Criticizing is one thing, designing an ingenious SOTA audio amplifier is another thing.
Do you understand how the circuit works?
If yes, please share your understanding with us.
CraigBuckingham said:
What I mean is I have repeatedly stated that that first schematic was drawn up in my simulation program to confirm I wasn't posting a non working circuit. I have also repeatedly stated that when a friend found I still had boards left from a production run many years ago, I started redrawing the schematic to match the board silk screen. This made the boards usable for any one interested. I have no idea why my simulation program chose to label Q13 as X13. I don't enter that information, I just pick the parts from the part list.
I might also point out that I have not claimed no improvement could be made to this circuit. Improvements have been made. They do not fit on the old board, so they are not shown in the schematic.
I have also repeatedly stated that I had intended this to be a learning experience for the DIYers who cared to follow along from concept through the evolution of the design. Clearly you did not come here to learn anything. I am not going to change everything in the design to what you think might work better. This amp has been built and tested repeatedly. It works very well and tests very well. I am not interested in speculation about what the distortion might be or why this doesn't work. I know the answers to that already. I will not wast any more time on it.
I am attaching the current schematic for the 50W amp. This is the one that has been matched to the existing board. Please stop using the oldest schematic available.
Steve-
This attachment is the match to the PCB's you sold? I will print it and put it with the boards for assembly next month.
If I recall you also had a BOM from Mouser for them. May I ask you to re-post it with this schematic?
This attachment is the match to the PCB's you sold? I will print it and put it with the boards for assembly next month.
If I recall you also had a BOM from Mouser for them. May I ask you to re-post it with this schematic?
traderbam said:The stingingly guarded exchanges in this thread remind me of the 1956 movie Forbidden Planet. If this were Dr. Morbius' amplifier, perhaps he might have said:
"The fools, the meddling idiots! As though their ape's brains could contain the secrets of the Krill!"
That is a great movie. That is the source of the name Krell. Krill was my humerus (to me) take off of that.
troystg said:
Here is the BOM. Some of the cap values have changed slightly to make the part lead spacing fit the boards. The correct lead spacing sometimes wasn't available in the value I gave originally. I have sourced all parts from Mouser and Digi-Key. You may buy wherever you choose.
If you can open an Excel spread sheet I can E-mail you the BOM. The part numbers are links to the parts.
I see the drivers are specified as the same devices as the output pair.
Is this a typo?
How much leeway is there to select a different Voltage Gain Stage (VGS)?
I'm wondering if the RMI three pair output stage could be modified to KRILL style and keep the RMI VGS. The RMI feedback just happens to be a hardwired lead that could be tapped into the VAS output.
The same modification might apply to the Leach Lo Tim which was designed to have an inner feedback route and a Global NFB route.
Steve,
add the links to your current PDFs to your signature.
That way you can keep us all up to date with changes and new models etc.
Is this a typo?
How much leeway is there to select a different Voltage Gain Stage (VGS)?
I'm wondering if the RMI three pair output stage could be modified to KRILL style and keep the RMI VGS. The RMI feedback just happens to be a hardwired lead that could be tapped into the VAS output.
The same modification might apply to the Leach Lo Tim which was designed to have an inner feedback route and a Global NFB route.
Steve,
add the links to your current PDFs to your signature.
That way you can keep us all up to date with changes and new models etc.
AndrewT said:
I don't plan to go anywhere. I do get tired of answering the same question for the sixth or seventh time because someone chose not to read the thread.
c2cthomas said:Hi Steve - Good to see you here! 😀
Q12 - should this be a KSC23160BU?
Yes it should. I have no idea where that 5 came from. The program repeats the part every time I click the mouse. If I put in three of the same part they should all be the same except for the schematic designation. One more thing to be on the lookout for.
AndrewT said:
Taking your questions in order:
That is not a typo. I changed the driver simply to make mounting to the board simpler. Once again I am trying to make the schematic match the board. The board was made to use the higher power part because it was originally a 100W amp.
You may use any voltage gain stage you like. The only requirement is that it be able to drive it's own feedback load in parallel with the load presented by the output stage. In most cases the feedback resistor will be the predominant portion of the load.
Good idea. I will do that.
Steve Dunlap said:
Yeah - when I cut and paste cells in Excel it has a nasty habit of increasing things by "1" (so 125 becomes 126). I'm always finding something that has be "automatically increased" for me.
Thanks for the update!
I just heard this soundbite from the Irish Rugby team captain, Brian O'Driscoll, and thought it relevant to this thread:
Steve - I'm sorry you took what I thought was a simple conversational comment as an insult or critcism. In post No. 2 you show a device with an operating Vce of (I think it was) 196mV. Someone designing "by the book" might not be comfortable with this, I personally don't care.
To quote ostripper who is actively designing and commenting (with regards to hfe):
>Well said, I thought about that (leakage currents) between bases. I found that slower , lower Hfe devices actually work
>better (several mA on the "idled" side).
>Using fast transistors (KSA1381/3503) ,the OPS still switches
>unless you raise the CCS's to 10mA +, with the Mje's
>7-8mA gives better results. It seems the choice of devices
>is critical with this topology.
ostripper - Thanks for the files, I wasn't blocked from downloading LTSpice 🙂 After removing a couple text errors in your models I got this changing only to 20kHz and 50W. ??
Circuit: * C:\Documents and Settings\swurcer\Desktop\krill\Krill_OPS.asc
Ignoring Unknown BJT level: 2
Ignoring Unknown BJT level: 2
Direct Newton iteration for .op point succeeded.
Fourier components of V(c)
DC component:-0.29089
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 1.917e+01 1.000e+00 -0.07° 0.00°
2 4.000e+04 1.772e-03 9.246e-05 51.10° 51.17°
3 6.000e+04 3.230e-02 1.685e-03 -163.42° -163.34°
4 8.000e+04 2.096e-03 1.094e-04 98.96° 99.03°
5 1.000e+05 1.498e-02 7.813e-04 -157.10° -157.03°
6 1.200e+05 1.430e-03 7.461e-05 109.33° 109.41°
7 1.400e+05 7.928e-03 4.136e-04 -151.85° -151.78°
8 1.600e+05 1.025e-03 5.348e-05 114.28° 114.36°
9 1.800e+05 4.794e-03 2.501e-04 -148.44° -148.37°
Total Harmonic Distortion: 0.192687%
To quote ostripper who is actively designing and commenting (with regards to hfe):
>Well said, I thought about that (leakage currents) between bases. I found that slower , lower Hfe devices actually work
>better (several mA on the "idled" side).
>Using fast transistors (KSA1381/3503) ,the OPS still switches
>unless you raise the CCS's to 10mA +, with the Mje's
>7-8mA gives better results. It seems the choice of devices
>is critical with this topology.
ostripper - Thanks for the files, I wasn't blocked from downloading LTSpice 🙂 After removing a couple text errors in your models I got this changing only to 20kHz and 50W. ??
Circuit: * C:\Documents and Settings\swurcer\Desktop\krill\Krill_OPS.asc
Ignoring Unknown BJT level: 2
Ignoring Unknown BJT level: 2
Direct Newton iteration for .op point succeeded.
Fourier components of V(c)
DC component:-0.29089
Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 1.917e+01 1.000e+00 -0.07° 0.00°
2 4.000e+04 1.772e-03 9.246e-05 51.10° 51.17°
3 6.000e+04 3.230e-02 1.685e-03 -163.42° -163.34°
4 8.000e+04 2.096e-03 1.094e-04 98.96° 99.03°
5 1.000e+05 1.498e-02 7.813e-04 -157.10° -157.03°
6 1.200e+05 1.430e-03 7.461e-05 109.33° 109.41°
7 1.400e+05 7.928e-03 4.136e-04 -151.85° -151.78°
8 1.600e+05 1.025e-03 5.348e-05 114.28° 114.36°
9 1.800e+05 4.794e-03 2.501e-04 -148.44° -148.37°
Total Harmonic Distortion: 0.192687%
Steve Dunlap said:
Steve, you keep putting your words in my mouth.
You say "doesn't anyone read the thread" but you do not read or maybe understand what I said in my replies. I think US and Australian English are pretty close.
I didn't claim that you claimed no improvement could be made.
I am not asking you to change anything.
I have learnt a lot from coming here but not about amplifier design.
I am not asking you to change anything in your design. I don't "think" - I know. That is why I said it.
If you said at the start of this thread or during it that you would not accept any opinions on the circuit design and that the thread was only for DIY kits and discussion about that then I would not have contributed.
I take it that is now your position so I will respect that.
I don't get particularly good results from ostripper's latest simulation, either. Mostly because the naive input and VAS contribute a huge amount of distortion before the output stage even comes into play.
People need to take a whole-system approach to implementing this thing. There's no point in ruining a perfectly good output stage with a passively-loaded input stage. Try an active (current mirror) load.
I also get better simulator results with paralleled devices in the output.
People need to take a whole-system approach to implementing this thing. There's no point in ruining a perfectly good output stage with a passively-loaded input stage. Try an active (current mirror) load.
I also get better simulator results with paralleled devices in the output.
- Status
- Not open for further replies.