I recently acquired a Sansui T-80 am/fm tuner. The sound was pretty good,
but I had some flaky performance issues: The tuner would periodicaly lose
its' quartz-lock and the "tuned" inducator would drift badly. Also, the
entire digital display would periodically dim and lose some or all
functionality.

I opened up the case and first off noted that the main filter caps on the
power supply were leaky. Well, I figured, might as well bite the bullet and
replace all of the electrolytics given that this tuner was built in 1981.
So last night I replaced all 44 caps. I had trouble finding one of the spec
caps, a 0.15uF 100V electrolytic, and so ordered and installed two film caps
of this rating instead. All replacement caps matched the ratings of the
originals EXCEPT those 0.15uF caps... I think a previous owner had run into
the same problem with availability of this cap in an electrolytic and had
substituted a 0.22uF 100V.

I cleaned up the boards, put the whole unit back together and powered it up.
Well, first thing I noticed was the sound... low volume, tinny, no base at
all. Whereas before I started 25% volume on my amp would shake the house,
now I can easily go past 50% and still can't feel any thump in the floor.
The tuner drift issue has gone away, I suspect a cap labelled "AFC bias" was
the culprit there. However, I still have the dimming issue with the digital
display; perhaps the +5VDC regulator that powers the display board is flaky?
It's in a tough spot to blast with heat or cold without affecting a whole
whack of other components.

I left the tuner on all night thinking maybe there is some sort of initial
"burn-in" period and I THINK it sounded a bit better than at midnight last
night, but who knows...

Can anyone tell me where I might start to diagnose the tinny sound problem?
There are but a handful of coupling caps in the FM signal path, some in the
IF amp section, some adjacent to the MPX decoder. I am inclined to toss
back in the 0.22uF electrolytics that I replaced with film caps; they are
output coupling caps just to see what happens.

Any assistance greatly appreciated.

Dave

In article Do0Sf.29637$M52.21092@edtnps89,
Dave wrote:

Can anyone tell me where I might start to diagnose the tinny sound problem?
There are but a handful of coupling caps in the FM signal path, some in the
IF amp section, some adjacent to the MPX decoder. I am inclined to toss
back in the 0.22uF electrolytics that I replaced with film caps; they are
output coupling caps just to see what happens.

Is there any chance that you might have mis-read the value on the
original capacitors? As was noted in the discussions following your
original query, electrolytic caps of that very-small indicated
capacity are extremely unusual.

If they were, for example, 22 uF rather than .22 uF, or 15 uF rather
than .15 uF, then the substitution of small-value filmcaps could quite
easily have resulted in a severe rolloff of the bass frequencies...
possibly all the way up into the midrange.

I just plugged some numbers into a spreadsheet, and they suggest that
this could very easily be your problem. The impedance of a .15 uF
capacitor is quite significant at audio frequencies. If your preamp's
input presents a 10k-ohm impedance to the tuner output (not
unreasonable) then it looks to me as if the system's frequency
response would be down by 16 dB at 20 Hz, 6 dB at 100 Hz, and 1.6 dB
at 500 Hz. The actual amount of rolloff would depend on the preamp's
input impedance and the actual size of the cap that you used. .22 uF
and a 4.7k input impedance is even worse.

"Tinny" would be a good description of the sound, I think.

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

"Dave Platt" wrote in message
...
In article Do0Sf.29637$M52.21092@edtnps89,
Dave wrote:

Can anyone tell me where I might start to diagnose the tinny sound
problem?
There are but a handful of coupling caps in the FM signal path, some in
the
IF amp section, some adjacent to the MPX decoder. I am inclined to toss
back in the 0.22uF electrolytics that I replaced with film caps; they are
output coupling caps just to see what happens.

Is there any chance that you might have mis-read the value on the
original capacitors? As was noted in the discussions following your
original query, electrolytic caps of that very-small indicated
capacity are extremely unusual.


Well, anything is possible. I'll double-check them but I am pretty
gosh-darn sure that the spec rating is 0.15uF... and equally sure that the
cap I pulled was 0.22uF (just based on size there's no way it's a 22uF...
the 470uF old caps were about the size of 2200uF caps today; the caps I
pulled were 1uF or smaller for sure). And yet... it would seem to make
sense.

I noted that the AM is no better than the FM (actually it IS better but only
because AM is normally so bass-y to begin with). So there are only three
caps in the signal path after the AM and FM join downstream of the IF amp
IC... two between the IF amp and MPX demodulator and those pesky film caps.
The problem MUST be with one of these caps or the decoupling caps around the
MPX unit. OR I baked an ancient IC with my iron during my rework. I don't
think this is so, though, as the MPX chip is driving my signal meter very
nicely.

I made a conscious effort when replacing the caps to look up the rating on
the schematic, locate the new part, locate the old part, remove the old one,
hold it up next to the new one, read the info on each, then put the new one
in and chuck the old one into a can on my bench. When I finished I had used
up all 44 caps.

If they were, for example, 22 uF rather than .22 uF, or 15 uF rather
than .15 uF, then the substitution of small-value filmcaps could quite
easily have resulted in a severe rolloff of the bass frequencies...
possibly all the way up into the midrange.

I just plugged some numbers into a spreadsheet, and they suggest that
this could very easily be your problem. The impedance of a .15 uF
capacitor is quite significant at audio frequencies. If your preamp's
input presents a 10k-ohm impedance to the tuner output (not
unreasonable) then it looks to me as if the system's frequency
response would be down by 16 dB at 20 Hz, 6 dB at 100 Hz, and 1.6 dB
at 500 Hz. The actual amount of rolloff would depend on the preamp's
input impedance and the actual size of the cap that you used. .22 uF
and a 4.7k input impedance is even worse.

So I try to match the impedence if I can, right? Is there a
non-expensive-obscure-tool way to measure impedence of the preamp inputs and
tuner outputs and size the output coupling caps accordingly? The answer is
probably "yes, if your'e an electrical engineer with a McGyver streak in
you".

Any ideas on the on-again, off-again digital display? You know it'll never
give me a problem while the test leads are hooked up.

"Tinny" would be a good description of the sound, I think.

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

In article 6K1Sf.85$Zf3.19@clgrps12,
Dave wrote:

I made a conscious effort when replacing the caps to look up the rating on
the schematic, locate the new part, locate the old part, remove the old one,
hold it up next to the new one, read the info on each, then put the new one
in and chuck the old one into a can on my bench. When I finished I had used
up all 44 caps.

You would't be the person (by a lot!) to either misinterpret a decimal
point in a component value, or to encounter a schematic which has a
misplaced decimal due to an error in printing or production.

That's one reason I tend to favor the European notation - the actual
multiplier is used where the decimal point would be placed, and
preference is given to using notations with the digits to the left of
the decimal. In this notation, the part value you actually used would
be a 220n (220 nanoFarad, or .22 uF) while I'm suggesting the use of a
22u, or at least something no smaller than 4u7 or so.

So I try to match the impedence if I can, right?

Not match, really, but make-a-reasonable-relationship-with.

To a first approximation, what you'll be looking at is something
like a series voltage divider. The impedance of the coupling cap and
the preamp's input impedance are in series, and the preamp only gets
to "see" the portion of the voltage which gets past the cap.

You want to use a cap which is large enough, so that its series
impedance at the lowest frequency you're interested in (20 Hz is the
conventional number) is significantly less than the input impedance of
the preamp.

If we assume a tuner input impedance of 4.7k (a fairly conventional
value), and 20 Hz, we'd find that a cap whose impedance equals this at
20 Hz is

C = 1 / (2*pi*20*4700)

which works out to about 1.7 uF. Using a cap of this value would
result in a rolloff of around 6 dB at 20 Hz.

It's possible that the caps you pulled out were actually 1.5 uF, and
that the tuner's designed to have some rolloff in the deep bass.

I personally wouldn't use anything smaller than a 4.7 uF in this
application and would probably go larger.

Is there a
non-expensive-obscure-tool way to measure impedence of the preamp inputs and
tuner outputs and size the output coupling caps accordingly?

The first thing I'd do is look at the preamp manual. The input
impedance of the line-level inputs is often specified.

The second thing I'd do would be look at the schematic, or trace the
circuitry from the input jack towards the switching circuitry. The
usual input would have a resistor from input to ground to set the
input impedance.

If the preamp is of the "direct coupled" design, you could simply
measure the resistance at the "tuner" input with an ohmmeter (with the
power safely off, of course) to "see" the value of this resistor...
it'll probably measure with reasonable accuracy. If there's an
additional DC-blocking cap in the preamp prior to the input resistor,
this trick won't work.

You can probably rely on the rule-of-thumb that old-style
glassfet/firebottle (vacuum tube) audio components usually had input
impedances in the 100k-ohm range, while transistorized equipment is most
commonly in the 4k7 - 10k range. If you use 4k7 as a conservative
assumption you'll probably be OK.

Any ideas on the on-again, off-again digital display? You know it'll never
give me a problem while the test leads are hooked up.

I tend to look for cracked/fatigued/"cold" solder joints in cases like
this, especially on pins where any component having significant mass
is soldered to the board.

Simply re-heating all of the solder connections associated with the
display (ideally, using a bit of solder-wick and liquid flux to
remove the existing solder, then re-wetting the joint with
high-quality solder) may very well fix the problem.

Gently wiggling any cable connectors might also find the problem.

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

"Dave Platt" wrote in message
...

You would't be the person (by a lot!) to either misinterpret a decimal
point in a component value, or to encounter a schematic which has a
misplaced decimal due to an error in printing or production.



Well, well, well... the only caps I didn't carefully look at prior to
installation were the film caps, as they were _the only_ film caps I was
installing. I double-checked the packing materials, 0.15uF, yup. But last
night I had a look at the caps and, lo and behold, they are clearly printed
"1n5J250". Eureka, they are 0.0015uF caps, NOT 0.15uF. Popped in the old
0.22uF caps and we are back in business.
Bad DigiKey. Bad.

So I try to match the impedence if I can, right?

Not match, really, but make-a-reasonable-relationship-with.

To a first approximation, what you'll be looking at is something
like a series voltage divider. The impedance of the coupling cap and
the preamp's input impedance are in series, and the preamp only gets
to "see" the portion of the voltage which gets past the cap.

You want to use a cap which is large enough, so that its series
impedance at the lowest frequency you're interested in (20 Hz is the
conventional number) is significantly less than the input impedance of
the preamp.

By, say, a factor of 10? Do I want it to be insignificant compared to the
preamp's impedence? I'm confused, because below you say that a) a typical
transistor preamp has an impedence of 4.7k, AND you say that a typical tuner
input impedence is 4.7k, yet above you say that the tuner impedence should
be significantly less than that of the preamp. ??? Read below before you
respond to this one.

If we assume a tuner input impedance of 4.7k (a fairly conventional
value), and 20 Hz, we'd find that a cap whose impedance equals this at
20 Hz is

C = 1 / (2*pi*20*4700)

which works out to about 1.7 uF. Using a cap of this value would
result in a rolloff of around 6 dB at 20 Hz.

It's possible that the caps you pulled out were actually 1.5 uF, and
that the tuner's designed to have some rolloff in the deep bass.

I personally wouldn't use anything smaller than a 4.7 uF in this
application and would probably go larger.

So, higher capacitance = lower impedence = less rolloff of bass frequencies.
Let's go towards the other end of the scale. Why wouldn't I put in a huge
cap to let everything through? And how did you come up with 4.7uF as
opposed to 2.2uF or 3.3uF which are also both commonly available and larger
than your minimum target of 1.7uF? Or would 1.7uF make it _the same_ as the
preamp and you're making it much less (significant) with the larger cap?
Hmm.....

Is there a
non-expensive-obscure-tool way to measure impedence of the preamp inputs
and
tuner outputs and size the output coupling caps accordingly?

The first thing I'd do is look at the preamp manual. The input
impedance of the line-level inputs is often specified.

The second thing I'd do would be look at the schematic, or trace the
circuitry from the input jack towards the switching circuitry. The
usual input would have a resistor from input to ground to set the
input impedance.

If the preamp is of the "direct coupled" design, you could simply
measure the resistance at the "tuner" input with an ohmmeter (with the
power safely off, of course) to "see" the value of this resistor...
it'll probably measure with reasonable accuracy. If there's an
additional DC-blocking cap in the preamp prior to the input resistor,
this trick won't work.

You can probably rely on the rule-of-thumb that old-style
glassfet/firebottle (vacuum tube) audio components usually had input
impedances in the 100k-ohm range, while transistorized equipment is most
commonly in the 4k7 - 10k range. If you use 4k7 as a conservative
assumption you'll probably be OK.

Any ideas on the on-again, off-again digital display? You know it'll
never
give me a problem while the test leads are hooked up.

I tend to look for cracked/fatigued/"cold" solder joints in cases like
this, especially on pins where any component having significant mass
is soldered to the board.

Simply re-heating all of the solder connections associated with the
display (ideally, using a bit of solder-wick and liquid flux to
remove the existing solder, then re-wetting the joint with
high-quality solder) may very well fix the problem.

Gently wiggling any cable connectors might also find the problem.

Thanks a lot for your help, I *think* I am learning something here.


--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

"Dave Platt" wrote in message
...

The first thing I'd do is look at the preamp manual. The input
impedance of the line-level inputs is often specified.

Nope, neither owner's manual nor technical manual mentions it.

The second thing I'd do would be look at the schematic, or trace the
circuitry from the input jack towards the switching circuitry. The
usual input would have a resistor from input to ground to set the
input impedance.

The center post of my tuner input has a 560k resistor to ground. The signal
goes through a 1.5k series resistor, then there's a 47k to ground (as well
as a 220pF cap) before the signal path enters the volume control. What does
this tell me?

If the preamp is of the "direct coupled" design, you could simply
measure the resistance at the "tuner" input with an ohmmeter (with the
power safely off, of course) to "see" the value of this resistor...
it'll probably measure with reasonable accuracy. If there's an
additional DC-blocking cap in the preamp prior to the input resistor,
this trick won't work.

You can probably rely on the rule-of-thumb that old-style
glassfet/firebottle (vacuum tube) audio components usually had input
impedances in the 100k-ohm range, while transistorized equipment is most
commonly in the 4k7 - 10k range. If you use 4k7 as a conservative
assumption you'll probably be OK.

"Dave Platt" wrote in message
...
If we assume a tuner input impedance of 4.7k (a fairly conventional
value), and 20 Hz, we'd find that a cap whose impedance equals this at
20 Hz is

C = 1 / (2*pi*20*4700)

which works out to about 1.7 uF. Using a cap of this value would
result in a rolloff of around 6 dB at 20 Hz.

Okay, by your logic... my amp (note my previous post) has an input impedence
of (I calculate) 44.634k. So using your formula

C=1/(2*pi*20*44634)

I get a value of 0.178uF. And my spec cap is 0.15. Pretty close I'd say.
My 0.22uF's that are in there should be good?