Hello Chris,
We have BCxxx, 2N3904, 2N5550, 2N5551, 2N5401, MPSAxxx, and most of 2SAxxx, 2SC such as 2SC9012, 9018, etc, asian tipical. No Zetex, and other. But the price are very cheap compared to online price. In Asia region, lots of parts are fake, maybe.
For resistors no problem, 1% tolerate easy to find.
For the mosfet I still use BUK455 from an ex. ABB's motor drive.
Regards,
kartino
We have BCxxx, 2N3904, 2N5550, 2N5551, 2N5401, MPSAxxx, and most of 2SAxxx, 2SC such as 2SC9012, 9018, etc, asian tipical. No Zetex, and other. But the price are very cheap compared to online price. In Asia region, lots of parts are fake, maybe.
For resistors no problem, 1% tolerate easy to find.
For the mosfet I still use BUK455 from an ex. ABB's motor drive.
Regards,
kartino
Hi,
Based on that reply I've no doubt you can tweak up your comparator a little.
I don't find the high voltage amplifiers make a great current mirror or input pair.
Just might fix up your resolution a little. I've tried the high voltage amplifier type from Zetex for that job in pspice sims and.... sloooow.
If you decide to tweak it a little I'd be interested in knowing your results either way.
Do you listen to it much?
Cheers,
Chris
Based on that reply I've no doubt you can tweak up your comparator a little.
I don't find the high voltage amplifiers make a great current mirror or input pair.
Just might fix up your resolution a little. I've tried the high voltage amplifier type from Zetex for that job in pspice sims and.... sloooow.
If you decide to tweak it a little I'd be interested in knowing your results either way.
Do you listen to it much?
Cheers,
Chris
OK,
I have no time yet to go to store.
I listen the amp rarely, because the amp still have no box. But the amp sound good enough, better that other DIY SS project. I can accept the minor, no problem for my ear.
Cheers,
kartino
I have no time yet to go to store.
I listen the amp rarely, because the amp still have no box. But the amp sound good enough, better that other DIY SS project. I can accept the minor, no problem for my ear.
Cheers,
kartino
It`s done, take a look.
I use: 2N5551/5401, BC547, BD139 (for 12V stab.). BD139 seems overkill, because it stay cool, so there can be used 2N5551 without problem. Output devices- STP55NF06 (18 mOhm, 60V), I don`t know why I put they on HS 😉
Output cap = 1u + 150n + 5n
On backside are SR510 (Shottky 5A/100V).
PCB design is from Kartino with additional 470uF+220n on supply rails.
All stays cool, no audible "swirling" from inductor, no pop at powerup, with shorted inputs dc offset is 1..3mV (great!). Supplies are +\-24V.
Good luck!🙂
I use: 2N5551/5401, BC547, BD139 (for 12V stab.). BD139 seems overkill, because it stay cool, so there can be used 2N5551 without problem. Output devices- STP55NF06 (18 mOhm, 60V), I don`t know why I put they on HS 😉
Output cap = 1u + 150n + 5n
On backside are SR510 (Shottky 5A/100V).
PCB design is from Kartino with additional 470uF+220n on supply rails.
All stays cool, no audible "swirling" from inductor, no pop at powerup, with shorted inputs dc offset is 1..3mV (great!). Supplies are +\-24V.
Good luck!🙂
Attachments
I was thinking that there would be lack of high frequencies, but no, it does not. Sounds good, comparable with traditional BJT- based class- AB, even on that funny power supply (old and dried 6x1000uF per rail). Very agressive, "fat" bass - BJT cl-AB can`t do that.
I was surprised 🙄
Get a problem: changes load from non- inductive 3.65 Ohm/3W/1% (in glass of water) to real speaker results that mosfets got hot. Yes, I know, 12 Ohm gate series resistor is TOO small and there is no clamps and Bucherot wirings... BUT it WORKS!!! Zero dead mosfets at first iteration - very good result, eh? 😉
PS btw, I was use 5 or 10% generic resistors from our asian friends 😉 without matching. And transistors was without matching too. 3mV max DC offset... There only 2 things, that was matched - heatsinks 😀
I was surprised 🙄
Get a problem: changes load from non- inductive 3.65 Ohm/3W/1% (in glass of water) to real speaker results that mosfets got hot. Yes, I know, 12 Ohm gate series resistor is TOO small and there is no clamps and Bucherot wirings... BUT it WORKS!!! Zero dead mosfets at first iteration - very good result, eh? 😉
PS btw, I was use 5 or 10% generic resistors from our asian friends 😉 without matching. And transistors was without matching too. 3mV max DC offset... There only 2 things, that was matched - heatsinks 😀
Hi, congratulation!
It heard that your result better than mine. I feel you hear the damn good bass, is it?
But my layout is too small and did not use dedicated software whn create it. Do you use the same size or you make it bigger.
Actually I make new layout that solve the problem, but I still hopeless to get good component in Jakarta. Hence the layout still inside the disk.
Do you use the same value from me or Chris's? You see that some value for resistors for my circuit and Chris's are difference. I think the last chris's post will have faster switching. But I see the C=1n for current mirror make the whole cycle slow.
Regards,
kartino
It heard that your result better than mine. I feel you hear the damn good bass, is it?
But my layout is too small and did not use dedicated software whn create it. Do you use the same size or you make it bigger.
Actually I make new layout that solve the problem, but I still hopeless to get good component in Jakarta. Hence the layout still inside the disk.
Do you use the same value from me or Chris's? You see that some value for resistors for my circuit and Chris's are difference. I think the last chris's post will have faster switching. But I see the C=1n for current mirror make the whole cycle slow.
Regards,
kartino
2Kartino:
I have use your initial pcb design (~nov 2005). That one, which was zipped in 3 *.pdf (copper, silk and components side). Just add holes for electrolytic caps on supply rails and for TO220/TO247 compatibility, slightly increased width of power traces.
Other changes: shottky across MOSFETs, 1N5819 as Baker clamps, SR108 (Shottky) as bootstrap diode, 330uF panasonic HFG series at 12V stab. on both hi and low sides. And I havn`t solder 22pF on input transistor`s bases yet. All other values is yours initial.
All transistors was mentioned in my previous msg. There is NO rare or exclusive parts, all is cheapest as possible. If you can`t buy them, try IRFZ44N for low power- they can be found in dead UPS.
Output filter is 30uF/1.155uF 😉 For inductor was used 30 years old soviet "alsifer" core (nearest analogue is "cool mu") - it much better then today`s chinese iron powder cores, which can be found in dead chinese ATX PSUs for free. 😀
I`ve try increase output cap to 2.155uF but sound was degraded
So there is one thing, that I can and will do - ajust Ton/Toff (i.e. deadtime) for better efficient without noticeable sound degradation.
Actually, I think that your previous pcb is great, very good performer with proper components. But where is new design? I don`t see `em in attach...
Let`s take it to discussion!
About software: I don`t like LTSpice, I think that micro cap7 is better sim. Minor changes in pcb was done in photo shop 😉
PS my name is Konstantin (or Constantine if you like), and yours?
I have use your initial pcb design (~nov 2005). That one, which was zipped in 3 *.pdf (copper, silk and components side). Just add holes for electrolytic caps on supply rails and for TO220/TO247 compatibility, slightly increased width of power traces.
Other changes: shottky across MOSFETs, 1N5819 as Baker clamps, SR108 (Shottky) as bootstrap diode, 330uF panasonic HFG series at 12V stab. on both hi and low sides. And I havn`t solder 22pF on input transistor`s bases yet. All other values is yours initial.
All transistors was mentioned in my previous msg. There is NO rare or exclusive parts, all is cheapest as possible. If you can`t buy them, try IRFZ44N for low power- they can be found in dead UPS.
Output filter is 30uF/1.155uF 😉 For inductor was used 30 years old soviet "alsifer" core (nearest analogue is "cool mu") - it much better then today`s chinese iron powder cores, which can be found in dead chinese ATX PSUs for free. 😀
I`ve try increase output cap to 2.155uF but sound was degraded
So there is one thing, that I can and will do - ajust Ton/Toff (i.e. deadtime) for better efficient without noticeable sound degradation.
Actually, I think that your previous pcb is great, very good performer with proper components. But where is new design? I don`t see `em in attach...
Let`s take it to discussion!About software: I don`t like LTSpice, I think that micro cap7 is better sim. Minor changes in pcb was done in photo shop 😉
PS my name is Konstantin (or Constantine if you like), and yours?
I will post the new layout in this thread soon. I will check it first since I've been long time not see the layout.
I will try to use smaller cap for output filter like yours. Since I have no scope I just rely on my intuition and my ear?!?.
Thanks to people who will do test teh circuit.
Kartino is my real name, just it. I am from Yogyakarta, a central region of Java island that most of people have short name.
Nice to meet you here. I am glad to see people do something useful in this forum.
BTW I have plan to use paralel output mosfet for bigger output, I will post the schematic soon. The simulation looked good.
Regards
kartino
I will try to use smaller cap for output filter like yours. Since I have no scope I just rely on my intuition and my ear?!?.
Thanks to people who will do test teh circuit.
Kartino is my real name, just it. I am from Yogyakarta, a central region of Java island that most of people have short name.
Nice to meet you here. I am glad to see people do something useful in this forum.
BTW I have plan to use paralel output mosfet for bigger output, I will post the schematic soon. The simulation looked good.
Regards
kartino
I think that 1 GOOD output transistor is better than 2 or 3 worse. And the main advantage of UcD is simplicity for diyers. If possible, do not use oldtimers like IRF540, etc.
I have no scope too, so what?! 😉
Btw, it is good idea to create pcb with D2PAK compatibility or even entirely SMD- based layout.
I have no scope too, so what?! 😉
Btw, it is good idea to create pcb with D2PAK compatibility or even entirely SMD- based layout.
Hi guys,
You're just not going to be able to unify final passive component values unless you do the same with the active components. Different gains, delays, parasitics etc will require different values.
Same reason why it changes from simulation to real world.
I don't think the schottkies across the mosfets are helping, and willing to bet you can remove them, should not even be required at this power level.
A single well selected mosfet would be best, as well as a single well selected filter cap.
I wouldn't bother with SMD unless you have a scope and in the least are willing to go the whole nine yards with it, dual layer pcb, fully optimised.
Regards
You're just not going to be able to unify final passive component values unless you do the same with the active components. Different gains, delays, parasitics etc will require different values.
Same reason why it changes from simulation to real world.
I don't think the schottkies across the mosfets are helping, and willing to bet you can remove them, should not even be required at this power level.
A single well selected mosfet would be best, as well as a single well selected filter cap.
I wouldn't bother with SMD unless you have a scope and in the least are willing to go the whole nine yards with it, dual layer pcb, fully optimised.
Regards
I agree with Chris,
The most important IMHO, is value for R1. It is the role for the gate control. I can not say how, but you can try simulate by yourself. The R1 value is very sensitive.
I always play the other passive component to the reasonable value, then lastly set the R1.
I have idea to use variable resistor 100hm and fixed 100ohm for R1 since the value need so perfect, and no resistor can fill it.
SMD? In my dreams, since I live in Jakarta!?!
The most important IMHO, is value for R1. It is the role for the gate control. I can not say how, but you can try simulate by yourself. The R1 value is very sensitive.
I always play the other passive component to the reasonable value, then lastly set the R1.
I have idea to use variable resistor 100hm and fixed 100ohm for R1 since the value need so perfect, and no resistor can fill it.
SMD? In my dreams, since I live in Jakarta!?!
Hi,
I do it the other way around. Since R1 sets the output current of the comparator ... it therefore determines the driver input current and will define how hard they switch.
I(R1) must equal the comparator source current to keep DC offset minimal, but I think your plan for a pot here could be useful too 🙂
Regards
I do it the other way around. Since R1 sets the output current of the comparator ... it therefore determines the driver input current and will define how hard they switch.
I(R1) must equal the comparator source current to keep DC offset minimal, but I think your plan for a pot here could be useful too 🙂
Regards
Hmm, another report: good news. Raises Rgate from 12 to 27..33 Ohm dramaticaly lower temperature of MOSFET`s - it slowly (20 minutes) raises to +60C and after then it increase ~+1C during another 5 minutes endlessly.
Hmm.. probably need to decrease Toff too.
Btw, BD139 (12V stab) got hot - it`s now about +55C (recently it was near room temperature). I don`t understand- why?
2Chris: body diode has Trr=75ns. I think, that Shottky (~5ns and Uf=0,3V) are much better. 1 piece of SR510 costs $0,18 for me - not so expensive.
2Kartino: variable resistor in UcD?! No way! Only for testing purposes. May be. Simulate, simulate, simulate...!
Hmm.. probably need to decrease Toff too.
Btw, BD139 (12V stab) got hot - it`s now about +55C (recently it was near room temperature). I don`t understand- why?
2Chris: body diode has Trr=75ns. I think, that Shottky (~5ns and Uf=0,3V) are much better. 1 piece of SR510 costs $0,18 for me - not so expensive.
2Kartino: variable resistor in UcD?! No way! Only for testing purposes. May be. Simulate, simulate, simulate...!
Kuzmenko said:
That's pretty hot. My mosfets stay cold to the touch, never above room temp, which is too cold, but given no scope I was happy with that and stopped there, as it took alot to get it to that point, you've got a way to go with it yet.
You can try decreasing the turn off resistor, you'll probably smoke your first set of mosfets within a few seconds.
Measure your driver 12V supply, I doubt it's 12V anymore.
There's too much parasitic inductance for the schottky to be of any use, the body diode will still saturate. You should try it without first.
What do you suppose the pot on the UCD400 might be for?
There you go 😉
Regards
- Status
- Not open for further replies.