Universal Tiger

[aged one]

In the late 70s, my friends and I assembled 3 "Universal Tiger" amplifiers degigned by Dan Mayer and sold by SWTPC.

After a couple months of operation, they all blew their output transisters. Recently I was browsing Dr. Leach's site and read his comments regarding the design."These amplifiers were very unstable, causing them to oscillate, overheat, and blow tweeters."

I find all this strange as, at the time, the Hirsch-Houck Labs gave the amplifier very high marks for sound quality and stability

Have you experts ever looked at the schematic to determine what could have caused the problem and how it can be resolved?

I would appreciate your comments.

I attach a copy of the schematic.

[destroyer X]

Your oscilation will stop.

Now, go removing and testing if it will oscilate.... try to remove the differential input first.

The one that likes to oscilate, normally is the second stage, considering both differential transistors as "input", and first stage.

If not stopped with that capacitors, try to put in the other transistors too.

Hehe, will stop.

regards,

Carlos

[destroyer X]

Will stop oscilate.

regards

Carlos

[carlmart]

Why ceramic capacitors and not polystyrene?


Carlos

[destroyer X]

The lower the value, the better the highs you will reproduce.

The higher the values, you will triangulate your sinusoidal waveform starting from audibles frequencies.

In practice, i never go up than 21 picofarads, as a rule for me, as i found problems in some circuits using higher values.

regards,

Carlos

[SY]

The Tiger amps had all sorts of failure modes, not just oscillation. Thermal runaway transmuted many a silicon chip into carbon, taking much of the rest of the amp with it. Audio Amateur had an article some years back (late '80s, early '90s?) on a rebuild that claimed to address the issues- I think it was by Poehland.

[analog_sa]

the Hirsch-Houck Labs gave the amplifier very high marks for sound quality

Did they ever bother to listen?

[MikeB]

Ouch, c8=220pF, and no cdoms ? It's a wonder that it does not
oscillate permanently...
I would reduce c8 to 22pF, and add cdoms (B-C) of ~68pF to
Q3 and Q4. But these modifications will definitely change sounding.
Also, it's very likely that the amp selfdestruct with heavy clipping.
No care has been taken, that the cfp-output breaks down when
overdriven by clipping.
And what the hell does c10/c11/c12 ? Has oscillation been stopped by
applying a permanent capacitive load ?
At least c11/c12 guarantee that this amp does not oscillate very long.
And biasing a cfp-output with 2 diodes+pot ? Big nono !
If the pot get's dust, the amp will blow.

My recommendation: Do NOT build this one !
It might sound good, but at least a bit of reliability is a must.
You can redesign it, but then it's some standard ClassAB.

Mike

[sajti]

Originally posted by MikeB
Ouch, c8=220pF, and no cdoms ? It's a wonder that it does not
oscillate permanently...
I would reduce c8 to 22pF, and add cdoms (B-C) of ~68pF to
Q3 and Q4. But these modifications will definitely change sounding.
Also, it's very likely that the amp selfdestruct with heavy clipping.
No care has been taken, that the cfp-output breaks down when
overdriven by clipping.
And what the hell does c10/c11/c12 ? Has oscillation been stopped by
applying a permanent capacitive load ?
At least c11/c12 guarantee that this amp does not oscillate very long.
And biasing a cfp-output with 2 diodes+pot ? Big nono !
If the pot get's dust, the amp will blow.

My recommendation: Do NOT build this one !
It might sound good, but at least a bit of reliability is a must.
You can redesign it, but then it's some standard ClassAB.

Mike

MikeB,

this amplifier use same output stage configuration as the ETI466/480. The output stage has voltage gain too. This is about 2. The C3/C4 make some phase compensation, and the C11/C12 do same at the highest frequencies.
The planner not use Cdom, because of the large local feedback at the VAS, which is reduce the open loop gain. This combination has some advantages, such as the peak output voltage is very close to the rail voltage.
The bias is not so critical, because of the CFP output stage, but there must be better to make some local DC feedback...

sajti

[PB2]

I built these amplifiers as a young kid and yes they did blow up often. They seemed to go unstable with no output load, sometimes with the input unplugged, perhaps picking up RF, perhaps if clipped too hard. Not always sure what did it. What I did notice is that often C3 and C4 (polystrene) were melted, hmm caps are not supposed to dissipate energy but also R12, R14, R15, and R16 were badly burnt and being physically close might have caused them to melt. Eventually, I did the analysis and these resistors should be over 1W each, 2W to be safe, they were 1/2 W in the original design. Higher wattage resistors here helped very much. The design has an output stage gain of 2 which means the output can be driven into saturation and "stick" before the voltage gain stage clips which is not a good thing.
Feedback from the voltage gain stage to the neg diff amp input, might help compensate the amp, but really some analysis should be done first. I would not just go tacking caps in here and there to try to stabilize the amp. I'd be really curious to hear if anyone has looked at this design in SPICE.
I doubt the designers ever did stability analysis of this design just trying to "tune" it in their prototype which did not seem to work out very well.
This design brings back memories and was advanced for its day with a fully complementary output stage using fast (for the times) very high power devices when many were doing the RCA (and Citation 12) quasi complementary design.