Simple DI box schem available

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I think I'll pick up all the parts for this one from the surplus shop when I head back to my old town (Colorado Springs) this weekend. Hopefully, I can find the appropriate FETs or equivalents. I'm sure they'll have everything else.
 
[quote author="Consul"]Yes, it definitely looks like it's run directly from phantom power.

I'm also counting three caps in the signal path at least (I can't tell right away with C4). You would want decent film types for those, I assume?

Looks like an easy build. Unfortuantely, neither Mouser, Newark, nor Digikey seem to have part PN5457 (the FET) in their database. Is there an equivalent?[/quote]

I wonder if the PN5457 is the same as the 2N5457, a very available part. Although the Vds of that device is rated at 25V only. None of my cross references list PN5457 though.
 
DI-5457.gif


This is a fixed-bias grounded-Source FET amp. Gate bias is provided by jacking the Source up about 0.6V with the diode. Power is supplied directly by Phantom.

The FET: http://www.fairchildsemi.com/ds/2N/2N5459.pdf

If you look at the Typical curves on page 3, at 0.6V bias the FET will pass a nice 1.5mA. But the 2N5457 is specified for a range of Vgs(off) 0.5V to 6.0V. Biased at 0.6V, the Drain current for any specific 5457 could be anything from 0.4mA to 8mA.

The gain depends on the FET, and also on the mike amp input impedance. Mike amp input impedance tends to be around 2K differential, but I have seen under 1K and up to 12K (you can't go higher with standard Phantom power).

5457 Gm varies a lot with Drain current. Or actually with Vgs(off).... devices with different Vgs(off) have nearly the same Gm if bias is set to give the same current. That suggests a current-source would be better than a fixed bias voltage (but not as simple).

At 0.6V bias, various 5457s will give Gm of 0.2mmho to 4mmho, or effective Source resistance of 5K to 250 ohms.

Picking a 5457 with low Vgd(off), and a mike amp with low (1K) input impedance, voltage gain from gitar to one side of the XLR is 500/5K or about 0.1.

Picking a 5457 with high Vgd(off), and a mike amp with high (2K) input impedance, voltage gain from gitar to one side of the XLR is 1K/250 or about 4.

The average 5457 will give Id about 1.5mA, Gm about 3mmho, Drain resistance about 300 ohms. Gain into one pin of 1K or 2K inputs is about 1.6 to 3.

So it will work with the average 5457, but not with the extreme 5457. A very-low-Vgs(off) part will be very weak.

But I'm not sure it is "safe". Drain-source voltage breakdown rating for 5457 is 25V. Phantom is 48V (and possibly 52V). We can subtract 0.6V for the diode, whoop-ee. The actual Drain voltage will be less by the drain current flowing in the 6.8K Phantom resistor. For the typical 5457 this is 1.5mA in 6.8K or 10.2V. 48V-0.6V-10.2V= 37.2V across the FET. In real life, most "25V" parts will stand this for a very long time, but there will be some instant-deaths, and possible long-term failures. A low Vgs(off) part will not only have low gain, but could see over 45V in some cases.

If you build it, you might want to select FETs, and you might want to be prepared to replace C4 D1 with a low-voltage current source to stabilize voltages and gain.

Since it has gain (typically 3 to 6), hot guitars may overload mike inputs.

I also suspect it will overload on very hot guitar, though the overload action may be musically interesting.
 
Okay, it's going to take me a few passes to digest all of that. :shock: Thanks very much for another great post, though.

I think at this point what I'll do is take the schem and the text of your post back with me to the Springs, run it by a friend of mine there, and then we'll hit the surplus shop.

I don't know yet what I want to build this DI for, but I figure, since it's cheap and easy, it will be useful for something. If I can get it built while I'm still in the Springs, I can run a bass guitar through it at my brother's place and try it out. Personally, I'm probably more interested in running synths through it than anything.
 
Wow PRR!
You keep on impressing me :shock:
i hope if i stay around here long enough, i might to get to 1% of what you know about electronics, that would probably be enough to start designing and building gear for a living too! :green:

Thanks for sharing this with us :thumb:

I might need to reread this a few times to understand it deeper and as English is not my first language, it's a bit more complicated, still step by step i understand more everyday, electronics as well as English. :green:

Seems like you followed in the footsteps of your dad? Must be cool to learn and discover this way.
May i ask what and where you studied?

Cheers, :sam:

Tony
 
The PN5457 is the same as a 2N5457 but the schematic has the pin config wrong... it should be D-S-G as shown in the Fairchild datasheet

regards, Jack
 
However, with many JFETs anyway, the drain and source can be interchanged
Yes, that's true but I try to not rely on that since the channel is not always symmetrical and the characteristics may differ though it works reversed... for instance, gate-source capacitance could be different than gate-drain. In the J202 the Cgs is 3.7pf and the Cgd is 3.5pf so they are basically interchangeable but in the J111, Cgs is 14pf and Cgd is 9.5pf. Small but possibly important, so I like to try to keep them oriented correctly... has everyone dozed off yet? :grin:

regards, Jack

ps: I just looked up the 2N5457 and it is 4.6p and 4p so it is fairly symmetrical and interchangeable... but I'm obsessive so I'd still keep them straight :shock:
 
I proto'd this one over the weekend... WOW this thing has a lot of gain.. I mean i had to turn my preamp pots ALL the way down to keep from clipping... even then it had crazy gain and noise. I threw in a T pad and it calmed it down to a useable level but started to kill my high end response.. so I ask now.. what would be a suitable way to lessen the gain on the output of this circuit?
 
Thanks Gyraf, however i meant something like a trim pot or something that could tailor the gain to the length of cable.

thanks!
 
can something as simple as a trimpot between the two output pins work for this? 48v potential stays the same so would i need a cap between the pot and one side of the signal?

I assume that since these are JFETs that they are being used in a less than saturated region? Since I work mainly with MOSFETs under full saturation and my designs are mainly to swing the gates between rails as fast and hard as possible, I am used to the gain being a function of the voltage rail and not the FET itself. can someone point me to documentation or explain how the JFET determines the gain?

thanks.
 
Two ways I would suggest to reduce gain: Pad on front end, or try inserting series resistance on pins 2 and 3. This would reduce the voltage to the FETs and possibly make them live longer. This is going to increase your o/p impedance so be warned of that...

HTH!
Charlie
 
still wondering how a Jfet determines gain..? padding the front end of the DI only attenuates the guitar, but does not lessen the self noise of the circuit. still haven't tried tying a rheostat between the differential outputs (like a mic pre?) to lessen the outputs, but tying a resistor between them does indeed work. a T attenuator seems to lessen the hi end of the audio. attenuation of the output both lessens the noise to a usable level and also lessens the guitar to a usable level.

more tests to come when i get time this weekend.

:guinness:
 
1k rheostat across the outputs works well for attenuation. around 300R is where the best level/noise is.

cheers! :guinness:
 
I'm going to try a set of PN4392 FETs for my next set of DIs. I really like the sound of this circuit for some reason.. Hard to put my finger on what it does to the sound but it's almost like it's cleaner than just plugging straight into the console... ? I've changed some part values and added others too if anyone is interested.

:thumb:
 
Back
Top