Without measuring them, the schematic says they’re +/- 79 volts.
Here is the first completed board minus the missing 6 18k resistors.
Even with the thick aluminum backing the devices were all easily soldered to the board with the exception of the four larger transistors. I was able to solder the three smaller legs with my regular tip no problem, but when it came to the larger pad the heat sank into that aluminum lol. I had to take out one of the bigger wedge tips. This one has a blue marker, so I want to say it’s an 840 Fahrenheit tip.
Soldered the large tabs no problem, the solder wetted to the pads in about 1 second. One more from directly above.
And still have to do the pins of course.
Dan
I see that the standard 0603 resistor is generally rated for 1/16 of a watt, the ones I picked out say 300 mW. Since I didn’t see anything specified I’m guessing that the 1/16 watt are okay. Figuring that there is less than 158v across the resistor chain and all of the through hole resistors I use are rated at 250v for the 1/4w I’m thinking one resistor would work. If it would be better to use a pair of 1/4w 27kohm resistors to hit the 54k exactly (as it was in the original BOM) then I could do that as well.
Dan
Dan
Okay great, thank you. I powered the amp up on a measured the voltages at pins 7 and 9 and they were pretty close at +/- 76 volts. I went ahead and just bought some 0603 resistors so that it could look correct. I didn’t like the way it looked with resistors hanging off of it, plus if I want to mount a heatsink from the front of the board those resistors would make it pretty much impossible.
Is there an 0603 that you recommend? I saw that they are generally rated at 1/16 watt, the ones I bought are rated 1/10 watt or 100 mW, I couldn’t find any of the 300mW I originally bought in the right value. So if you say that I’m running at 460mW then this string of 3 isn’t going to cut it correct? Is this made to run on lower rail voltages? I’m guessing no. I know the wattage rating of resistors can get different sometimes, surprised when I bought some 1/2w through hole and when I got them they were physically smaller than some 1/8w through hole I had. Not sure how something so tiny could dissipate 1/2w. Should I be okay with these 1/10w resistors? Do you recommend another value still? Like maybe three 20 kohm or 22 kohm instead of the 18?
Regardless, I have another fault in the amp to track down, but this board definitely showed me the original was damaged, probably damaged from the fault in the amp. Thank you so much for this project!
Dan
Okay, great thank you, so these 1/10w 0603s I picked up should be good to go then. I appreciate your answer.
Dan
Could you please share the project of the KIСAD clone STK3102 on an aluminum substrate or correct it because instead of 3 18 kOhm resistors series in the 0603 package, I want to put one in the 2512 package 56 kOhm or 4 resistors 13 kOm in 0603 package . I need this clone for repair luxman L435. I check power disipation in Microcap on this resistor 0.36W. Thanks and grateful in advance! For some reason I couldn’t find how to send you a private message!
Last edited:
Can you tell me the supply rail voltages in your Luxman?
This resistor chain can be modified at higher rail voltages to keep dissipation under control, while still having sufficient operating current.
I have a supply voltage range of +-60V, but I need to give a margin of +15% according to our supply network standards. Maybe use 0805 instead of 0603. Can you send me a board design for Kicad and I can move the elements there myself? Thank you in advance!
alvlmoskalchuk@gmail.com
alvlmoskalchuk@gmail.com
Last edited:
Saabracer23, I have some question! How clone STK work? Not problem? You placed 3x0603 18kOm or?
For that supply, we need to emulate the STK3152. I will think about the necessary changes to achieve 60V nominal, 90V max working.
The module data sheets do not give any clues what transistor types were used or resistor values.
One issue is that a general purpose 0603 resistor is rated at 75V max.
Some of the amplifiers pushed these STK modules really hard
I will remake this project in KiCad 8, for the JLCPCB process with a 1.5mm aluminium plate, 0.1" dielectric.
My old files were based on an alumina ceramic, not the same thing.
It would be really helpful if users of these modules can put up photos and measurements of their old modules and how they are mounted.
Is the module clamped to a heatsink?
Are there any mounting ears?
My old files were based on an alumina ceramic, not the same thing.
It would be really helpful if users of these modules can put up photos and measurements of their old modules and how they are mounted.
Is the module clamped to a heatsink?
Are there any mounting ears?
One would have an overview of all models with this STK hybrid IC (like the very good sounded Luxman R5030), many users who no longer assumed that a repair would still make sense could now replace it for resonable costs.
And I know that the STK versions mentioned under
https://www.diyaudio.com/community/...-iv-v-looking-for-reliable-substitute.396173/
were used in a very wide range of brands and models from the 80s.
I would check the possibility of substituting a plate amplifier with a BIG heat sink, and wiring it with ribbon cables to the PCB where the STK module was mounted, it is an effective solution.
The issues with heat dissipation are addressed, and you can use a higher voltage rated amplifier, some assemblers like Sony ran them too close to the limit.
I use the term assembler, because the circuits were designed by the STK maker, and only minor changes in values could be done by various OEM purchasers.
To close to the limit meant high failure rates.
If you do use a SMD module, please think of adding an inrush current limiter to prevent incidents at start up.
The issues with heat dissipation are addressed, and you can use a higher voltage rated amplifier, some assemblers like Sony ran them too close to the limit.
I use the term assembler, because the circuits were designed by the STK maker, and only minor changes in values could be done by various OEM purchasers.
To close to the limit meant high failure rates.
If you do use a SMD module, please think of adding an inrush current limiter to prevent incidents at start up.
The current in the longtailed pair is about 2mA, the output drivers about 6mA. This should be OK up to 70V rails, with appropriate driver transistors.
The current through what was identified as a 51k resistor, setting the current through the bias diodes, is proportional to the supply voltage.
This means that at higher supply voltages, more like 100k is suitable. This would be 3 x 33k resistors in my version.
This brings the dissipation in these under control at 75 mW each.
The current through what was identified as a 51k resistor, setting the current through the bias diodes, is proportional to the supply voltage.
This means that at higher supply voltages, more like 100k is suitable. This would be 3 x 33k resistors in my version.
This brings the dissipation in these under control at 75 mW each.
Hi,
I built your -c version using FR4 and chose the thinner dielectric thickness. If the dielectric is the same thickness I do not see a reason to use the aluminum backed tech vs regular fr-4.
On the fr-4 design you could remove the solder mask on the back side, have it tin or even gold plated. Removing the SM would allow for better cooling.
For the network that is used to bias the ips and vas you could replace a resistor with a constant current source such as S-102, chose resistors for about 20v across the CCS diode
I also found when I reviewed the design and soldered it up that I think you could get away with using 0805 sized parts but I would need to do the layout to verify if it’s possible.
I have yet to test the version you built along with a stk-0050-II that recently designed.
I built your -c version using FR4 and chose the thinner dielectric thickness. If the dielectric is the same thickness I do not see a reason to use the aluminum backed tech vs regular fr-4.
On the fr-4 design you could remove the solder mask on the back side, have it tin or even gold plated. Removing the SM would allow for better cooling.
For the network that is used to bias the ips and vas you could replace a resistor with a constant current source such as S-102, chose resistors for about 20v across the CCS diode
I also found when I reviewed the design and soldered it up that I think you could get away with using 0805 sized parts but I would need to do the layout to verify if it’s possible.
I have yet to test the version you built along with a stk-0050-II that recently designed.