Leach but larger

[taj]

Attached is my variation of original Leach 4.5.

I hope to build it into an existing chassis and power supply that provides ~72vdc rails and TO-3 heatsinks that accommodate 4-pairs per channel. Yes, it'll pack a nice punch.

So, I've swapped out the transistors with somewhat newer, higher power-rated ones, and added more output devices. And so my questions are:

Do I need to modify any circuit elements to make this work? I'm just learning electronics, so I am a bit ohms law challenged still. I'm not sure if the voltages or currents need adjusting to suit, or whether my transistor changes have screwed up any circuit operations.

Also, is my choice of devices suitable/acceptable? Better choices?

(I have the 2sa970/2sc2240s on hand, which is why I've used them in the input section.)

Comments/advice?

..Todd

[taj]

part 2 of schematic ...


NOTE: I don't know what C19 & C20 (in the output section) do, so I've left them located as in the original circuit, but maybe they should be moved to the last pair. (Just guessing.)

[jaycee]

Spotted one error - the value of R45 R47 R145 R147 is not 0.33K ohms

You may need to use more than 10pF Miller capacitors C10 and C11. I'd go with 100pF to start with and perhaps drop down to 47pF if it's stable.

[taj]

Originally posted by jaycee
Spotted one error - the value of R45 R47 R145 R147 is not 0.33K ohms

Good eye. Thanks.


Regarding the Miller compensation, would that advice that be due to my transistor choice? This is Leach's original value.

..Todd

[taj]

I've also just increased the value of R13 & R14 to 3.9k, as instructed in Leach's documentation for when you alter the rail voltage.

..tj

[AndrewT]

Hi,
I disagree with changing the VAS Cdom.
Instead experiment with the three stage feedback C8 & C9 to improve stability if needed .
Alternatively put in a trace from the bottom of C10 to the bottom of R4 (two stage feedback) and insert in this route room for a small cap and in parallel a series pair of cap + resistor (=C//[R+C]). This location for the feedback is reputed to sound nicer and does not reduce the slew rate by sucking LTP current to charge Cdom.

Add the missing resistor to the protection sensor. See Jens' schematic http://www.delta-audio.com/Leach_Clo..._7.4.6_Sch.pdf R68 & R69.
I suggest you move C17 & C18 over to next to Q10 & Q11 bases. Then insert a resistor between the tops of C17 and R28 and between C18 & R29. This allows you to adjust the time constant of the transient detection/triggering point and make it higher than the DC triggering point that would apply to a long term overload condition eg shorting the output.

The output base resistors are quite high @ 10r. Be prepared to make these as low as 0r0.

[taj]

Originally posted by AndrewT
Hi,
I disagree with changing the VAS Cdom.
Instead experiment with the three stage feedback C8 & C9 to improve stability if needed .
Alternatively put in a trace from the bottom of C10 to the bottom of R4 (two stage feedback) and insert in this route room for a small cap and in parallel a series pair of cap + resistor (=C//[R+C]). This location for the feedback is reputed to sound nicer and does not reduce the slew rate by sucking LTP current to charge Cdom.

Add the missing resistor to the protection sensor. See Jens' schematic http://www.delta-audio.com/Leach_Clo..._7.4.6_Sch.pdf R68 & R69.
I suggest you move C17 & C18 over to next to Q10 & Q11 bases. Then insert a resistor between the tops of C17 and R28 and between C18 & R29. This allows you to adjust the time constant of the transient detection/triggering point and make it higher than the DC triggering point that would apply to a long term overload condition eg shorting the output.

The output base resistors are quite high @ 10r. Be prepared to make these as low as 0r0.

Hi Andrew. Thanks for the response. I love it when you go getten all technical on me.

Seriously, I sort-of understand everything except the output base resistor comment. Can you explain what these resistors affect and what I'm going to look for to know how to tune them to the right value?

Much appreciated!

..Todd

[taj]

I am also wondering about the VAS/Pre-driver transistors. (2sa1360/2sc3423) I've got no-load rail voltages of about +-74vdc. Do these transistors 'see' both rails. They are 150v rated I believe, so that may be courting disaster methinks.

..Todd

[AndrewT]

Hi,
the pre-drivers should use the same voltage selection criteria as for the driver and output devices.
I think your selection is excellent, except that +-74Vdc will rise when mains voltage is at maximum tolerance.
However, the devices never see the full 2*Vrail during operation, but more like 90% to 95% of that voltage (when pushed to near clipping with high impedance or open circuit load values).
There will also be a tolerance on the device Vce0. So they should not blow up just because you hit 151Vce once in a while. But if you are into disco music AND run them hot then there is more chance of hitting their limits.

You could build a regulator for the front end and just run the drivers and outputs from the unregulated.

I wonder just how low the Vrail droops when the pre-driver is trying to send maximum signal to the drivers? Theoretically you can never hit even 95% of 2*Vrail with an output load connected. Then add 6% mains tolerance and you should be well inside Vce0.

Most designers seem to select the output base resistor to 0r0. A few mention that stability is improved if set to a lowish value (2r2) but that distortion is compromised, particularly with lower load values. Stability is the MUCH more imporatant criteria, if it cannot be ensured elsewhere in the circuit. Leaving pads and traces on the PCB is good practice just in case. I am a believer in putting all the optional pads into the PCB and using only those that are needed. It does not cost anything and usually does not waste space/area either.

I love it when you go getten all technical on me

I'm so glad we are on opposite sides of the globe, otherwise folks might get the wrong end of the stick.

[taj]

Originally posted by AndrewT
Hi,
the pre-drivers should use the same voltage selection criteria as for the driver and output devices.
I think your selection is excellent, except that +-74Vdc will rise when mains voltage is at maximum tolerance.
However, the devices never see the full 2*Vrail during operation, but more like 90% to 95% of that voltage (when pushed to near clipping with high impedance or open circuit load values).
There will also be a tolerance on the device Vce0. So they should not blow up just because you hit 151Vce once in a while. But if you are into disco music AND run them hot then there is more chance of hitting their limits.

I wonder just how low the Vrail droops when the pre-driver is trying to send maximum signal to the drivers? Theoretically you can never hit even 95% of 2*Vrail with an output load connected. Then add 6% mains tolerance and you should be well inside Vce0.

Good information. Thanks. I think I should plug in the transformers and measure their no-load voltage to get a better idea of their true output. Who knows what (North American) mains voltage they had in mind for their 105VCT output rating. Could be anywhere from 110vAC to 120vAC.

You could build a regulator for the front end and just run the drivers and outputs from the unregulated.

That's a good idea. Probably not necessary based on my listening habits, but good to remember in case I decide to try it out in a concert rig.

Most designers seem to select the output base resistor to 0r0. A few mention that stability is improved if set to a lowish value (2r2) but that distortion is compromised, particularly with lower load values. Stability is the MUCH more imporatant criteria, if it cannot be ensured elsewhere in the circuit.

More good info. Thanks again.

I'm so glad we are on opposite sides of the globe, otherwise folks might get the wrong end of the stick.

Not to worry. I'm happily married (to a female).

..Todd