[amb]

Quote:

Originally Posted by Pars I was hoping amb or Dan might pop up with their opinions .

LOL

Pars is correct. If you could take a measurement quickly and consistently each time without appreciably heating up the transistor, then you don't need to heatsink it. However, putting them on a heatsink won't hurt either. The key is whatever you do, do the same thing for each one.

 

[bryerd]

I've tried both routes (or really all three of the below).

1. Tried measuring the current draw on loose transistors (no heat sink). Found that the currents changed rather quickly and could not get an accurate reading by simply counting "1 Mississippi 2 Mississippi...". This was enough to drive me bonkers.

2. Decided to record the current draw using a stopwatch at 10 seconds, 30 seconds, and after 60 seconds. While the current draw did increase for each successive time, once plotted, the curves for all three time durations were very similar in shape (see attachment). Conclusion: actual time was not that important for matching, but using a stopwatch was very helpful (so was the hold function on my DMM). However, 60 seconds was excessive to wait when measuring numerous transistors, and they get quite hot!

3. Went back and remeasured with the transistors mounted to a heat sink. All I did was use a clothespin (wooden kind with the spring) to keep the back of the transistor tight against a piece of metal (I used my actual heatsink with a piece of Sil-pad, but a 1/8" thick 1x1" aluminum angle should work too). Upon applying the voltage, the current quickly converged and did not increase or fluctuate at all. No stopwatch was needed, and yet good, repeatable results were obtained within seconds. The clothespin also made it very easy to change transistors as well.

Good luck,
Dave

 

[Pars]

Quote:

Originally Posted by bryerd I've tried both routes (or really all three of the below). <snip> 3. Went back and remeasured with the transistors mounted to a heat sink. All I did was use a clothespin (wooden kind with the spring) to keep the back of the transistor tight against a piece of metal (I used my actual heatsink with a piece of Sil-pad, but a 1/8" thick 1x1" aluminum angle should work too). Upon applying the voltage, the current quickly converged and did not increase or fluctuate at all. No stopwatch was needed, and yet good, repeatable results were obtained within seconds. The clothespin also made it very easy to change transistors as well. Good luck, Dave

Good idea!

 

[Kruemelix]

Hello,

finally got it to match all the transistors. I checked some of the 1358/3421 a second and third time with the same method quoted above. Around 90% of the transistors were in the +/-1 range, the last 10% in the +/-2 range. IMHO that should be accurate enough. Furthermore I doubt that you will get a much better result with the clothespin method without thermal paste (what's the correct word for?). Can just someone post me a expedient constellation via copy/paste for four boards?

2SC3421
134
136
137
138
139
139
140
141
141
142
143
143
143
144
144
144
144
144
144
145
145
145
147
147
148
150
151
151
151
153
154
154
154
154
154
155
155
156
156
156
158
159
160
161
161
162
166
169

2SA1358
195
196
197
197
199
199
199
200
200
201
202
203
203
203
203
203
204
204
204
205
205
205
205
206
206
206
206
206
207
207
208
210
211
211
211
212
212
213
213
213
213
213
214
214
215
216
223

2SA1145
145
146
146
146
146
147
147
147
148
148
149
149
149
150
151
151,5
152,5
153,5
154,5
155,5
167
170
172
173
173,5
174
175,5
176,5
177,5
177,5
178,5
179,5
180,5
181
185
186

2SC2705
150,5
153
155
155
157,5
158,5
159
160
162
162
163
163
164,5
164,5
165,5
165,5
170
170
170
172
172
172,5
173,5
173,5
175
178
179
181,5
181,5
182,5
183
183

2SC3381
left-right
397-397
409-411
415-415
416-415
420-415
421-421
435-433
447-447
453-451
457-453
533-521
591-581

2SA1349
left-right
477-474
477-479
484-479
489-494
493-499
494-504
496-499
511-521
529-529
546-551
574-579
584-584

Thanks,

Thomas

 

[dip16amp]

2SC3421 154, 154, 154, 154
2SA1358 206, 206, 206, 206
2SA1145 181, 180.5, 179.5
2SC2705 182.5, 181.5, 181.5
2SC3381 397-397
2SA1349 477-479

2SC3421 150, 151, 151, 151
2SA1358 205, 205, 205, 206
2SA1145 178.5, 177.5, 177.5
2SC2705 179, 178, 175
2SC3381 415-415
2SA1349 496-499

2SC3421 144, 145, 145, 145
2SA1358 203, 204, 204, 204
2SA1145 174, 173.5, 173
2SC2705 173.5, 173.5, 172.5
2SC3381 421-421
2SA1349 529-529

2SC3421 144, 144, 144, 144
2SA1358 203, 203, 203, 203
2SA1145 155.5, 154.5, 153.5
2SC2705 155, 155, 153
2SC3381 447-447
2SA1349 584-584

 

[DigiPete]

Yeah,

I just did the 15 seconds on a stopwatch (no heatsink) method.
It worked just fine, and saves a bit of time not having to mount transistors for the check.

Pete


Quote:

Originally Posted by bryerd I've tried both routes (or really all three of the below). 1. Tried measuring the current draw on loose transistors (no heat sink). Found that the currents changed rather quickly and could not get an accurate reading by simply counting "1 Mississippi 2 Mississippi...". This was enough to drive me bonkers. 2. Decided to record the current draw using a stopwatch at 10 seconds, 30 seconds, and after 60 seconds. While the current draw did increase for each successive time, once plotted, the curves for all three time durations were very similar in shape (see attachment). Conclusion: actual time was not that important for matching, but using a stopwatch was very helpful (so was the hold function on my DMM). However, 60 seconds was excessive to wait when measuring numerous transistors, and they get quite hot! 3. Went back and remeasured with the transistors mounted to a heat sink. All I did was use a clothespin (wooden kind with the spring) to keep the back of the transistor tight against a piece of metal (I used my actual heatsink with a piece of Sil-pad, but a 1/8" thick 1x1" aluminum angle should work too). Upon applying the voltage, the current quickly converged and did not increase or fluctuate at all. No stopwatch was needed, and yet good, repeatable results were obtained within seconds. The clothespin also made it very easy to change transistors as well. Good luck, Dave

 

[Kruemelix]

Hello,

I just realized that I measured the 1145/2705 only at 2,5 mA - will this be a problem as they are working at 1 mA and 15 mA?

Thanks,

Thomas

P.S. is dip16amp's suggestion correct this way? Or are there better possibilities?

 

[Pars]

Should be OK as these transistors are quite linear and won't change much at a higher current.

Dip16amp has built a few of these, and his recommendations seem solid without extensively going thru your list... I don't see any other suggestions anyhow.

Are you planning on socketing the transitors/FETs? The FETs in particular aren't easy to desolder/remove/swap, so at a minimum I would socket those. Not sure whether you can socket the output transitors or not, but the rest of them I would if it were me doing it. Its educational at the least to try to predict what will happen with offset if you put a certain hFE value in a certain position, and allows you to quickly verify the effects on offset to bring them down as close as possible to 0.0mV without the servos, and without cranking the pots. This will result in an amp that is as linear as possible.

 

[j4cbo]

Hmm...

I attempted to measure my 2SA1145/2SC2705s, using a fixed 470k resistor and two meters. The collector current ran through an analog VOM, and I simultaneously measured the voltage drop across the resistor with a VTVM. (My test equipment is kinda old.)

The 2SC2705s wound up in the 110-137 gain range, and the 2SA1145s are all between 151 and 157.

Did I mess up the measurement somehow?

 

[amb]

Quote:

Originally Posted by j4cbo The 2SC2705s wound up in the 110-137 gain range, and the 2SA1145s are all between 151 and 157.

These agree with my own experience with these transistors, the PNP devices seem to have higher Hfe on average. Don't sweat it too much, it's not too bad.

 

[j4cbo]

Quote:

Originally Posted by amb These agree with my own experience with these transistors, the PNP devices seem to have higher Hfe on average. Don't sweat it too much, it's not too bad.

Gotcha. Should I match low with low and high with high, or just not bother?

(I did do a rough match on the LEDs.)

Next comes the drive transistors, and there's a problem. BDEnt sent me 2SC3421s all in the "Y" gain variety, but all but six of the 2SA1358s are the "O" type. So I've got six pairs of Ys, and another fourteen 2SA1358Os.

I'm thinking I'll go through all the 2SA1358Os, do a quick gain test, and pick the highest two. Fortunately I've got the PNP/NPN split difference working on my side in this case.

 

[j4cbo]

Anyone? :/

 

[Pars]

Well, you've got two options: 1) you can send the "O"s back to bdent to have them replaced with "Y"s, or 2) you can try it. I haven't ordered from bdent before (always been lucky... and I do mean LUCKY on gain ranges with mcm), but from what I understand, you really need to call them when you order and make sure they send matched gain ranges. If you can quickly check them and get a baseline idea of where they are at compared to your PNPs (can't remember which is which... SA=PNP/ SC=NPN or the other way around), and they match up OK, I'd probably just go with it. Many of the types there is a considerable difference between the NPN/PNPs in terms of where the gains fall.

 

[Edwood]

Anyone considering building a Dynahi or two, or need lots more parts?

I'm giving up on the Dynahi (before I even started).

I bought enough parts to build almost two of them. Right down to the Case, Heatsinks, and all the electronics parts.

If anyone is interested I'm posting them up for sale. Let's just say my home has to be more than ROHS compliant because of a pregnant wife.

http://www6.head-fi.org/forums/showthread.php?t=203666

-Ed

 

[dhp]

Quote:

Originally Posted by Edwood Anyone considering building a Dynahi or two, or need lots more parts? I'm giving up on the Dynahi (before I even started). I bought enough parts to build almost two of them. Right down to the Case, Heatsinks, and all the electronics parts. If anyone is interested I'm posting them up for sale. Let's just say my home has to be more than ROHS compliant because of a pregnant wife. http://www6.head-fi.org/forums/showthread.php?t=203666 -Ed