Some ideas for a clean balanced line driver?

[jasperoosthoek]

Hi Guys!

I haven't been posting here for quite a while as I haven't been building anything homegrown but only using them ;-). Here's the evidence, some shameless self promotion ahead:http://www.youtube.com/watch?v=GigKfcOsZdw Well, that's until now as I want to record my band live and I'm in dire need of some new stuff. All DIY of course  .

What I want to do is to record my guitar directly before the effects] and of course record the amp with an SM57. The idea is to get an idea of how my guitar sounds live and be able to re-amp the signal and play around with it. I don't have my amp at home and it's way to loud so there's no use tweaking around here. Also I want to collect material to make a live CD or get better sound quality on youtube videos.

So my first project will be a combination of a clean booster and a balanced line driver in one stomp box. The clean booster drives the effects the line driver connects to the audio interface. The clean booster is a Newtone FlowerBooster: http://www.newtone-online.nl/downloads/NewtoneFlowerBooster.pdf

The line driver uses a 600 ohm to 600 ohm transformer. In the schematic it is a cheap Xicon transformer. In reality I'll be using a vintage Beyerdynamic that I bought of ebay with the same impedance ratio and an excellent frequency range. The transformer is there to eliminate ground loops. Maybe it doesn't need to be balanced but if you hook up a normal guitar cable it works line an unbalanced connection anyway. And the free inputs on my interface are balanced so I'd better make use of it.

So my real question is this: Now I'm using a bipolar transistor DC biased to the mosfets drain as the biassing is very easy. Would it be better if I use another BS170 mosfet or a jfet to drive the transformer? Headroom and noise are my biggest concerns. Basically I would like it to be studio quality :-D

Would it be a good idea to use trimpot to set the bias for a fet?

Any ideas and design suggestions would be greatly appreciated!

[R.G.]

First, what you have shown there will not be clean. The Newtone Flower Booster is a copy of the Zvex MOSFET booster, and it has the typical distortion of a no-feedback single device. Likewise, a single emitter follower will not be particularly clean on its own.

It is really hard to beat an opamp used below its clipping points for a clean driver, especially if you have ambitions of "studio quality". An LM833 is specified for driving a 600 ohm load cleanly on +/-15V power supplies. This is going to be much better for your stated purposes than what you have shown. An NE5532 opamp driving the transformer from +/-15V power supplies would also be another good way to go. Studios love the NE5532. And the big power supplies would give you more clipping room to keep it clean.

So my answer to your real question is "don't do it that way."

[rockhorst]

Have a look at the buffer/booster stuff at AMZ, different components all togetjer, but would get the job done I think (and could be used for reamping as well). link

[ubersam]

Here's my idea: build a buffer/splitter like the AMZ Jfet buffer splitter but with a dual opamp instead of jfets, use an LT1054 to get your +/-15v. One buffer feeds your pedals, the other goes to the x-former.

[head_spaz]

Texas Instrument's DRV134 is hard to beat.
It's a single chip solution with performance
specs comparable to a good transformer.

[Mike Burgundy]

Also good stuff:
http://sound.westhost.com/project51.htm
http://sound.westhost.com/project87.htm

[jasperoosthoek]

Thanks for all the suggestions, you guys are great!

So, I've spent or maybe waisted an hour today to build the Newtone Flower Booster (without the line driver). And indeed as R.G. pointed out, it's not clean at high gain settings. With the gain turned up the voltage on the drain is only 0.8 volts, with the gain turned down it is ~6 volts. Turning the knob changes the bias and the charging of Q2 upsets the output. You get a scratchy sound and the sound momentarily dies out. So maybe I'll go for the AMZ Mosfet Booster instead which doesn't have this issue: http://www.muzique.com/schem/mosfet.htm First I'll test the Flower Booster on stage in a few days to see if it's a keeper.

Comparing the suggested opamps the NE5532 and LM833 are low noise but need a higher power supply voltage and therefore a different power supply. The DRV134 works on 9 volts (nice!) but has 20dB more noise that it's competitors. And altough it can drive a balanced line, it doesn't act as a ground lift (afaik) so I'd still have to use a transformer to connect it to the interface. And what happens if I connect it to a mic input with phantom power? Does it fry the chip?

I'll have to think if I want to build a separate +/- power supply for the line driver or stick with something which works on 9 volts. Higher voltage makes things easier I've learnt today... Maybe I'll put them in separate box: the buffer/splitter in between the effects and the line driver somewhere tucked away. So many ideas, so little time .

So for the line driver, it's going to be an opamp  . I don't really think headroom will be a big issue as all my effects run on 9 volts. Do you guys have any suggestions for a low noise opamp (~5 nV/sqrt(Hz) noise) capable of driving 600 ohms and working on 9 volts supply?

[asatbluesboy]

Is an OPA2134 overkill?

[PRR]

> need a higher power supply voltage and therefore a different power supply.
> I would like it to be studio quality :-D

What is "studio quality"?

Some pro gear can put out more power than a Champ. (Broadcast studios sometimes ran +30dBm on outside lines using 5W-10W amps for cleanliness.)

9V supply with normal circuits can't make over 3V which into 600 ohms is 'only' +12dBm. "Studio" gear outputs will do 7V unbalanced and 12V balanced.

OTOH _any_ studio line input should be happy with -10dBV into 10K ohms. Part of the purpose of a studio system is to take assorted inputs of various strengths and deliver a uniform program level. That's what half the knobs do!

> what happens if I connect it to a mic input with phantom power?

If you have mike inputs, build a Direct Box. That's what they are _for_. Don't re-invent wheels. You are not the first to try this. Study, copy, plagiarize.... and see what Beringer offers.

[liquids]

Is an OPA2134 overkill?

Not if you need a high input impedance such as jfet input op amps offer.
Otherwise, stick with the others - lower noise if you don't need the high input impedance.



RG - you mention the LM833.  I've had no issues with them and bought some at times because you seemed to admire them.  But others (PRR?  was it you?) have said they're a lesser-quality or more 'primitive' op amp than the NE5532 for such a situation...the TL072 vs LF353 comes to mind (I like both, the 072 is sometimes  cheaper, the 353 is more 'primitive,' I do like how it clips whencompared to other op amps... for example...not that I do that beyond the breadboard.  

Anyhow....LM833 & NE5532 ...discuss?  

[PRR]

> (PRR?  was it you?) have said they're a lesser-quality

I doubt I ever said that.

I probably pointed out that they are less popular than the venerable 5532. You should know that Link Wray was less popular than Frank Sinatra.... so what does that prove.

Either is more-capable in 600 ohms (if you have any 600 around) than the otherwise fine TL072.

Either or any is a more-sensible design than a couple switching devices working naked on low voltage; i.e. "don't do it that way." -R.G.

> (~5 nV/sqrt(Hz) noise)

Why do you need that number? 0.4 microVolts across 7KHz? The guitar itself is likely over 2uV.

[Mike Burgundy]

It always makes me smile when someone knows Link. Thanks Paul.
FWIW: a great many, MANY studio desks of great renown use both the 072 and the 5532.
072 where high input Z is wanted, 5532 anywhere where it's possible to drive it with low enough impedance and low noise is wanted.
These include SSL, DDA, Neve and D&R.
Doesn't mean there isn't anything better out there, but it does tell you that with careful designed circuits these chips are definitely good enough for the higher end studio stuff.

[merlinb]

Anyhow....LM833 & NE5532 ...discuss?  

AFAIK the LM833 and NE5532 are identical- same part. Am I wrong?

[jasperoosthoek]

Thanks guys for all the suggestions and discussion so far!

I'll post a new schematic in a few days when I made up my what to do. But it looks like I'm going to use mosfet buffer with a fixed gain of one working on 15 volts and a NE5532 working op +/- 15 volts to drive the transformer. I'll make two identical units, one on the input of my board an one on the output. Far away from stomping feet.

If I'm going to keep the FlowerBooster it will be as a separate unit in my effect loop, not in the box with the line drivers.

[liquids]

PRR - the comment I was recalling was more even handed than I recalled. It did seem to suggest that the 833 is a 'why bother' unless you have them on hand, as compared to a 5532, as 'the same or better.'
   
At the time of the post, I had been 'riding' comments by RG about how he liked the LM833 and that it was designed for audio, which seemed like something good in principle.  But again, recolection adds interpretation to (mis)perception. 

Clearly, both kinds of chips are potentially good, especially for lowest added self-noise, and potentially good fidelity (as is sometimes desirable), when they 'see' a source signal that is low output impedance.  Likewise, they are good in that they may be found cheap by the dozen - well, that is, good if cheap isn't an inherent indication of inferiority, as it is to some   

And yes, of course a passive guitar pickup is not a good source signal for these chips - unless, well, you like that that sound...and some do.  A buffered and/or clipped, low output impedance audio or guitar signal is the 'better' place to utilize this kind of op amp...if you are not trying to use the characteristics of these op amp (particularly the lower input impedance as compared to jfet input types) as yielding a low-fi, some say 'amp-like' audio effect as many chip-swappers do, without realizing why....such a chip 'seeing' a passive guitar pickup & volume knob, due to the low input impedance, makes some chip-swappers think these (and other similar kinds of) op amps are the ones to use for buffering, because the low input impedance offers an pleasing way to avoid 'too much high end' yielded by jfet input op amp buffers and it "keeps that amp like response when you dial back the guitar's volume knob," but it still is a buffer...once the signal is at the output.  The 4558 is somewhere between the jfet input type op amp and these chips for input impedance spec, and for the low-fi buffer effect, it can be a happier medium for those who want less high end and gain preserved than a jfet buffer offers but more than these chips offer.

But yeah...for guitar....even a low output impedance, clipped signal (lots of harmonics) doesn't often benefit noticably from these chips over a jfet input op amp...most of the time.   The ability to drive 600 ohm loads is really it's best feature for a pedalboard usage...and that is an infrequent need.  But a 5532 does drive headphones well 'enough' for my own personal headphone amp, for when I am traveling and on vacations....and it is cheap...keep meaning to pop something more pricey in there as originally intended, but it does fine without s more expensive, 'superior' chip driving the headphones.

But anyhow, heres the post I was recalling, if  recall correctly 

(note that my comment about prefering the 'jfet input lm833' in the 'ideal op amp' thread was followed by a '... oops  ' - twas a mind-bender/joke)

http://www.diystompboxes.com/smfforum/index.php?topic=81472.msg679033#msg679033

[R.G.]

AFAIK the LM833 and NE5532 are identical- same part. Am I wrong?

They are different. The NE5532 is the old-standard darling of the audio world, having lower distortion and higher output drive than its contemporaries. However, it suffers from a remarkably low input impedance as compared to later designed opamps. The LM833 was designed specifically for audio use, instead of DC accuracy, or some of the other things that opamps are designed for. It has a high input impedance and residual distortion as low or lower than the 5532 (or 5534) depending on the frequency, gain, test conditions, etc. 

Walt Jung's "Audio IC Op-Amp Applications" has some good comparisons with other devices to the 553x. It's a bit disappointing that the LM833 is mentioned, but not thoroughly compared. However, the 553x and the TL08x, the OP-27/37 family, and others get detailed work.

In actual use, I've always found the LM833, NE553x, and OP-27/37 family to be audibly equivalent unless there were issues with input impedance or bias. The performance is closer than my ears can distinguish.

As an aside @jasper, the book I mentioned has a design for a universal line amp, using one of a number of the opamps mentioned, Class AB bipolar current boosters, and a transformer-isolated output. This comes closer to what a studio would need.

The JRE-990 discrete design also has its ardent supporters, and is admirably DIY. You can make this thing up for huge power supplies if you want to tinker. +/-30V is reasonable with cheap transistors, and should be enough headroom for most line driver uses. 

[PRR]

> AFAIK the LM833 and NE5532 are identical- same part.

'5532 is NPN input.
http://www.ti.com/lit/ds/symlink/ne5532.pdf  page 2

LM833 is PNP input... in my printed copy (see below). The current PDF omits an internal schematic.
http://www.ti.com/lit/ds/symlink/lm833.pdf  
I don't believe the printed LM833 schematic is as detailed as the '5532 schematic, but it is clearly simpler and fundamentally different.



The '5532 topology is convoluted. The 2nd stage is differential. There's one more stage for pull-up than pull-down. Three comp-caps.

The LM833 is much more orthodox. Almost looks like a '741, without the stacked input devices, but with a more elaborate AB-bias string.

Skimming the published specs: not identical, not more than a hand-span apart.

(One occasionally major diff: the '5532 has always had a 44V limit, LM833 like other "40V process" parts is specced 36V max. Some pro gear does run '5532 above +/-15V or +/-18V rails. LM833 is not rated to take such use.

I am not sure, but suspect: some audio geek inside National saw that the '5532 "was getting old", saw that the new/improved processes allowed the same results with a simpler design, and got the project through development but _didn't_ sell-through to customers. LM833 is not majorly better, most '5532 buyers could afford a few extra cents (presently $0.19 vs. $0.55), and the audio market is fragmented and often sticks with what works.

That's what it was. I note that now both chips are fronted by TI. It is remotely possible they could merge the two products on one mask.

[jasperoosthoek]

Hi guys!

Just to freshen up your minds: I'm building a buffer and line driver that connects my guitar to a line input.

I've been very busy since my last post. I've built a passive 10 channel mic/line level splitter for recording all the mics and line out of the bass amp of my band. I put it in a chunky flightcase together with a 14 channel audio interface (Tascam US1641) and a laptop. I tested it last sunday during a gig and it works just fine.

So now I can return to this project as it connects directly to the same interface. The available inputs of the interface have a switch to select either -10dB or 4dB, no other volume control. Therefore I put a volume control in the driver circuit. Also I like to keep the voltage requirements simple and use a single supply. In the schematic below I used 15 volts but it could also be 20 volts. Therefore I decided to use the NE5534 instead of the NE5532 as it is happy working on voltages as low as 6 volts.

Any suggestions would be appreciated!

[PRR]

'5534 needs external compensation for low gains (as when your POT1 is dialed to zero).

'5532 is really the same thing, will be happy on 15V, except the compensation is built-in (not enuff pins on the DIP).

Do you really need 5 Meg input?

What is R3 for?

[markeebee]

@Jasper - sorry, a bit OT - what have you found to be the best software for your setup?