STK3102 driver, replace with TO-220's...could it be done?

[Bassblaster]

Im about to be feeling really stupid...i looked at the schematic for the sansui...this STK chip doesnt look like the driver...

The original thread i made in the Sansui forums NO ONE said anything about kQ2 or kQ3. watch a $1 transistor be my problem and all the effort that went into this "new" STK IC be wasted.
im going to swap those transistors with the other channel first thing tomorrow. 2:45AM for me right now

 

[sregor]

Q1 base us the input. Q2 base is the input for the feedback. Q3 and r1 goes to +V. Usually decoupled with about a 100 ohm resistor and 100.01 uF capacitor. R3/R2 connect to -V with the same decoupling. I would try R2 to set current through input pair to about 2 mA (70/2 would be about 35K ohms) then R1 should be about .6 volts with 1 mA so 680 ohm. R4 + R3 are picked to set the current through the bias circuit. 10 mA would be R3 +R4 = 7K, but that is a bit high. Q3 would have to dissipate 700mW at idle. 5mA would be 14K for the sum of the 2. Something like a KSA1381 or mje350 or would be a better Q3. Need to do more research for values of R3 and R4, and size of bootstrap cap.

 

[sregor]

Im about to be feeling really stupid...i looked at the schematic for the sansui...this STK chip doesnt look like the driver...
View attachment 940582
The original thread i made in the Sansui forums NO ONE said anything about kQ2 or kQ3. watch a $1 transistor be my problem and all the effort that went into this "new" STK IC be wasted.
im going to swap those transistors with the other channel first thing tomorrow. 2:45AM for me right now

Odds are its the IC anyway but you never know.....

 

[Bassblaster]

Okay i didn't swap them first thing today but i did right when i got home, left channel is still bad with the right channels outputs and now the right whatever k02 and k03 are for. all thats left is the STK pack. i checked before the the single going INTO the STK is clean on both channels. Okay. im going to work on update the schematic of this new pack

 

[Bassblaster]

Okay i didn't swap them first thing today but i did right when i got home, left channel is still bad with the right channels outputs and now the right whatever k02 and k03 are for. all thats left is the STK pack. i checked before the the single going INTO the STK is clean on both channels. Okay. im going to work on update the schematic of this new pack

This is where im at now

Everything look good with this design?

 

[sregor]

This is where im at now
View attachment 941025
Everything look good with this design?

Yes except q4 and q5 are functional (i.e. not necessary) duplicates of kQ2 and kQ3 in the amp. The collector of Q3 and the free end of R4 should go to the bias circuit and drivers on the amp board. (pins 10 and 11 of the original IC. Base of Q1 (input is same as pin 15, and the base of q2 should be connected to pin 14 (feedback) not back to q3.

 

[Bassblaster]

OOOO I SEE NOW. This should be just about right. is the C1845's good for Q1 Q2? not much left to do i feel. I need to somehow figure how i'm going to fit these parts on a 2"x2" blank PCB board.
You said that Q3 will be dissipating 700mW at idle, I dont think TO-126 packages can dissipate that much heat on its own. i might have to mount that to something. maybe a 1/16 inch aluminum flat bar cut to length would work.
Now the wattage of the resistors are probably good at 1/4W since there in the mA range 1/4W should be more than enough

Now you earlier said for each transistor we cut out the lower the sound quality will be. i understand that but by how much. Say at a normal volume would i be able to go "wait a minute, this doesnt sound right".
OR what about power output? the working right channel with the STK can put out a CLEAN 122W into 8 Ohms. could power output drop with this simple 3 transistor board?

 

[sregor]

My bad - no connection between 5 and 7.
C1845 should be good for q1 and q2
Power is more a function of supplies than topology- it should be about the same output.
My guess is that listening to it, you won't notice the difference, but if you A/B'd it with the chip, you should here slight difference. It's the difference between .1 and .01 distortion.

 

[Bassblaster]

My bad - no connection between 5 and 7.

I Thought that looked a little off

C1845 should be good for q1 and q2
Power is more a function of supplies than topology- it should be about the same output.
My guess is that listening to it, you won't notice the difference, but if you A/B'd it with the chip, you should here slight difference. It's the difference between .1 and .01 distortion.

Ill order a thing of 20 C1845's since they're so common. i still have 15 or so A992's for other projects
for the resistors i like Yageo's MFR series. 1% metal film and are cheaper than a potato chip

 

[srinath]

These seem to be available on fleabay - and not too $$ either. Am I missing something ? Are those fake ?
Also these are in the Yamaha CR **40 series too aren't they ??? Or the R series ??
Cool.
Srinath.

 

[Bassblaster]

These seem to be available on fleabay - and not too $$ either. Am I missing something ? Are those fake ?
Also these are in the Yamaha CR **40 series too aren't they ??? Or the R series ??
Cool.
Srinath.

i know eBay has STK3102 packs. but if it can be replaced with a new board, why not? im not too sure if the 3102's on eBay are fake or not. the 3102 is world's less common than say the 0050 so its less likely to be counterfeit.
but this 3102 is used in some Yamaha and Akai gear that i know of.

 

[Bassblaster]

for the resistors i like Yageo's MFR series. 1% metal film and are cheaper than a potato chip

Apparently Mouser doesnt have many 35K resistors AT ALL. will using a 34.8K resistor change much with the circuit?

 

[Eastham]

Interesting thread. Can't wait to see how it works out.

 

[Bassblaster]

I Wonder if i want to get a PCB made for this...
Well anyway how should we go about finding out the value of the bootstrap cap and R3 R4?

 

[sregor]

Start with measuring the voltage drop across kR10 (39 ohms) on the good channel. Use this to determine the current through R3+R4. I'm guessing 200 to 400 mV which is 5 to 10 mA.

 

[Bassblaster]

R10 above C9 on the driver board the voltages are 1.04V and 1.3V

 

[Bassblaster]

i updated the schematic again. its more messy now because if i want to get a PCB made everything has to be connected together.
Will it be ok to change R2 and R5 35K to 36K? loads more options available with 36K

 

[sregor]

R10 above C9 on the driver board the voltages are 1.04V and 1.3V

So voltage drop across the resistor is 260 mV or 6.6 mA of current. So R3+R4 need to be about 10K. The ratio of the 2 in various circuits (hk Citation 12, some old Sansui and Panasonic amps) is about 2 to one, so use R4 is 6.8K and R3 is 3.3K should work. power dissipated by R3 is around 325mW so 1/2 would be okay but 1 watt would be better. The bootstrap cap should be 47 uF at 50 or more volts. Its one of those cases where bigger is better, but if you make it too big, start competing with DC servo and power supply filtering.

i updated the schematic again. its more messy now because if i want to get a PCB made everything has to be connected together.
Will it be ok to change R2 and R5 35K to 36K? loads more options available with 36K
View attachment 944007

The value is not that critical. 36K should be fine.

One more missing piece. There should be a small ceramic cap between base and collector of Q3. 30 pF should be okay, but some designs went higher. (check for oscillation with load.) .

 

[Bassblaster]

This is what i got now. Man i wish i know all this stuff. i would be out here making replacements for all the STK modules.

 

[Bassblaster]

Well i was going to order the parts but of course a couple of them are backordered :/
That chances of this working good should be pretty high. if it works i want to order PCB's for it. its going to be a pita to mount all of that on a blank ratshack board.