LJM L20V10 tips/mods etc

Almost forgot to say:
Today I have replaced all KEC output transistors because these Korean transistors quite easily burn (*), at increased voltages. I have replaced them with original Toshibas purchased from Mouser. 2SC5200N/2SA1943N are replacement for KTC1047/KTB818.
2SC5200N in place of KTC1047 and 2SA1943N in place of KTB818. Carefully, do not reverse PNP and NPN types. This will increase reliability and sound. Output transistors are buffer and they dont have to be matched. hFE values are irrelevant, almost.
Keep the remaining transistors.

(*)
Let me explain why I've replaced output transistors during the initial testing - I have schedduled that for the end. Two output transistors burned as soon I have increased voltage to ca. +/- 45. Luckily, I have worked with double protection: variac+bulb. So, no serious damage, but waste of precious time.

EDIT: Right now I was in my workshop to test my progress while burning in: Toshibas really make difference. Signifficant.
EDIT 2: 2SC5200N/2SA1943N are out of stock at Mouser, but Farnell has lots in stock.
 
Last edited:
  • Like
Reactions: Enterpryse
I am running at 45-0- 45 with no problems, however I would like to jack the voltage up to 65-0-65 so I can use my linear PSU.

Really appreciate you sharing your findings. I'll put those tranny's into kicad when I get chance and update my schematic ( I'm very much a kicad beginner).
 
Items with an XXX need confirming.

1745407335469.png
 
  • Thank You
Reactions: Berlusconi
@Enterpryse
Great schematics indeed.
Meanwhile I had a hard day - lots of work to be done.
Regarding the biasing resistor R16: after the burn-in I have meassured the potentiometer: 3620 Ohm. I have replaced the potentiometer with 3K3 + 330R in series. Bias voltage is arround 20mV and there are no crossover artefacts at all. Bias voltage should be meassured between the output and emitter of any output transistor.
More tomorrow, now I'm rather exhausted.
Where you've ordered, at Farnell? They have both 2sc5200N/2SC1943N.
EDIT: Note that Toshibas have much higher (230V) Collector-emitter breakdown voltage and much higher fT value.
 
Last edited:
I'll check the value of the pot once I have installed the new components on the board and set up the bias, hopefully it'll be around the 20mV you've measured.

The O/P tranny's I ordered from Mouser, these ones are easier to find from more suppliers.

I'm struggling a little with Kicad as even with the auto numbering turned off it still seems to change the component references. I thought I had it nailed but then noticed some annoying changes.

The emitter resistors on my original board are 0.1 ohm. On the schematic I left them as 0.2. which are you actually running?
 
OK, latest from me.

Bias resistor set at 3.9K which equates to about 3mV across the 0R1 emitter resistor (not the 0R2 you are using). If I raise it to 20mV the amp starts cooking and the bias voltage drops all on its own.

Also, my current power supply is only 30-0-30 but will be 45-045 shortly and I hope move to 64-0-64 finally (if it doesn't go pop). So I'd expect this to have an effect on the bias, I can't say by what though!

All for now.
 
@Enterpryse
Hi,
Today I have replaced all small transistors with these purchased at Mouser, the outputs are already Toshibas. This weekend I plan to go further with replacing the drivers with MJE15034 and MJE15035. I did this because I've suspected that some unreliable component has caused that my several output transistors have burned, already twice at higher rails.

Now, after replacements, square wave response is signifficantly better at 20KHz - almost ideal. I will finish them off - all, to remain at safe side and with better SQ. Later on I will publish exact list of transistors along with their hFE values

If you decide to do the same be carefull - NCC5551 has different pinout: 2nd and 3rd leads should be reversed if using normal 2N5551

Do you still have that empty board? Have you got transistors separated from the board. I have destroyed them all during desoldering so I can't measure them for comparison. Could you please meassure their hFE values.
 
Thanks @eneterpryse, but I have realized that there is no need to look back into delivered semiconductors - below is the list of appropriate replacements. Otherwise, all hFE values of the replacement components are rather higher, example: BC546BTA hFE 316.
Now, the board doesn't blow above +-45, has great square wave response and there are no visible crossover artefacts even at very low bias. This board may operate adequately even at nearly zero bias with quality components. And speakers are safe in case something nasty happens randomly.
Interestingly, the other board, still unchanged, has rather good performance except the bias has to be increased to 20mV to reduce crossover distortion.
I will complete this phase of testing now and proceed with FFT tests sometimes in may. If it passes, I will keep it.

D1047-> 2SA5200N (Toshiba)
B817-> 2SA1943N(Toshiba)
2SC4793 ->MJE15034G(ON)
2SC1837->MJE15035G(ON)
2SD669->TTC004BQ(Toshiba)
BC546->BC546BTA (ON) (hFE 316)
BC556->BC556BTA (ON)
A92->MPSA92 (Diotec)
NCC5551->2N5551BU (ON)
 
  • Like
Reactions: Enterpryse