Comp Diodes: 1N914 vs 1N4148

[Steve Mavronis]

I'm a stickler for trying to get exact as possible part value specs when cloning a pedal. Well I'm working on a Dyna Comp clone and the specs call for three 1N914 diodes. Right now I only have 1N4148 diodes and today and Radio Shack their packaging was labeled 1N914/1N4148 but they were all the 4148's like they are interchangable. I'm reading the specs are 'kinda' similar. Should I wait to find some 1N914 diodes to stay on my Comp spec layout or just use the diodes I have? The only other part variation is a silly 0.047uF cap that should really be 0.050uF, but that's within 5% tolerance so I'm fine with that.

[Scruffie]

Use them, they share the same specs.

[Steve Mavronis]

Yeah the more I read about this I see they are functionally equivalent and should be fine for this pedal clone application. A lot of tech sheets say they are direct swaps. Someone correct me if that is wrong. When I update my as-built schematic and PCB layout I'll make note of the original part spec change just for reference.

Found this article comparing several common diodes: http://www.cliftonlaboratories.com/diode_vf_vs_if.htm

The plot below shows measured and LTspice predictions for these four diodes. I do not have a SPICE model for the 1N270 diode, so I instead used a similar 1N34A Germanium point-contact diode in the simulation. Also, the 1N4148 and 1N914 SPICE models are essentially identical, so their plots overlay each other and only one is visible.



Analysis

1N914/1N4148—It's easy to dispose of the 1N914/4148 diodes. The agreement between SPICE, measurement and the datasheet is quite good. We can look at a three data points, 0.1, 1 and 10 mA, for example. Except at 10 mA, the measured data matches the datasheet as close as I can read it. The SPICE simulation runs 4-7% lower than the measure values, which is likely close enough for almost all simulations.
  Forward Voltage (millivolts)

Forward - Current Measured - SPICE - Datasheet
0.1mA 500 480 500
1mA 620 585 620
10mA 746 694 730

[CynicalMan]

That graph does show a difference, but it's rare in guitar effects for signal diodes to pass more than 1mA of current. Also, I'd suspect that the differences between the diodes are either just individual differences, or from the diodes being from different manufacturers.

[Steve Mavronis]

Yep, looks like the 1N4148 is the recommended replacement for the 1N914 so I'm good to go!

[brett]

Hi
I'm just guessing, but it looks like the 914 vs 4148 differences in the graphs are due to part-to-part variation, not model-to-model variation.
Individual SS diodes are likely to be all over the place when run near a deadly maximum current (ie very hot, very active junction).  Things like ambient temperature and air circulation in the room could be affecting the results quite a lot.
IMO anyone exploring 914 vs 4148 needs to get a girlfriend/boyfriend/better hobby/life.
cheers

[amptramp]

I did see a difference in one databook showing the 1N914 having 4 pF junction capacitance and the 1N4148 having 2 pF junction capacitance but not all manufacturers had the same values.  Other than that, they are functionally identical.

[Mark Hammer]

My sense is that, apart from component-specific differences in forward voltage, differences between types of diodes with the same family tend to be relevant only to things occurring well past the spectrum of relevance to musical instruments.  So, to use an example, a junction capacitance of 2pf vs 4pf probably does make a difference if one is using them for switching at very high speed.  But for signals under 15khz (and generally under 8-10khz), such differences are moot.

Remember, we use those components, but the manufacturers didn't make them specially for us.

[amptramp]

When you consider a tube screamer has 51 pF across the anti-parallel diodes, the 2 pF and 4 pF diode versions make no difference that would be noticeable even to a "golden ears" listener with a full range of hearing.

[Mark Hammer]

As I seem to repeat here several times a year, there are things that legitimately "matter" in a substantive way in certain contexts, but not in others.  The mistake made by many is to mistakenly assume that because something matters in context X, that it also matters equally in context Y, Z, and others.

Of course, if something doesn't really matter all that much, making the "wrong" choice has little impact.  But why worry about something that doesn't require it, right?  And, as researchers who study expertise will tell you, much of the difference between an expert and a novice, when problem-solving, lies in experts being able to ignore all those things they know don't matter, which would take up a great deal more of the novice's attention and time.