Anybody else prefer discrete semiconductors over opamps?

Frequently asked question on transistors: types, substitutions, how to test, use and misuse them.
User avatar
RnFR
Old Solderhand
Information
Posts: 4879
Joined: 09 Jan 2008, 06:02
my favorite amplifier: Traynor YBA-III, Fender Super Six
Completed builds: custom fuzz.
Location: Inner Earth
Has thanked: 139 times
Been thanked: 133 times
Contact:

Post by RnFR »

great answer! moved to FAQs.
"You've converted me to Cubic thinking. Where do I sign up for the newsletter? I need to learn more about how I can break free from ONEism Death Math." - Soulsonic

Blog-APOCALYPSE AUDIO

User avatar
JakeAC5253
Resistor Ronker
Information
Posts: 324
Joined: 26 Oct 2010, 00:59
my favorite amplifier: '81 JCM800 2203
Completed builds: FET-Opamp based OD
JFET clean boost/OD
Extreme Fuzz for doom
Has thanked: 4 times
Been thanked: 1 time

Post by JakeAC5253 »

marshmellow wrote:
JakeAC5253 wrote:Little bit of a different angle than the initial post. What do you prefer for pre-post gain buffering, JFETs, BJT Opamps, or JFET opamps? I am ruling out discrete BJTs because they are so common, and furthermore, I just don't like them as buffers. I almost think that from a design standpoint, you could more easily tweak the output impedance on the discrete JFET buffers than you could using an Opamp buffer, though strictly mathematically I think the JFET opamp would make the most efficient buffer because the output impedances are usually measured in hundreds of ohms. Of course efficiency doesn't always sound good...
For output buffering a BJT would be in most cases be preferable, because you will more easily achieve a lower output impedance than with a JFET, if this is what you aim for. Of course the preceding stage has to be taken into account as well for what might be the better choice. For any opamp the output impedance will be a fraction of an Ohm, so typically a resistor is used to isolate the output from a possibly capacitive load. Still, I don't see much merit for guitar circuits to optimize a circuit in that respect, an output impedance of a few kOhms is perfectly fine in that environment.
JakeAC5253 wrote:Discrete JFET buffers can be used at 9v, but with a JFET opamp like the TLO72 (a JFET input opamp that I like) the specs say that it needs to be powered closer or exceeding a single supply rated between 18v-24v. BJT opamp buffers are lower input impedance, so it wouldn't really be the ideal part for an electric guitar to see first in the chain, although they have a lower output impedance than even a JFET opamp, which may be better for the gain part of the circuit to see... I read a post from RG that said that using discrete JFET input/output buffers would increase the amount of noise and parasitic crosstalk between traces due to the high input impedances... how bad is this typically, and do JFET input opamps exhibit this as well?
Too many, not necessarily true, generalizations for my taste. You just can't look at one specific part without taking the closed loop that is surrounding it into account. As for noise, the specific circuit has to be analyzed to see which part will perform better.
JakeAC5253 wrote:Talking about discrete JFET buffers, does the buffer become closer to unity gain as the voltage is raised higher? What is the relationship there, because I know a 9v supply is not ideal in most, if not all, instances of the circuitry we often play with.
No, the supply voltage doesn't matter for output impedance (assuming it is reasonably high for the circuit to work at all). Simplified, the amplification of the source follower will be A = gs, so in order to minimize the loss a JFET with higher transconductance would be the better choice. But I don't really see the point for most circuits where you almost always have a preceding stage with gain that will easily compensate the small loss. So it won't really matter that much if your output buffer has an amplification of 1 or 0,9 or 0,8.


Thanks so much. These "JFET losses" you mentioned, is that due to the higher output z of the device? Or a different property? Incidentally, I have recently replaced a BJT input buffer with a JFET buffer having made no other changes, and thought that the circuit as a whole sounds a bit quieter now. I could be overthinking it.

Post Reply