Technics SU-8044 - Measurements

[IAmAlwaysSerious]

I recently picked up a Technics Integrated Amp to power a pair of bookshelf speakers. A THX 789 has been powering them just fine for now, but i'd like to move it back to my desk to power some headphones again.
As much as i like the glow of the tube in the Schiit Vali 2 i've been using, it does audibly distort. The 789 also has a much better volume knob, in my opinion.

Since i have some okay equipment lying around, i decided to make a few measurements of the amp, especially after the recent measurements of the JBL SA600 posted recently.
The SU-8044 was, if the internet is to be believed, released in 1979. It's not quite as old as transistors, but it's much older than my 789 and at probably half a decade older than the speakers i plan to power with it.
I'm not sure when exactly this unit was built. If anyone knows where to find a manufacturing date on these, i'd appreciate knowing.

All the measurements here are high-passed at 10Hz and low-passed at 20kHz because everything outside of that is likely unimportant. This amp might be DC coupled, so i ought to check and see if there's any DC offset. If i have time, i might look it over with a scope tomorrow.

The first thing i decided to test was the usual 1kHz, here's the FFT:

Input 1 is the left channel and input 2 is the right channel. The DAC used for the test was a THX Onyx which measures slightly worse than the one that Amir tested some time back (SINAD in the ballpark of 102dB, i think. THD around -110dB). Because of this, it's possible that the distortion in the amp is being slightly exaggerated. The DAC where i intend to use it is probably significantly worse, so errors here don't bother me too much.
Measurements were made from the headphone output with a E1DA Cosmos ADC set to 10Vrms, without any additional load. Once i get my hands on a decent test load, i may re-evaluate it(I didn't use speakers because i didn't feel like making that much noise today). In any case, i expect the headphone output to be relatively similar in quality to the speaker outputs. I'm pretty sure both are single ended.

I'll say, the next measurements surprised me. I have no clue what is causing the hump, but i don't think it's the ADC and i don't think it's the DAC either:

One channel clearly suffers worse than the other, but they both show similar faults and to similar degrees. The weird drop off in distortion confuses me since i don't think this thing has any sophisticated gain staging. I should probably point out that this amp appears to be in good condition but is completely unrestored. None of the caps appear to be bulging or burst from what i can see through the top grate.

Re-running that section of the sweep didn't seem to change anything, so i decided to run the sweep just on the THX onyx, Full scale here is 2.7Vrms:

There's a bit of weirdness, but nothing like that hump. No easy answers to this mystery then.

Now, the info i could find on the internet suggests that this thing should have a pretty good phono input. I decided to run an intermodulation sweep (3005 Hz & 4462 Hz 1:1) with the THX Onyx powering it, but brought the voltage range of the sweep way down, capping out somewhere around 180mVrms instead of 2Vrms. The results look good, considering that this is the combined noise and distortion of the amp, phono pre-amp and THX Onyx:

(the second one is the right channel, i appear to have wrote left twice)

Noise is clearly dominating for most of it, as expected.

It occurred to me while writing this that i should probably have an FFT of the 1kHz tone into the Phono input, so i went and hooked everything back up for just this graph. And while i said THD+N was dominated by noise, i didn't realize just how bad it was. The distortion is actually about the same as the THX Onyx output, so i think we can pretty much discard it in the following graph, but noise here is... Well, it's bad:

It seems like the numbers were slightly worse this time. I think that next time i test a Phono input, it'll be with the THX Onyx set to full volume and then i'll use the 789 as a pre-amp to bring the voltage down to 5mV. That might get me a better THD, and maybe even slightly less noise. It could be that a decent amount of the noise is from the THX Onyx, being amplified by ~44 dB by the pre-amp and amp.

I listened to this thing with my larger living room speakers, and it seemed to power them pretty well when playing music from my phone and the THX onyx. I'd bet money it'll handle the bookshelves just fine; in a way, i have bet money on it.
Realistically, this is on par with plenty of modern stuff and even beats some expensive amps of today.

P.S.:I don't actually intend to use it with a record player, but i might give it a shot sometime just to see how it is.
P.P.S.:I don't think that at any point i saw a difference of more than a dB between the fundamentals of each channel, even if the distortion was quite different between the two, so i don't think the difference between the two channels will be at all audible.

 

[solderdude]

I'm not sure when exactly this unit was built. If anyone knows where to find a manufacturing date on these, i'd appreciate knowing.

The first 2 numbers of the serial number are the year of production.

You should also measure the FR. IME not all Technics amps measure 'flat' with the tone control potmeters in the centre position.
Speaker amps should be measured under load. They might perform fine without load (all are class-A basically) but distortion might be much higher under load.

Headphone Rout = 340 ohm (which is quite high) so there's that as well. Measuring under actual loads may give completely different results.

 

[IAmAlwaysSerious]

The first 2 numbers of the serial number are the year of production.

The Serial number is mostly letters, but the first and second actual numbers in it are 8 and 2. 1982 sounds right, since i believe it released in 1979.
Thanks for the info.

You should also measure the FR.

I had quickly done that with periodic white noise and the speaker outputs on friday, and i recall it appearing flat, but i'll retest after work tomorrow. I figured it would be boring so i didn't save the measurement. hah

Speaker amps should be measured under load.

I do plan to set up a little something for this, but i'm not able to just yet. I could use speakers but i don't want all the noise.
I suppose an 8 ohm resistor that's able to handle twenty or forty watts is likely to be sufficient, right? It won't offer much in the way of reactive load, but i bet it'll be fine.
i'll probably order something soon.

At least going through the headphone jack tells us it would make an acceptable headphone amp but probably a bit noisy for IEMs, eh.

 

[Rja4000]

My very first amp.
I bought it new in 1979, indeed.

Nice looking meters and buttons.
2 x 38 W under 8 ohm, if I remember correctly.
Headphones output has a quite high impedance, as @solderdude confirmed.

I didn't have enough money for a decent pair of speakers back then. So I was listening to music with a Stax SR44, the "cheap" electret Stax.
I still have the Stax.

 

[IAmAlwaysSerious]

2 x 38 W under 8 ohm, if I remember correctly.

You've got it exactly. Just the perfect amount of power to have on tap with the speakers i plan to use. I'll be hooking it up to a pair of Technics SB-F30 i found for cheap a couple weeks ago.
Interesting, the meter on the front shows up to 100W for peaks.

You're right about the knobs all feeling nice, too. That solid metal feels much better than the plastic buttons on the (slightly) newer amp in my living room.

I didn't have enough money for a decent pair of speakers back then.

I suppose that's the great thing about listening to music today; i can find a decent pair of thirty-five year old speakers for a few bucks and and amp for a little more.

So I was listening to music with a Stax SR44

Funny, i almost bought a pair of these a few months ago as a way to try out STAX. I ended up with a pair of Stax SR Gamma instead though, for fairly cheap. The sound alright to me, but it would be fun to see them measured properly.

 

[pma]

I'll say, the next measurements surprised me. I have no clue what is causing the hump,

Looks like sudden change in bias current of the output stage. They use those STK0039 modules

and this is the distortion plot from the amplifier sheet. Their measurement shows high dependence of distortion on frequency.

 

[solderdude]

Looks like sudden change in bias current of the output stage. They use those STK0039 modules

unlikely in this case as the test was not loaded. So the STK039 would always be in full class-A.

 

[restorer-john]

The Serial number is mostly letters, but the first and second actual numbers in it are 8 and 2. 1982 sounds right, since i believe it released in 1979.
Thanks for the info.

Just take the cover off and look at the date codes on a bunch of the interior components. That tells you.

 

[pma]

unlikely in this case as the test was not loaded. So the STK039 would always be in full class-A.

Yeah, but plots look like Gm doubling point or oscillations.

 

[solderdude]

The amp is just an LTP + 1 gain transistor + bootstrap (low capacitance value as well).
I don't remember these things oscillating but when I serviced those it was 30 years ago and they were considered 'old' back then already ..

 

[IAmAlwaysSerious]

Just take the cover off and look at the date codes on a bunch of the interior components. That tells you.

It's like the tech equivalent of counting rings in a tree.

I actually looked for a date on the two STK0039 in there through the grill on top when i bought it. Couldn't make out anything like a that though.

I suppose opening it would have been the better bet.

The amp is just an LTP + 1 gain transistor + bootstrap

Seems like might be able to answer a question I had while looking through the top grill of the thing.
Just near each of the power amp chips, I see a big 0.47 ohm resistor with three poles, it seems to be where the two outputs on an STK0039 are summed for its channel.

My best guess is that it's there as part of the power output sensing circuit, but that seems wrong to me. Any idea what it could be?

 

[solderdude]

These are emitter resistors and are essential for maintaining idle current.
In most amps they are also used to adjust bias current but the STK039 are 'fixed' (so determined by the emitter resistors basically)


All Panasonic/Technics model devices (at least from that era up to at least 1997 and most likely beyond that date) had the production year in the first 2 numbers.
I worked for the Dutch Technics/Panasonic service centre in the Netherlands for about 6 years or so.
Its the only correct way to date Technics/Panasonic.

 

[Doodski]

Service manual is here. The STK0039 has a equivalent circuit that appears to be a Darlington arrangement. STKs are usually not as clean sounding as a properly designed discreet amp. I'm not experienced first hand with the STK0039 but maybe somebody here is and can chime in on it's integrity.

 

[solderdude]

I have replaced quite a few of these STK0039 back in the days. The 0039 was the most used one of this particular series.
Quasi complementary output stage with fixed idle current (so no potmeters that could go faulty).
Later ones were almost complete amplifier circuits that only required some external parts.

The ones you can get these days are even poorer than the originals.

 

[IAmAlwaysSerious]

I've had some interesting results from the Phono input with a real world load on the left channel. It's possible i made a mistake when i was playing with this before, but i'm not sure where it would have been.
Here's the result of a 5mVrms 1khz tone into the SU-8044:

The output meter on the amp read about 5 watts, i must say the 1kHz tone playing out of the speaker was quite hard on the ears. I had a pillow blocking the sound and it was still much more sound than i would have liked. Speaker was a Technics SB-F30 on the 8Ohm posts.
Full scale here is 10Vrms, i did measure the signal out of the amp from the headphone output, but i suppose i can try measuring in parallel to the speaker if we aren't happy with this.

The thing i find most interesting here is that there's hardly a difference between the channels even though one of them is driving a speaker. In fact, most of that difference is present in the THX Onyx i used to generate the signal and the 789 attenuating that signal.
Here's a measurement of the Onyx+789 generating the 5mVrms i input into the amp:

It's not quite as good as i hoped, but i'll take it. Full scale in this image is, i believe, 2.7Vrms.

 

[IAmAlwaysSerious]

You should also measure the FR. IME not all Technics amps measure 'flat' with the tone control potmeters in the centre position.

Looks like this is one of the good ones. All pots set to their center detents, "loudness" and "high filter" both set to off, using the tuner input, i got what looks to be very flat across the board. I set volume fairly low because i don't really know how much power these old tweeters can handle.

Note that the scale here is extremely exaggerated:

This is on the left channel, which is currently powering an 8 Ohm Speaker, right is substantially similar. Total range is exactly 1dB, when i check with the cursor in the measurements. (ignore the "distortion" legend at the top, we're only looking at the fundamental.)

It occurs to me that a little wiggling of the knobs might get that a bit closer to perfectly flat, but this will easily be good enough where i plan to use it.

 

[solderdude]

Sometimes the pots have to be set slightly 'offset' from the center. 1dB lift is audible already, certainly when it concerns a large part of the frequency response when critical listening is done.

 

[IAmAlwaysSerious]

I messed around a little bit with it, but wasn't able to get a flatter response than turning treble down by one. It's still roughly 1dB of variation, but with bass being slightly higher than trebble.

Considering the small speakers it's being used with, that's unlikely to hurt the sound in any noticeable way.

Anyway, the amp is set up in its final resting place. I'm pleased with the performance and the measurements seem to support that.
I'm almost surprised with how good a cheap forty year old amp can be.

 

[microwave]

Hey one question, so I have the same amp and I'm gonna use it with the qudelix 5k as my dac/preamp and I'm wondering how loud should I set the q5k to not blow anything ?

 

[bvotje]

It's like the tech equivalent of counting rings in a tree.

I actually looked for a date on the two STK0039 in there through the grill on top when i bought it. Couldn't make out anything like a that though.

I suppose opening it would have been the better bet.


Seems like might be able to answer a question I had while looking through the top grill of the thing.
Just near each of the power amp chips, I see a big 0.47 ohm resistor with three poles, it seems to be where the two outputs on an STK0039 are summed for its channel.

My best guess is that it's there as part of the power output sensing circuit, but that seems wrong to me. Any idea what it could be?

Hi, the big 0.47 ohm three pole resistor you refer to, any idea where I can buy these things? Mine seems broken and I cant find them online. To be honest, my knowledge in this field is minimal.