LuDEF

Member
Joined 2019
Paid Member
^ If ZM doesn't get back to you, I'd start here.

https://www.mouser.com/datasheet/2/609/lt3092-2036966.pdf

My suggestion would be to first read the sheet and determine what those segments of the part number denote.

Then, I'd choose the package size that fits the boards you have.

Then, if you're curious about at least one other piece of the puzzle, maybe do an internet search re: what you're looking for. If that is a bit tricky, I found this in <5s.

https://ez.analog.com/power/f/q-a/163816/temperature-grade-reliability

Then, if a solution is not certain, maybe post back.

I think that X package with X grade will work well, and I am fortunate to have found it in stock, so I have picked X part linked below. Is that okay? Or... order a kit for certainty.

Just suggestions. It's fun (for me) to learn what some of the mumbo jumbo means on spec sheets and parts numbering nomenclature, maybe it will be for you.
 
Official Court Jester
Joined 2003
Paid Member
dunno, don't care for suffix
what I'm looking regarding LT3092 is : preferably SOT223 case, if not available then TSOT-23-8 case

recently I resorted to make all pcbs containing ref. CCS with 2 options - instead of onboard CCS, there is small Daughterboard soldered with 2 fat pins to main board , and then it can be either LT3092 (normal case and ultrasmall SMD) or LM334

LM334 still much easier to obtain than LT3092

frankly, dunno where I posted all related info, but Codyt's build is with Daugterboards, just go back few posts to see his pics

onboard LT3092 reserved strictly for future Babelfish XA252, nasty xconductance of pucks needing every ounce of predictability in biasing circuit, at least when my OCD is deciding factor :rofl:
 
Official Court Jester
Joined 2003
Paid Member
found few posts with daughterboards presented

https://www.diyaudio.com/community/threads/the-marriage.390965/post-7140688

think I'm going to take another trick from Pa - need to have spreadsheet documenting what I made, recently is happening that I'm seeing new things in my old posts :rofl:

last night I was searching through one ("my") thread on Baby DiyA, just to see renders of line stage I was working month or so back, and totally forgot about :devilr:
 
After installing more caps to cap-mx output, and running only one channel at a time, same thing (though i guess not unexpected).

I went to comparing spice and real pcb thermals. Spice says irf510 and bc139 should be dissipating 500mW each, that would get kind of toasty. Yet mine i cannot even tell are even warm (30C or less).
Maybe there is reason for suspicion here even though i get ~-1.345V on Lu gate that spice says
1668785665698.png
 
Member
Joined 2019
Paid Member
I must be missing something

Definitely.
IRF510 and BD139 should be hot and require heatsinks if CCS operates as it should at 30 mA.

Did you measure R18 on your PCB (R9 from the sim) without de soldering? There is a BC550 base to emitter junction that could skew measurement. Try with reversed positions of ohmmeter probes.

What is voltage across R11 on the PCB?
 
Definitely.
IRF510 and BD139 should be hot and require heatsinks if CCS operates as it should at 30 mA.

Did you measure R18 on your PCB (R9 from the sim) without de soldering? There is a BC550 base to emitter junction that could skew measurement. Try with reversed positions of ohmmeter probes.

What is voltage across R11 on the PCB?
Ok, so i took off pcb and de-soldered and replaced bc139 and bc550 for different one (expect this to make 0 difference but desparation)
While they were off measured this R9, now without bjt Rbb effect and still 29 ohms, but given how critical it is i changed it and now it is 24.5 ohms (ok, not perfect, 27mA instead of 30, but if it's that it should show some improvement at least and then i know where problem lies)
another check of voltage shifter Iq

your LTSpice pic : check voltage across R8

it must be (0V65/22R)*61R

(where did you found 61, dunno)

I remember my value is 68R :rofl:
Measured this one at 66..
1.73V measured across this..ok pretty much correct, (0.65/24)*66
On gate of LU now is 1.172V
DC offset needed retuning, so hopeful now things are different

Thank you guys for all help
As you maybe remember my previous forum name was Zbunjen, maybe i should have kept it (and you know why, we are all same region) :rofl:

EDIT: no change, though now the IRF is warmer, but ok, this was worthy sidequest
It is actually interesting but output curve is identical down to minute detail
Maybe its just factory reject Ali LD1014's, what else could it be?
 
Last edited:
both channels showing the same?
Yeah, azure is new modified right channel, red is old unmodified left.
In reality they are perfectly overtop each other, i just clicked "separate traces" in REW so you can even see two.
1668799130420.png

Cue my embarassment when amp is fine and it is somehow recording setup..but then makes no sense everything i ever measured was flat, except with this.
 
Member
Joined 2019
Paid Member
EDIT: no change, though now the IRF is warmer, but ok, this was worthy sidequest
All corrections that were made were necessary. CCS had too small current (cool semiconductors) and you didn’t add capacitors at the Cap mx output for nothing. Check this if you haven’t already:
https://www.diyaudio.com/community/...-at-the-capacitance-multiplier-output.389938/

In reality they are perfectly overtop each other, i just clicked "separate traces" in REW so you can even see two.
That’s weird. Both channels measuring the same, mean two semi-defective LD1014D in a row.
If LU's are main suspects, I can send you for free a matched pair soldered on aluminum PCB adapters. They have nice triode curves but that is DC measurement. Who knows how they work at AC.