Those Radio Shack Mini Amps ...

[atmasphere]

That is indeed an SA-175 -- the first version had switches on the bottom, below the knobs.

View attachment 1244859

SA-175B and SA-175C moved the switches to the top, with the knobs below, like in Blast's pic at post #15.

I have an original SA-175, as well as a C version. The switches are above the controls, not below. The C model has good bandwidth to 60KHz; the original barely makes it to 20KHz and really isn't flat, owing to its coupling transformers in the output section and germanium construction. The date code on my original suggests it was made in October of 1969.

In case anyone wonders what the appeal is, the form factor of the SA175 is very nice and gives it a mid century modern appearance that isn't seen in most of the other Realistic amps. Sony made an amp and matching tuner that were upright and also have a mid century look, but they are much larger. I like the small form factor as the amp is strictly for use in my bedroom. I'm using it with a set of Optimus Pro 5s with the Linaeum tweeters and a sub. I'm never ever close to clipping and due to the single-ended nature of the design, with all that 2nd harmonic its sounds rather sweet, like a tube amp. It probably would not do that at full power, but it will never be driven that hard.

 

[Stuart Pedaso]

I have an original SA-175, as well as a C version. The switches are above the controls, not below. The date code on my original suggests it was made in October of 1969.

Thanks for pointing that out as I had incorrectly implied top positioned switches began with SA-175B. This despite having privately noted the change occurred earlier when gathering data. So first catalog year (1969) models have switches below the knobs. 1970 and '71 catalogs show them positioned above -- prior to the "B" variant which started in 1972.

Yours manufactured late '69 would've been part of the second production wave; stock for the upcoming 1970 catalog year.

In case anyone wonders what the appeal is, the form factor of the SA175 is very nice and gives it a mid century modern appearance that isn't seen in most of the other Realistic amps.

I've always appreciated the clever design choice here of building to depth rather than width. Though largest of Tandy's "minis," at just 7" (~18cm) wide SA-175 and companion components effectively displace less usable shelf space than even SA-10!

 

[atmasphere]

I just picked up an SA-10 on a lark. I replaced all the electrolytics in the power supply and one channel so I could suss out how it behaves. Bandwidth was good to well over 150KHz- a bit unexpected. I expected the tone control to create a rolloff up high but it didn't. But the bass was an issue- it got into trouble by the time I got down to 30Hz; doubling the output coupling capacitor value had no effect. I'm investigating other component changes to see if I can extend if further. Also its four ohm power was double its 8 ohm limits- so apparently it can behave as a proper voltage source with many loudspeakers.

This amp, the SA175B and C, the SA-100C and SA-102 to my eye are the best looking of the minis and are also the best sounding. The reason seems to be the relatively primitive output circuit, which runs single-ended all the way to the output transistors. So the single driver transistor in any of them has to be heatsinked, and drives one of the output transistors directly. The output of that is the speaker signal, but is also used to drive the other output device. IOW the output section is the phase splitter circuit as well. Overall the power amp section employs 4 transistors and 3 coupling caps. I've seen tube amps like this; none very impressive. But in the case of a smaller solid state amp of this topology the primary distortion component is the 2nd harmonic. So to get a taste of what these amps are about, any one of these models will give you an idea, although I believe the SA-10 be less representative. I was surprised by the SA-102 in that with just one channel driven did over 6 watts with no trouble- I was expecting only 3. At any rate I suspect the popularity of these currently has a lot to do with that 2nd harmonic. Some of the later versions used output ICs; these ICs are considerably more sophisticated and don't make as much distortion and I suspect thus less appealing.

It would be interesting to compare an IC-based version against a refurbished amp of this type.

I'm not aware of any other manufacturer making amps this small that are meant to be hifi. Lafayette made the Stereo 10; I'm guessing there that the B version is the one to get but technical information is scarce. I can see that in the 10 interstage transformers are employed. I'm thinking that type is to be avoided....

 

[mhardy6647]

 

[Blast]

Atmasphere, as you may have seen from post #15 I am a (LONG time) owner of the aforementioned models SA-100B, SA-175 and SA-175C (and their matching tuners).

One of my first lessons in electronics and transistors was understanding how and why the model SA-100B produced and ran on a negative (-25v) DC supply to operate the amp. Before absorbing the technology (any better than I have) I mistakenly thought that perhaps the guy that drew the schematic had just drawn the diodes in backwards.

Unfortunately, I have never listened to (much less even operated) models SA-10 and SA-102. I keep thinking about adding those to the collection but it hasn't happened yet.

But, I enjoyed your write-up and opinions on the amps that I do own and was very interested in your work on your model SA-10. I would like to hear your comments and opinion regarding any audible differences between both the discrete (31-1982 and 31-1982A) and IC (31-1982B) output versions.

Brian

 

[Stuart Pedaso]

I just picked up an SA-10 on a lark. I replaced all the electrolytics in the power supply and one channel so I could suss out how it behaves. Bandwidth was good to well over 150KHz- a bit unexpected. I expected the tone control to create a rolloff up high but it didn't. But the bass was an issue- it got into trouble by the time I got down to 30Hz; doubling the output coupling capacitor value had no effect.

Surprising results. SA-10 seems shy of "high fidelity" aspirations from its marketing perspective. Catalogue ad copy promotes it as a "starter" component suitable for "den, office, children's" use. User manuals don't bother to rate these below 60 Hz. Though published HF bandwidth fares better for the discrete vs IC amp version:

Some of the later versions used output ICs; these ICs are considerably more sophisticated and don't make as much distortion and I suspect thus less appealing.

The move to IC amps was almost certainly a cost cutting measure rather then performance based. When SA-102 (31-1963) adopted the LA4440 IC (31-1963A) US price dropped from $79.95 to $59.95.

SA-150 is a repackaged SA-102(A). It's price held at $59.95 throughout its 9 year run.

SA-155 is an SA-150 with slight I/O changes. Its price held at $59.95 for at least another 9 years.

SA-10 enjoyed similar price stability, due at least in part to its move to IC amps. It retailed for $29.95 throughout its 20 year run, with the last 10 years employing chip-amps. Compare mhardy6647's 31-1982A image to the barren landscape of my 31-1982B below it:



Anyone experienced in manufacturing will see obvious production cost differences contrasted here.

I'm not aware of any other manufacturer making amps this small that are meant to be hifi. Lafayette made the Stereo 10; I'm guessing there that the B version is the one to get but technical information is scarce. I can see that in the 10 interstage transformers are employed. I'm thinking that type is to be avoided....

Lafeyette also had the Stereo 20. I had always thought of Stereo 10 as roughly competing with SA-100x / SA-10 and Stereo 20 with SA-101 / 102 / etc. Stereo 20 was also available as an assembled PCB. User supplied power supply, cabinet, switches, I/O and knobs. Got one around here somewhere.

 

[atmasphere]

I'm guessing that the low frequencies were cut off with intention as the amp makes so little power. It was certainly designed before the era of the desktop! The problem is such a cutoff will have audible phase shift to frequencies well above 60Hz! The SA-10 has temporarily replaced a 5 watt class D amp I built. The class D has much better low frequency bandwidth and despite my desktop 3" full range drivers being somewhat limited in that department the difference is clearly audible. However the power is adequate, so it seems worth it to me (since the class D has horrible offsets at low power, making for something that sounds like bad crossover distortion, something that is no worries in the SA-10) to see if I can get it to go a bit lower.

 

[atmasphere]

Since I need to make a trip to get parts for the SA-10, I decided to suss out an SA-101 that I had also picked up recently, since my efforts to find a clean SA-175 for the basement system have been frustrated. This amp employs much the same circuit but makes about 3 watts. I just discovered that the source of the rolloff on the bottom end of this amp is because the line level input signals are first given inverse RIAA equalization and then fed to the magnetic phono preamp! This apparently was done to reduce the cost of switching. Even though I'm a vinyl fan, I am thinking to bypass all that as this amp will likely never be used with vinyl.

 

[Stuart Pedaso]

... I decided to suss out an SA-101 that I had also picked up recently, ...
... I just discovered that the source of the rolloff on the bottom end of this amp is because the line level input signals are first given inverse RIAA equalization and then fed to the magnetic phono preamp!

I found that quite curious myself. It's the same for early (discrete) SA-102, which is just a repackaged SA-101:



Beginning with later SA-102, all LA4440 based amps have inputs other than Phono bypassing the phono preamp and RIAA equalisation. But an inverse RIAA circuit is still used at the "Phono" inputs when switched to "ceramic" position:

 

[atmasphere]

I found that quite curious myself. It's the same for early (discrete) SA-102, which is just a repackaged SA-101:

Beginning with later SA-102, all LA4440 based amps have inputs other than Phono bypassing the phono preamp and RIAA equalisation. But an inverse RIAA circuit is still used at the "Phono" inputs when switched to "ceramic" position:

That's actually compensation for the fact that ceramic cartridges have a variable reluctance pickup characteristic which almost compensates the RIAA curve. I'm impressed that they bothered to put that in there; most cheap stereos treat a ceramic cartridge as it its flat, which it isn't.

My guess on the SA-101 why they pumped all the signal through the magnetic phono section is they wanted more gain for the output section. But that gets the amp in trouble if there's deep bass; the phono section overloads before the amp reaches full output in the bass due to EQ errors in the phono section and resulting overload. The switch easily allows you to bypass the phono stage with the phono input hardwired directly into the phono section (so I didn't have to abandon it after all) without any modification to the board or the switch. This got rid of the distortion present in the bass and now I just see a gentle rolloff at about 60Hz which is what I was hoping for. Switching to a 16 ohm load put the rolloff an octave lower and since the output coupling cap is outside of the feedback loop, it appears to be the main culprit. This really reflects my personal experience working as a service technician at Radio Shack in the 70s- they frequently left performance on the table, often easily fixed, if it saved them 25 cents.

 

[atmasphere]

Turns out the bandwidth limitation in the SA-101 is the input coupling cap to the power amp. Its a 0.056uf device; the cutoff frequency is about 21Hz. Since I was unconcerned about limiting low frequencies for power reasons I replaced those parts with a 0.1uf. I replaced the output coupling caps with 2000uf units, the main power supply cap with 3000uf and the secondary part with 1000uf. I also doubled the value of the coupling caps feeding the top output transistor. The amp makes about 6 watts with one channel driven into 4 ohms, so is able to behave as a voltage source with most speakers. As with my SA-175, making it listenable meant disconnecting the loudness contour.

 

[Stuart Pedaso]

The switch easily allows you to bypass the phono stage with the phono input hardwired directly into the phono section (so I didn't have to abandon it after all) without any modification to the board or the switch. This got rid of the distortion present in the bass and now I just see a gentle rolloff at about 60Hz which is what I was hoping for. Switching to a 16 ohm load put the rolloff an octave lower and since the output coupling cap is outside of the feedback loop, it appears to be the main culprit.

Turns out the bandwidth limitation in the SA-101 is the input coupling cap to the power amp. Its a 0.056uf device; the cutoff frequency is about 21Hz. Since I was unconcerned about limiting low frequencies for power reasons I replaced those parts with a 0.1uf. I replaced the output coupling caps with 2000uf units, the main power supply cap with 3000uf and the secondary part with 1000uf. I also doubled the value of the coupling caps feeding the top output transistor. The amp makes about 6 watts with one channel driven into 4 ohms, so is able to behave as a voltage source with most speakers. As with my SA-175, making it listenable meant disconnecting the loudness contour.

You've just posted a tutorial for boosting performance of an SA-101!

Now to find where I stashed mine ...

 

[atmasphere]

You've just posted a tutorial for boosting performance of an SA-101!

Now to find where I stashed mine ...

Here's more, and applies to all of the mini amps that don't employ chips or driver interstage transformers:

The circuit is very similar in all of them. The SA-10 is the simplest, having only a voltage amplifier in front of the power amp. But in all of them, bandwidth is curtailed on the bottom end due to distortion that otherwise arises. That distortion is caused by small electrolytic caps (two in particular); partly a cost issue and probably also a method of getting the amp to seem like it has more power since the bass is where the energy is, and smaller amps like this are often on smaller speakers.

The first electrolytic to be questioned is the input coupling cap. The formula for calculating its cutoff frequency is f= Rx C x 6.28/1,000,000 People who are familiar with this formula will have noticed that usually the divider is 1, but here its a million so that the actual cap value in uf and the resistor value in ohms can result in Hertz. Essentially the input bias network values are in parallel since they go to ground or Vcc+ which has a low impedance with respect to ground. We can see right away that this cap is rolling far too high; I recommend setting it about 4x lower. I've ignored the impedance of the base of the transistor BTW in the above formula. This isn't rocket science; I just need to be able to measure and hear improvement with real bass.

The second cap is in the emitter circuit of the input transistor to the power amp- its an emitter bypass cap. Its low end timing constant means that anything below that frequency is rolled off. Now if you simply put in a bigger output coupling cap, the result will be that there is less feedback below the cutoff of that emitter bypass cap, so the amp makes more distortion. On top of that, the coupling cap to the base of the top output transistor has a timing constant too. If the amp is seeing frequencies below the cutoff, the output waveform will be asymmetrical! Now the amp won't see any of these frequencies unless either the caps in question start to fail, or if you install bigger values at the input and output of the amp.

But it would be nice if the amp made more bass- especially in the case of the SA-10, if you're running a desktop. So you have to replace all five electrolytics in the power amp section. The output coupling cap should be 2200uf if you really want a proper timing constant with relation to an 8 ohm load. That puts the -3dB point at about 9Hz which is still a little high, but it will mean less phase shift over much of the audio band so the amp will be more 3D. In the case of the SA-10, this means that all the caps have to increase in value by about 10:1. In the cases of the other amps its more like 4:1.

Also, to me the amps tend to sound a bit dull with digital sources. This is not due to bandwidth, as all of these amps have bandwidth to 60KHz and beyond. Its due to the loudness contour. I don't like loudness contours because they are really hard to get right- there will be only one input level, one speaker efficiency and one distance from the speaker in a room of the right liveliness for the contour to be correct! So I pull the loudness parts out (a cap and resistor tied to the tap on the volume control). This gives the amp more gain and greater liveliness without being bright.

Now there is one spot in the circuit that I'm a bit curious about. The input voltage amplifier to the power amp is the transistor (usually a 2SC373 or the like) with the emitter bypass cap. Its biased on, but the bias circuit provides no feedback. The issue is that when you look at its output it can be quite distorted; its output is a composite of the input signal and the feedback correction. So it makes sense (to me, anyway) that this stage needs to be low distortion on its own. So instead of the existing bias network, I'm thinking to pull out the network and replace it with a single feedback resistor from the collector, which is set to provide the base with the correct current. In the case of the SA-10, SA-101 and SA102, this value would be 100K with the regular bias network removed. I suspect the designer didn't do this on account of the circuit needing gain, but with the increase in gain from no loudness contour, this seems like it should work just fine. I've not tried this bit yet but plan to very soon.

 

[atmasphere]

I did the mod as described just above.

Distortion is slightly lower, gain is the same. so this works but is a minor improvement. BTW 100K proved to be too low a value; the circuit was happier with 150K.

I found that at low frequencies, part of the waveform develops problems in the stock unit; increasing the emitter bypass capacitance helped that a little bit. Still chasing that one- its one thing to have a rolloff, its another to have distortion take off with that rolloff! Back in the desktop the 101 does seem to sound better for all these mods- more defined, but smooth without brightness.

 

[Gnesen]

I too have one of the newer SA-10s, but I think it's fully discrete. It's a fun toy, but not very useful beyond a cheap desktop system.

Perhaps a little off topic, but I have a couple of the little STA-20 receivers and think they actually sound quite nice. Certainly good enough for background tunes. And the tuner performance is surprisingly good.

 

[atmasphere]

Further news in the upgrade the mini amp dept:

As mentioned before, IMO the classic versions of these amps are the kind that have no interstage transformer or output IC. They employ four transistors- an input voltage amplifier and driver direct-coupled to one of the pair of outputs.

The voltage amplifier has an emitter bypass capacitor; feedback is applied to that cap and is thus coupled to the emitter of the transistor. I'm guessing that this is a practice brought forward from the tube era, as a lot of tubes will perform better using a cathode bypass. But if you bias the base by using a feedback resistor off of the collector (I'm currently using a 332K resistor for that), then the emitter bypass capacitor and the resistor it bypasses are no longer needed and can be jumpered out of the circuit with no downside. This is because that network is needed only if the bias network on the base of the transistor is stock (with a resistor to ground and one to Vcc+).

https://www.radiomuseum.org/images/..._usa/realistic_solid_state_stereo_2020029.png
(this is from the SA-10; the circuit is identical in the other amps)

The advantage of getting rid of that bypass cap is now there are no timing constants in the feedback loop- it will accept input right down to DC, so you don't lose feedback as frequency goes down in the bass region. At first I did this to only one channel, but it worked so well I did it to both. So far with all these mods I've not cut or removed any traces- the circuit board is stock. The upside is that the amp (in this case the SA-101, which I got for $30.00) retains the smooth quality it had to start with (due to a 2nd harmonic) but its more detailed and transparent without additional brightness. It plays bass better too.

It appears that this mod can be performed to any of the amps that employ this circuit; SA-10, SA-175B and C, SA-100C, SA-101, SA-102.

I don't see why not to perform this mod on all of these, since people pimp them out with new electrolytics anyway, which are required on any electronics this old.

 

[Stuart Pedaso]

It appears that this mod can be performed to any of the amps that employ this circuit; SA-10, SA-175B and C, SA-100C, SA-101, SA-102.

This just keeps getting better.

Glad I haven't shot that order off to Mouser yet. Waiting for needed parts will also allow time to bang through a couple of projects already queued up, to clear bench space for the 101.

Fun to be had here!

 

[Blast]

Fun to be had here!

Always!

 

[atmasphere]

If you google '4-transistor amplifier' you will see many circuits very similar to what is used in the Radio Shack mini-amps (specifically those that do not employ an interstage transformer or output IC). One thing that is different is that the mini-amps use an additional electrolytic cap in the base circuit of the top power transistor.

This is not a coupling cap- its a smoothing cap. The amp will run perfectly well without it, except at or near full power. At those power levels, a 60Hz hum waveform is superimposed over the output. This causes IM distortion; to eliminate it the additional cap was installed as a bypass to prevent the sawtooth present on the collector circuit from being also in the base circuit bias voltage- a nice touch, and nice that this part is even found in the little SA-10. This reduces the hum and thus IM associated with it at higher power levels. Of course if you have boosted the main filter capacitance to a higher value (3300uf is easily supported by the board spacing for the old filter cap and the power transformer) then this issue is reduced. But reducing IM is always a good thing, so this part should be in good shape to really hear what these amps are about.

 

[atmasphere]

Apparently my idea about the voltage amplifier used in most of the mini-amps isn't new. Here is a schematic I drew up of the SA-175c.

The voltage amplifier has feedback from its collector which also biases the base with the 2.2M resistor. The 5.6K feedback resistor is only used to linearize this stage. Note that it takes the feedback off after the output coupling cap.

The main feedback is the 43K resistor which provides the collector voltage for the driver. The 470uf cap in the base circuit of the top output device is used to kill buzz. It takes the noise waveform and applies it to the emitter of the transistor so that the transistor can compensate for the noise. When the transistor saturates, the hum waveform will be seen on the output.

The SA-10, SA-100c and SA-101 use variants of this circuit. They usually do not included the output coupling cap in the loop, and often the cap used to kill buzz is tied to the output of the amp rather than the base of the top output device.

Schematics of the SA-175 don't seem to exist in any form on the web so I thought I'd supply it here.