Resurrection: Old Parts - New Amp.

[MJL21193]

I don't need any more amps.

BUT...

I have these parts that are bothering me - mocking me, daring me to use them. They were once in the thick of things, members in casts of hundreds brought together to reproduce music. Their implementation was less than ideal, somewhat undeserving of their (near) excellence.
I liberated them from the bonds of mediocrity, ill repair and flat-line status to one day be used in a project that could be a more deserving purpose. Could be a new beginning... a revival from the dead.

Here are some:



Old solders, scarred and rough. We see sergeant Toshiba, lieutenant Panasonic, colonel Sanken and marshal NEC. Also in attendance is corporal Mitsubishi. All battle hardened veterans, but with many years of service left to offer.

I have some ideas, some plans to try and make the best of these. Putting them into modern uniforms with new support troops, playing out a modern war on a bright new battlefield.


I'm working with a schematic that I'll post shortly.

[MJL21193]

Heatsink and transformer:



The heatsink is from an old Pioneer and the transformer came out of either an 80's era Fisher or a Kenwood, i can't recall. It's a 40-0-40, probably 300VA or better.

I'm still working up a schematic but have had some distractions. I want to use the 2SA798/C2291 for a symmetrical complementary differential input stage. These are monolithic pairs with 5 pins (emitter is common). 50V, so I'm looking at a cascode on each to work at the higher voltage.

A look at the proposed front end. Suggestions?

[MJL21193]

Wow, not much interest...

It only looks like a pile of junk
These parts cost me next to nothing - most are out of old non functioning amps and receivers that I picked up cheap or had given to me for nothing. The objective is to build a high quality 2 channel amp from these parts, without using anything new (except electrolytic caps and resistors).

Full schematic below. The input pairs are the monolithic 2SA798/c2291 and for the cascode and CCS I will use original NEC 2SC1845/A992. I have plenty of the C1845 but I'm short 1 A992. I'll need to do a bit more looking.

For the VAS, I've picked the 2SA949 and the 2SC2229. These are fast, low Cob and high enough voltage for the task. They are running at ~2mA, so dissipation will not be high.

[SQLGuy]

Looks like a fun project.

In the amp I'm working on now, though, I think I am dealing with some transistors that test fine and still amplify, but are damaged and causing annoying levels of noise.

I would be concerned the some of the pulls you have among your parts may be similarly damaged. Do you have enough extras to swap parts in and out to get a "good" set, or do you have a good way to test them more comprehensively?

[ostripper]

It only looks like a pile of junk

Glad to see somebody is helping to clean the planet besides me.

I am ADDICTED to junk , have all the kids bring it to me for
extra candy money/cell phone minutes. Trannies are suspect
unless unit is newer and had just low voltage issues.

If newer, Caps and trafo's are gold ($$$$) , more money for
better caps and LM4702's.. which by the way ,are great for
using your typical HT reciever "guts" (darlington or big sanken)
here is my latest find// $1.00 for "candy"..

BTW ..zener and decouple your cascodes ...15v zener to ground
and 10u to rails...

OS

[MJL21193]

Originally posted by SQLGuy
Looks like a fun project.

In the amp I'm working on now, though, I think I am dealing with some transistors that test fine and still amplify, but are damaged and causing annoying levels of noise.

I would be concerned the some of the pulls you have among your parts may be similarly damaged. Do you have enough extras to swap parts in and out to get a "good" set, or do you have a good way to test them more comprehensively?

Hi SQLGuy,
I have quantities of several with a few exceptions: The NEC outputs are fairly rare and I haven't found any others. I do know these are working well though as they were pulled from an amp that was working fine. I already noted the 2SA992's, and that I'm one short (need 8 in total). I have since found 5 more, so I have some spares. These I have tested with my DMM just to get an idea if they are viable. I could create a "testbed" to run these transistors on, but I think I'll roll the dice and try them in the final circuit. If there is a problem, It should be easy enough to spot.

More device talk: I chose the 2SC2590 and the 2SA1110 for the CFP "slave" stage and the 2SA1490/2SC3854 Sankens (no spec sheet links available) as the drivers. I have these running hot and contributing to the output.

[MJL21193]

Originally posted by ostripper


Glad to see somebody is helping to clean the planet besides me.

BTW ..zener and decouple your cascodes ...15v zener to ground
and 10u to rails...

OS

Hi OS,
I think it's something that more should do here (rather than endlessly chat, dream, argue and pose) - you can easily strip enough out of a receiver or two to build a very good amp for next to nothing. I have excellent new parts that I could use to build with, but this will be a nice challenge and a way to justify my childhood tendency to take everything apart...

You are forgetting that I don't like the evil zeners in that position. A voltage divider like I have used for Patchwork works very well.
I did fail to put a decoupling cap on my voltage divider to ground though, and thanks for reminding me. I will use a 1uF MKP or ceramic.
Speaking of which, why are we using silver mica for Cdom when a good quality ceramic will do the job? Is there really that much of a difference?

[ostripper]

why are we using silver mica for Cdom when a good quality ceramic will do the job? Is there really that much of a difference

Here is what I found out , it is not the material of the cap that
is the issue , it is the construction that matters. multiple layer ceramics are the same as silver mica's , just cheaper..

The single 2 plate ceramics sometimes exhibit resonance
at certain frequencies (just like 2 parallel PCB traces).
Miller compensation is a resonate circuit , so adding another
pole might not be good.

You are forgetting that I don't like the evil zeners in that position

Then use some precision Vreg IC's (dirt cheap) ,or even 78/9 -15's (scrap)..
OS

[MJL21193]

Originally posted by ostripper

multiple layer ceramics are the same as silver mica's , just cheaper..

Then use some precision Vreg IC's (dirt cheap) ,or even 78/9 -15's (scrap)..
OS

I made a mistake on a Digikey order last year and got 100pF ceramics instead of 100nF bypass caps. I ran out of silver mica to finish my six channel amp and tried the ceramics - they worked fine, no problems. Sounds the same to me too(that doesn't mean much).
I up for trying it here also.

Why do you think I need to complicate the cascode supply? The voltage regulators you mention will not operate (for long) at my supply voltage.
My voltage divider gives ~28VDC for the cascode at 1.2mA - very good and simple.

[jaycee]

Also the ceramic dielectric must be the high stability type. A lot of them vary capacitance widely with voltage and temperature.

The only thing that would give me the willies about using parts like this, is what happens if they break? OK, in the case of the transistors, you could sub modern parts. I have some nice 2SK389 dual FET's from the Teac amp I gutted, but am cautious about using them because they are pretty much irreplaceable.