I'm considering ordering Sanken 2sa1494 and matching pair (2SC3858) in bulk to match. I need them to repair an old vintage amp whose transistors blew up.
To avoid fakes and have better chances of properly matching them (I'll buy a Peak DCA75 for that...) I would buy them from digikeys which probably can sell them from the same lot and I know they won't have been mixed around.
How many do you think are necessary for a good match, I'll need 8 each npn/pnp cause the amp has 4 pairs per channel.
And do you think this is a good idea in the first place ?
I have found original transistors for the amp (from 81/83 era) which would save me modifying the tracks of the amp (different pin order!....) but they are rare and known to go out.
What are your experiences with this ?
Cheers and Merry Christmas
To avoid fakes and have better chances of properly matching them (I'll buy a Peak DCA75 for that...) I would buy them from digikeys which probably can sell them from the same lot and I know they won't have been mixed around.
How many do you think are necessary for a good match, I'll need 8 each npn/pnp cause the amp has 4 pairs per channel.
And do you think this is a good idea in the first place ?
I have found original transistors for the amp (from 81/83 era) which would save me modifying the tracks of the amp (different pin order!....) but they are rare and known to go out.
What are your experiences with this ?
Cheers and Merry Christmas
Bellow is the driver stage (B) and power stage (C). The originals are OD503A-Q, the amp is a Technics SE A3. These were supposedly a revision of the OD503A-P when as far as I could get information blew up even more frequently.
There is a page about this common problem here:Adrian-Kingston.com - Technics SE-A3
I also found out someone (on Audiokarma) with past experience (although very long ago) that has used Sanken 2SA1494/2SC3858, instead of the Toshiba 2SA1987/2SC5359.
History goes like this, amp was going into protection and I found I was getting 70V on the collectors of the right channel transistors (both on the + and - rails), heatsink on that side also got pretty hot, a bit hotter than the other side and I think hotter than normal, although this amp runs on the hot side I'm told.
Being the smart ****** I am I decided to switch the driver board of the bad channel to the left. Result: a spark, blown fuses and blown transistors that were previously good (as measured on circuit) on the left channel.
I've also detected two practically open resistors on the right driver board (354 and 357 see schematic), maybe those were the problem, although being fused resistors they might have gone open after the transistors got blown. Rest of the components on the driver and power boards seem ok (compared both channels in circuit and only think bad were the fusible resistors on the right side driver board...).
I don't understand much more but I notice that the power transistors have emitter resistors which help share the current between transistors. They're 0,47 ohm, but how much matching does that imply ?
I'll also have to decide if I want to cut the tracks on the power board or extend and bend the legs of the new transistors, I'm more inclined for the second using some heat shrink or one of those tubbed protections for the legs....
Guess that's more or less all...
HERE'S DIRECT LINK TO IMAGE AS SITE IS SHRINKING IT A BIT:http://i74.photobucket.com/albums/i...d_Power_Stages_Better_CloseUp_zpsrwlqba5t.jpg
THanks
There is a page about this common problem here:Adrian-Kingston.com - Technics SE-A3
I also found out someone (on Audiokarma) with past experience (although very long ago) that has used Sanken 2SA1494/2SC3858, instead of the Toshiba 2SA1987/2SC5359.
History goes like this, amp was going into protection and I found I was getting 70V on the collectors of the right channel transistors (both on the + and - rails), heatsink on that side also got pretty hot, a bit hotter than the other side and I think hotter than normal, although this amp runs on the hot side I'm told.
Being the smart ****** I am I decided to switch the driver board of the bad channel to the left. Result: a spark, blown fuses and blown transistors that were previously good (as measured on circuit) on the left channel.
I've also detected two practically open resistors on the right driver board (354 and 357 see schematic), maybe those were the problem, although being fused resistors they might have gone open after the transistors got blown. Rest of the components on the driver and power boards seem ok (compared both channels in circuit and only think bad were the fusible resistors on the right side driver board...).
I don't understand much more but I notice that the power transistors have emitter resistors which help share the current between transistors. They're 0,47 ohm, but how much matching does that imply ?
I'll also have to decide if I want to cut the tracks on the power board or extend and bend the legs of the new transistors, I'm more inclined for the second using some heat shrink or one of those tubbed protections for the legs....
Guess that's more or less all...
HERE'S DIRECT LINK TO IMAGE AS SITE IS SHRINKING IT A BIT:http://i74.photobucket.com/albums/i...d_Power_Stages_Better_CloseUp_zpsrwlqba5t.jpg
THanks
Last edited:
HERE'S DIRECT LINK TO IMAGE AS SITE IS SHRINKING IT A BIT:http://i74.photobucket.com/albums/i...d_Power_Stages_Better_CloseUp_zpsrwlqba5t.jpg
I bought 25 each Mj1521194 and 95, to replace 20 MU15024 and 25. These were from the US subsidiary of Farnell and probably were genuine ON semi.
I used a 12 V power supply to match Vce. I used a 10 ohm resistor to collector, and 100 ohm from PS to base. Most were measured .14 or .13 v Vce. There was one or two .12 v and were a few .11 v. I used the oddball .11 v ones in another project. I used a heat sink in the test and let them stabilize a minute before wrting the results on them with a sharpie marker.
.47 ohm resistors IMHO in your project make matching Vce a little less critical than lower value emitter resistors. My PV-1.3k has 0.5 ohm emitter resistors, and Peavey states that the transistors bought from them directly require no further matching. I would not of course mix old MJ15024 and new MJ15194: the thermal behavoir of the different processes would be quite different.
I used a 12 V power supply to match Vce. I used a 10 ohm resistor to collector, and 100 ohm from PS to base. Most were measured .14 or .13 v Vce. There was one or two .12 v and were a few .11 v. I used the oddball .11 v ones in another project. I used a heat sink in the test and let them stabilize a minute before wrting the results on them with a sharpie marker.
.47 ohm resistors IMHO in your project make matching Vce a little less critical than lower value emitter resistors. My PV-1.3k has 0.5 ohm emitter resistors, and Peavey states that the transistors bought from them directly require no further matching. I would not of course mix old MJ15024 and new MJ15194: the thermal behavoir of the different processes would be quite different.
Here's a better one , the original image you uploaded, just right click the link and open in another tab, so you can go back and forth without closing windows.
http://i74.photobucket.com/albums/i...tages_Better_CloseUp_zpsrwlqba5t.jpg~original
Next time uoload to imgurl or tinypic which give you a link which will open inside DIYAudio
Back to the amp:
1) build and use a lamp bulb current limiter.
Since regular filament bulbs are dissapearing, I suggest a "pencil" 100W quartz lamp, with the proper fixture of course.
So you avoid burning more expensive stuff
2) remember that now you don't only have to worry about dead powerb transistors but a damaged driver board too.
3) seeing the schematic, matching is not *that* critical as you may think, getting them from same batch should be enough.
Just specifically *ask* for them, otherwise the Mouser employee might pick the last 8 from one box and fulfill the order with others from a different one.
I guess in the order you can specify "please pick all same type with same manufacturing code, as in *all* from he same box"
FWIW I buy transistors in bulk (50/100/200/500/1000) depending on type, and with modern ones just measure a few from any new batch, all others same code are *incredibly* close.
Wonders of modern precise and automated manufacturing.
, the original image you uploaded, just right click the link and open in another tab, so you can go back and forth without closing windows.http://i74.photobucket.com/albums/i...tages_Better_CloseUp_zpsrwlqba5t.jpg~original
Next time uoload to imgurl or tinypic which give you a link which will open inside DIYAudio
Back to the amp:
1) build and use a lamp bulb current limiter.
Since regular filament bulbs are dissapearing, I suggest a "pencil" 100W quartz lamp, with the proper fixture of course.
So you avoid burning more expensive stuff
2) remember that now you don't only have to worry about dead powerb transistors but a damaged driver board too.
3) seeing the schematic, matching is not *that* critical as you may think, getting them from same batch should be enough.
Just specifically *ask* for them, otherwise the Mouser employee might pick the last 8 from one box and fulfill the order with others from a different one.
I guess in the order you can specify "please pick all same type with same manufacturing code, as in *all* from he same box"
FWIW I buy transistors in bulk (50/100/200/500/1000) depending on type, and with modern ones just measure a few from any new batch, all others same code are *incredibly* close.
Wonders of modern precise and automated manufacturing.
JdAo, I know this thread is >2 years old, but perhaps you still hang around... How did it go replacing the transistors ? did you get the amplifier to play again ? I am in the middle of restoring an Se-A3 and while the OD503AQs in it seems to be fine I wonder if I should just go ahead and replace them while I have her open.
Thanks
Correct. I made a simple test fixture with a few resistors, designed to make the transistor conduct at around 2 Amps through a load resistor. Measuring the voltage drop across the load resistor with a good multimeter was good enough to pair up power transistors - even from differing batches.
That's right - you need to test with at least at half an amp. collector current on a large heatsink and under the same temperature conditions if you seriously want to match power transistors within tight limits. That's really not easy to do and it's simpler and much easier to buy in tubes or as they come from the same tube, tray or whatever smallest package unit the manufacturer supplies them in, and avoiding matching if you can. Sanken's quality is really first class and consistent so I doubt you'll have problems this way, so long as your supplier understands what you want.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.