New sx-636 Refurb Issues and Questions

[fromnowon]

Thanks again!
I didn't lift a leg to test, but with the amp off the original D1 and D3 measure 651 and 660 with both legs soldered in.
The 4148s measure 681 and 682, so they are actually a little higher Vf's than the originals.

But, do I need to lift a leg to get an accurate comparison? It's no problem to do that or to remove one altogether.

Thinking ahead, would this be a good choice for the Schottky diode for D1 and D3 if it's required?
https://www.mouser.com/ProductDetai...=sGAEpiMZZMtQ8nqTKtFS/D9SVzsgHTKGnR5ayPNy9/c=

I really appreciate the help.

 

[LesE]

Normally you would want to lift one leg but it's not always necessary. Based on what you're seeing, it's probably not worth trying the 1N4148.

The BAT46 that you linked to is the one I'd try.

 

[fromnowon]

Upon further consideration of the Schottky diode, I expect that it's much lower Vf may also necessitate the replacement of the trimmer as well as R21 with a larger resistance value.

Ok thanks, I'll try the Schottky BAT 46's, will have to order those so I can't report back for a while.

Also I'll order 200 ohm trimpots just in case needed:
https://www.mouser.com/ProductDetai...=sGAEpiMZZMvygUB3GLcD7uBcHvh8FSVun6ip4y0mgtU=

I have a good assortment of resistors already.

In the meantime I can check resistors as Larry suggested and see what I find.

Thanks again for the ideas, it would be great to be able to adjust to 20 mV and it's a good learning opportunity. (It sounds VERY nice even as-is at 40 mV, I've been listening a little even though maybe I shouldn't.)

 

[fromnowon]

There is a Tuning Fork on amp adjustment that's interesting - I attached it, not that I can understand it all, but it does explain the problems with both too low and too high idle current.




So now I sort of understand why it's important to correct it in my amp even though, honestly, it sounds fantastic at 40 mV at the volumes I listen! I'll keep working on it.

 

[fromnowon]

oh, and I checked DC offset:
Left: 19.4 mV DC
Right: 24.0 mV DC
I know it's not adjustable, so glad it's at least in the acceptable "ok" range.

 

[fromnowon]

Hello again! I'm trying to get back to this and finish up my SX-636, hopefully this weekend. I really appreciate all the advice you guys have given me.

To restate my problem from above - it's the amp adjustment: The lowest I can get (after replacing all the transistors and outputs) for Left Channel (21 & 17) is 42.0 mV. The lowest I can get for Right Channel (8 & 11) is 31.7mV. That's all the way CCW. Of course, they should be adjusted to 20mV​

Update: I finally did get two 511-BAT46s; the Forward Voltage Drop on these is 271, vs. original D1 of 748 and D3 of 765. I also got some 3362P-1-201LF Borns trimmers, and I have a bunch of resistors. So now I can do some experimenting.

First I might put the old output transistors back in one channel, and try the adjustment, to see if it's my MJ15015G/MJ15016Gs that are causing out-of-range.

I'm curious how the designers arrived at 20 mV. I've been trying to read about "quiescent operating point or Q-point" but it's complex. Why does this circuit need 20mV? Could my transistor substitutions result in a different Q-point in this design?

Is this voltage divider biasing we're talking about? I could take 20 mV at face value but I'm interested in understanding it better if I can, just for learning purposes.

 

[LesE]

I could take 20 mV at face value but I'm interested in understanding it better if I can, just for learning purposes.

A Class-B amplifier characteristically exhibits high THD at low signal levels as the result of crossover distortion. Introducing some bias (idle) current dramatically lowers THD at low levels but it come at the cost of output transistor power dissipation. As bias is increased, one reaches a point of diminishing returns where the THD improvements becomes smaller while the output transistor power dissipation continues to increase. The 20mA idle current spec is likely a compromise between low level THD performance and heat dissipation.

The following illustrates the relationship between THD, output power and bias current as well as how the output transistor power dissipation is affected by bias.

 

[fromnowon]

Wow, that's really helpful. Thanks LesE! That's awesome that you were able to find those plots for an SX-636.

Interesting that they plotted a 40mV bias curve, and it has the lowest THD of those 5 samples, really low at low power output levels.

I've been listening to this receiver a lot over the past month, with the bias set at about 40mV. Sometimes all day long and sometimes I've tested pretty high volume levels with speakers. It doesn't get hot at all, barely warm - I still have the covers off, and the heat sink and outputs are almost cool to touch. It sounds GREAT. Crystal clear, punchy deep bass especially with headphones. It's addictive!

I wonder, could I leave the bias at 40mV? Maybe I like it like this.

 

[LesE]

That's awesome that you were able to find those plots for an SX-636.

Actually, these plots weren't found. I extracted this data from an LTSPICE circuit simulation for the SX-636 amplifier. Although this isn't empirical data, it's useful for illustrative purposes.

Interesting that they plotted a 40mV bias curve, and it has the lowest THD of those 5 samples, really low at low power output levels.

Not a coincidence. I ran the simulation for 40mA knowing that this is the bias you are currently using.

I wonder, could I leave the bias at 40mV?

As long as it's not running overly warm, you're probably safe leaving the bias where it is.

 

[fromnowon]

I really appreciate that you went through the trouble to run that simulation. Thanks LesE! That's great information to have.