Hello all...

I've got a few Texas Instruments speaking learning toys...the old Speak &
Spell, Read and Math. I recently picked up two more, mainly for the add-on
ROM word cartridges they came with. One is a Speak & Read from around 1987,
the other a Speak & Spell from 1984.

These two units don't work properly. They both fail in pretty much the same
way. When powered on, they work normally for a moment or two and then the
speech output starts failing. Symptoms range from starting to say a word
normally and being unable to finish, to random noise and "hash" sounds
coming from the speaker. Every now and then the speech will "correct" itself
and try to continue, only to fall down again.

The Speak & Spell is much worse. It will actually start corrupting the
vacuum fluorescent display and power off (presumably it's crashing) after
some random length of time.

I decided to look at what I could, and found a schematic online. It's of
rather poor quality and doesn't cover the "passive" components in the
circuit.

http://lnk.nu/casperelectronics.com/20mv.jpg

There is a power supply board in the units that outputs multiple voltages
from a single battery supply. I did checks on this compared to my working
units. It appears to be OK.

So I turned to the only other idea I had. I checked the clock oscillator
circuitry to see what it was doing, using the frequency counter range on my
multimeter (the only method I have). What I saw was strange...the clock
frequency seemed to be wandering all over the place. Doing a cross check
with my working unit produced another difference. While probing it, the
voice shot up dramatically in pitch any time I touched either of the clock
pins on the speech synthesizer IC. That never happened in either of the
broken units.

I think this is trying to tell me something, but I'm not familiar enough
with the type of oscillator used or the circuit in general to know what. I
asked a friend of mine, who said "that looks simple" but he's very busy and
hasn't gotten back to me. I'd greatly appreciate any thoughts.

William

Well, if the clock is wandering, isn't that a good place to start
troubleshooting?

Is the oscillator crystal-controlled, or a simple RC circuit?

It's an RC oscillator circuit. The schematic shows an adjustable resistor.
None of my units have that--there is a fixed resistor. I cannot tell what
the value of the capacitor should be...guessing 60pF?

I forgot to mention--in my working toys, the clock still appears to wander,
but much less severely. It seems to fluctuate based on whatever the
microcontroller is doing.

Lacking in-depth experience with these circuits, I also want to be sure
whether I'm on the right track or not.

William

I'm guessing...

The circuits are probably not asynchronous. That is, the clock has to be
within a certain tolerance range for the chips to "work well with each
other". If the clock wanders, the chips may lose synchronization.

On Sat, 28 Apr 2012 13:58:22 -0500, "William R. Walsh"
m wrote:

These two units don't work properly. They both fail in pretty much the same
way. When powered on, they work normally for a moment or two and then the
speech output starts failing. Symptoms range from starting to say a word
normally and being unable to finish, to random noise and "hash" sounds
coming from the speaker. Every now and then the speech will "correct" itself
and try to continue, only to fall down again.

I've fixed a few of those:
1. Rotten battery contacts.
2. Dead batteries.
3. Replace every electrolytic you can find, even it if looks ok. If
you have an ESR tester, use it.
4. Some units have a tantalum capacitor on the power supply board.
Replace it.
5. Dirt and filth UNDER the big chips. Alcohol and a brush.

The Speak & Spell is much worse. It will actually start corrupting the
vacuum fluorescent display and power off (presumably it's crashing) after
some random length of time.

That's the symptom of a dead tantalum on the phenolic power supply
board. What's weird is that I tested the original tantalum for
leakage and ESR. It was fine, but I replaced it anyway. That was the
problem.

Argh... gotta run....
--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

William R. Walsh wrote:
It's an RC oscillator circuit. The schematic shows an adjustable resistor.
None of my units have that--there is a fixed resistor. I cannot tell what
the value of the capacitor should be...guessing 60pF?

I forgot to mention--in my working toys, the clock still appears to wander,
but much less severely. It seems to fluctuate based on whatever the
microcontroller is doing.

Lacking in-depth experience with these circuits, I also want to be sure
whether I'm on the right track or not.

That dithered clock is a technique we use to make it
appear that our designs are 'quieter' than they really
are, RF noise - wise.

http://www.maxim-ic.com/app-notes/index.mvp/id/1995

Since all parts of the circuit see the same clock, everything
remains synchronized.

Relativity in action!

--Winston

William R. Walsh wrote:
Hello all...

I've got a few Texas Instruments speaking learning toys...the old Speak&
Spell, Read and Math.

BTW, if you get a chance, power one of your working units from
a lab supply. There will be a voltage at which it says:

"Say 'l[owai er=09 e0rj'"

You will say "l[owai er=09 e0rj".

Then it will say:

"No. Say 'ro4w nwubyrmvto u'"

Indicating of course that this is what it expected to hear.

It is quite wonderful.

--Winston