What are some good 2W or so complimentary BJT SMD devices that work up to 100V? I'm looking for something to take 300mW-1W continuous that is relatively fast.
PS - I don't believe everything I hear about mental decline with age. There was some research done that shows that older people are slower NOT because they are in decline, but simply because they have more experience and information to process when they make a decision. Of course, a person that doesn't want to grow will largely stay the same. It takes a certain amount of personal effort and suffering to really change yourself, which is a lot to ask.
PS - I don't believe everything I hear about mental decline with age. There was some research done that shows that older people are slower NOT because they are in decline, but simply because they have more experience and information to process when they make a decision. Of course, a person that doesn't want to grow will largely stay the same. It takes a certain amount of personal effort and suffering to really change yourself, which is a lot to ask.
I know the through-hole parts all very well. I'm investigating something that can be assembled at the factory with little extra labor needed.
I found these:
http://www.mouser.com/ds/2/115/ZXTP25100CZ-75501.pdf
http://www.mouser.com/ds/2/115/ZXTN25100DZ-95102.pdf
But I want to see more options. Finding something with good Vcesat behavior would be a plus.
I found these:
http://www.mouser.com/ds/2/115/ZXTP25100CZ-75501.pdf
http://www.mouser.com/ds/2/115/ZXTN25100DZ-95102.pdf
But I want to see more options. Finding something with good Vcesat behavior would be a plus.
Hi dvv,
I also discovered that most "talked up high end" products were often substantially inferior to products produced by major players. Their top and near top of the line special products are better designed and built compared to the garage guys. No surprises there.
One thought I have on the superiority of J-Fets has to do with the input signal being pre-distorted before it even enters the amplifier. The rest may have more to do with the transfer characteristics of a J-Fet, but you can lower the gain of a BJT stage through degeneration. Those can sound pretty good also. In the end, for me it is just easier to use J-Fets. There are also studies on the susceptibility of BJT stages to RF energy and the fact that they naturally rectify some of this noise. That's difficult to quantify in real home use situations. I think your 170DC uses a uPA68H dual J-Fet.
Replace the ceramic and polyester capacitors in the signal path with mica where temperatures are higher, polystyrene in the input section. Use film / foil capacitors to replace the low value electrolytic capacitors.
Replace the degeneration resistors with matched metal film types. Use higher rated power metal film resistors for the feedback loop resistors.
Match the BJT transistors in the signal path. If complimentary, try and match the beta. Otherwise you should be able to match beta and Vbe (Vbe is not important in my view).
Check the match on the uPC68H, replace with a pair of J-Fets in the usual way.
Clean up solder joints, but you probably got them all when removing the parts for testing / replacement.
You should now hear a dramatic improvement in sound quality. I kid you not. That little commercially produced 170DC is now a giant killer. Now, you can really enjoy the music. The 3250 responds well to the same treatment.
-Chris
I came to that same conclusion many years ago also. I had the advantage of having top line equipment crossing my bench every day, so I could compare amplifier sound quality easily. The more interesting ones I took home to compare with my 300DC. In this way it was fairly easy to discover general similarities as I also had the schematics / manuals.
I also discovered that most "talked up high end" products were often substantially inferior to products produced by major players. Their top and near top of the line special products are better designed and built compared to the garage guys. No surprises there.
That's easy to answer. You were focused on other aspects.
One thought I have on the superiority of J-Fets has to do with the input signal being pre-distorted before it even enters the amplifier. The rest may have more to do with the transfer characteristics of a J-Fet, but you can lower the gain of a BJT stage through degeneration. Those can sound pretty good also. In the end, for me it is just easier to use J-Fets. There are also studies on the susceptibility of BJT stages to RF energy and the fact that they naturally rectify some of this noise. That's difficult to quantify in real home use situations. I think your 170DC uses a uPA68H dual J-Fet.
You sold your 170DC, didn't you? If you still have it, it is time to sort it out with all the knowledge from today.
Replace the ceramic and polyester capacitors in the signal path with mica where temperatures are higher, polystyrene in the input section. Use film / foil capacitors to replace the low value electrolytic capacitors.
Replace the degeneration resistors with matched metal film types. Use higher rated power metal film resistors for the feedback loop resistors.
Match the BJT transistors in the signal path. If complimentary, try and match the beta. Otherwise you should be able to match beta and Vbe (Vbe is not important in my view).
Check the match on the uPC68H, replace with a pair of J-Fets in the usual way.
Clean up solder joints, but you probably got them all when removing the parts for testing / replacement.
You should now hear a dramatic improvement in sound quality. I kid you not. That little commercially produced 170DC is now a giant killer. Now, you can really enjoy the music. The 3250 responds well to the same treatment.
-Chris
George,
Thank you very much for that insight about starting with something simple and your recommendations. I always seem to learn something when you get involved with the conversation. I will get my hands on a mini-dsp and see how much I can screw things up! I've built some nice oscillating networks in analog before, why not blow something up digitally, just kidding of course but it will be par for the course as they say. Hands on learning most of the time seems to be more effective than just learning theory and not understanding the real applications and limitations.
Vacuphille,
I just don't see why you would want to develop a dsp solution to a crossover application and then revert to an analog solution? There is no way around the effects of discrete capacitors and coils in a passive or even an active xo once you revert to classic crossover design, I can easily go back to those concepts, they have been beat to death by now, so many know issues that have no analog solutions. Doing any type of time alignment or phase correction in the analog domain seems to be wrought with issues that really have no practical solutions. I never thought I would come to these conclusions but there is just no way to disregard current development, either at a very high level by someone like Scott W could easily implement or many other amateurs here on this website.
I just have to keep my head in the books and reading all I can and also as George just suggested start off with something simple and learn and evolve from there. I am also not a believer that as we age we stop learning or become less effective than a younger person. What I will say is that when we are young we typically don't have the background knowledge and have not become tainted by prejudices in our thought process.
When I was young I say I was to ignorant to that fact, I had no fears of what you learn later can be a disaster and was willing to take a chance on something that made no sense, you just went ahead and tried something, you learned a lesson of what no to do again or discovered that what you were told wasn't necessarily gospel, it was just opinion. I know I learned more by making mistakes than when everything just worked and you could forget about trying something new. Disasters and discovery where the best teachers in fact.
Thank you very much for that insight about starting with something simple and your recommendations. I always seem to learn something when you get involved with the conversation. I will get my hands on a mini-dsp and see how much I can screw things up! I've built some nice oscillating networks in analog before, why not blow something up digitally, just kidding of course but it will be par for the course as they say. Hands on learning most of the time seems to be more effective than just learning theory and not understanding the real applications and limitations.
Vacuphille,
I just don't see why you would want to develop a dsp solution to a crossover application and then revert to an analog solution? There is no way around the effects of discrete capacitors and coils in a passive or even an active xo once you revert to classic crossover design, I can easily go back to those concepts, they have been beat to death by now, so many know issues that have no analog solutions. Doing any type of time alignment or phase correction in the analog domain seems to be wrought with issues that really have no practical solutions. I never thought I would come to these conclusions but there is just no way to disregard current development, either at a very high level by someone like Scott W could easily implement or many other amateurs here on this website.
I just have to keep my head in the books and reading all I can and also as George just suggested start off with something simple and learn and evolve from there. I am also not a believer that as we age we stop learning or become less effective than a younger person. What I will say is that when we are young we typically don't have the background knowledge and have not become tainted by prejudices in our thought process.
When I was young I say I was to ignorant to that fact, I had no fears of what you learn later can be a disaster and was willing to take a chance on something that made no sense, you just went ahead and tried something, you learned a lesson of what no to do again or discovered that what you were told wasn't necessarily gospel, it was just opinion. I know I learned more by making mistakes than when everything just worked and you could forget about trying something new. Disasters and discovery where the best teachers in fact.
Anatech,
Sacreledge! That you should even mention such a possibility in my presence! No way is that amp ever going out of my home as long as I am alive. Nor is its smaller integrated brother, the 1152DC (practically a mix of the 3250b and a slightly less powerful version of the 170DC). Those are gems to be sherished and a little bit revered.
Thanks for the tips, I might do that but with the integrated amp only, I want the separates to remain as true to the original as possible.
And of course the king of mass produced integrated amps, the one and only H/K 6550. That one is a story unto itself, one I still have trouble believing myself A few hard core tube friends commented on it by saying it's the only SS unit they have heard which makes music with an easy 10/10 mark at anywhere near its price point in its day.
Sacreledge! That you should even mention such a possibility in my presence! No way is that amp ever going out of my home as long as I am alive. Nor is its smaller integrated brother, the 1152DC (practically a mix of the 3250b and a slightly less powerful version of the 170DC). Those are gems to be sherished and a little bit revered.
Thanks for the tips, I might do that but with the integrated amp only, I want the separates to remain as true to the original as possible.
And of course the king of mass produced integrated amps, the one and only H/K 6550. That one is a story unto itself, one I still have trouble believing myself A few hard core tube friends commented on it by saying it's the only SS unit they have heard which makes music with an easy 10/10 mark at anywhere near its price point in its day.
Source noise is intermodulated by dynamically modulated circuit stage noise spectrum modulation.
IOW feedback introduces complex higher order modulation and intermodulation terms.
BQP's reduce system total IMD.
Noise is information.
Dan.
Last edited:
Sorry squadron leader, just don't understand your banter today (Python)
Those are good too.
Another question. If you're getting boards assembled by the PCB fab, what do you have to watch out for to keep cost down and not do anything unnecessary?
Read it ten times.Sorry squadron leader, just don't understand your banter today (Python)
There is no such thing as a noise free microphone generated signal, therefore music contains embedded circuit noise.
Circuit intrinsic noise is modulated by instantaneous signal level which includes preceding stage noise.
Recursive (feedback) systems introduce an extra layer of modulation (intermodulation) of the signal (and embedded noise).
Pure tones emerge relatively unscathed, but noise is logarithmically enlarged.
IOW due to time delays/phase skewing in recursive systems noise cannot be perfectly controlled/cancelled and is in fact dynamically magnified.
Dan.
Cites? Maths? I'm particularly curious about the "logarithmically enlarged" noise (which would mean, when compared to a deterministic signal (?) it grows more slowly? Or did you mean to say "exponentially enlarged"?)
Note as well that the very good paper of Boyk and Sussman (see http://www.its.caltech.edu/~musiclab/feedback-paper-acrobat.pdf) on IM distortion, although based on idealized device models, deals with very simple local feedback, not multistage affairs.
- Status
- Not open for further replies.