Hi Sajti
I was interrupted yesterday so I didn't finish my reply.
That is an excellent number, my compliments, your transistors are a little quieter than the BC550/560 and the circuit looks nice, low impedance* and fairly low gain.
Best wishes
David
*Except in the current mirror, where it is correctly implemented NOT low!
Last edited:
Dear Dave,
thx, ok, I too have the 3rd edition of this must have book so I know the tables of last available bjt's ... sad to see the near future unobtanium list...
I tend to design a ltp using parallel transistors as D. Self showed in his APAD 6 book to further reduce the noise.
BTW: the latest design of single mosfet IXYS powered SA2017 amplifier has a real life (!) measured 99db SNR at 2.83V rms output 20k bw or > 101db A-w so your extremely sensitive horn speakers should be dead silent...
BR, Toni
thx, ok, I too have the 3rd edition of this must have book so I know the tables of last available bjt's ... sad to see the near future unobtanium list...
I tend to design a ltp using parallel transistors as D. Self showed in his APAD 6 book to further reduce the noise.
BTW: the latest design of single mosfet IXYS powered SA2017 amplifier has a real life (!) measured 99db SNR at 2.83V rms output 20k bw or > 101db A-w so your extremely sensitive horn speakers should be dead silent...
BR, Toni
Nice!
I haven't used those IXYS Mosfets yet but I think this could be the project for them
One chunky IXYS Mosfet can replace several "ordinary" Fets so at least the circuit will look simpler, as well as simpler to build.
A class H with outer and inner IXYS in the OPS should be OK for maybe ...>400W?
Have not reached the point to sim that yet but this could be one of the simplest hi powered efficient amps that is doable.
Or maybe an IXYS fet for the outer transistor but a smaller BJT for the inner?
Probably depends on the relative potential of the inner and outer rails, more to sim.
You have any opinion?
Best wishes
David
After a moments reflection I now think that the ones Toni and I worked on are overkill for the inner transistors.
There's probably better IXYS options for the inners.
They also have a VTO of 4 V so they waste ~5 V off your rails in a Self style OPS.
No problem for an OStripper style Class H (I think? More study required.)
Best wishes
David
Last edited:
It's a trade-off of course, the smaller dies have lower Cob but more Rbb.
I don't have measured data for the parts you mention but the ZTX618/718 are ~20 pF and <10 ohm Rb.
The ZTX450/550 are <15 pF and still <10 Rb.
The point is that there are plenty of parts that are available and have much lower Rb than the BC550/560.
Best wishes
David
Last edited:
I didn't know that website, it looks like excellent work, thanks.
The data looks very consistent with Art of Electronics but has several transistors I didn't know about.
The ZTX857 looks excellent, also we can build low noise preamplifiers for 300 V !!
The BC550 is not even mediocre, it's bad, even worse than the model.
For most people - no.
But why not have truly exceptional performance when it can be done with no cost - except attention to detail?
It is not a commercial product where my time is costed, and I enjoy to make these details the best possible.
Best wishes
David
I didn't know that website, it looks like excellent work, thanks.
The data looks very consistent with Art of Electronics but has several transistors I didn't know about.
The ZTX857 looks excellent, also we can build low noise preamplifiers for 300 V !!
The BC550 is not even mediocre, it's bad, even worse than the model.
For most people - no.
But why not have truly exceptional performance when it can be done with no cost - except attention to detail?
It is not a commercial product where my time is costed, and I enjoy to make these details the best possible.
Best wishes
David
Pay attention that BC550 is from Philips in this measurements, I use Onsemi and don't have bad experience with it.
BR Damir
Yes, but even if it's 4 times better than the Philips (to match the model's ~170 RB) it is still 100 times worse than the Zetex.
(OK the noise is only 10 times worse!)
No plausible manufacturers variation will fix that.
Best wishes
David
Last edited:
zetex 817/807 .... used these on one of my 20 IPS's. A1 performance.
Most likely my choice for this amp , as well.
Mouser $.20C ... cheap SOT23.
I concede , the super-pair is beat by the Hawksford. (below)
Damn , sub ppm up to 5K , even at 100v p-p.
OS
Attachments
So have you substituted the cheap SOT23 into a Simulation yet?
This thread is a real classroom for learning all the nuances of amp design and is truly a delight to mind and hopefully soon the ears.
The trip was time for you to sharpen the saw and gain a lot of insights.
So have you substituted the cheap SOT23 into a Simulation yet?
This thread is a real classroom for learning all the nuances of amp design and is truly a delight to mind and hopefully soon the ears.
The trip was time for you to sharpen the saw and gain a lot of insights.
Funny thing about simulations .... swapping most models has very little effect.
The base design determines the output.
The LT/OEM models won't show some of the "errata" , that would most likely
occur in the real circuit.
The KT-Cordell models seem to be quite truthful in this aspect.
Temperature modelling sucks. But , I have had all these circuits before me
in reality and know how the red led's negate the stages.
On my IPS , I will have 2 sets of led pads , if the VAS co-efficient is positive
one could add a extra red led to negate it. The "spooky" IPS was (slightly) negative as temperature rose over 30C (with 2 led's). Same VAS.
itself thermally , so this Hawksford might need the 2 led's (no super-pair).
Same with this output stage.
I saw a NAD class G with same arrangement of outputs. 2 diodes added
to the EF3 Vbe multiplier should do the trick. But, I will add extra pads
on the prototype to "fine tune" the end result.
Insights ?? Wait till you see this 3rd generation of layout.
OS
Yes, of course. My own circuits and simulations do not have such series resistors to contribute noise.
As you say - not fixable otherwise.
Also yes.
Best wishes
David
There was a 1k series input resistor in the schematic from post #1 and onward - R1.
At least one of the schematics I downloaded had an additional 220 in the input, probably left over from a test.
Post #81 for instance.
Best wishes
David
Sorry if I was too heavy handed in my criticism of the BC550/560, I had the start of a cold and was a bit unsettled.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.