BA-F1 restore

The owner and I discussed the whole situation, we decided that rather than spending a lot of money on OP transistors that didn't exactly match the original good set anyway, it would be best to replace all for new.

All the attributes of the proposed replacements that Oilmaster mentioned above made sense to me as well as a much bigger SOA. So you won't always be using the amplifier with your heart in your mouth, with an ever dwindling supply of expensive original OP transistors in the background.

As regards 'square pin round hole' - AK member Kale solved this using gold plated round pins soldered to the TO-3P transistor leads. However I have seen a Youtube video on AK where this was not thought necessary and the square pins were rounded with a combination of pressure, and some careful filing, I am still thinking about this aspect.

The thread that AUD101 referenced is the kind of thread that convinced me that the TO-3P's could be the perfect replacements. :thumbsup:

However this thread gives the exact type numbers:-
http://www.audiokarma.org/forums/index.php?threads/au-919.291461/page-4 - see post #76

One piece of good news is that no cutting of the transistor tabs will be needed due to the design of the heatsink in the BA-F1. ;)
 
Last edited:
Yes indeed, Z doing another nice job there.

What it boils down to, IMO, is that 2SA1117 and 2SC2608 are rather standard devices (too poor for this Sansui) with limited bandwidth (fT),poor beta-drop and much higher Cob.

2SA2223A & 2SC6145A are so-called high-fT devices, with low Cob, and of top of that a very linear Beta-Ic curve, making them VERY good output devices for which this Sansui was designed around.

I would say here something.... this is picture from Sansui service manual:
35501787051_91cba8a2f3_o.jpg


And as you can see, Sansui recommended 2SA1116 / 2SC2607 as substitute to those NMA1012 NMC1012 combination, so it seems to me that NMA1012 is Non Magnetic version of 2SA1116 and NMC1012 is Non Magnetic version of 2SC2607. Sanken 2SA1117 and 2ASC2608 are the same at all technical data as 2SA1116 / 2SC2607, just 2SA1117 and 2ASC2608 has higher max.Ic, so because of that, higher max. power too.

So at the end… ft of 2SA2223A is 35MHz and ft of 2SA1116 or 2SA1117 is 20 MHz, maybe that would be good, but maybe would not… such us I would not use Sanken 2SA1117 and 2ASC2608 in some AUX1 amplifier or in some AU20000 (at first amplifier maybe they will be to slow and at second amplifier maybe they will be to fast…)
Here it is test signal of one BAF1 with installed 2SA1117 2ASC2608, look frequency of that square signal (taken at speakers terminals)!
PICT1275.jpg

The only one not good thing for those 2SA1117 and 2SC2608 is what they are not cheap...
 
As regards 'square pin round hole' - AK member Kale solved this using gold plated round pins soldered to the TO-3P transistor leads. However I have seen a Youtube video on AK where this was not thought necessary and the square pins were rounded with a combination of pressure, and some careful filing, I am still thinking about this aspect.
I still have those 2SC3263Y / 2SA1294Y what I made to TO3 socket :) BTW they are with fT 60MHz!! I made them in time when I didnt have those Sankens for AUX1...
2SC3263.jpg
 
I would say here something.... this is picture from Sansui service manual:
35501787051_91cba8a2f3_o.jpg


And as you can see, Sansui recommended 2SA1116 / 2SC2607 as substitute to those NMA1012 NMC1012 combination, so it seems to me that NMA1012 is Non Magnetic version of 2SA1116 and NMC1012 is Non Magnetic version of 2SC2607. Sanken 2SA1117 and 2ASC2608 are the same at all technical data as 2SA1116 / 2SC2607, just 2SA1117 and 2ASC2608 has higher max.Ic, so because of that, higher max. power too.

So at the end… ft of 2SA2223A is 35MHz and ft of 2SA1116 or 2SA1117 is 20 MHz, maybe that would be good, but maybe would not… such us I would not use Sanken 2SA1117 and 2ASC2608 in some AUX1 amplifier or in some AU20000 (at first amplifier maybe they will be to slow and at second amplifier maybe they will be to fast…)
Here it is test signal of one BAF1 with installed 2SA1117 2ASC2608, look frequency of that square signal (taken at speakers terminals)!
View attachment 1059623

The only one not good thing for those 2SA1117 and 2SC2608 is what they are not cheap...

I stand corrected Sir !
Indeed a VERY impressive amp response/performance !!

That's the kind of counter-arguments we like to receive :beatnik:
 
I used to use the 2607/1116 as replacements years ago doing BA-F1 and AU-919 restorations all the time. I could never tell any difference in performance or measurements using them vs. the originals. I also used these in one channel, and left the originals in the other in the same amplifier allot of times, and there was absolutely no difference that I could ever see or hear. The 2607/2608 and the 1116/1117 are not standard TO-3 devices.

I have also gotten exactly the same results using the 2ST2121/2ST5949 in this amplifier, making these excellent and perfect replacements as well for the NMA1012--NMC1012, and the 2SC2607/2SC2608--2SA1116/2SA1117 devices. Problem with the ST devices is finding them though, and Kale is the only trustworthy show in town for proper replacement devices right now if you are not ready to go to another package.

If you like me, are a fan of the search function, this has all been hashed out before.

I have never seen a datasheet for the NMA/NMC parts, so all I have ever had to go on was the service manual for the AU-919 listing those substitutes. Please share where you got your information about the original parts Oilmaster.

Thanks
 
Last edited:
I have never seen a datasheet for the NMA/NMC parts, so all I have ever had to go on was the service manual for the AU-919 listing those substitutes. Please share where you got your information about the original parts Oilmaster.
Thanks

Well I mentioned above already that I stand corrected, but go ahead.... :whip:

From my various Japanese databooks I got hFE 20~30, ft ~ 20Mhz, which hich I don't qualify as high ft devices.
Cob was deduced based on low hFE/fT, but incorrectly, but I am pretty sure that the Beta-Ic curve surely ain't that flat as the 2SA2223A & 2SC6145, simply because they were by far not capable of manufacturing straight linear beta-Ic power transistors at the time.

Leading to my (too) quick conclusion that the 2SA2223A & 2SC6145A would be superior devices to put in.
 
So many questions could be here, for example about hfe; we have only min. Hfe data for 2SA1117 which is 20 (min. Hfe for 2SA2223A is 40) and about fT... 2SA1117 has 20MHz, 2SA2223A has 35MHz... Cob for 2SA2223A is 500pF, for 2SA1117, I dont know... for me, the most important thing is here what Sansui recommended...
BUT, if you wish to do upgrade, and if you wish to spend less money and still get good final result… in that case you can use any modern LAPT TO3P Sanken combination…. I think that continuing debate about this transistor topic here, in BAF1 restoration thread is not good… maybe in some other thread, but here, I am waiting news and new interesting pictures from Hyperion :thumbsup:
 
The replacement OP transistors arrived today - so I tested one of each on my 'Peak Tester' result - vbe's within 10mV and hFE the same! at 100 - This is unusual in my experience, (in a good way I think), usually there is a significant difference between the hFE of a (complimentary pair) PNP and that of an NPN transistor.
 
Last edited:
Here is a general view of the lower chassis - showing the 'meter drive and protection board' in the centre.

IMG_2291a.jpg

And now refreshed and ready to be re-installed.
IMG_2297a.jpg

I just finished getting one of the driver boards stripped of old parts, cleared of 'glue' - and washed clean of dust. :)
 
Last edited:
I have a question regarding C17 & C21 on the driver PCB - should they be 1000pF (Polystyrene) as per the schematic (and parts list) or 1500pF - a clumsy replacement (on the track side) by the previous owner of this BA-F1 ?

I am rounding up the very last additions to a parts order, so confirmation of this would be really helpful. ;)

Edit Update
I'm leaning seriously towards 1000pF on this one.
 
Last edited:
Back
Top Bottom