STK-0050 replacement for SX-780 and others

[rcs16]

Their has been a bit of interest for a replacement of what was the original Sanyo STK-0050 module used in a SX-780 receiver. Many question the integrity of what can be obtained these days as STK-0050 modules.
I decided instead of sending these files to individuals or trying to make a buck from folks needing them, I thought that it would be better served if I make a contribution to the community.
The design was simulated, so I will include the ltspice schematic. If someone wants the .asc file to play with I can provide that as well.
There were some earlier threads where I introduced this design, you can follow them for some background and pics of one implementation.
The pcb fab files use pcbway naming conventions. You can change them if you chose to use a different fab.
.GTL Gerber top layer
.GTO Gerber top overlay
.GTS Gerber top solder
.GBL Gerber bottom layer
.GBO Gerber bottom overlay
.GBS Gerber bottom solder
.GKO Gerber keepout layer
.DRD Excellon drill file or rename to .XLN for JLCpcb (rev A1 is for JLCpcb)

BOM:
863-MJE15032G MJE15032G ON Semiconductor Q3
863-MJE15033G MJE15033G ON Semiconductor Q2
863-NJW1302G NJW1302G ON Semiconductor Q1
863-NJW3281G NJW3281G ON Semiconductor Q4
80-R82DC3470AA60K R82DC3470AA60K KEMET C2
594-SFR16S0003929FR5 SFR16S0003929FR500 Vishay R9
594-SFR16S0004708JR5 SFR16S0004708JR500 Vishay R1,R5
594-5073NW4R700J PR01000104708JR500 Vishay R3,R4
594-SFR16S0003901JA5 SFR16S0003901JA500 Vishay R8
594-SFR16S0001211FR5 SFR16S0001211FR500 Vishay R6
833-RL202-TP RL202-TP Micro Commercial Components (MCC) D1,D2
534-9902 9902 Keystone Electronics Screw for Q1-4
534-9901 9901 Keystone Electronics Screw for Q5
534-4692 4692 Keystone Electronics for Q1,4
594-SFR16S0007509FR5 SFR16S0007509FR500 Vishay R2
951-SPK10-006-00-104 SPK10-0.006-00-104 Bergquist Company Ins for Q1,4
512-BD13910STU BD13910STU ON Semiconductor Q5
652-3362P-1-501LF 3362P-1-501LF Bourns R7
534-3049 3049 Keystone Electronics For Q2,3
567-120-SA 120-SA Wakefield-Vette
594-K221J15C0GH5TH5 K221J15C0GH5TH5 Vishay C1,3
532-53-77-4G 53-77-4G Aavid Ins for Q2,3

Mouser BOM upload:
863-MJE15032G|1
863-MJE15033G|1
863-NJW1302G|1
863-NJW3281G|1
80-R82DC3470AA60K|1
594-SFR16S0003929FR5|1
594-SFR16S0004708JR5|2
594-5073NW4R700J|2
594-SFR16S0003901JA5|1
594-SFR16S0001211FR5|1
833-RL202-TP|2
534-9902|4
534-9901|1
534-4692|1
594-SFR16S0007509FR5|1
512-BD13910STU|1
652-3362P-1-501LF|1
534-3049|2
567-120-SA|1
594-K221J15C0GH5TH5|2
532-53-77-4G|2

Link to shopping cart:
https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=976621ac80

The docs and fab data are in one zip file, stk-0050-8.zip

You are responsible to build it all yourself. I am here only to help, answer questions. I encourage others to help the ones that are doing this for the first time.
I did think about adding the protection circuitry (mtf mentioned it one time) as was done in the sx-880, but that I will leave for now as it requires further mod to the design (new pcb) and some extra wiring to match the sx-880 design.

Good Luck
Rick

Index: (thanks to Zeb)

Al Plate Page3, post #54
Using the sub to replace the pair of STK0080's. Page 6, post# 107
Bias procedure for a .22 ohm emitters. Page 6 post #115.
Alternate outputs, Would FJP1943 / FJP5200 work? Yes, page 7 post# 123
Is there a way to set the idle current with these new modules? Yes
Procedure for idle setting on DBT. Page 23 Post# 451
Making the heatsinks ready for component install. Page 26 Post# 518
Will it replace the STK-0050II? Page 31 Post# 607
Fisher CA-800 sub install. Page 7 Post# 134
Fisher RS-2010. Page 29 Post# 576
Pioneer SX-680: Post #428, #571 in http://audiokarma.org/forums/index....lacement-for-sx-780-and-others.721181/page-29
Technics SU-V2 (STK-8040). Page 30 Post# 569
Speaker fuse size. Page 7 Post# 140 – 142
Fusing the circuit. Page 21 Post# 419
Current protection. Page 11 Post# 203
Other Output subs. Page 13 Post# 248
Alternate trimmer. Page 28 Post# 552-555
Q2 alternates. Page 25 Post# 495
R2 alternates. Page 13 Post # 259
Initial checks. Page 14 Post# 280
Installing the transistors. Page 15 Post# 281
Bias setting for the SX-780. Page 16 Post# 301
Does the sub work for a STK-0059? Yes. Page 17 Post# 329
More about the sub circuit design from Rick. Page 18 Post# 359
Technics SU-V4? Will it work? It should. Page 20 Post#386 and Post #410
Use in a SX-580. Page 21 Post# 417 and Post# 424 - 428
Some extra assembly notes: Page 48, post #490

Update: Nov 22,2019
I uploaded the A1 revision, you can order from jlcpcb.com using this file
STK-0050-A1-3.zip
Add link to youtube video

Add drawing for lead forming dimensions
Add rev A4 info

 

[Watthour]

Thanks for sharing. That might come in handy some day.

 

[zebulon1]

I would like to try an experiment with your project. No set to try it on.
I'll start looking for a wreak.
What about amp'ing it up to a 0080 or 0150?

 

[rcs16]

What about amp'ing it up to a 0080 or 0150?

It is already a super charged STK-0050 using 150 or 200W o/p devices.
Even if you have working STK-0050 modules, replacing with them with these is kind of like insurance and a upgrade.
Still if you short the outputs, they will probably blow unless you add a fuse in the supply lines link to the pcb instead of the wire bridge = more cheap insurance.

 

[tsd71]

Great post,a big help with the 780 issues, I have a few here that still work and a few that dont, I'm gonna try this out, where to get the PCB's?

 

[rcs16]

try pcbway!! I ordered a power supply pcb from them this past winter, using the $13 for 10 offer of 100x100mm 2-sided boards. I got 11 pcbs and they even came in gold plate even though I asked for the default process of, green solder mask, HASL plating, 1oz copper.

 

[qprhooligan]

The price of pcb fabrication. Would probably drop with quantity ordered. Maybe we could do a group buy?

 

[rcs16]

the board is 2450x1425mil or 96.46x56.1 mm so it qualifies for the $13 for 10 pcs pcbway special plus shipping. Do you need to get any cheaper than $1.30 ea One of the cheaper parts on the BOM.
For a GB, someone would have to step up and do the org, order, shipping to each who signed up.
BTW
R2 should be 75 ohm 1/2w SFR16S series
R6 should be 1K2,
R7 can be 500 or 1K trimpot. A multiturn is not necessary.

 

[punchback]

I have an order "under review" right now. Pcbways has to review the gerbers before you can pay. They review from 1am to 4pm GMT. It looks like it will be 10 for $10 plus shipping by HK Post for $15. I have a couple of units with the STK0050 modules and this could be real nice if the need arises.

 

[punchback]

I probably jumped the gun on this a little bit. If a couple of people want a blank fab to assemble and test let me know and I 'll send one out.. Based on a couple of YouTube videos about pcbways it'll be a couple of weeks.

 

[rcs16]

For HK post it could be 3-4 weeks from my experience. It is the economical choice if you are in no rush.
In the mean time you could be fabricating the aluminum adapter plate!

But if original Sanyo STK type devices are available, why not use them?

Because you can get much better performance out of the discrete stk-0050 module. One example is that you can set the bias current correctly and increase it to have a larger class A region. Increased bias lowers the crossover THD by a considerable amount. SOA is increased dramatically, so that means less failures. Easier to repair. Modern day devices. I could go on
I agree if you are hard up for $, then your STK modules are the economical choice. that is why Pioneer used them to begin with. But you have to ask yourself why did they not use a STK's in the sx-880,980,1080,1280? I think the answer is performance.
During the listening evaluation, the tester did not notice a difference however.

 

[punchback]

I saw your complete module in another sx780 thread. What did you use for the heat spreader? Is that .25 aluminum? Do you have a dimensioned drawing for that? I know it would be a little fiddly getting everything to line up. I guess you could make up the module, then lay it on the aluminum and mark drill and tap.

 

[rcs16]

That module that you are talking about was built by someone else and I believe it was 0.25" Aluminium plate.
The idea is to make it close to the thickness of the orig stk module so that you can use the original screws to mount the plate to the existing heatsink.
You can use the provided layout drawings as your template or use the pcb with devices lead formed but not soldered, as your guide. you want all the devices to seat on the same flat surface of the al plate.
this project tests out your mechanical and electrical skills.

 

[qprhooligan]

is there any issues with not having a bias diode like a stv-4? Does the sx-880 use a bias diode in its design?

 

[markthefixer]

is there any issues with not having a bias diode like a stv-4? Does the sx-880 use a bias diode in its design?

Two general types of circuits are used for this circuit function: the bias diode and the single transistor Vbe multiplier.

The diode is usable when the voltage drops of the circuits are tightly controlled and predictable. Which sometimes causes us problems with replacement transistors in amps...

and yes, the sx-880 uses an stv-4h diode.

 

[rcs16]

hi guys,
the transistor or bjt in the Vbe mutiplier can be in a few different configurations. I chose the most common one, the single bjt, similar to the original. basically copied an earlier Sansui design and it was also suggested in B. Cordell's audio PA design book.
In this design, it would be beneficial to characterize the Vbe multiplier to see how it performs. Basically measuring o/p stage bias current change vs device temperature. The goal is to keep a relative steady bias current vs temp, so when it gets hot it does not run away or be starved of bias current at idle.

 

[HBrown]

I have been running the original prototype modules in my unit for several months almost nonstop in my shop with no issues. They sound really good.

I have been exploring a cost effective option for the heatsink plates. The original plates I made in my home machine shop from 1/4" aluminum plate and they were labor intensive to make a nicely finished product- so much so that fabricating them to sell at a reasonable cost is not possible by hand means. Since then I had a batch of them fabricated by cnc punching from 3/16" aluminum. They had slight distortion from the punching process and required subsequent filing and lapping to be usuable. I am now looking into having them lazer cut.

As Rick stated above, a lot of things have to come together from both an electrical and mechanical standpoint to get these things installed, but, the finished product works very well and is discrete and can be rebuilt or repaired if ever the need arose. That was the main impetus for Rick & I working through this exercise- to get away from having to roll the dice with sub-par STK replacements.

 

[HBrown]

A few pics of the modules. [/ATTACH]

 

[punchback]

Did you ever consider rotating the Q1 - Q4 180 deg and scooting the their through holes towards the edge of the pcb? This would make it fit in the same footprint as the original STK. I'm using a gerber viewer that has a measuring tool and it seems that there is enough room. You could even add a second set of mounting holes to allow this possibility. I don't have access to gerber editing software or I'd try it myself. I may try on the pcb's I ordered by drilling out in the bare FR4 and sky wiring.

 

[HBrown]

You could probably do that. Don't forget large enough holes through the board itself to allow the device mounting screws to be installed from the top down through the board. I believe you would have to re-route the whole board to accomplish that.

We looked at several different units I own that used the STK. None of them had space limitations on the original heatsinks that made it an absolute requirement that the original footprint be strictly maintained. Real estate on that board gets pretty tight when you start actually routing it- standard 'board design rules' and all that. Some of the tracks on the board carry considerable current and adequate size needs to be maintained. We considered surface mount components and a smaller layout, but that began getting away from the 'ease of building and serviceability' aspect of it.

I was originally intending to 'sky wire' the whole thing to test the design, but that quickly became a major headache.