I now have, thanks to my earlier failures, 7 kits now... which also means that I have 14 of each transistor to choose from, which implies (thanks to the magic of the mathematics of permutations and combinations) that I'm almost certain to have a couple of well-matched pairs of each.
So, I spent 3 rather tedious hours last night pulling all the trannies to be matched out of my boards (made somewhat easier by using my gas iron's heat-knife as a giant soldering pen, big enough to touch all 3 trannie pins at once), and putting them in a big grid I created to allow matching pairs to be found. I then put the 2 best pairs of each back into the 2 boards I'm working on, and left it at that for the night (since 2am seems like a bad time to turn on the power).
This arvo I rechecked everything and powered up one amp, and I'm pleased to say that all 4 emitter resistors now have voltages within 2mV of each other. I'll test the other as well, and keep an eye on their temperature for a while, then do a new listening test tonight.
So, I spent 3 rather tedious hours last night pulling all the trannies to be matched out of my boards (made somewhat easier by using my gas iron's heat-knife as a giant soldering pen, big enough to touch all 3 trannie pins at once), and putting them in a big grid I created to allow matching pairs to be found. I then put the 2 best pairs of each back into the 2 boards I'm working on, and left it at that for the night (since 2am seems like a bad time to turn on the power).
This arvo I rechecked everything and powered up one amp, and I'm pleased to say that all 4 emitter resistors now have voltages within 2mV of each other. I'll test the other as well, and keep an eye on their temperature for a while, then do a new listening test tonight.
That's a very good match, Jeremy! I aim for around 2mV deviation at a bias voltage of 60mV across 0.47R.
You learn real quick......
Hugh
You learn real quick......
Hugh
I have a good teacher. 😀 BTW I've replaced the 0R33's with 0R22's now that I've got better matching, and the voltage is still stable.AKSA said:
As I write this, my wife and I are listening to Nina Ferro with Joe Chindamo (maybe Australia's best jazz vocalist and jazz pianist), and my wife just commented that it sounds better than she's ever heard. So that's a good sign!
3% spread (2mV in 60mV) in Ic at the same Vbe is very good.jp_howard said:
0.2mV in 10mV (2%) is even better. Well done.
SC ULD2
"my wife just commented that it sounds better than she's ever heard. So that's a good sign!"
Jeremy
That is always a very good sign !
Many wives would just say "That's nice." then walk away.
Well done !
So the much maligned SC ULD2 isn't such a dog after all ? 🙂
SandyK
"my wife just commented that it sounds better than she's ever heard. So that's a good sign!"
Jeremy
That is always a very good sign !
Many wives would just say "That's nice." then walk away.
Well done !
So the much maligned SC ULD2 isn't such a dog after all ? 🙂
SandyK
Jeremy,
You probably need to own an amp for a few weeks, and listen through many sessions and genres to form a solid opinion of its qualities.
That said, the SC ULD 200W amp is a good amplifier, and I think quite competently designed, with just one basic flaw in the output stage which of course is highly visible.
It can be very simply tweaked for much more sonic quality, however. The majority of designers consider only the engineering, aiming for zero distortion, without too much concern for the hearing and perception issues. I have found these issues to be very important. Amps are designed for human beings, not arcturan megadonkeys, and psychoacoustics are important.
We shall discuss this over numerous coffees in the future.
Hugh
You probably need to own an amp for a few weeks, and listen through many sessions and genres to form a solid opinion of its qualities.
That said, the SC ULD 200W amp is a good amplifier, and I think quite competently designed, with just one basic flaw in the output stage which of course is highly visible.
It can be very simply tweaked for much more sonic quality, however. The majority of designers consider only the engineering, aiming for zero distortion, without too much concern for the hearing and perception issues. I have found these issues to be very important. Amps are designed for human beings, not arcturan megadonkeys, and psychoacoustics are important.
We shall discuss this over numerous coffees in the future.
Hugh
Re: SC ULD2
You don't have enough money to pay for my design 😀
sandyK said:
Perhaps Mihai is in the business of selling his own amplifier designs too?
SandyK
You don't have enough money to pay for my design 😀
Re: Re: SC ULD2
In that case.
I'll start with this.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=111756
Very nice😀
One Question if I may...
You biased yours for 2Watt class A, then the rest B, (AB).
How much difference was there in biasing the circuit for 1Watt class into 4ohms?
roender said:
In that case.
I'll start with this.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=111756
Very nice😀
One Question if I may...
You biased yours for 2Watt class A, then the rest B, (AB).
How much difference was there in biasing the circuit for 1Watt class into 4ohms?
1W of ClassA into 4r0 is the same bias a 2W of ClassA into 8r0.
BTW,
Roender's choice of 200mA/device for 20mVre across Re=0r1
gives a total bias current of 600mA. This gives ~5.7W of ClassA into 8r0 and ~2.8W of ClassA into 4r0.
The earlier version using 17mVre gives ~4W of ClassA into 8r0 and ~2W of ClassA into 4r0.
BTW,
Roender's choice of 200mA/device for 20mVre across Re=0r1
gives a total bias current of 600mA. This gives ~5.7W of ClassA into 8r0 and ~2.8W of ClassA into 4r0.
The earlier version using 17mVre gives ~4W of ClassA into 8r0 and ~2W of ClassA into 4r0.
Thanks for the kind words. It's very useful to get feedback like this about what measurements to aim for.AndrewT said:
(BTW, since I'm using 0R22's, I'm running at 18mV - so I guess that's about 1%. 😎 Since Jaycar is open today I can get a bigger heatsink and hopefully then run at 20mV with the 0R22's.)
For those of you that are interested ...
Here is a view of the PCB that better shows the traces. Red is top side and Green is the bottom layer. The PIC is larger than the actual PCB. The actual PCB measures 3 by 7 inches.
Here is a view of the PCB that better shows the traces. Red is top side and Green is the bottom layer. The PIC is larger than the actual PCB. The actual PCB measures 3 by 7 inches.
An externally hosted image should be here but it was not working when we last tested it.
That's a nice board Hugh. Is anyone else is interested in a group buy, with Hugh's permission of course.
Could anyone provide some info for me please.
1 Could we increase the rails to +-70 as is or would we need component changes and what power output would it achieve.
2 What would we have to reduce the rails down to to get this amp stable into 2 Ohms.
I have always thought that a kit amp able to meet multiple needs would be good.
Thanks
Could anyone provide some info for me please.
1 Could we increase the rails to +-70 as is or would we need component changes and what power output would it achieve.
2 What would we have to reduce the rails down to to get this amp stable into 2 Ohms.
I have always thought that a kit amp able to meet multiple needs would be good.
Thanks
Hey Phoenix358,
That ain't my board, that's Carl's!!
With 70V rails, you are looking at roughly 520W into 4R. 56V rails will give you around 320W into 4R.....
Raising rail voltages is problematic, it impacts stability, and takes considerable design care.
But why would you want so much power?
Cheers,
Hugh
That ain't my board, that's Carl's!!
With 70V rails, you are looking at roughly 520W into 4R. 56V rails will give you around 320W into 4R.....
Raising rail voltages is problematic, it impacts stability, and takes considerable design care.
But why would you want so much power?
Cheers,
Hugh
Re: SC ULD2
Or it's a dog with a nice bark... (Particularly given the price point). In fact, I had planning to install new heatsinks today, but I've spent all arvo after getting back from Jaycar listening to my new amp. Maybe I need 2 sets of amps, so I can listen to one while I work on the other... 😉sandyK said:
Apologies to Carl, mixed up Hugh and Huff. 50% eyesight left, and counting down 
Sometimes we need power. My passion is designing loudspeakers and have done so professionly. I am now working on a design with 8 Morel Neolin 200mm drivers per side.
Why power....cause we can!😎
Sometimes we need power. My passion is designing loudspeakers and have done so professionly. I am now working on a design with 8 Morel Neolin 200mm drivers per side.
Why power....cause we can!😎
Transistors
The mpsa18's recommended by SandyK weren't readily available in Melbourne on an Easter Saturday, but I did notice that bc550's seem similar, so I bought a few from Jaycar and found their Hfe over double the bc546's in the kit. (The bc550's have Hfe around 550-600 in the few I've checked).
Would it make sense to replace the bc546's with bc550's?
How about the 2SA1016's that were suggested - what other options (if any) are there for them?
The mpsa18's recommended by SandyK weren't readily available in Melbourne on an Easter Saturday, but I did notice that bc550's seem similar, so I bought a few from Jaycar and found their Hfe over double the bc546's in the kit. (The bc550's have Hfe around 550-600 in the few I've checked).
Would it make sense to replace the bc546's with bc550's?
How about the 2SA1016's that were suggested - what other options (if any) are there for them?
SC ULD 2
Jeremy
The best part is that after all the electrolytics finish forming, and the tweaks are all in place, it will sound even better.😉
SandyK
Jeremy
The best part is that after all the electrolytics finish forming, and the tweaks are all in place, it will sound even better.😉
SandyK
Hugh,
I don't need quite that power level but some do, why not meet the need.
I do need an amp capable down to 2 Ohms. Each amp will run 2 pair of the Neolins in isobarik mode. Other wise I need twice as many amp modules, that would probably provide more headroom.
Pheonix
I don't need quite that power level but some do, why not meet the need.
I do need an amp capable down to 2 Ohms. Each amp will run 2 pair of the Neolins in isobarik mode. Other wise I need twice as many amp modules, that would probably provide more headroom.
Pheonix
Transistors
Jeremy
If you can get a pair that are as closely matched for both VBE and HFE as the BC546, give them a try.The BC550 are lower voltage rated than the BC546, but in this area of the circuit, it doesn't matter. Are you sure that you wouldn't rather listen for a little while longer, before doing any further mods other than the heatsink changes ? (not forgetting to slightly countersink/deburr the holes for a better mounting area,as Hugh recommended)
SandyK
Jeremy
If you can get a pair that are as closely matched for both VBE and HFE as the BC546, give them a try.The BC550 are lower voltage rated than the BC546, but in this area of the circuit, it doesn't matter. Are you sure that you wouldn't rather listen for a little while longer, before doing any further mods other than the heatsink changes ? (not forgetting to slightly countersink/deburr the holes for a better mounting area,as Hugh recommended)
SandyK
- Status
- Not open for further replies.