Discrete DI-Box

[NewYorkDave]

guitarists did not want distortion

That's the conventional wisdom nowdays. But I say, listen to some late '40s/early '50s Bob Wills and His Texas Playboys (Fender's first "celebrity endorsers") and see if you still believe that :wink:

Here's one guitar slinger from Wills's crew who liked to rock on overdrive:

Junior Barnard

 

[ulysses]

Sure, the pioneering pioneers adpated quickly and embraced the sound. But you also have to remember that these amplifier topologies, and their associated distortion, were around a decade or two before Leo wrapped them in tweed.
Also, I think it's important to remember that even though we may like the sound of the early distortion, that still doesn't mean it was intentional. (I'm thinking further back than Bob Wills here, like the distorted vocals on early jazz, blues and country recordings.)

 

[PRR]

> guitarists did not want distortion. Rock and Roll was invented ... 3rd harmonic ... only needed two fingers instead of three

There is some truth in that.

Many old-school players did not want the amp to go BLATTT when hit too hard. Of course that did become "the" sound. And a long long time ago. I just found an early Les Paul where he uses clipping for emphasis (not as his sole trick).

No, what I'm saying is a gentle rise with signal level, before you hit "max" and go to clipping.

There is musical justification. What is the difference between a yell, a shout, and a scream? There may be little difference in acoustic power. The difference is the singer (or yeller) pushes up the harmonic overtones. Common speech and mild singing drops above 800Hz, a good holler or operatic climax brings up the 2KHz band on the same nominal pitch.

Most orchestral instruments also have a rise of harmonics at high level. (In the extreme, a flute will "jump" to a harmonic at high drive level, which is both a way to extend the instrument's range and a royal pain if you want to play a low note loud.)

The string/pickup combination for a solid guitar also changes harmonic balance at high plucking level but in a different way. I suspect that a "good" set-up aims to minimize this, and let the amplifier add the scream.

Against this interpretation: some very soulful players use very souless amps. A lot of it is in the touch.

> listen to some late '40s/early '50s Bob Wills and His Texas Playboys

"Noel Boggs ...In 1946 he met Leo Fender while working with Spade Cooley at the Santa Monica Ballroom. He became the proud owner of Fender's first steel guitar and an important endorser and promoter of Fender's equipment. The friendship between Noel Boggs and Leo Fender was such that Leo Fender was the godfather of Noel's daughter Sandy."

 

[Samuel Groner]

Typically there is a negative input Z at some frequency. A small (few pF) cap to common is usually sufficient (gate to common that is).

Thanks, that's it. I need a mere 0.2 pF to make the response top flat in simulation. I guess it is save to assume that a practical implementation will not need this capacitor (at least if some close ground plane is used)?

Samuel

 

[Samuel Groner]

I updated the scheatic linked on the first page. Switched back to 2SK170 for lower noise, cascoded though this time. Reduced output stage quiescent current (possible due to the higher turns ratio of the new output transformer) now gives an ample 30 Vpp input headroom. PSRR is very good (an unrealistic 150 dB in simulation) above 10 Hz.

Samuel

 

[bovox]

Hey Samuel, nice DI-"ding".Couldn´t Q2 be a 5457 and Q6 a 550C with high beta? Cheers Bo

 

[bcarso]

[quote author="PRR"]

> listen to some late '40s/early '50s Bob Wills and His Texas Playboys

"Noel Boggs ...In 1946 he met Leo Fender while working with Spade Cooley at the Santa Monica Ballroom. He became the proud owner of Fender's first steel guitar and an important endorser and promoter of Fender's equipment. The friendship between Noel Boggs and Leo Fender was such that Leo Fender was the godfather of Noel's daughter Sandy."[/quote]

Good old Spade Cooley. I remember how shocked I was when he went up on the murder charge.

I'm not sure I had ever seen his picture before, but in the above he bears a eerie resemblance to Bill O'Reilly.

 

[Samuel Groner]

Couldn't Q2 be a 2N5457 and Q6 a NC550C with high beta?

The 2N5457 should basically work, though we are a bit close with respect to breakdown voltage and the higher Idss rating (2N5459) would be preffered for higher pinch-off voltage.

I realise that I have two Q6's in the schematic--the MPSA18 should read Q7. Which one are you referring to? The MPSA18 has very high beta (higher than the BC550C).

Samuel

 

[bovox]

Hej Samuel, I was referring to the Q7.
Cheers Bo

 

[Samuel Groner]

Q7 can be a BC550C, though I would set R10 = 120k and R11 = 1.1M then to accomodate the higher base current.

I update the schematic to correct the doubled Q6.

Samuel

 

[gnd]

What is the reason for changing output trafo to LL1576?

thnx

 

[Samuel Groner]

What is the reason for changing output trafo to LL1576?

Per Lundahl recommended the switch without giving further explanation. A look at the datasheet however suggests that the LL1576 will accept much higher levels which might be important for some applications.

Samuel

 

[Samuel Groner]

Just a quick update on this one: I ordered the parts I don't have at hand and hope to build this thing ASAP--it's at position 3 in the projects queue. The schematic got a little update with respect to overload protection.

Samuel

 

[clintrubber]

[quote author="Samuel Groner"]I updated the scheatic linked on the first page. Switched back to 2SK170 for lower noise, cascoded though this time. [/quote]
No criticism but just wondering about the addition of cascode Q2. Since output is taken from Q1-source there's no Millering of the Q1-capacitance so only a few tenths of pF will be around I expect, and likely even overruled by the cable capacitance + C3 (10pF).

Best regards,

Peter

 

[Samuel Groner]

It's not for reducing input capacity per se but rather to minimise distortion with high source impedances due to modulation of input capacity. Often referred to as (one form of) common mode distortion in opamp design. See two posts from Brad on the first page as well.

Samuel

 

[clintrubber]

I see, reducing cap not for loss of BW but for THD. Sounds :thumb:

 

[clintrubber]

[quote author="ruffrecords"]Personally I cannot see the point of the transformer, unless it is for a particular sound and it seems an expensive way of providing phantom power. It provides no isolation in the topology used and that's an important property of a DI box.[/quote]
I saw the Bo-DI does a bit to separate grounds, see the components on the right in the figure:
http://web.telia.com/~u31617586/Active DI box 1975.jpg

 

[recnsci]

[quote author="Samuel Groner"]It's not for reducing input capacity per se but rather to minimise distortion with high source impedances due to modulation of input capacity. Often referred to as (one form of) common mode distortion in opamp design. See two posts from Brad on the first page as well.

Samuel[/quote]

This is tricky part. First fets need to be matched, and if they are not things
could get worse than no cascode. Problem is that dCgd/dVgd gets higher for
smaller Vgd.

cheerz
urosh

 

[clintrubber]

[quote author="recnsci"]
[quote author="Samuel Groner"]
It's not for reducing input capacity per se but rather to minimise distortion with high source impedances due to modulation of input capacity. Often referred to as (one form of) common mode distortion in opamp design. See two posts from Brad on the first page as well.

Samuel [/quote]

This is tricky part. First fets need to be matched, and if they are not things
could get worse than no cascode. Problem is that dCgd/dVgd gets higher for
smaller Vgd.

cheerz
urosh [/quote]


FWIW, this circuit (posted by Jim Zuehsow) does the cascode-thing as well (and also extends the first stage to a CFP-like topology & generates balanced signals right away):

BTW & semi-OT, apart from eventual lesser headroom, any disadvantages for this 2*JFET+BJT-threesome ?

 

[bcarso]

[quote author="clintrubber"][quote author="Samuel Groner"]I updated the scheatic linked on the first page. Switched back to 2SK170 for lower noise, cascoded though this time. [/quote]
No criticism but just wondering about the addition of cascode Q2. Since output is taken from Q1-source there's no Millering of the Q1-capacitance so only a few tenths of pF will be around I expect, and likely even overruled by the cable capacitance + C3 (10pF).

Best regards,

Peter[/quote]

Note that we often think that Miller is gone with just tying the drain to the rail, and it is true in a way---but you still get Cgd (times 1). For the SK117 this is about 3pF @ 10V, twice that for the SK170. With a bootstrapped drain this goes away almost completely (up to a certain frequency, when things deteriorate switftly!). Note also that using a second FET as the part in cascade yields less that complete cancellation owing to its own finite source lead impedance, which is at least 1/gm (and larger with voltage swing at its drain). Still, something like a 2N4391 in cascade with a 170 is pretty good, and gives a decent drain voltage without fuss for the lower device.

Another consideration: look at the curves of gate leakage versus drain-gate voltage for the short-channel parts like the SK170. For most of our apps it remains small even approaching breakdown, but for very high-Z sources like condenser mic capsules with gigaohm resistors it becomes important. With active drain drive this can be held in check.