Problem with home made guitar amp

[dj_death]

Hi guys this summer i built the 100W Guitar Amplifier (MKII) project at ESP (http://sound.westhost.com/project27.htm). I designed pcboards for the pre-amp,power-amp and a small supply board. After building and testing the boards i put them in a big wooden box. After many mistkakes and debugging. I built the amp completely and works very nice. But there is a little problem when i set the pre-volume pot(gain) at max (even at lower positions) and set the Master Volume pot at high settings the amp begins to scream. Actually i can't describe it, it's a high frequency sound that is added with the signal. I think that the problem is in my preamp because i used signal from other source and the power amp performs very well even at full volume. I checked all the wires jacks and connections but i can't find any mistakes. At medium Master Volume settings the amp works very well. The preamp board, transformer, bridge rectifier and the supply board are mounted on the wood. The back of the box is a metal piece in which i mounted the poweramp and the heatshink. Do you think that is a shielding problem , any idea . Please help!!!!

[Joep]

Hi,

It looks like you have some oscillation in your amp.

Try moving around wires a bit to see if the problems stops/changes.  Use shielded wire for the wires carring the signal and keep them as short a possible.

This can be nasty problems to debug.

Hope this helps a bit.

Joep

[Peter Snowberg]

Hi DJ,

I think Joep is correct. Some power is feeding back somewhere. If his suggestions don't yield results, you may want to try adding more power supply bypassing around the preamp. To start, try adding a 0.1uF cap across pins 4 and 8 of the opamp with leads as short as you can make them. If that changes things but doesn't kill the squeal, try other values.

You might also want to try adding another RC power supply filtering stage just in front of the zener diodes.

If you can, try a test with the pre running off it's own power supply. That should tell you much right there.

You may end up having to give the preamp it's own box for shielding. Oh yea, also try keeping the pre as far away from the other pieces as you can. You could even have some magnetic coupling with the power transformer.

The standard cheat in these cases is to start adding small caps to the signal chain that will bleed off supersonic stuff to ground while affecting the guitar as little as possible. Ideally they should be 22pF or smaller. As these get larger, the sparkle goes away.

I'm assuming you don't have a scope. This is a classic place where they pay for themselves. Very cool tools!

Whoops.... one last thing... Do you have any compact fluorescent lamps by your rig? They put out a good deal of supersonic interference. I'm playing with an OD pedal right now and as I put my hand towards one such lamp, the hum in my amp all disappears as the supersonic signal eats up all the available power. If I strike a chord and hold my hand by the lamp the chord becomes almost totally muted. It works somewhat like a volume input on a crude Theramin.

Good luck!
-Peter

[Rob Strand]

I think the problem is either the wiring of the power supply ground,or  coupling induced oscillation because you haven't got good shielding.

On your layout the ground side of R6 and R2 should connect near each other.  The ground to the two large power supply caps should connect to R26 first, this ground track shouldn't pass through the R6/R2 ground track. Make sure you have the 10ohm resistor in the ground for the preamp power as shown in Fig 1B and make sure you connect both the signal and the ground of the preamp to the signal and ground of the power amp - connect this ground near the R6/R2 gorund point.

If you take all these precautions and you still have the problem then it's likely to be a coupling issue.  Firstly make sure you speaker leads are well away from you preamp, you preamp wires and especially the input socket.  Try twisting the wires to each of your pots.  Make sure you use shielded cable in the input and between the preamp and power amp.

Do you have a jack for the speaker output?    That circuit must use an  isolated ground jack for the speaker output.

[dj_death]

Hi everybody, Joep and Peter thank you very much for your advice. I have checked many things last month and after your suggestions i re-checked some things. I moved all wires and parts on the boards with some changes to the "scream" sound which made it worse than
before( not permanent of course ). I have put two supply 100n caps (for + and - )as close as possible to the ic when i first built the preamp. I put them after building the board so i just made 4 holes.I put the caps and i connected them to the ic with small wires. I don't know if putting bigger caps reduce the problem. I also put capacitors at my signal chain as you advised me Peter. When i used 33pF and 100pF capacitors wasn't any effect except of some noise, but when i use a  1n and a 2.7n cap
the screaming was reduced, but not at the highest Master volume settings. With a 4.7n cap the screaming was disappeared completely but as you imagine many guitar frequencies has been cut too.Also there was too much noise at the Master Volume Settings when using nF caps. I found that when i was getting closer my hand in some jacks and potensiometers the screaming was getting worse. I believe and i hope that there isn't any magnetic coupling with the transformer. I also want to mention that i had many problems when i etched the preamp board. Many vias where conducting with its other so i cut them off with a knife and i checked them with my polymeter for any shortcircuits. The DCvoltages
also seem ok. But do you think that rebuilding the preamp in a new board whould be the solution to the problem ? I designed a much better and completely new layout with supply caps close to the ic. Is it worth trying or there will not be a result? Thank you foryour suggestions again !!!

[Peter Snowberg]

I would try Rob's suggestions next. Start replacing signal lines going and coming form the pots, jacks, and board interconnects with shielded cable like RG-174, but ground only ONE end of the shield. Avoid making any ground loops.

I've had speaker wires cause havoc more than once. Try to keep everything as spread out as you can.

Since you can make it go away by adding a big cap, the next question is where can you put that cap to silence things. That at least tells you where the problem is existing in order to track things down further.

REALLY good amps (IMHO) use PCBs over point to point so that these issues are all eliminated in the design phase. Point to point makes for a lot of variablility. You may have to bite the bullet and two layout #2 to get the cure.

Good luck!
-Peter

[dj_death]

Hey Rob thank you for responding. Actually i don't use an R2, i didn't want to use the HI LO input that they suggested. When i saw the schematic i thought that when you use the High input the other end of R2 (100K) was not connecting anywhere. Do you think that the R2 must be there it just makes a voltage divider with R1. I have already connected R26 directly to the 4700uF filter caps ground point but i connected it the preamp's ground point to filter caps ground point. I fixed that and the problem remains. I think that my wiring is ok compared to fig 1B.  I used normal cables with plastic around at input and between pre-amp and poweramp. My wooden box is designed like a real guitar amp cabinet so the speakers are (2 loudspeakers& 1 tweeter) inside the box. The power amp is connected directly to the speakers so i don't use a jack for speaker output. So what do you think wiring is the problem? What about rebuilding the preamp board and rewire everything (pots, inputs, outputs etc.)?Is it going to solvethe problem ? It's hard to do all that again but i really want to make this amp work nice. Do you think that the problem can be caused by the Zener Diode supply for the preamp?

[Rob Strand]

Actually i don't use an R2,

Actually the R2 I was referring to was the one on the poweramp (It seems the pre and power amp desiginators use the same numbers ie. they are for each individual section).  With that in mind you might want to re-read my post regarding the ground connections between the preamp and power amp - just in case something has been overlooked!

> I think that my wiring is ok compared to fig 1B

OK, if you follow it by the letter, pareticularly regarding the grounds and 0V connections you should be safe.

So what do you think wiring is the problem? What about rebuilding the preamp board and rewire everything (pots, inputs, outputs etc.)?Is it going to solvethe problem ? It's hard to do all that again but i really want to make this amp work nice.

I understand where you are coming from.  I built stuff  in wooden boxes when I was a boy and had similar problems.

You might have to tell me a bit more about how you have constructed the unit.  Are the pots mounted into wood or have you got them on a metal panel?   What about the input socket?   Are there metal panels anywhere?

Whatever the case, the metal pot cases should be connected back to ground at the preamp.  With the wooden box case, solder a wire chain between all the pot cases and return this to the preamp ground.

Also, make sure the ground of your input socket returns back to the mains ground.  This is actually shown in Fig 1B.

Do you think that the problem can be caused by the Zener Diode supply for the preamp?

Not likely at this stage.

[dj_death]

Ok Rob sorry for the misunderstanding i checked it again without results. I mounted everything input, effect loop and potensiometers into wood. This is realy bad i guess. I new it wasn't good but i thought i wouldn't have such problems.I have a metal panel at the back of the amplifier in which the power amplifier and the heatsink are mounted, something very simillar to the prototype's photos at ESP.This metal is grounded (i used a
hole for this) but is insulated with a white paint.I couldn't find a sheet of metal without this paint only rusty pieces. The pots i used are plastic, i can't find pots in metal case easily. The 2 electronic shops in my small city haven't wide range of parts.For instance One of them has only 1/2 watt resistors and upwards (no 1/4 or 1/8). But i will try to find some of them if it is so important. I also have to mention that i built a small supply
board with the filter caps on it the loop breaker and the zener diodes / resistors which provide the supply for the preamp. The preamp board was powered by wires from this little board. It is a must to place the resistors/zeners the preamp board? I designed a new board for the preamp, because the current one isn't a nice board at all. Is there something that i have to be careful with, in my new layout (i.e. This part has to be close to this, this away from this ,this connection away from the other etc.)? Is there a way to show my layout at  this forum (.bmp)? If you check it's parts placement i will be very happy.

[Rob Strand]

The preamp board was powered by wires from this little board. It is a must to place the resistors/zeners the preamp board?

No, but you should have some caps across the supply rails *at the preamp*, 100uF would be nice but 10uF would probably do.

Is there something that i have to be careful with, in my new layout

The basic layout show on the ESP site should be safe.

Please understand though, it's hard for me to spot a layout problem from these posts, which might otherwise be obvious if I saw the real thing.  I do have a good feel for you problem and where you stand but there might be something you have overlooked - it would realy stink if you redo the layout and a small problem is still lingering.

I would definitely try to get the current layout to work before you go ahead and change things, even if it's a precarious set-up, because at least then you have the structure for a solution.

The thing I'm not sure about now is how long the leads are from the PCB to the pots.

Assuming the problem is coupling I would be trying:
-  add a *grounded* metal panel where the pots are. Locate it on the inside face of the front panel and have the pot pertrude through it.
- add a grounded metal panel under the preamp PCB.  If possible extend this out past the outline of the PCB - particulary on the edge where the pot  leads connect.

The basic theme is to add judicious shielding to reduce the coupling.

[dj_death]

I Have to Thank you Rob for your concern.I suppose you convinced me Rob, i will try to reduce coupling. I have a metal panel large enough to mount the pots on them but it's insulated with this white paint at both sides,but i can ground it of course. This should be aproblem for shielding right? I also found a small sheet of metal for shielding the preamp. I'll try this as soon as possible. Thank you again Rob and all of you guys, in the forum.

[Rob Strand]

I have a metal panel large enough to mount the pots on them but it's insulated with this white paint at both sides,but i can ground it of course. This should be aproblem for shielding right?

Code Select Expand



Given your pots are plastic that won't matter.  What you have to do is scrape the paint off at one point then connect a wire to that point and return it to the ground at the preamp.  It's hard to reliably solder to the panel so your best bet would be to drill a hole at one point,  clean the paint off around it, screw on one of those solderable lugs, and solder the wire to that.

The presence of paint  won't effect the shielding but you will need to clean the paint off to make an electrical connection for the wire. You *have to* connect a wire to the panel and return it to ground on each shield you use - an ungrounded shield is a disaster and can actually make things worse.

Regarding the plastic pots.  Some plastic pots have small metal lugs which are soldered into the PCB for mechanical stability/strength.  You can solder wires to these and return them to the preamp ground- it might help a bit.


[dj_death]

Thank you for clarifying this Rob. I'm going to try this although it's a very difficult task at this point i hope it helps. Thanks again everybody for your help. I'll report anything new.

[dj_death]

Hi again, i collect metal sheets for shielding and i prepare to mount the panel for the pots. Actually at the back metal panel of the box (there i mounted the poweramp, the heatsink and the power connector) I originally grounded this metal with the Earth (ground). I think that it was recommended at Fig.3 at ESP. I should connect another ground to this metal panel and i should connect all the metal sheets together? I mean the pot panel, the back panel and the preamp metal sheet. I also plan to put a grounded metal sheet under the transformer which is not a toroidal, just a conventional 150W transformer. What ground i have to use there? Finally i would like to ask how is the shield in a shielded wire ? It is just a bunch of copper fiber around the wire or a thin metallic shell. I went
to an electronics store today, i asked for shielded wire and he wasn't sure about what wire i wanted. He has only wire of the first type i mentioned.

[Rob Strand]

You want the earth (from the mains) to go to the transformer first.  Putting a shield under the transformer probably won't do much.  The important thing is the transformer should be earthed, make sure the transformer metal is actually getting connected to the mains earth, often the lacquer prevents this.

Connect a wire from the heatsink earth back to the mains earth.
Connect the input socket ground back to the mains earth.

Any other additional shield panels can be connected to the mains earth point, or, to the a ground point on the PCB that you are trying to shield - the latter method has minor benefits for coupling based oscillation problem.

Audio coax is usually has bare strands for the shield around the main plastic covered conductor, good stuff has a braid instead of the stranded stuff.  Multicore cable often has an aluminium foil around the main plastic covered conductor, some of these come with a runner wire which connects to the shield and you can use this for manually wired corrections - some cables don't have the runner wire which makes them unsuitable for manual wiring (these are designed to be crimped in a particular connector).

You can use normal coax but for some of the pots this will be combersome so you can use some of that multicore stuff - *BUT* only do this on a pot by pot basis, do not combine multiple pot wire into the same multicore cable - this can cause oscillation in some instances. For example if you have four core + shield cable you would use three of these for the pot leads then connect the shield to the circuit ground, the unused core is also connected to circuit ground.  In this case where the shield is jst a shield you don't have to connect the other end of the shield to anything.  For input and connections between boards you generally should connect the shield to both ends (unless a wiring diagram specifically shows not to). Try to find a combination which is practical for you to wire up.  Often you can get away without using shielded cable on the pots *provided* you tightly twist the wire for each pot together (ie. bunches of threes).

When you use shielded cable you must connect the shield to the circuit ground.  Also watch out the cables don't move around and cause the shield to hit and short something.

[dj_death]

Thank you again for your patience Rob. I'm planning to place an aluminium panel at the front of the box. So i cut a rectangular piece at the front panel of the box and i will mount the pots and the input/effect jacks directly to this panel. The question is how the jacks will be isolated from the chassis as ESP states. I plan to use an aluminium sheet with this white paint i mentioned but i think that i can't avoid an electrical contact there. The metal part of the jack (not the nut) will conduct with the exposed metal of the hole that will be drilled. What should i do with this ? Also i whould like to clarify something to me. The mains earth is the earth  at the power plug not the connection between the filter caps which is the main zero volt point. And one last thing if i understood well i can use the grounded shield of the shielded wire to connect a pin to ground. In example i can use the shield to connect the ground of the input jack
to preamp's ground point or i need an extra wire for this? I hope i don't become annoying with all these question.

[dj_death]

I forgot to mention that i use the loop breaker. ESP states that in thise case i must not connect the input socket to mains Earth, is that correct? I made all these questions before in what ground should i use in each shield, because you told me to connect the front panel with preamp's ground.

[Rob Strand]

question is how the jacks will be isolated from the chassis

Going for an isolated audio ground system is a matter of choice, it pretty much eliminates the chances of ground loops - a lot of my stuff is built this way.  The problem with this method is it requires more care if the shielding is to work properly (if at all).  In general, I don't agree with the complete isolated configuration suggested by ESP, see below.

If you go for an isolated system then, yes, isolating the input socket is a pain with normal sockets.  You need to put a plastic isolating washer/grommet around the contact area - these are hard to find.  An easier way is to use isolated sockets, the black plastic ones are easiest to use and find.

The mains earth is the earth at the power plug not the connection between the filter caps which is the main zero volt point.

Yes, exactly correct.  In an isolated system the two don't connect.

grounded shield of the shielded wire to connect a pin to ground.

Yes.  In your case the shields should connect to the preamp ground (preamp 0V) not the mains earth.  (In some systems you will see the shield connected to mains earth, sometimes this is "more correct" somettimes not.

In example i can use the shield to connect the ground of the input jack to preamp's ground point or i need an extra wire for this?

It's fine to use the shield like this.

I hope i don't become annoying with all these question.

No problem.  This stuff can be quite tricky.

Getting back to the completely isolated issue: Completely isolating the circuit from main earth can often result in clicks and pops cause by high frequency interference.  I suggest connecting the circuit ground to the mains earth at one point via a 100nF ceramic capacitor (an option is to add a 10 ohm resistor in parallel with this cap)- the best place to do this is at the input socket. This ensures the circuit has an RF path to mains earth.  For safety, the front panel should be connected to mains earth, not to the circuit ground like I mentioned before.

[dj_death]

I've already connected the circuit ground to the mains Earth through the 10Ohm resistor in parallel with the 100nF cap as stated at fig.3 at ESP.I 've connected the mains earth to circuit ground as shown in this figure, so i'm going to connect it to the input socket instead of connecting between filter caps. Although i have much experience in theoritical,
analog and digital electronics, this is the first time i'm involved in shielding,removing interference, and noise. All these suggestions were very helpful and i want to thank you all at these forum. I'll shall post any improvemnet.

[dj_death]

I checked your note about R2 and R6 at the power amp again and i saw that that their track is connected directly to the filter caps ground point. Where to connect the ground point at the power amp? Let's say to C6 or C7  ground? Is there a misunderstanding, did you just meant "don't connect R26's ground to the track of R2,R6 but you"? Sorry for bothering again , i have new questions all the time.