Here is a link to the results of adding the LF01 to my top MyRef build.
http://www.diyaudio.com/forums/chip-amps/54571-my-audiophile-lm3886-approach-340.html#post2835237
http://www.diyaudio.com/forums/chip-amps/54571-my-audiophile-lm3886-approach-340.html#post2835237
Hi Siva,
after buiding with great success the MyRefC i would like to test on a perfboard your very well reviewed LF-01 module. Can you tell me wich is the last BOM for this tiny circuit or if you changed something to the BOm from the last version you published ? Thanks !
You may use the same BoM listed here:
http://www.diyaudio.com/forums/anal...discrete-hybrid-opamp-module.html#post2656231
The experimental values are:
R1: 330R..1k, depending on how much current you'd like to source from the LM318 to bias it into Class-A. About 470R seems to work fine.
R2, R3: 10R, 4.7R or 22R, 6.8R. Both pairs seem to be about the same sonically.
The main thing to look out for on a perfboard is to bypass the LM318 fairly close to the pins without excessive trace lengths. 6.8 to 10nF, preferably a high-quality film/foil like a polystyrene, FKP2, FKS2, FKC2 or similar gives the best results, though you can manage with an MKP2 or C0G ceramic in a pinch.
Vaguely related: You may also want to check out my LF03 dual fully-discrete opamp module in the Analog Line Level section. I'll give an update shortly, but prototype #2 already seems to be sonically comparable or better than a pair of OPA627AU SOIC8 opamps mounted on a dual SOIC-8 to DIP-8 adapter, when used as the I/V converter in a Marantz CD-4000.
Pardon ,a noob Here .
Can we tweak R4 and get rid of that diode 4148 with something exotic like Bs170 fet ? (even I don't like complexity but curious about possibilities)
Can I expect some rough values of specs like slew rate , input noise voltage .
Wish if I could tweak it as amp for my hd448 headphone (maybe i am going too far).
Can we tweak R4 and get rid of that diode 4148 with something exotic like Bs170 fet ? (even I don't like complexity but curious about possibilities)
Can I expect some rough values of specs like slew rate , input noise voltage .
Wish if I could tweak it as amp for my hd448 headphone (maybe i am going too far).
That diode is needed to (loosely) mirror the Vbe drop in the PNP output device. There are other possibilities, but they tend to complicate the design or reduce bias stability. One possibility is an NPN transistor which is driven (with a suitable level shifter) by the opamp. I haven't exhausted all possibilities in simulation (obviously), but went with the simplest that gave good Class-A sonics.
Slew rate: hard to estimate - it's constrained by the opamp (for slow-enough opamps) and by the output stage (for fast-enough opamps).
Noise: mostly defined by the opamp LTP.
Headamp: Sure, that is one of the intended applications. You probably need a low-noise opamp like the LT1028 (and several others) and beefier output transistors like the 2sb649/d669 pair or maybe just the BC639/640 pair.
I have also noticed and read the thread related to LF03 but fully discrete scares me a bit. Let's see how i will succeed with LF01.
Regarding LF01 do i need to have some matched transistors? I was thinking to use some smd resistors.
If you're making two units (one for each channel), then it helps to match devices at each location *across* channels. Resistors to within 1%, and DC beta of transistors to within a few percent. SMD is fine, if you can manage to solder them on perfboards.
To build, it is pretty simple and I would prefer to make it without a PCB as a proto.
But I do wonder what all could be the side effects for improper placements but still quite confident as there are no inductive or capacitive components.
q3 and q4 with be tied together as a 4 legged and q1 ,q2 along with resistors and diodes will form a five legged insect.
c1 will be soldered directly under the IC. and remaining things is as simple it is to be connected to the ic base.
regarding the success and quality we have to trust Linuxguru.!!
If you mean the transistor package pinout, it helps to have the triangular pinout (usually available only on loose bagged TO-92 transistors) for assembly on the LF01 PCB. However, the inline pinout (usually found on both taped and loose bagged transistors) is fine for perf boards and similar - there's no impact on performance. You may have to bend the leads according to your perf-board layout.
Likewise, the C1815, A1015 and C1845 transistors are merely suggested devices that I've already tried. In practice, you can go for generic small-signal devices that are easier to find in the bc54x/55x series, 2N5401/2N5551, mpsa42/92, etc. Some/all of these may not use the ECB pinout of the Japanese devices, and may not have comparable performance, but they're usable in this application.
The pins (circular cross-section Berg strips) are available from EBay vendors in lengths of 40 pins.
learn to use the instruments you have.
Learn to build accurate instruments that can be relied on for the accuracy you have proved you can build into them.
eg,
precision stepped dB attenuator.
precision stepped decade resistance bank.
precision Hamon divider. http://conradhoffman.com/HamonResistor.html
precision (simple resistive) Wheatstone bridge.
using a resistor to monitor current flow.
etc.
etc.
BTW,
combining a Wheatstone Bridge, with a Hamon divider on one side and a stepped precision resistance on the other side allows very accurate resistance measurement over a vast range that a bench DMM cannot approach for accuracy. eg 0r01 to 100M to <0.01%
Learn to build accurate instruments that can be relied on for the accuracy you have proved you can build into them.
eg,
precision stepped dB attenuator.
precision stepped decade resistance bank.
precision Hamon divider. http://conradhoffman.com/HamonResistor.html
precision (simple resistive) Wheatstone bridge.
using a resistor to monitor current flow.
etc.
etc.
BTW,
combining a Wheatstone Bridge, with a Hamon divider on one side and a stepped precision resistance on the other side allows very accurate resistance measurement over a vast range that a bench DMM cannot approach for accuracy. eg 0r01 to 100M to <0.01%
Last edited:
Ever wondered What an electronics enthusiast would do in weekends when rest of his friends are hanging out in pubs and parties..
I did simulation in Ltspice..(newbie here)
Used basic components here to be quick.
first is 200khz and second in 2mhz.I hope pictures are self explanatory .
what all reasons are attributed for decrease in amplitude..
I did simulation in Ltspice..(newbie here)
Used basic components here to be quick.
first is 200khz and second in 2mhz.I hope pictures are self explanatory .
what all reasons are attributed for decrease in amplitude..
Attachments
Q4, the output-stage PNP, is incorrectly connected - the collector and emitter are reversed. LTSpice doesn't know or care about this - it will apply whatever model it has for the transistor (Gummel-Poon or Ebers-Moll), regardless. You'll probably get a very low-gain (hFE ~=1) transistor in inverted operation, and that's probably why it shows attenuated output.
My Bad , I Did not notice the emitter of pnp in negative. when I saw it worked in first run I assumed circuit is alright. (shockingly) it showed me working perfect in 200khz range.I will try again correcting it.
BTW does that implies Lt spice is flawed? I was thinking of designing some filters.
Any other better software ?
BTW does that implies Lt spice is flawed? I was thinking of designing some filters.
Any other better software ?
Last edited:
LTSpice (or any other simulation program) has no way of knowing if a particular circuit configuration was deliberately intended or an unintentional error. They're all going to simulate it the way it is drawn, but may throw up errors or warnings in some cases. As simulation SW goes, LTSpice is actually very convenient and probably the easiest to use. TI's TINA is another popular alternative.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.