Better to change all of them, because they might fail in a not so nice way.
Meaning the entire lot has reached end of life.
Meaning the entire lot has reached end of life.
When one transistor in a parallel bank fails you change them ALL. Works that way on any amplifier. Cheap out at your own risk.
Question is that now the originals are nor available, so which ones to use?
A total conversion to Mosfet / Bipolar is essential, or a chip amp which matches the rails.
A total conversion to Mosfet / Bipolar is essential, or a chip amp which matches the rails.
I say give it a try.Thank you Naresh, I think we've established that. The question is:
"is there any disadvantage to doubling up the emitter resistor from .47 ohm to 1 ohm to control that excess current flow and allow me to bias that entire side up?"
Thanks!
Thanks, Nelson. It makes sense to me, but there's a lot that I don't know, and I wasn't sure if there was a specific known reason for why equalizing the current from that one stream would somehow affect things. It makes sense that in a production environment, it's more efficient to manufacture identical boards and match transistors in a test rig than to control current flow with individual emitter resistors, but ultimately, the goal seems to be to have equal current flow from each stream into the river.
"When one transistor in a parallel bank fails you change them ALL. Works that way on any amplifier. Cheap out at your own risk."
WG, you might read the rest of the thread and realize that that is exactly what we're working on here. We would love to hear your suggestions about what, exactly, we should change them ALL to...? 🤔
I got a strong nod towards the Exicon MOSFETs today from a helpful contributor today. The Vishay IRFP9140PBF still looks promising. Found a used oscilloscope that should get me started down this road with some real data.
Interesting story: I got the "new" version Model 6A boards installed and with much anticipation installed the FrankenAmp into the Mark Levinson/Magnepan 3.6R /system. Half way into the first piece of music, I had already arrived at my conclusion - the magic was gone. 😥
Over the next few hours, the data started to settle: the gain was significantly higher, but the sound was just so much more sterile and defined. And this was with the Silmic IIs installed in the input stage. I could definitely play it louder than the Model 6 boards before I started to hear distortion, but everything up to that point was just kind of "unemotional".
It occurred to me that Wayne had mentioned:
"A few things to note are it has no global feedback and uses Mikes variation of error correction like Cordell and Hawksford."
...when he provided the schematic of the 6A board. I wondered if popular demand had caused Michael to re-design more for distortion than purity of sound - On close inspection, you can see that the boards are definitely "physically" different with different circuit layouts. And the Model 6, wired with the 76V rails, did get a bad rap from reviewers. So Michael would have some motivation to make some changes.
So that's what spawned this new line of questions. I removed those boards after a few days of listening (and keeping them running on 6 ohm resistors overnight, just in case the new Silmics needed to "burn in") and pulled every IGBT out of every board I had. I was able to get a near perfect match-up for both Model 6 boards, except for that one transistor... (there's always one in every crowd, isn't there... 🙄) which brought us to this point.
An assortment of 3W carbon film resistors should be arriving today, so we'll plug it in and give it a try. I would rather have that silky sound back, with less volume. It'll help me concentrate on the MOSFET project. 😉🙂
"When one transistor in a parallel bank fails you change them ALL. Works that way on any amplifier. Cheap out at your own risk."
WG, you might read the rest of the thread and realize that that is exactly what we're working on here. We would love to hear your suggestions about what, exactly, we should change them ALL to...? 🤔
I got a strong nod towards the Exicon MOSFETs today from a helpful contributor today. The Vishay IRFP9140PBF still looks promising. Found a used oscilloscope that should get me started down this road with some real data.
Interesting story: I got the "new" version Model 6A boards installed and with much anticipation installed the FrankenAmp into the Mark Levinson/Magnepan 3.6R /system. Half way into the first piece of music, I had already arrived at my conclusion - the magic was gone. 😥
Over the next few hours, the data started to settle: the gain was significantly higher, but the sound was just so much more sterile and defined. And this was with the Silmic IIs installed in the input stage. I could definitely play it louder than the Model 6 boards before I started to hear distortion, but everything up to that point was just kind of "unemotional".
It occurred to me that Wayne had mentioned:
"A few things to note are it has no global feedback and uses Mikes variation of error correction like Cordell and Hawksford."
...when he provided the schematic of the 6A board. I wondered if popular demand had caused Michael to re-design more for distortion than purity of sound - On close inspection, you can see that the boards are definitely "physically" different with different circuit layouts. And the Model 6, wired with the 76V rails, did get a bad rap from reviewers. So Michael would have some motivation to make some changes.
So that's what spawned this new line of questions. I removed those boards after a few days of listening (and keeping them running on 6 ohm resistors overnight, just in case the new Silmics needed to "burn in") and pulled every IGBT out of every board I had. I was able to get a near perfect match-up for both Model 6 boards, except for that one transistor... (there's always one in every crowd, isn't there... 🙄) which brought us to this point.
An assortment of 3W carbon film resistors should be arriving today, so we'll plug it in and give it a try. I would rather have that silky sound back, with less volume. It'll help me concentrate on the MOSFET project. 😉🙂
OMG... it's baaaaaaak 
I installed the newly transistorized Model 6 boards (not 6A) back in the FrankenAmp and biased the hot spots on the heat sinks up to about 25 degrees (C) above ambient (It's about 10 degrees cooler in the house right now than it will be in 4 months).
Magic. It really is amazing how much difference a circuit can make. And the lesson is, when you find something you like, stay with it! 🤗
I installed the newly transistorized Model 6 boards (not 6A) back in the FrankenAmp and biased the hot spots on the heat sinks up to about 25 degrees (C) above ambient (It's about 10 degrees cooler in the house right now than it will be in 4 months).
Magic. It really is amazing how much difference a circuit can make. And the lesson is, when you find something you like, stay with it! 🤗
I tried to explain to my wife last night why I had been spending so much time in the tech room over the last few weeks, when we had a perfectly good Model 5 in the rack just a month ago.
How adding transistors would add current to the circuit, but transistors need to be matched, and by reading the voltage at the emitter resistors we can add current to the output by biasing up into class A, which would prevent the switching distortion into class B - but only up to the point where thermal runaway might pose a problem, and changing the coupling capacitors to a silk-impregnated hemp would dampen the....
...and (as Arlo would say) she STOPPED me... RIGHT.... THERE - and said: "It sounds like butter again".
Huh. Must be the Teflon. 🤔 😅
How adding transistors would add current to the circuit, but transistors need to be matched, and by reading the voltage at the emitter resistors we can add current to the output by biasing up into class A, which would prevent the switching distortion into class B - but only up to the point where thermal runaway might pose a problem, and changing the coupling capacitors to a silk-impregnated hemp would dampen the....
...and (as Arlo would say) she STOPPED me... RIGHT.... THERE - and said: "It sounds like butter again".
Huh. Must be the Teflon. 🤔 😅
Last edited: