G-4500 Volume Wierdness

Now we are on the same page. Those critters are circled in green and are the input differential pair. The two sets of wires circled in red are your input wires to the main amplifier section of the amp. We are interested in the voltages in mv-ac, thought is to measure the value of ac they are carrying referenced to ground. Cool place to use your scope probe and signal generator. Set the signal generator to 1khz and 750mv output, probe the output from the signal generator with your scope to verify the scope settings match the sig gen (neat way to learn scope grid markings too :), then attach sig gen to aux input. You should be able to measure the ac voltage at those two white wire connection points with the receiver turned off, if not there are a few extra steps to add, let me know how you make out.marked.jpg
If the reading rises too quickly add a resistor to each white wire to reduce the voltage appearing at the input stage of the amp. Unusual to have to do this. This would be the same thing as turning down the gain knob on a stand alone power amplifier. Next to the input pair, see the orange cap, that is the dc coupling cap, as long as you are in that part of the circuit would not hurt to replace it value for value with a Nichicon muse series bipolar or if the value is 1uf or below a even nicer Wima film capacitor. A 22k resistor would be a good starting point, change one and try and then decide if the reduction in signal is too aggressive and adjust accordingly.
Finally, the other thread that has a new member discussing his G4500 replied and states his unit also has a rapid (sensitive) volume control rise curve so we might just be chasing our tails as mentioned earlier in the replies but hey why not tailor the volume control to your liking. Can't hurt a thing, maybe the Sansui engineer that designed the small G series was hard of hearing lol.
-Lee
 
Fantastic stuff Lee. I'll report back as soon as I have something to report. You know the lower one of the two areas in that picture is one of the places where the excessive glue is. You can see it close up a few posts back.
 
Done reorganizing my bench and I've got the signal generator now. Trying to setup for your tests above and ran into a problem which I think I can sort of ignore but wanted to throw it out there anyway.

Set the generator for 1 K, dialed the amplitude to approximately .750 mvac on the scope but when I checked it against my meter it was 2.3 mvac. So I set the voltage on the generator using my meter (pictured) and that produces about .250 mvac on the scope. The meter is reading higher by a factor of 3. All of the settings on the scope are what I would term normalized. The probe is on 1x. The Horiz Time setting is at .1 msec, the vert volts setting is at .1 volts DC (it should be on AC but the reading is the same). So I'm reading the scope as 1 msec and .250 mv. Does 1msec correspond to 1khrz? I can't get that right in my head.

The scope has a 3v test point and it checks out on the nose with the scope. But when I check it with my Fluke or my old analog meter, I get 1.5v. (So instead of being higher on the meters like before it's actually lower by a factor of .5!) As an additional check of the meters, I set my bench power supply to 3v and that's what I got on both meters.

So it seems something is not quite right in scope land or maybe I'm doing something fundamentally wrong. I believe I can probe the G-4500 anyway and will be looking for differences. I also spent some more time with the schematic and noticed that the input side for the Aux ports is pretty much straight through the tuner board with only two resistors that I already checked out. But now that I know this I want to measure resistance from the jacks through the resistors to the Loudness/Balance board to see if I find the imbalance somewhere along the way. Then I'll do your tests and I'll probably end up removing the Loudness/Balance board again so I can trace the signal through it (it's in a terribly inconvenient location). I should also mention I got the 8 ohm resistors you discussed earlier.

So I'll try all this stuff and report back with what I find. I'm okay with modifying the output to make the system more user friendly but I'lI need to find the imbalance first and most importantly. Any feedback you could provide about the scope readings I would hope would be beneficial.

Thanks again!


upload_2017-3-12_11-51-3.png
 
So I did encounter the problem you thought I might. I could not get a reading at the white wires leading to the amp. However I did check a lot of the circuit leading up to the bass treble section before I lost the signal and will try to convey those readings.

Measuring voltage with the meter from the Aux jacks through the loudness/balance board to the volume pot on both sides. Starting with .729 mvac at the jack, still .729 at first resistor R97/98 after the input selector switch, 709/665 (higher number is vol off, lower number is full vol) after R97/98, 689/602 after second resistor R107/108, same entering Loud/Bal board and Amp board. Resistance change from volume pot is not linear. Resistance doesn't start to drop on vol pot until 4.5 or 11 o'clock on knob when I see finally 687, then it drops off faster until zero. 689 to 0 at signal in on vol pot. Signal at wiper is reversed with 0 at 0 and 599 at full vol. Balance lead seems correct on each side at 599 at center and staying 599 for whichever side is being tested and dropping to zero on other side. Loudness switch was on the whole time.

Checking with the scope at vol pot signal in I saw .21 with vol off and .18 with vol full (remember I'm not seeing same voltage reading with scope as meter). Same on both sides and not at all linear. I also saw a flattening of the bottom of the wave on the upswing side. On the wiper I had 0 at 0 and .18 at full vol.

So it seems as though the left/right imbalance I was measuring before directly on the board without an introduced signal does not show itself when measuring an introduced signal. At least not up until it enters the amp side where I lose it.
 
pick up a pair of 8ohm 20watt cement resistors which will be ok as long as you dont max out the output for long. Hook them across the speaker teminals, also at the same time hook your dmm across the 8ohm leads, set the meter scale to ac and measure how fast the voltage ramps up. Watch to see where the volume control position is when you hit the 1vac level and if you are at 0vac when the control is fully counter clockwise.

I will try this next after my old grey brain has a rest.
 
So far so good. We are looking for curve symmetry around the center horizontal line of the scope, by centering the curve trace so it crosses the center line (equal amount both above and below the center line) you then will be able to read directly off the graduated vertical lines the amplitude of the wave. Also,I find it helpful to bring the total curve length together by decreasing the time coefficient i.e. more towards 0.5 milliseconds rather than towards 1 sec. Remember that you are measuring ac values so another good practice is to set scope to dc coupled and ensure you have the probe ground attached to a ground point on the receiver frame. There are times when you have to add a small test clip extension to do this which is ok for this series of tests. I would expect to see @0.0vac on the white input wire as circled above with loudness and tone off and volume at minimum with unit on, and @0.3vac by 9 oclock on the volume knob, remembering that these are input values to the amplifier stage. Conversely with a dummy load across the speaker outputs you could measure ac by touching the probe to one of the speaker load wires and look for symmetrical rise, in this example I would look for 0.0vac at minimum volume and about 1vac at 9 o'clock on the volume dial.Nice bench setup you have now, won't be long and you will be teaching us a thing or two :)
-Lee
 
So that seemed worthwhile. Shows the imbalance and rapid rise. Left side started smoking the resistor almost immediately at full volume.

upload_2017-3-12_19-9-5.png
 
Well, that shows us a couple of things. One, the volume control slope is really rapid which verifies what you brought up in your initial post. Two, not good to peg the volume control, that forces the amp into hard clipping and will lead to meltdown. To measure at full roar you will need to invest in a couple of 100watt non inductive resistors, 200watt is better yet. Those poor 20 watt resistors were used for low output testing, poor things. If your unit had a mute control that would quickly solve the whole thing so we are going to play instead with the input signal level at the amp board instead.
When I test an amp for output levels at clip I use large dummy loads and scope and ac setting on multimeter to look for the point where the top and bottom of the sine wave start to flatten out noting the ac value on the meter at the same time. You should not stay up there long, get the reading and back things back down unless you are trying to stress test the amp to its limits. Everything gets real hot in a hurry. Even when taking distortion measurements I am only at full clip for a very limited time.
Simple balance control fix, pull the knob off and replace it one click in the opposite direction of the error, remember this level of receiver does not contain the more esoteric pots the higher ones use in the Sansui food chain.
 
Scoping the speaker outputs, when I got to the 4 position on the dial with the left side (11:00) (4.25 on right), the wave form started to distort (clipping?) with the top falling off to the right and bottom rising up to the right (symmetrically). Measuring 6.5 and 6.1 volts respectively. That would be about three clicks further on the dial than detent 14 in the chart above where I was already measuring 11-12 volts on the meter. So maybe an actual 14-15 volts. Which is pretty close to a factor of three difference from scope as seen before.

upload_2017-3-12_20-4-16.png
 
Simple balance control fix, pull the knob off and replace it one click in the opposite direction of the error

The audible balance position is at about 3:00 so that would be a bit too much fooling myself to consider.

I very much understood the dangers of pushing it and staying there. You implied as much in your initial post about using the resistors. Although I thought you did want me to get a number at full volume. My bad. I put it there just long enough for the meter to give me a number and then backed out. It was tough to get the scope picture I just posted too. Basically I set up the shot and the amp, scope etc. then turned on the amp, took the picture and turned off the amp. But there it is.

Your new avatar pic is sweet!
 
Amp should provide @18vacrms at clip. However that 20w resistor is probably not happy at this point so would not push it further. Secondly, you need a true rms dmm to get more accurate vales. Finally, thats excellent scope work and yes the wave form is indicating the clipping point. Btw you cannot generate relevant data if you had the dut (device under test) on a dim bulb tester, takes full mains to really get after it :)
 
you cannot generate relevant data if you had the dut (device under test) on a dim bulb tester, takes full mains to really get after it :)

Dim bulb tester?

So I am clipping at less than half volume (control wise). :eek:

What do you think of the scope issues I'm having? :confused:
 
Scoping the speaker outputs, when I got to the 4 position on the dial with the left side (11:00) (4.25 on right), the wave form started to distort (clipping?) with the top falling off to the right and bottom rising up to the right (symmetrically). Measuring 6.5 and 6.1 volts respectively.

Okay, now I'm quoting myself! I just realized that when I took these speaker scope measurements I had the probe set on 10X and you can see in the photo that I was set on 2 volts on Channel A. So that means 60 volts measured not 6 right? How could that be possible. I double-checked it and tried 1X at 5 volts and although it was off the screen I could adjust the vertical to see I was covering 12 divisions or in other words 60 volts. What the heck! :rant: The other measurements I took on the input signal were all done at 1X. I switched to 10X because it was going off the screen but forgot to factor in the probe.

Tomorrow I'll scope the signal wire to the amp.
 
Remember that rail voltage is +40vdc/-40vdc. Driving the amp to rail (well into clipping) could possibly give you that high a reading. I have never pegged an amp to see what happens so kind of new territory for me. Once you get larger dummy loads you might repeat the test noting on the scope at what point the curve just starts to flatten and observe the meter reading at that point. My math is pretty rusty these days however (and may be way off base here) AC=DC/0.636 so 60vac=x/0.636 or about 38vdc which is close to rail voltage before efficiency losses. Others free to chime in on this one, do not want to present incorrect theory.
-Lee
 
Several interesting notes there. And of course more questions here. First I should mention that I haven't been taking readings at full volume on the scope if that was implied. Only at the point where it starts to clip which is at the 11:00 position on the dial. It's at that point that I'm calculating 60 volts DC. When I created the table above with my Fluke I did go to full for a second to see what number would appear (29 VAC). I could have sworn my scope was giving me the same reading regardless of whether it was on AC or DC. I'll try again. Also, I'm going to double-check some of my other readings.

I'm not interested in trying to blow this thing up, only in trying to fix the imbalance and calm the volume ramp. All the number stuff is just me trying to understand and trying to provide you and others with the right critical data to help.

So if I read what you're saying above correctly, there should be a difference between AC vs DC readings? I don't recall that part In the books I've read so far but I do understand how a DC current can be half of an AC current depending on the transformer power supply setup.

I've also been wondering if using a different scope probe could be making a difference besides the obvious 10X issue. I've got several different probes so I'm going to test them against each other to see what there is to be seen.
 
Checking with the scope at vol pot signal in I saw .21 with vol off and .18 with vol full (remember I'm not seeing same voltage reading with scope as meter). Same on both sides and not at all linear. I also saw a flattening of the bottom of the wave on the upswing side. On the wiper I had 0 at 0 and .18 at full vol.

Once again having to correct myself; These measurements I took at the pot with the power off were right in my notes but wrong when I wrote them up. I see 2.1 with vol off and 1.8 at full vol. So at least this does jive with the scope readings being 2 to 3 times higher than the DMM. So at least that now makes sense even if it doesn't completely make sense. I also tried switching from DC to AC on both the scope and meter and saw no difference.

I have also now checked all my scope probes and they all produce the same results though some are 10x and some are 1x and some are both. I saw some very slight differences and some wobbly signals but all were essentially the same.

I'll now try probing the white wire using the resistors at low volume. I have to look at your notes again. And I'll also try using a 100mvac signal instead of 750 to see what happens.
 
Test pre-amp white wires at amp still at .750vac:
Left:
Scope .3 vac at 9:00
Meter .3 vac at 11:00 .126 vac at 9:00

Right:
Scope .3 vac at 9:00
Meter .3 vac at 11:00 .109 vac at 9:00

-Scope numbers are fairly accurate but dial position could be slightly off, maybe 1 click to the positive.
-Meter numbers at .3 vac were not exact because of detents and clock position is not exactly the same right to left. I believe right was a bit higher.
-Meter numbers at 9:00 are exact comparison. Looks like we found the imbalance?
 
Last edited:
Back
Top Bottom