Thanks, I already replaced those transistors, and all but chip resistors in the dc servo circuit. I found a faulty weld and hardwired pins 8 and 9 to R157. Also, the dc offset oscillation at the speaker terminals is just a few mV which I think is expected given the higher voltage going to the dc servo in the right versus the left channel: U101(pin 1 WRT ground = 300mV); U1(pin 1 = 15mV). What I'm trying to understand is if that higher voltage and the higher Offset voltage at U101 pin 8 and 9 = 150mV, is just fine, might cause issues with other components, or is a sign something is awry somewhere else.
Last edited:
DC Voltage differences within the servo circuits only indicate that they're doing their job of forcing a DC balance so that there is there is virtually zero DC at the output, as you find. Matching of the LTP pair is usually done for hFE or gain balance, which is a separate issue that affects the signal regardless of DC levels.
There's a manufacturing problem with matching SMD semis though, since they are auto-attached from reel tapes of thousands of components to the boards during assembly and you can't assign particular pairs to any location that way. That's why I doubt the LTPs were a matched pair unless they were hand-matched and fitted after the board was finished and tested - perhaps in a special test/handwork jig. That's possible but probably increases the costs more than it bestows any real benefit since matching between any two transistors taken in production sequence from a wafer, is nowadays going to be as good as it gets. Final testing will show up discrepancies that may need correcting and so long as there aren't too many, the cost could be absorbed.
In other words, if you're repairing something like this, you probably need to address the same part/sequence matters by buying "cut tape" quantities and selecting pairs by their sequence, rather than anything more.
There's a manufacturing problem with matching SMD semis though, since they are auto-attached from reel tapes of thousands of components to the boards during assembly and you can't assign particular pairs to any location that way. That's why I doubt the LTPs were a matched pair unless they were hand-matched and fitted after the board was finished and tested - perhaps in a special test/handwork jig. That's possible but probably increases the costs more than it bestows any real benefit since matching between any two transistors taken in production sequence from a wafer, is nowadays going to be as good as it gets. Final testing will show up discrepancies that may need correcting and so long as there aren't too many, the cost could be absorbed.
In other words, if you're repairing something like this, you probably need to address the same part/sequence matters by buying "cut tape" quantities and selecting pairs by their sequence, rather than anything more.
DC Voltage differences within the servo circuits only indicate that they're doing their job of forcing a DC balance so that there is there is virtually zero DC at the output, as you find. Matching of the LTP pair is usually done for hFE or gain balance, which is a separate issue that affects the signal regardless of DC levels.
LTP hFE1/hFE2 are at 99%, Vbe1/Vbe2 are around 95%. If the LTPs are mismatched then we would have more offset, correct? The dc servo voltages increased with the new LTPs, I mistakingly attributed the dc servo voltage increase as an indication of how well-matched the LTPs are. Mixing apples and oranges so to speak. Got it 🙂 The dc servo is doing its job just fine and this has been a reminder to revisit fundamentals of electronic circuits And the scientific method before replacing components. Thanks again for help.
Last edited:
I don't have the engineering confidence to confirm everything there but I think you summed up pretty well, what's been discussed so far. I'd continue with reassembly, sticking to a general rule of picking semis that are paired in mirrored or balanced circuit formats, in their manufactured sequence. Loose semis may be a problem but simple resistors and caps are surely easy enough to measure and match if its really needed. Keep hands off when taking measurements and test that you aren't inadvertently bridging pads with solder - easy to do with handwork, once your parts format shrinks down as small as it can be now.
Yes very small. I frequently use the magnifying app on my smart phone to check for any shorting. I think it would be exhausting finding a better matched complementary bc856b/bc846b than hFE1/hFE2=99%, and Vbe1/Vbe2= 95%, which I put into the right channel. We discussed how wether or not better matched complementary LTPs can affect sound quality, and I've read reports that either matched or poorly-matched LTPs can reduce or increase 2nd or 3rd order harmonics in certain types of circuit designs. To test this theory on my particular input stage I ran a 200 Hz violin waveform (available at - https://meettechniek.info/additional/additive-synthesis.html) from my mac mini -> dac -> Destiny -> headphone output -> mac mini line in. I used Audacity to run the spectrum analysis. The results are exciting: Except for a louder 60 hz signal in the right channel which I can faintly hear when I put my ear to the speaker, I found L/R symmetry in the amplitudes of the harmonics!
1. The semis I replaced in the right channel did not affect the 2nd and 3rd order harmonics.
2. Creek Audio probably wasn't able to match the stock LTPs as closely as I did in the right channel.
3. Replacing the semis in the right channel had no affect on sound quality based this analysis.
This might be a decent tool to analyze for harmonic distortion.
1. The semis I replaced in the right channel did not affect the 2nd and 3rd order harmonics.
2. Creek Audio probably wasn't able to match the stock LTPs as closely as I did in the right channel.
3. Replacing the semis in the right channel had no affect on sound quality based this analysis.
This might be a decent tool to analyze for harmonic distortion.
That's an interesting spectral display comparison - might look at trying it and a few variations myself. Thanks for sharing the details.
On the music content side of things, I've been listening to a recent concert recording of (of all things) the score for an electronic game, but a game like no other..... I'm a bit of a (gentle) gaming enthusiast and noticed that Bethesda studios recently sponsored a 10th anniversary concert for their immortal blockbuster production "Skyrim". It features the LSO and choir at full tilt in a full, live recording session.
If you like orchestral music, soaring voices and epic themes, this is one mighty work - brings me and apparently many others to tears with the beauty and joy of the music and production which is a full length, formal attire occasion. Anyways, the comments tell the merits of this production better than I can, for anyone interested in a real sonic tour de force of a modern classic.
I may have overstepped the bounds of links in posts here but you should be able to search a link under "10th anniversary concert...etc." if necessary.
On the music content side of things, I've been listening to a recent concert recording of (of all things) the score for an electronic game, but a game like no other..... I'm a bit of a (gentle) gaming enthusiast and noticed that Bethesda studios recently sponsored a 10th anniversary concert for their immortal blockbuster production "Skyrim". It features the LSO and choir at full tilt in a full, live recording session.
If you like orchestral music, soaring voices and epic themes, this is one mighty work - brings me and apparently many others to tears with the beauty and joy of the music and production which is a full length, formal attire occasion. Anyways, the comments tell the merits of this production better than I can, for anyone interested in a real sonic tour de force of a modern classic.
Last edited:
Thanks for sending music. I enjoyed listening on my Destiny. I've never heard of Skyrim but plan to try and play it at some point.
'Glad you enjoyed it! I imagine it adds something to the rewarding feeling of getting your Destiny amp back to work too. I'm still overwhelmed by the rich choral harmonies and I'll probably over-indulge, as usual, in listening until tired of it but that's how it goes.
Idle DC offset spikes in the right channel were finally fixed after resoldering several SMD resistors including R101 and R102.
Correction… idle dc offset spikes at this very low magnitude ( a few mV) are caused by drafts over the input transistors. Simply blowing through the heat sink vent caused a 5mV voltage spike. it appears that less dc offset stability in the right channel can be attributed to more airflow over the right channel input transistors, which are situated next to the left/right channel heatsink gap. Left channel input transistors are near the front panel where air seems to be more stagnant.