Updating the Dynaco Stereo 120

[djoffe]

I took an old Dynaco Stereo 120 Power amp, and made some PCB's to retrofit LM3886s. I've attached a picture of the first completed version. I haven't listened to it yet, but the quick measurements so far:
*a bit more than 60 Watts into 8 Ohms before it clips
*harmonics of 1 kHz better than 80 dB below the fundamental below clipping
*3 dB frequencies of 10 Hz and 100 kHz
*about 1.1 Vrms input for full power output

The insides look a bit nicer than they used to, I think...Hopefully I can set it up for some listening in the near future. (picture is just a quick snap from my cell phone camera).

[lanchile]

I would use a "real" heat-sink for those LM's. I do not think what you are using as a heat-sink will be enough to cool down those chips ( unless you do not push them) The LM3886 can get really hot if they do not have an adequate heat-sink. Just my thoughts!!!

[indianajo]

The heat management of the ST120 was to me it's biggest failing, worse than the crossover distortion it would produce at volumes lower than a clock radio. I never listened at volumes that low anyway, but after a 3 hour choir rehearsal at probably 10 W minimum/ channel, mine melted the solder on the snap-in output cap (not the original dynaco with the holes in the tabs) and produced a 12" fireball from vaporized solder when the solid core wire pulled away. The junior choir was very impressed with the fireworks. After a needed break, the rehearsal finished on the ST70.
My original output ST120 with the DJoffe quiet bias increase pcb, will get hotter than bathwater at the transistor flanges after 4 hours idling or 1 V output service. I'm satisfied with the sound, so rather than change the output transistors, I put "Mouse Ears" on it. Two 3" fans salvaged from PC switcher power supplies blow through the cover on the two output transistor mounting flanges. They are screwed to a steel channel sitting under the amp, and are powered by a 9VDC wall transformer separate from the St120 power supply. As the whole hi-fi is on a power strip, this is not an operational problem. At 12 hours, the output flanges are cool to the touch.
I have plenty of surplus heat sinks but haven't figured a way to cram them in and still put the cover on. PA amps use fans pretty universally, anyway. On into the 40th year of this tough beast. It is now amplifying the output of my HDTV tuner on PBS network music shows.

[jez]

Are you using a unipolar PSU and output capacitors there? Different. Useful if there's a DC fault though as it will save your speakers!

[djoffe]

Yes...the original supply is regulated 72 Volts, so I'm using that single supply and a 3300 uF output capacitor

[djoffe]

I thought lanchile's caveat about heatsink size was worth looking into. My conclusion is that for any kind of "normal" listening, recycling the Dynaco heatsinks should be adequate. Here's the data I took:

I drove one channel into 8 Ohms (didn't have two power loads), and found the following temperature rise data for the driven channel:

Ambient temperature is 25 C.
Temperatures measured with a thermocouple on the case of the driven LM3886.

These measurements (except #4) were made with the perforated top removed:

1. Temperature measured with a thermocouple on the case of the LM3886.
Power up the amp, with no signal, case temperature of LM3886 stabilizes at 39 C (e.g. human body temperature).

2. If we rate the amp at 60 watts into 8 Ohms, then the old FTC 1/3 power pre-conditioning test calls for pushing 20 Watts into 8 Ohms, or an output voltage of 12.65 volts RMS. The case temperature was 100 C in this condition.

3. Calculations show that on a 72 Volt supply, 16 V RMS is about the worst-case internal dissipation. In that case, we have worst case case Temperature of 108 C.

4. Same test conditions, but with the top in place, case temperature increases to 112 C.

Some thoughts...This is probably a test much rougher than any normal listening. Even in this case, we're beneath the limits assuming 25 C ambient.

Thermal resistance from junction to case is 1 C/W. Worst-case internal dissipation is about 32 W, meaning the junction temperature will be 32 C above the case temperature. That says with the given heat sinks, driving 32 Watts continuous into 8 Ohms (a worst-case condition), makes the junction temperature around 140 C. The part is rated to Tj=150 C, with an instantaneous SOA temperature of 250 C, per note 9 on the data sheet.

So...for normal listening...heatsinks probably adequate...for high ambient temperatures, or very compressed music at very loud levels, you might want to use finned heat sinks, as lanchile recommends.

[djoffe]

Here's a close up of the bare PCB that I used to make "modern" amplifier modules to replace the original ones in the Stereo 120.
For reference, the other picture is the PCB side of one of the original modules.
I'll try to upload a stuffed board picture in a day or so.

OK...I added a third and a fourth picture...the old zobel networks and the output caps with the wound 'round inductors...just because it almost looks like art!

[djoffe]

just starting to load the board.
all but the power op amp.
side view
mechanical detail
finished assembly (one channel)

[GlassAudio]

Do you have a schematic for the mods?

[djoffe]

I'll try to post the schematic in the next day or so...It's done in a kind of heavy duty system that isn't appropriate for posting...I'll try to make a clean pdf to post.