transistor matching in Kenwood KR-5030

robthomp

New Member
Hi,

I just bought a Kenwood KR-5030 from a guy for cheap because it "didn't work". ( I'm new to amplifier repair/restoration, but I've been an electronic hobbyist on and off over the years.) So far I've fixed the power switch - a well known issue with this amp, cleaned the controls, and cleaned up the relay in the protection circuit which was too crudded up to function. I replaced the caps in the power supply, including the large filter caps. It's working pretty well and sounding good so far.

My question: I've got non-zero DC offset - 15 mV in one channel, 35 mV in the other. I bought a bag of 2SA750's from the auction site, that's the transistor in the input differential pairs, and I'd like to find two tightly matched pairs and put them in. Do I try to match VBE or hFE, or both?

I have tried to read up on this question before posting, but opinions seem to vary a lot and the answers seem to depend on the situation.

Thanks,

Rob
 
I'd probably throw the eBay transistors in the trash. Most likely counterfeit. When the time comes for transistors use a modern substitute available form an authorized distributor like Mouser, Digikey, etc.
 
I recently got good results on matching the inputs and mirrors something like +/- 2mv DCO on a non-adjustable machine.

Change the inputs 2SA750 (Q1-4) with KSA992FTA and the 2SC1775's (Q5-8) with KSC1845FTA

Get a small pile of each (like 20 cents each or less when 10+ ordered) and match 8 trans, 4 KSAs, 4 KSCs as close as possible. The F suffixes are pretty high gain as the old ones specced (E or F suffix) were also.

In the DC offset thread IIR 25 mv is the cutoff between dont worry and maybe perhaps look into it. So you have both cases. Matched, with the bias set this amp will sound as good as it will ever. To be thorough, Id replace all the 40 year old small caps in the audio path as well
 
I'd probably throw the eBay transistors in the trash. Most likely counterfeit. When the time comes for transistors use a modern substitute available form an authorized distributor like Mouser, Digikey, etc.

Yeah, I was wondering about that. The order hasn't shipped yet, I'm going to try to cancel it.
 
I recently got good results on matching the inputs and mirrors something like +/- 2mv DCO on a non-adjustable machine.

Change the inputs 2SA750 (Q1-4) with KSA992FTA and the 2SC1775's (Q5-8) with KSC1845FTA

Get a small pile of each (like 20 cents each or less when 10+ ordered) and match 8 trans, 4 KSAs, 4 KSCs as close as possible. The F suffixes are pretty high gain as the old ones specced (E or F suffix) were also.

In the DC offset thread IIR 25 mv is the cutoff between dont worry and maybe perhaps look into it. So you have both cases. Matched, with the bias set this amp will sound as good as it will ever. To be thorough, Id replace all the 40 year old small caps in the audio path as well

Thank you for the detailed suggestions. I see those transistors are readily available and inexpensive. I'm going to add them to an order I'm putting together with caps for the audio board.

When you match in this situation, do you match the Vbe or the hFE?

Rob
 
If you bought the 2SA750 from LF Components, then they are most-likely the real deal. However, the KSA992 will do the job considerably cheaper.
 
Thank you for the detailed suggestions. I see those transistors are readily available and inexpensive. I'm going to add them to an order I'm putting together with caps for the audio board.

When you match in this situation, do you match the Vbe or the hFE?

Rob
In my experience I have found that a lot of new modern digikey/mouser sourced stuff, tends to differ more %-age wise in the HFE than Vbe. Meaning lets say those FTA trans I gave you are in the 400-500 range, if I pull in HFEs that are only a few single digits away, the Vbe difference is negligible. Im sure the reason why is in my old college notes but the converse, if Vbe matches can I discard Hfe? I don't believe is true at all. If Vbe is dead on exact and Hfe1 is 300 and Hfe2 is 100, there will be an issue. (I do recall that particular lab where we generated the Hfe 'curves' from like 20 data points to just prove a point I am sure.)

I stupidly line them up on the desk until I saw a post where someone tests them in the 'cut-tape' and writes the hfe on the tape. brilliant!
 
In my experience I have found that a lot of new modern digikey/mouser sourced stuff, tends to differ more %-age wise in the HFE than Vbe. Meaning lets say those FTA trans I gave you are in the 400-500 range, if I pull in HFEs that are only a few single digits away, the Vbe difference is negligible. Im sure the reason why is in my old college notes but the converse, if Vbe matches can I discard Hfe? I don't believe is true at all. If Vbe is dead on exact and Hfe1 is 300 and Hfe2 is 100, there will be an issue. (I do recall that particular lab where we generated the Hfe 'curves' from like 20 data points to just prove a point I am sure.)

I stupidly line them up on the desk until I saw a post where someone tests them in the 'cut-tape' and writes the hfe on the tape. brilliant!
Thank you! This is extremely helpful with the hFE vs VBE question. I love claims that are supported by both theory and empirical evidence.
 
If you bought the 2SA750 from LF Components, then they are most-likely the real deal. However, the KSA992 will do the job considerably cheaper.
Yes, it is LF Components. Your comment is very helpful, I will keep that in mind. It's interesting, the KSA992 are about one-quarter the cost. So I wonder why there is a market for the original SA750, assuming they are authentic. I suppose some people find that they do sound different, with a more original and authentic sound? I wonder if I would be able to tell? Also, I know that many people are committed to NOS and accurate reproductions on principle.
 
So I wonder why there is a market for the original SA750, assuming they are authentic
Many techs, including some supposed 'pros', know little to nothing about transistor substitution and if the original part is not available they have no idea where to turn (with the possible exception of an NTE substitution book).
 
Yes, it is LF Components. Your comment is very helpful, I will keep that in mind. It's interesting, the KSA992 are about one-quarter the cost. So I wonder why there is a market for the original SA750, assuming they are authentic. I suppose some people find that they do sound different, with a more original and authentic sound? I wonder if I would be able to tell? Also, I know that many people are committed to NOS and accurate reproductions on principle.
for the same reason people use interconnects only made from copper mined in new zealand and not austrailia....
 
...I stupidly line them up on the desk until I saw a post where someone tests them in the 'cut-tape' and writes the hfe on the tape. brilliant!...

Same here- I used to use ice block trays for Hfe sorting until I saw handwriting on the 'ammo belt' tape. Doh. :)
 
Well, the caps and transistors arrived from Digi-Key. I recapped the audio board and also replaced the bias adjustment trimmer pots. The originals were very disfunctional and I couldn't really set the bias.

With the new pots I was able to easily adjust the bias. The amp sounds fantastic. I do not consider myself to be an audiophile, and my hearing has issues (probably from too many rock concerts when I was young). But even I can tell that there is a day and night difference between now and the way it sounded before I recapped the power supply, audio board, new trimmer pots and cleaned everything. Great sounding amp, and so much power!

Now, a little bad news - after recapping the audio board and replacing trimmer pots, the DC offset went up a little. It was 15mV/35mV, now it's 20/40mV. So next step will be to relplace the eight transistors discussed above.
 
when you do, for grins, and this is busy work so feel free to say no, as you pull each, write down its location and hfe and even Vbe if you are feeling froggy. then write down the same for the stuff going in. when done, measure DCO and the difference we can equate to Hfe differences and actually come up with a differential equation to describe the drive section as a black box. I know, math geek BS but its a learning experience.
 
when you do, for grins, and this is busy work so feel free to say no, as you pull each, write down its location and hfe and even Vbe if you are feeling froggy. then write down the same for the stuff going in. when done, measure DCO and the difference we can equate to Hfe differences and actually come up with a differential equation to describe the drive section as a black box. I know, math geek BS but its a learning experience.
Sure, why not, I'm a big fan of collecting data. Would it be more informative to do this first for Q1-Q4, measure the DCO, then proceed to do Q5-Q6, or should I just do all Q1-Q8 at once?
 
Back
Top Bottom