Sony 2SD625 line output transistor

[space_charged]

I'm working on a Sony KV9000UB with a blown line output transistor. I'm trying to find an equivalent of the 2SD625 line output transistor. I have tried a few high voltage/power darlington transistors and the one that gives the best results is a BU931T but it runs hot. I have a working KV9000 and the base waveform is quite a bit different with the BU931T in place of the 2SD625.

2SD625 is available on line but quite expensive, so a modern equivalent would be nice.

[Maarten]

I don't have my equivalent books handy now, so will check them tomorrow. Meanwhile, did you try the BU806?

[space_charged]

Nop, haven't tried a BU806. Looks similar to the BU931T I'm getting some results from.
To make it work I've had to put a 1K pot in series with the base and tweek that for good waveform and reasonable current. The BU931 gets too hot and shows signs of runaway as in increasing current. The base waveforms from the working original 2SD625 and the BU931 attached. The YELLOW trace is the BU931, the purple is the original Sony 2SD625.

I'm thinking that assuming it is getting saturated (turned on fully) then excess heating could be due to rise/turnoff time being too long so maybe my BU931 doesn't have the speed needed. A BU806 sounds like a line output transistor. My BU931 is aimed at automotive ignition switching - I think.

Thanks for your suggestion which I will try and also for consulting your books tomorrow.

C

[Maarten]

I hope to get to my databooks this evening. Meanwhile you could try bridging the 1k pot with a fast diode (UF4007, maybe even 1N4148) anode to base to get it to turn off faster if that is indeed (part of) the problem.

[space_charged]

Yes indeed - I'll try that. I "play" a fair bit with switch mode supplies so I have a box full of UF4007 fast diodes.

[space_charged]

Tried it but it provokes lots of nasty ringing, see 'scope photo.

Perhaps the problem is the opposite, that the BU931 turns off faster than the original transistor? But then why would it be getting so hot?

[Maarten]

For some reason, my ECA guide doesn't think the 2SD625 is a darlington. It gives BUY49 as a direct equivalent and BU406...409 and 2SC2767 as replacements with different outlines.

According to the circuit of the KV-9000E I found, it's driven from a transformer. Maybe it isn't a darlington? The 2SD626 seems to be the only one in that series.

[space_charged]

Interesting. The data I've got on 2sd625 gives its hfe as 120 which is on the low side for a darlington but high for a strait power transistor.

I have also already tried a high gain single transistor as a replacement. I used a KSC5027OTU. This functions to some extent, but not enough to produce a picture. It shows symptoms of lack of drive really. It also gets hot, which is reasonable if its not saturating when on.

I have the service data on this set but only from the radio and Television books. If you have a better service sheet for it I'd love a copy.

Chas

[Maarten]

I downloaded a rar file with a few scans from elektrotanya.com.

I didn't mention, but ECA classifies the 2SD625 as an audiofrequency / general switching transistor. That may explain the higher hFE compared to dedicated horizontal output transistors.

[space_charged]

OK, thanks. I've downloaded the service info.

As to the 2SD625, if its classed as audio/general then its not going to be that fast.

Might try a BUY49 and BU406

C

[Studio263]

If the transistor gets hot it is because it isn't being turned fully on, or is being turned off too slowly. I would try something like a MJE340 and reduce the value of R514 from 3.3R to perhaps 1 or 1.5R and see if it runs cool. Sony's own transistors have higher current gain (hfe) than normal ones which is why increasing the base drive often helps. Also check C520, which will distort the line drive waveform if it is 'tired'.

This is a low voltage / high current line output stage - the line pulses on the transistor's collector are only 175V. Darlington transistors are not suitable here as they are too slow.

[space_charged]

Yes all very much as I thought, not saturating or rise/fall time too long.

I thought the 2SD625 WAS a darlington. The data I have on it suggests it has a MINIMUM hfe of 120 which is going it some for a power transistor.

The BU931 does work, but gets too hot. As its high gain, I'm assuming it will saturate. I guess you are really confirming my worries about it not being fast enough.

Thanks for the tip about C520. I had my eyes on it anyway.
Just to clarify, I started off trying to repair a KV9000 that had a blown (b/e open) 2SD625. As it was expensive and hard to get I tried some modern transistors. The only one that actually produced a picture was the BU931.

I then switched to my WORKING KV9000 in case there was some other fault confusing the transistor substitution experiments. The results on the the otherwise fully working set were exactly the same as the original non-worker.
To do these experiments to find a reliable substitute LOT, I removed the 2SD625.

If I put it back, then the set runs normally and 2SD625 runs only warm.

I have sourced a 2SD625 and it has now arrived so maybe this is a wild goose chase. Still it would be nice to know what is going on and also to have a suitable modern equivalent.

I will try your idea of lowering R514 and see if I can get it to work with lower gain transistors.

Thanks for the thought that 2SD625 is NOT a darlington. I assume the hfe is lower than the 120 (min) quoted by a well known web site on transistors.

Off to try an ordinary high voltage/current non darlington with reduced base drive resistor R514, and thanks for all the replies and help
C

[space_charged]

OK, well the set does work, a bit, with the KSC5027OTU as the LOT. I have to reduce R514 to about 1 ohm and also put a 2uF (non-electrolytic) across C520. The waveform is a lot better with none of the nasty ringing the darlington produced. I don't think the transistor is high enough gain, not saturating. It gets too hot, but a bit less quickly than the darlington did. It is on a heat sink, and one considerably bigger than the one the Sony 2SD625 was on.
The KSC50270 is listed as having a minimum hfe of 20.

[HamishBoxer]

I am afraid Sonys need their 100% correct transistors. Very critical I found over the years.

[Maarten]

Or 100% correct replacements. Sometimes they also provided replacements themselves, soetimes with small modifications. The likely reason for this is that Sony didn't really look at what parts were available when starting with a new design, but made an optimal design and only then looked for parts that would work in the design or probably only at that stage makes small compromises in the design. They sometimes used selected versions of existing parts, but in that case they were usually marked as such.

[Jac]

According to my Japanese data-book, the 2SD625 has an fT of 60 MHz.
Vcbo = 250V, Ic = 4A, no hfe in the table.

Hope this helps a bit.
Jac

[HamishBoxer]

Yes I meant that Maarten,Sony originals I should have stated that.

[space_charged]

Thanks for these thoughts. Yes it seems like the circuit is tightly designed around the 2SD625. Love the thought that it might also have been a specially selected 2SD625 - eeeek!
I do seem to remember from my holiday job in a radio and TV repair workshop that if it was a Sony then only the right/original part would do. I learned a lot from the experienced service engineers.

Thanks for the characteristics of the 2SD625. I got some data from the net and in particular from a well known site but I'm doubting the accuracy of some of what I read. In particular a minimum hfe of 120. Might have to resort to a characteristics plot.

Any thoughts on whether all the transistors in the set will be as critical?

[Maarten]

It may not have been specially selected in this case. Just a high gain high speed type which is rather tame compared to some later types. I don't have hands on experience with those earlier Sony models, but in later models the selection was indicated by a printed suffix (for example BU508AS instead of BU508A - the slightly better BU2508A worked just fine if I remember correctly). Also in that case, if you replace it with a later production example or with a slightly better equivalent, chances are you won't have a problem with it since it would meet the original selection criteria.

Also, I don't think every transistor is as critical. Sony may have had optimised designs, but they weren't completely daft. Horizontal output transistors tend to be somewhat critical regardless of the brand, anyway.

[space_charged]

Got the BU807 (high voltage fast darlington) today. Decided to give it a try, despite no longer being sure the 2SD625 has an hfe of 120. The BU807 is quoted as hfe of >100.

BIG surprise, it works. It gets very modestly warm and the overall current drawn by the set is down to 2.2A which is exactly as when the 2SD625 is fitted. The base waveform shows no signs of the parasitic ringing I've seen with all the other darlingtons I've tried.

So thanks to all for help and suggestions. Looks like a BU807 is a very good replacement for 2SD625, at least in my KV9000.

All this has been done on a previously working set. Now I'm going back to the non-working set and will fit a BU807. I have swapped the line output transformer on the non-working set with the working one so now I know both transformers are OK.

The non-working set has a field timebase fault. Both frame transistors get hot and the oscillator isn't running.