Bill Whitlock's paper at Jensen).
Transformer balanced - could use a big discrete op amp like a Hard 990C or a 2520 to drive a transformer.
High performance differential line driver like the THAT 1646
Differential output using a dual monolithic opamp.

Power supply could by handled, for example, by one of Joe Malone's JLM Audio linear power supply kits which provide user selectable + and - rails and +48V phantom.

I could envisage racking these up for a pristine 8 channel preamp, or putting 2 into a small enclosure for a portable preamp.
 
Ah!
Found the schematic, yes it is just lifted from the TI data sheet and I am sure it is a good design but I have some reservations?

No RF proofing is shown. You need at least 100R in each leg and 100puff to pin 1 close to the XLR. Better would be a dual wound toroid. Phone "chirrup" is very hard to keep out.

Self does not refer to the INA217 specifically but is not a fan of "instrumentation" amplifiers as they rarely give the noise performance promised, at least not at usable gains (work best at ~60dB+). The chip is also very expensive compared to the NE5534 at around $5.00 a pop.

Edcor traffs? Can't find a microphone type? As for ratios well, the bigger the stepup the better the noise performance of course but past about 1:3 performance suffers unless you can pay a lot of money. But 1:3 is a 10dB lift and that is better (iirc) than the benefit of discrete transistors over the 5534 which is about 4dB.

'Sall good fun!

Dave.
 
Have a look at the data-sheet for the THAT 1512 for an alternative circuit - with RFI management.

With no disrespect for Doug Self, I might just note that preamp chips have found their way into a number of highly regarded products. The DAV Broadhurst Gardens 1, for example.
 
PRHunt said:
Have a look at the data-sheet for the THAT 1512 for an alternative circuit - with RFI management.

With no disrespect for Doug Self, I might just note that preamp chips have found their way into a number of highly regarded products. The DAV Broadhurst Gardens 1, for example.
Click to expand...

None taken (on his behalf!). The BG1 gets a very good write up in SoS 2003 but it is something of a Decca "clone" and the review had this comment about noise " The noise figure is given as less than 25µV, which equates to better than -110dB relative to a +20dBu output level (specified with a 10Hz-18kHz measurement bandwidth, 150(omega) source impedance and 32dB of gain in use) — more than adequate for most purposes."

Note, "more than adequate" not, "phenomenally low"? Then, at well under £500 for two channels (then) it is stonking value. Self is an engineer all the way and thus has to be aware of costs. This is especially true in mixer design where using $20 chips in bucketloads would price you out of the market....Stooodio people do NOT pay Russ Andrews prices! The balanced feedback discrete transistor design achieves very low noise at low cost.

I have a couple of THAT 1512s somewhere. I shall dig them out and offer them free to first shout,first get. I shall not, for personal reasons get into much DIYing for some time to come.

When/if I do I shall be looking at the LM4562 chip, the only IC of reasonable cost that performs better in almost all departments than the NE5532. These will be coupled with OEP traffs as I want a a pair of low noise pre amps for wildlife work. (sounds rather grand and "BBC" don't it? It ain't, it's my backyard!)

Dave.
 
ecc83 said:
The balanced feedback discrete transistor design achieves very low noise at low cost.
Click to expand...

Is the design you mention the Graham Cohen double balanced topology, or one of Self's designs?

I really want to get a copy of Self's book "Small Signal Audio Design". Tossing around whether to get a paper edition or e-book.

When/if I do I shall be looking at the LM4562 chip, the only IC of reasonable cost that performs better in almost all departments than the NE5532. These will be coupled with OEP traffs as I want a a pair of low noise pre amps for wildlife work. (sounds rather grand and "BBC" don't it? It ain't, it's my backyard!)
Click to expand...

Hmmm. Input and ouput traffs?
 
PRHunt said:
Is the design you mention the Graham Cohen double balanced topology, or one of Self's designs?

I really want to get a copy of Self's book "Small Signal Audio Design". Tossing around whether to get a paper edition or e-book.



Hmmm. Input and ouput traffs?
Click to expand...

Not seen the Cohen schematic so cannot comment but I suspect they are both derivatives of the "classic" topology that has been around for years?

The problem with the simple design is that there is no NFB to the transistors and therefore distortion gets unacceptably high at high signal inputs. Self solves this by introducing a second IC stage (the humble TL072) and applies feedback to both emitters.

Output traffs? Hmm, not thought.
The output is actually going to drive a "balun" that puts the signal onto a CAT5e cable along with the video signal from a camera so there really is no need of a transformer. I do however have a couple of OEP Z1604s kicking about so WTH not? I also have the lesser specc'ed OEP 1200 which would actually be fine since the bandwidth is stopped at about 200Hz and MOL will be well under a volt.

I am presently using a Behringer X802 for the task and it is well up for it but obviously a bit of a waste when all I need is a pair of mic amps. The mics are cheap ass "karioke" types (well, one is a quite decent 20quid Maplin cardiod) modded to XLR balanced and wrapped in cling film. They seem to last about 6months depending upon the weather!

Get the paper version! I keep mine handy en route to the bog!
Dave.
 
I think that the Cohen topology is somewhat different. Here's some light reading: Graeme Cohen's Page

"Classic" circuits that I think of (for example, Jensen white papers) unbalance the signal with a transformer, amplify it with one or two stages, then re-balance for output. Which is not to say that's a bad thing. John Hardy's preamps are almost universally loved, and these are based on the "simple" Jensen circuits. Conceived, of course, with infinite attention to detail, hand selected parts, and using what are probably the most expensive and most non-distorting transformers available.

And Dave, I do heartily agree about the need for good reading material for visiting the throne!

Paul
 
PRHunt said:
I think that the Cohen topology is somewhat different. Here's some light reading: Graeme Cohen's Page

"Classic" circuits that I think of (for example, Jensen white papers) unbalance the signal with a transformer, amplify it with one or two stages, then re-balance for output. Which is not to say that's a bad thing. John Hardy's preamps are almost universally loved, and these are based on the "simple" Jensen circuits. Conceived, of course, with infinite attention to detail, hand selected parts, and using what are probably the most expensive and most non-distorting transformers available.

And Dave, I do heartily agree about the need for good reading material for visiting the throne!

Paul
Click to expand...
Heh!
Have be quick, got to visit wifey in hosp v soon.

That balanced mic amp is essentially the same as Self's in that it uses emitter degeneration but it is in fact an instrumentation amp topology and DS is, as I say not keen on them for audio. Great claims for CMRRs must be taken with a shovel full of salt since OP amps are not the limiting factor here, matching resistors is and we are talking better than 0.1% to even approach the CMMR of the chips themselves.

However I shall concede that Mr Cohen's work is at the cutting edge of what is possible, cost little object Whereas Duggy gives excellent designs at reasonable cost and component availability. Mind you! He does give circuit for a balanced input, unity gain line in amp that uses no less than 12 sections of NE5532 PER channel! This give THE lowest noise possible. (Well, unless you went to 24 sections!).

See yous soon.
Dave.
 
Been looking at Self again and the AD797 is about the lowest noise IC around being some 10dB quieter than the already very good NE5534. It costs 10 times as much however being £8.18 +vat here from RS.

If that noise improvement can be realized in a practical circuit it would beat the best of the hybrid transistor/IC designs it seems? I might jeeeeust push the boat out for my birdy amps!

The THAT chips are very clever and there is at least one commercial pre amp on the market using them but their main claim to fame is I understand their very high CMRR? Not of great interest to the home studio bod unless he is plagued with very high interfering fields.

Oh! The wife Paul! Thanks, yes she was very much brighter today, seems she is on the mend after some very tough abdominal surgery.
Dave.
 
You must log in or register to reply here.
Back
Top