What's this Amplifier Design

[john curl]

Well, DUH! What is the size of the heat sink?

 

[gni]

Heat Sink Details

4.25" wide x 2" high fins x 1" deep fins -- 19 total Black unit
[around 67 sq. in.]

plus

5.25" x 1" x 1/2" sub fin Silver unit

I've run the unit bridged mono into 32 ohms at 56 V (98W) fins were
untouchable so ran a high capacity fan to keep it cool. No bad smells
and still works. Undersized heatsink it is - - TIPs, Vbe transistors, and thermistor
all mounted (TIPs with screws and pads, Vbe pushed close with compound, and
thermistor pressed between heatsink and PCB with compound and foam.

65W max per TIP transistor = 260W total max needed to dissipate.

 

[gni]

Measurements with 60Hz sine wave

Right
Input: 473mV [Vgain=15.64dB]
Output: 2.863V (8 ohms resistive)
R223 & R224: 34.6mV [emitter]
R221: 666mV
R219: 1.4mV [output base resistor]
R220: 1.5mV [output base resistor]
R217: 1.158V [Vbe bias]
R218: 627mV [Vbe bias]
R213: 2.445V [NFB]
R214: 437mV [NFB=16.33dB]

Left
Input: 460mV [Vgain=15.82dB]
Output: 2.843V (8 ohms resistive)
R123 & R124: 32.5mV [emitter]
R121: 663mV
R119: 1.4mV [output base resistor]
R120: 1.4mV [output base resistor]
R117: 1.154V [Vbe bias]
R118: 622mV [Vbe bias]
R113: 2.717V [NFB]
R114: 438mV [NFB=16.24dB]

 

[AndrewT]

replace r118 & r218 with a 5k multi turn pot, wired as a variable resistor.

Change the heatsink for one with 4times the dissipation capacity.

try turning up the bias voltage by reducing the VR value (start at 5k). Monitor the Vre and the output device temperatures.

 

[gni]

Working! Need to see what I got in the box.

 

[gni]

Nothing. . . .no change in voltage across emitter resistors! Ran it up to max.

Less is better - 1.8k created rising voltage

2K is around 5mV

 

[gni]

Installed 2 kohms (two 1 kohm) instead of 2.7 kohms.

Bias was 0 at 2.7 kohms.

Now bias is less than 7.5 mV (34 mA through 0.22 ohms).

Thank-you Andrew!

PS - will have to fan cool instead of bigger heatsink . . . the TIPs are just too close together!

 

[john curl]

15 mV is optimum.

 

[gni]

I tried 1950 ohms which gave around 10mV, but the undersized heatsink was
around 100 F instead of 82 F at 7.5mV(68 F room temp). Once I set up the
forced air cooling I can up the bias. . . 15mV (68mA).

Why 15mV may I ask. . . seems there is a diminishing return. . . is this the point
between high quality reproduction and "I don't want to listen to this crap any more."

 

[Mr Evil]

...Why 15mV may I ask. . . seems there is a diminishing return. . . is this the point
between high quality reproduction and "I don't want to listen to this crap any more."

You might want to read the essential D. Self article on power amplifier distortion. It contains some information on the optimum bias for an output stage.

 

[john curl]

We discussed it on the Blowtorch thread months ago. Optimum transition between class A and class B region.

 

[gni]

Thank you Mr. Evil and John Curl. Slowly making into Blowtorch thread (17,570 replies). Will try the "essential D. Self article on power amplifier distortion."

Thank-you for the sources.

John, you seem to know a fair amount about circuit design. You and AndrewT
should get together and design a new amp. . .

 

[PH104]

Look at and around post #15302 in Blowtorch thread.

 

[gni]

Thank you for the correct post. . . would have taken me a week. . . Blowtorch Preamp thread = John Curl = John Curl in this thread? Sorry. . . little slow today.

 

[AndrewT]

I am not a designer.
I am not even qualified in electrical or electronics.
Do not put me on the same page as our esteemed designers.

 

[gni]

Yet you know what you are talking about.
(I just made the connection of John Curl in last post).

Checked bias - Room Temp 61 F (16 C)

Right - 4.5mV
Left - 4.3mV

Last Night the room was up to 71 F (22 C) and the right channel was
up to 9.1mV and left was 8.6mV.

The bias was double with a 10 degree F change (6 C change). Will the bias
continue to change - 18mV at 81 F (27 C) and 36mV at 91 F (33 C). At some
point the heatsink won't really be radiating the heat anymore.

Thermal runaway?

Maybe with the original value of 2.7k ohms will cause the bias
current at +100 F (38 C). to slightly rise.

Designed to not self-destruct due to the under-sized heatsink.

 

[AndrewT]

Change the heatsink for one with 4times the dissipation capacity.

I forgot to mention that the Vbe multiplier must monitor the temperature of the output stage.

 

[ostripper]

Strange, Radioshack underbiased this amp just so they could
use a cheap , undersized heatsink..
As far as the dual differential amplifier topology being RIP..
BS , many good designs (the leach included) are based on
this method and are stable and musical.
I would not really rate this one , as economics were more
of a factor in this case..
OS

 

[wg_ski]

Strange, Radioshack underbiased this amp just so they could
use a cheap , undersized heatsink..

You say that as if it's the first time you've ever seen a major manufacturer do that. I see nothing "strange" about it.

 

[Mr Evil]

...As far as the dual differential amplifier topology being RIP..
BS , many good designs (the leach included) are based on
this method and are stable and musical...

I like symmetrical LTPs, but why use one and then go and use only resistors for tails instead of current sources? It should be possible to get better performance for the same component count from a single LTP + current source than symmetrical LTP + resistors.