Hi Mike,
I am very familiar with Greg Ball.
Getting leads for the 'scope is a top priority for you right now. You absolutely need to use it, and if leads are the only thing holding you up .... Even just one for now. Do it this weekend, or as soon as you can. Avoid the switchable X1 X10 probes. The switch moves easily and may wreck your 'scope if it switches to X1. You need to use the X10 probe as it greatly increases the bandwidth of the probe and reduces loading on the circuit under test. A X1 probe will often make things misbehave. Better not to have this right now. I have glued some of mine in the X10 position.
Mike, looks like you are now off to a good start. Get your probe and warm up your gear!
-Chris
I am very familiar with Greg Ball.
Getting leads for the 'scope is a top priority for you right now. You absolutely need to use it, and if leads are the only thing holding you up .... Even just one for now. Do it this weekend, or as soon as you can. Avoid the switchable X1 X10 probes. The switch moves easily and may wreck your 'scope if it switches to X1. You need to use the X10 probe as it greatly increases the bandwidth of the probe and reduces loading on the circuit under test. A X1 probe will often make things misbehave. Better not to have this right now. I have glued some of mine in the X10 position.
This may affect the compensation too. Again, the 'scope is your best friend here. Look for ringing while driving the amp with a squarewave if you can. Remember that adjusting the feedback to reduce gain means that you are increasing the feedback used. The value of the compensation cap is determined by both the part and the PCB layout. You may not be able to copy directly from Peter's design.
A good experiment. Keep in mind that Tantalum caps have no tolerance for being reverse biased at all. They tend to short completely. They are also known to have very high distortion, and yet they were touted as the wonder cap in the later 70's. This will be a great chance for you to see if you can hear any difference between them.
Mike, looks like you are now off to a good start. Get your probe and warm up your gear!
-Chris
Thanks Chris.
A couple of q's:
What is meant by "ringing"?
What are you referring to by the term "compensation cap"?
Rgds.
Mike.
A couple of q's:
What is meant by "ringing"?
What are you referring to by the term "compensation cap"?
Rgds.
Mike.
Ringing looks like this:
A compensation capacitor is a capacitor that is used to keep a circuit stable. Typically these are found across feedback resistors. The term "Miller" capacitor is also used, where this refers to a capacitor placed between the base and collector of a transistor for stability purposes (by reducing the Miller effect, hence it's name)
I wouldn't bother with the Caddocks on an LM3886. Just good quality metal films, perhaps 0.1% tolerance if you want to get clever.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
A compensation capacitor is a capacitor that is used to keep a circuit stable. Typically these are found across feedback resistors. The term "Miller" capacitor is also used, where this refers to a capacitor placed between the base and collector of a transistor for stability purposes (by reducing the Miller effect, hence it's name)
I wouldn't bother with the Caddocks on an LM3886. Just good quality metal films, perhaps 0.1% tolerance if you want to get clever.
Hi Mike,
There is no way I could have explained this any better than Jaycee has. His post is spot on, and so is his advice to you.
Once you can look at waveforms with your 'scope, things will start to fall into place. You can take a digital photo of a trace and post it for answers to your questions.
-Chris
There is no way I could have explained this any better than Jaycee has. His post is spot on, and so is his advice to you.
Once you can look at waveforms with your 'scope, things will start to fall into place. You can take a digital photo of a trace and post it for answers to your questions.
-Chris
Chris , Jaycee.
Your advice is gratefully recieved.
Can anyone tell me exactly what C22 is doing? This is the cap I intend to experiment with using the Tants and Ruby ZLH I have to hand. What differences should I expect with these caps and why?
Also with uprading resistors am I right that the ones in the siganl path will have a greater effect on the SQ than the resistors that go from signal to ground?
Thanks in advance,
Mike.
Your advice is gratefully recieved.
Can anyone tell me exactly what C22 is doing? This is the cap I intend to experiment with using the Tants and Ruby ZLH I have to hand. What differences should I expect with these caps and why?
Also with uprading resistors am I right that the ones in the siganl path will have a greater effect on the SQ than the resistors that go from signal to ground?
Thanks in advance,
Mike.
Another question:
It has been suggested that I solder the decoupling caps directly to the chips pins. If I do this then would it be a further improvement to then run a new ground wire from these caps to the ground on one of the main smoothers - or even the ground on the trafo? I have some good solid copper wire to hand which would surley be better than the thin traces on the PCB?
Mike.
It has been suggested that I solder the decoupling caps directly to the chips pins. If I do this then would it be a further improvement to then run a new ground wire from these caps to the ground on one of the main smoothers - or even the ground on the trafo? I have some good solid copper wire to hand which would surley be better than the thin traces on the PCB?
Mike.
Hi Mike,
C22 breaks the DC ground in the feed back loop. Because there is no DC return to ground, the gain at DC is 1. If this capacitor were to short, the gain at DC would increase to the mid frequency AC gain. That means that any DC offset voltage referred to the input would be amplified by the same amount as the AC signal. This means much greater DC offset voltages across the speaker. If that capacitor were to open up, your AC gain would drop to 1. The signal coming out would be very close to the same level as the signal going into the chip input pins.
C22 is in a sensitive location in your amplifier. Any distortion caused by this part will not be reduced by feedback, but rather amplified by the amplification factor of the circuit. This is one location where you will hear differences between capacitors if any exist.
-Chris
C22 breaks the DC ground in the feed back loop. Because there is no DC return to ground, the gain at DC is 1. If this capacitor were to short, the gain at DC would increase to the mid frequency AC gain. That means that any DC offset voltage referred to the input would be amplified by the same amount as the AC signal. This means much greater DC offset voltages across the speaker. If that capacitor were to open up, your AC gain would drop to 1. The signal coming out would be very close to the same level as the signal going into the chip input pins.
C22 is in a sensitive location in your amplifier. Any distortion caused by this part will not be reduced by feedback, but rather amplified by the amplification factor of the circuit. This is one location where you will hear differences between capacitors if any exist.
No. Any component in the signal path can afect the sound quality. If you have a shunt resistor from signal to ground, it is in the signal path directly.
-Chris
Hi Mike,
Connect your bypass capacitors between the supply pin on the chip to the existing bypass caps. If you are also installing an intermediate value, tack that across the existing bypass caps. Better to be safe than sorry!
-Chris
Don't do much beyond replacing parts in this amp. Adding wire is an invitation to creating a short circuit and possibly destroying the amplifier. The same holds true for modifying the current grounding scheme. It works now, so don't change anything.
Connect your bypass capacitors between the supply pin on the chip to the existing bypass caps. If you are also installing an intermediate value, tack that across the existing bypass caps. Better to be safe than sorry!
-Chris
And I thought you were getting along nicely now!
Mike,
What each of us can hear is down to our own ear training, familiarity with the test music & system and a willingness to accept there can be changes. Nobody really understands this area completely so don't be offended if others can't hear cap or wire changes and you can.
I will respect other people's right to say they hear no difference between a silver wire and a silver-plated wire or that the impedance of their mains supply doesn't affect their amplifier's sound etc etc. You should only be irked when someone tells you what YOU can and can't hear.
Simon
Mike,
What each of us can hear is down to our own ear training, familiarity with the test music & system and a willingness to accept there can be changes. Nobody really understands this area completely so don't be offended if others can't hear cap or wire changes and you can.
I will respect other people's right to say they hear no difference between a silver wire and a silver-plated wire or that the impedance of their mains supply doesn't affect their amplifier's sound etc etc. You should only be irked when someone tells you what YOU can and can't hear.
Simon
Hi Simon.
My message above was somewhat argumentative yes, but I mean no antagonism.
Its morning and I havent had my Coffee and Cigarette yet! LOL!
My message above was somewhat argumentative yes, but I mean no antagonism.
Its morning and I havent had my Coffee and Cigarette yet! LOL!
Good morning Mike!
Your post could be interpreted as antagonistic, but I hope it won't be now.
I think you're a bit like me in that your understanding of circuits is just at a modest level so it's easier to leave that to the original designer.... and to look at component swaps for improvements. On the other hand some people understand circuits well and because their grounding is in well understood technical concepts they have a hard time with mysterious ones (sound quality) because it's difficult to measure and so personal.
Simon
Your post could be interpreted as antagonistic, but I hope it won't be now.
I think you're a bit like me in that your understanding of circuits is just at a modest level so it's easier to leave that to the original designer.... and to look at component swaps for improvements. On the other hand some people understand circuits well and because their grounding is in well understood technical concepts they have a hard time with mysterious ones (sound quality) because it's difficult to measure and so personal.
Simon
The resistors going to ground are just as much in the signal path as those that don't. There are some strange notions held by more than a few audiophiles about the way currents flow...
Hi Stream.
I had to buy 10 x MUR860 OnSemi Diodes so if you want 4 let me know. I also have some Mundorf M-Lytics coming to replace the Cambridge Audio 10,000uF main caps so you can have those if you want? (You could upgrade to 'SE' buy fitting these)
PM me if your interested.
Mike.
I had to buy 10 x MUR860 OnSemi Diodes so if you want 4 let me know. I also have some Mundorf M-Lytics coming to replace the Cambridge Audio 10,000uF main caps so you can have those if you want? (You could upgrade to 'SE' buy fitting these)
PM me if your interested.
Mike.
Hello everyone.
I spotted a free mod last night!!!! 🙂
There are PCB links near the volume and balance pots which can be used to bypass the whole tone circuit and the traces that loop round via the 'direct' switch. This would eliminate about 40cm of PCB tracking from the signal path and also ensure the signal lines dont go near the trafo.
If you look closely at the picture below you will see 4 links just above the balance pot on the right. The bottom 2 are L/R output from the opamp. In between the the Volume and balance pot are 2 more links (hidden from view just under the Vol pot). These are the input to the volume pot. All thats needed is to remove these links and join up the signal traces with some good copper wire (like CAT 5 wire for example, or even locate some MKP DC blocking caps here).
Mike.
(I spotted a simillar mod on the LM3886 output to the speakers which will again eliminate a large length of PCB trace. Unfortunately there are no links for this one so it will involve cutting traces and soldering jumpers across. More details to follow plus an image to illustrate the mod)
I spotted a free mod last night!!!! 🙂
There are PCB links near the volume and balance pots which can be used to bypass the whole tone circuit and the traces that loop round via the 'direct' switch. This would eliminate about 40cm of PCB tracking from the signal path and also ensure the signal lines dont go near the trafo.
If you look closely at the picture below you will see 4 links just above the balance pot on the right. The bottom 2 are L/R output from the opamp. In between the the Volume and balance pot are 2 more links (hidden from view just under the Vol pot). These are the input to the volume pot. All thats needed is to remove these links and join up the signal traces with some good copper wire (like CAT 5 wire for example, or even locate some MKP DC blocking caps here).
Mike.
(I spotted a simillar mod on the LM3886 output to the speakers which will again eliminate a large length of PCB trace. Unfortunately there are no links for this one so it will involve cutting traces and soldering jumpers across. More details to follow plus an image to illustrate the mod)
Is there any reason why I should avoid mounting the output resistor inside the coil?
The reason I ask is that Im thinking of ways to improve on the long lengths of PCB trace and shorten the signal path as much as possible. Its getting obsessive I know but if I mount the resistors inside the coil I lose roughly an inch of PCB trace per channel. Every little counts!
The reason I ask is that Im thinking of ways to improve on the long lengths of PCB trace and shorten the signal path as much as possible. Its getting obsessive I know but if I mount the resistors inside the coil I lose roughly an inch of PCB trace per channel. Every little counts!
I'm not keen on the resistor inside the coil.mikesnowdon said:
The resistor impedance may be affected by the magnetic field and the inductance may be affected by the materials of the resistor. If either of these become non linear then they add to the distortion. I have not experimented with both to hear if it's audible, but others may have results to report.
Thanks for posting Andrew. I have seen resistors inside the coil before but like you I have no idea if it affects SQ adversely. Those possible impedance/inductance issues are a concern.
Opinions seem divided on if its ok to remove these components alltogether. The Gainclone kits dont have them but they can help stop weird speaker/cable loads causing stability issues apparently...
Hopefully someone can give a definative answer on this?
Mike.
Opinions seem divided on if its ok to remove these components alltogether. The Gainclone kits dont have them but they can help stop weird speaker/cable loads causing stability issues apparently...
Hopefully someone can give a definative answer on this?
Mike.
- Status
- Not open for further replies.