New Transistor variability marantz 18

[Douglie]

I bought some new transistors for the model 18. They are the ones that precede the final audio power transistors. I got a batch of three NPN and five PNP because that’s how they were sold. There is a very large variation in the hfe values determined by my Peak semiconductor analyzer.

PNP-2N5195 hfe= 67, 92, 124, 133, 164
NPN-2N5192 hfe= 34, 48, 53

This seems unreasonable but I know little about transistors.

I also got six KSC3503 for the scope. They measured hfe= 82, 82, 83, 83, 83, 84. This seems much more reasonable and I installed four of these and they fixed the scope problem.

I also bought two sets of the power transistors. The NPN were 84 and 98. And the PNP were 95 and 120. I used the 98 and 95. This variation seems high but it seems I lucked out. They’re installed and sound great and look perfect on the oscilloscope. The PNP I replaced was 149, the NPN was dead.

I’m afraid to use the high variability ones especially since it’s running good. I’d like to get input from the group before I go back to the vendors.

Thanks.

 

[ChrisMarantz]

Transistors are typically grouped into "Gain Ranges" - Back in the day, they would be sorted into Color Groups (ie. Red 1 - 10 Blue 11 - 20 Green 21 - 30) There would be a dab of paint of the Transistor Case to indicate their group. Later suffix Letters we used and printed on the Transistor Case.

Typically, as long as you are choosing Transistor pairs within the same Gain Range - they will be fine. If you are choosing 2 pairs - 1 for each channel - each pair should be in the same Gain Range. Most service manuals will have a recommended Gain Range for their application.

According to the Onsemi Data Sheet for both or the Transistors you mentioned (with "G" suffix) - the Hfe is shown as a min. of 20 and max of 80 - a fairly wide range. If you choose the middle ground of 50 as you target - you may need a larger group Transistors to sort through

Chris

 

[Douglie]

Thanks, Chris. That helps immensely. I couldn’t find hfe values on the data sheets I looked at. Are the pieces that are >80 out of spec?

Also I replaced two little FETs on the scope board. They had a different pinout than what was in there. (I got the ones that are listed on the forum page dealing with model 18 part numbers). I bent the legs so they would fit and they work fine. However the little round green thing that fits on the legs fits a little cockeyed now. What is that? Thanks, Doug

 

[Douglie]

Chris - Update - the hfe was on the data sheet I was looking for and the range is 20-80 like you said. Seems like I should be able to exchange the >80 ones, right? Three of them aren’t even close.

Also, for what it’s worth, one FET I replaced had a white dot on it, the other had a yellow.

 

[ChrisMarantz]

Seems like I should be able to exchange the >80 ones, right? Three of them aren’t even close.

Are they OnSemi Transistors or are they mystery Transistors? If they are mystery Transistors, you can't know what their Hfe range is suppose to be. Just because OnSemi makes only 1 type of these Transistors, it isn't to say that another manufacturer did make them in a selection of gain ranges - Chris

 

[Skywatcher]

I bought some new transistors for the model 18. They are the ones that precede the final audio power transistors. I got a batch of three NPN and five PNP because that’s how they were sold. There is a very large variation in the hfe values determined by my Peak semiconductor analyzer.

PNP-2N5195 hfe= 67, 92, 124, 133, 164
NPN-2N5192 hfe= 34, 48, 53

This seems unreasonable but I know little about transistors.

I also got six KSC3503 for the scope. They measured hfe= 82, 82, 83, 83, 83, 84. This seems much more reasonable and I installed four of these and they fixed the scope problem.

I also bought two sets of the power transistors. The NPN were 84 and 98. And the PNP were 95 and 120. I used the 98 and 95. This variation seems high but it seems I lucked out. They’re installed and sound great and look perfect on the oscilloscope. The PNP I replaced was 149, the NPN was dead.

I’m afraid to use the high variability ones especially since it’s running good. I’d like to get input from the group before I go back to the vendors.

Thanks.

Where did you get those transistors? I hope not FleaBay because it is flooded with counterfeits. A knowledgeable seller usually (but not always) doesn't sell in "odd" quantities like that, especially for stereo audio. Although a lot of the counterfeit peddlers have caught on to the "in pairs" game. Those 5195 with over 100 hFE are a red flag.

 

[Douglie]

Yes they are mystery transistors from the bay. Another lesson learned. Where is a good place to buy? The KSC3503 I mention above are Onsemi and very close in hfe.
Question: on the Onsemi datasheet they have different “hfe classification” categories. Do you know if they make a big batch and then sort by hfe? Thanks

 

[Watthour]

In my experience with large batches of OnSemi (Fairchild), Siemens, and random lots of other devices, even within a manufacturing run of devices within the same gain classification, measured gain can vary over 30%. Many reel/tape packaged lots have varied this way on the same reel, and very randomly, and I presume that these are all within the same production run of devices.

I find it better to measure devices and match them as closely as possible (within a few points) to maintain balance. Since meters can vary widely due to test voltage and frequency conditions, the absolute number is not as critical as the relative number. Any two devices tested on the same meter and yielding similar hFE numbers should be a safe "matched pair" for practical use.

 

[Douglie]

Thanks very much for sharing your experiences. My transistor knowledge has increased exponentially. The devil’s in the details....