LM4780 problem - need Help

[pacificblue]

You advise me to measure AC on outputs?

..to find out, if there is still hum, without connecting the speakers.

I don't know if DC at the input consistent with the output, because I didn't use or connect any input source still, only input caps are soldered.

There is always an offset voltage. That is amplified. With a gain of 21 and 630 mV DC at the output, you should have 30 mV offset at the input. The question is, where does that come from? That is why you should try without Cc.

You recommend to turn back to National schematic? Using Ci caps with this configuration won't be wise I guess.

You should return the amplifier to its last working state, which was the National PA-100 schematic according to you. Then make the changes towards the CarlosFM version step by step. Control the success after each step. That way you know, which step is causing the trouble and then you can investigate, why.

 

[AndriyOL]

Well guys, I've found what cause excessive current consumption. It was input caps. When I disconnected these caps from the input of the amp it starts work without this problem. However DC offset is high enough. At first pair of chips which had 32mV and 200mV became 117mV and 84mV respectively without input cap and any source connected. At second pair of chips which had 41mV and 630mV became 100mV and 83mV respectively without input cap and any source connected. The amp (2 chips) works with 40W bulb hardly glowing it. I have never had such phenomenon with this amp.
When I connected source (sound card) to input (without input caps) and play some music DC offset became 596mV - 635mV (for chips which had 32mV - 200mV respectively) and 623mV - 612mV (for chips which had 41mV - 630mV respectively).
Does this DC offset goes from source?
When input caps were still there with excessive current consumption with 630mV DC offset my nichrome wire became hot, but now with input connected without input caps and with 635mV DC offset even with music playing wire is deadly cold. So, I can't get it!
Do you know how to lower DC offset without placing Ci caps?

Thanks.

 

[AndriyOL]

Btw, I measured AC offset which was 0.020VAC without input connected. During music playback AC offset was varied from 0.08VAC to 1VAC. Perhaps it could be more, but I didn't adjust more loud (After all it's 0.6V DC offset).

 

[pacificblue]

I am at a loss now. Where do those Ci come from? You were alway telling us, you didn't use them in this version, but that you had no problems with the PA-100 version, where you did use them.

Anyhow, if the removal of Ci solved your problem, you should find out, what is or was wrong with the capacitors. Wrong connection? Broken? Pins shorting to others?

The datasheet states that the input offset voltage can be up to 10 mV. With a gain of 21 you can expect up to 210 mV at the output without Ci.

You cannot measure DC offset with music playing, because you will measure the music signal as well. And there is no such thing as an AC offset. What you measure is ripple. Music is AC!

If you want to leave Ci out and bring DC offset down, use DC servo circuits. Search for AN-1192. There is an example in it.

 

[AndriyOL]

I can't understand you. I didn't tell that I use Ci caps now. I use schematic drawn by myself, based on carlosfm schematic. I only told that I used Ci caps when I had National parallel schematic according to datasheet of LM4780 before I redesigned my amp to schematic based on carlosfm schematic.
So, I asked how to lower DC offset on schematic based on carlosfm schematic without placing Ci caps!

Btw, I told that removal of input caps solved my current consumption problem, but not Ci (feedback) caps.

If music signal is AC than my DMM shows half of that value (630mVDC). Is this the DC offset or no? Because I don't know.

Thanks.

 

[pacificblue]

I told that removal of input caps solved my current consumption problem, but not Ci (feedback) caps.

I misunderstood you there.

If music signal is AC than my DMM shows half of that value (630mVDC). Is this the DC offset or no?

No. An ideal amplifier would have 0 V at in- and output, when no signal is present. A real amplifier however has a small DC voltage due to tolerances, impedances that are neither zero nor infinte or even unsymmetric rail voltages. That small DC voltage is called offset, because it offsets the signal from the middle between the rails. You can only measure it, when no signal is present.

Why don't we want offset. For one thing it limits the available output swing, because the zero volt signal is slightly offset to one rail. Imagine rail voltage of 10 V that gives you in theory 7,1 V AC swing. Now with an offset the signal will have only 11 V to one rail, but only 9 V to the other. 9 V gives you in theory only 6,4 V AC. Now take into account that the voltage influences the output power by the square. Into a 4 Ohm speaker 7,1 V give 12,5 W. 6,4 V only give 10,1 W.

The other reason we don't want offset is that it is DC, which is unhealthy for the speakers.

To fight offset you bias the input, i. e. you connect a DC voltage that is exactly the opposite value of the offset voltage. E. g. the offset is 50 mV, then you connect -50 mV to the input and as a result you get 0 V offset. So, in theory you can connect a trimmer between the rails, its wiper to the input pin and adjust the offset voltage to zero with that.

The offset voltage of such a power op amp however is also influenced by external components and by the working temperature, because impedances change with the temperature. Look at the light bulb tester, when it gets hot the resistance rises, thus the voltage at the transformer primaries falls. The same thing happens on a smaller scale at the amplifier input. Therefore it is better to have an automatic offset regulator that reacts dynamically to any offset changes. A DC servo is such a device.

 

[AndriyOL]

So, DC offset of my amp are 117mV - 84mV and 100mV - 83mV. But 630mV DC is a ripple AC music signal. However this value (630mV DC) is substantially stable and appears when source is only connected (without Cin caps).
This is it I have told initially about 0.5-0.6V DC offset without Ci caps.
Do you know is it safe to use the amp with such offset and without Cin caps?
What do you think about Cin caps problem described above?

If I replace chips offset will be better (people told that they had 30-40mV DC offset with such configuration)?

Thanks.

 

[AndriyOL]

Btw, DC offset of my sound card without music playing is 29.9mV(1ch), 30.7mV(2ch). DC offset with music playing is from 28mV to 34mV for both channels.
That's mean, as you explained me, that DC offset of my sound card amplified and became 623mV and 612mV respectively. If so, than I can't understand where there is DC offset of the amp? I guess DC offset from outputs must be bigger than DC offset of sound card amplified by 21 of gain.

If I set preamp will it lower the DC offset?
Likewise could you tell me what is the best preamp now?

Thanks.

 

[pacificblue]

So, DC offset of my amp are 117mV - 84mV and 100mV - 83mV.

That sounds reasonable.

But 630mV DC is a ripple AC music signal. However this value (630mV DC) is substantially stable and appears when source is only connected (without Cin caps).

You are mixing things up.

• - A music signal is AC. If you measure DC at the output with music playing your readings are meaningless, because half of the music signals adds up to your readings and your DMM is to slow to follow the fast changes in the signal.
  - Ripple is also AC, but has nothing to do with the music or the DC offset. Ripple is, what AC remains on the rails after rectification and smoothing. Most of that will be rejected by the IC, but something remains in the signal. Nobody wants ripple, but some will always remain. So it is also useless to measure that. The reason I had asked you to measure AC was to find out without speakers, if there was hum present.
  - DC offset of the amplifier is measured with a 100 Ohm resistor across the input. You can solder a 100 Ohm resistor into an RCA connector and plug that in instead of your source.
  - DC offset of the source is measured at the source output with the source powered, but not playing.

Do you know is it safe to use the amp with such offset and without Cin caps?

No. The amplifier won't mind. Your speakers will.

What do you think about Cin caps problem described above?

You have a grounding problem. Without Cin the compensating currents flow through the amplifier input, and being DC you can't hear them. With Cin you block them, so they take a different route, cause voltage drops along the way and thus hum.

• - Check the polarity of your soundcard output and amplifier input. Is ground on the same side?
  - If the ground in your amplifier is earthed, put Cin back in and disconnect the ground from Earth. Make sure that the case remains connected to Earth during the test. If there is no hum, insert a ground loop breaker between Earth and ground. There are several ways to make them discussed here in the forum. You will find them with the search function. Or connect the soundcard ground with a very thick wire to the amplifier ground.

If I replace chips offset will be better (people told that they had 30-40mV DC offset with such configuration)?

Lower offset is always better from a technical perspective. It may not change your hearing impression however. And 100 mV offset are not uncommon with chipamps either.

That's mean, as you explained me, that DC offset of my sound card amplified and became 623mV and 612mV respectively. If so, than I can't understand where there is DC offset of the amp? I guess DC offset from outputs must be bigger than DC offset of sound card amplified by 21 of gain.

No. The offsets are in parallel. They don't add up.

If I set preamp will it lower the DC offset?

No. It will add a DC offset of its own into the equation. And it will amplify the DC offset from your soundcard, so the overall DC offset will be higher.

Likewise could you tell me what is the best preamp now?

"The best preamp" does not exist. You can use a preamp for a certain task. Then there is the possibility to choose according to the technical specifications that are best suited to that task. Or you can choose a preamp according to its influence on the sound. Then it is a question of taste. In the forum the SOAP is discussed at the moment. Use the search function.

 

[AndrewT]

I am confused about what happens when.

Andriy,
could you go back to stage 1 and disconnect everything from the transformer?

Plug in your bulb tester.
Plug in your transformer.
Set your meter to 600Vac.
Switch on.
Measure the open circuit secondary voltages and the mains voltage.
Post the results.

 

[AndriyOL]

My mains voltage ought to be 220VAC, but sometimes it become 200, 210, 230 VAC.
My secondary toroid voltage under 220VAC condition are 25VAC-25VAC (twin secondaries) for both transformers. However on one toroid my meter shows 24VAC between secondaries (disconnected from each other), another meter shows 2VAC at the same point. Nevertheless an ohmmeter of my DMM with 40MOhm value shows nothing between secondaries (they are not shorted) and toroid works without problems.
Besides, I tested my PSUs independently with 40W bulb and with 4.7K 2W resistor load across voltage rails (60VDC) and there wasn't any problem - toroids were deadly quiet, bulb wasn't even glow.

 

[AndrewT]

find out why one DMM reads 2Vac.
If it's broken throw it away and don't confuse this thread with it's readings.
If it's right then find out why it differs from the other meter.

Solve this first.
Two isolated secondaries will "float" when connected to a battery powered or isolated meter.

Are you trying to read the difference between two floating secondary windings?

 

[AndriyOL]

- Check the polarity of your soundcard output and amplifier input. Is ground on the same side?
- If the ground in your amplifier is earthed, put Cin back in and disconnect the ground from Earth. Make sure that the case remains connected to Earth during the test. If there is no hum, insert a ground loop breaker between Earth and ground. There are several ways to make them discussed here in the forum. You will find them with the search function. Or connect the soundcard ground with a very thick wire to the amplifier ground.

If I would misconnect soundcard output and amplifier input there won’t be any sound, but I played music (at low levels of course) and the sound was gorgeous. Ground is on the same side. I use standard RCA connectors.
I don’t have Earth in my mains voltage rails. Chassis is not grounded.
Btw, before when I disconnected Cin caps and DC offset became 84mV-100mV VAC on outputs was 0.020VAC and there was no hum when speakers were connected (without Cin caps and input source) to amp. After I played music (1hour) DC offset remains the same but VAC on outputs became 0.130VAC – 0.160VAC (max.) with noise (hum) in speakers. When source is connected – hum disappear.

 

[AndriyOL]

find out why one DMM reads 2Vac.
If it's broken throw it away and don't confuse this thread with it's readings.
If it's right then find out why it differs from the other meter.

Solve this first.
Two isolated secondaries will "float" when connected to a battery powered or isolated meter.

Are you trying to read the difference between two floating secondary windings?

Yes.
It's strange.
My friend told me that this is magnetic leveling and nothing to worry about. Maybe it's because I wind secondaries in opposite direction (however I'm not sure about).

 

[AndriyOL]

If the polarity of soundcard and amplifier is ok, no ground loop, than what's wrong with input coupling caps? When I disconnected them, excessive current consumption disappeared. When I connect them again 100W bulb starts to blink without source connected (before 40W bulb did it).
What if the input impedance it too low and chip pick up and amplifies all surrounding garbage?

 

[AndriyOL]

Well guys, I found the problem of current consumption which derived from improper placement of input coupling caps. These caps were placed directly on Zobel caps. So, I slightly hefted them and they work ok now. However, with input caps sound is little worse, not such live, rich and natural as without these caps.
There is only high DC offset left which could only be eliminated, as pacificblue was kind enough to explain me, through feedback coupling or DC servo adjustment.

Is there any other solutions available? Maybe resistor replacement to make it suitable for amp needs in way to nullify offset?

THX.

(post modified at members request)

 

[AndriyOL]

If there is no other solutions,
than many thanks you guys for your contribution in building my experience and my peculiar construction.
This is a good forum.
Cheers.