Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 01:05:36 PM

Login with username, password and session length

Recent Posts

 

Serious Blocking Distortion

Started by Littlewyan, July 05, 2013, 10:54:13 AM

Previous topic - Next topic

Littlewyan

Brand new, electrolytic by the way, the .22uF is a Polyester I think. Its Orange and Rectangle shaped

Littlewyan

#151
How would I lower the distortion in this circuit? Is it just a case of trying different JFets? Or if I increase the IDSS Current then the JFET will be able to handle more before distorting?

phatt

Hey Littlewyan,
               I just had a listen of those 2 clips and I've heard far worse from some brandname amps so considering it's only one fet you have done well so maybe quit while you are ahead. :-X

You have some good basic sound building blocks you will need more stages if you want to go further into tone shaping.

I don't know why you want cleaner as it seems fairly clean in places.
Bare in mind even clean electric guitar has some clipping and done right it can enhance the result.

Maybe go find some classic Shadows hits, most will assume it's all super clean electric guitar but if you listen carefully you will notice there is an edge to a lot of the notes played.
The classic triode rattle as the signal slightly deforms at higher levels.

Bare in mind that most guitar pups made today are far more powerful (bigger output swing) than days of surf music and that won't help if you are trying to get a super clean sound.
Phil.

Roly

Swapping FET's around is shooting in the dark; better to persist with a given FET and try to get the biasing and gain where you want it.

Quote from: LittlewyanHow would I lower the distortion in this circuit?

Well first up it would be helpful to know where the distortion is arising, and why.

For example, is it the FET stage that is clipping, or is the gain of the FET stage causing the chip amp to clip?

Looking at the TDA2050, what input signal level would cause it to clip?

Assuming a supply voltage of 30 volts (adjust the following to suit the actual supply voltage);

The output will clip when the signal level is 30 volts pk to pk, or 15 volts pk.

The approximate gain is set by the ratio of the 22k feedback resistor and the shunt to ground, 680 ohms;

22000/680 = 32 times

So the input that will cause it to clip is the output voltage divided by the gain;

15/32 = 0.46875 volts peak

So it is already pretty sensitive and we would expect a guitar signal at this point to be able to fully drive it, however we have the volume control to hold this under control.

According to my sim of your "AMP - final" circuit the FET stage provides an output of about 2 volts peak for about 0.5 volts peak input, so it has a gain of about;

2/0.5 = 4 times

This is with the bypass cap set to zero, the centre position of your three position switch.

Again according to my sim this stage won't clip until the input is around 1.8 volts peak, which is a fair bit more than I would expect from a guitar pickup, so with the switch in the middle position this first stage really shouldn't clip.

I would however expect the chipamp to clip with the volume set anywhere above about 2/10.

With the switch in the 0.68uF position the FET gain is quite a bit higher at 440Hz and the input clipping level falls to around 250mV peak, which ain't much at all.

Traditionally you would explore this gain distribution using a signal generator and oscilloscope, but LTSpice should give you a reasonable approximation.

If you want to reduce the gain of the chipamp then you can either reduce the 22k feedback resistor, or increase the 680 ohm shunt.

HTH
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

#154
Thanks Phil, means a lot :).

Roly, as usual very informative. I will say that the amp did sound best to my ears when the capacitor wasn't in the circuit with the volume turned up, I guess thats because it isn't driving the chip quite so hard. Now if I change the 22K resistor to 18K then that will lower the gain to about 26.4 which will then give me 0.56818 volts peak, however I'd probably lose a bit of volume so I'd then increase the supply voltage to perhaps 32volts to make up for it so then I'd have 0.6 volts peak before the chipamp starts clipping. How does that sound? Am I going in the right direction?

Hang on a second, according to the datasheet in the single supply circuit if I decrease the size of the 22K resistor then that increases the gain? So either the datasheet is wrong or we're using the wrong equation to work out the gain.

Ah no apologies Roly, whoever wrote that datasheet got it wrong. Checked one of my books and checked the TDA2030 Datasheet and decreasing the 22K resistor does decrease gain.

So looks like if I decrease the gain slightly and then up the supply voltage by 2 volts then I could reduce the clipping on the input of the chipamp.

Roly

Quote from: Littlewyanhowever I'd probably lose a bit of volume so I'd then increase the supply voltage to perhaps 32volts to make up for it

Taking it backwards, the supply voltage to the chip amp determines the available voltage swing on the output, thus for a given speaker load impedance, the amount of power that the chip amp can deliver before clipping (keeping in mind the other limiting ratings of whatever chip is being used).

By increasing the feedback (by reducing the feedback resistor or increasing the shunt resistor) we reduce the gain.  But seeing that the output swing is determined by the supply voltage, the overall effect is to increase the amount of input signal required for full output - the power amp becomes less sensitive.

Now in your case the chip amp is set perhaps to be a bit too sensitive, so by reducing the chip amp gain a bit we will come closer to the classic "one volt" signal level that has been the traditional input level for main amps since Noah built a stereo.

You don't need to compensate for reduced gain with more supply voltage because your basic problem seems to be excessive gain along the chain leading to unwanted distortion.

At the moment it looks to me like you have too much overall gain even when the Source bypass cap is switched out, going from clean to clip at perhaps 1 or 2/10 on the volume.  In a stereo we don't want distortion so the gain is set up so that clipping starts at perhaps 9/10, but a guitar amp is a different animal and we want a range of both clean and clipped tones, so a reasonable place for clipping to start (assuming the mythical "average" signal) is say around 5/10.

Another way of looking at what we are doing now is adjusting the scale of the volume control to give you a nice spread between clean-ish below 5/10 with a nice transition into dirt at higher settings.  At the moment it seems you are getting into dirt pretty much right off the stop - it's a bit "Formula 1", just touch the accelerator and you have wheel spin.

Even with the overall gain reduced by several times, say 4:1, you are still going to have ample gain for dirt up high, and even moreso with either of the caps switched in.  There will be no lack of raw horsepower, just that with reduced gain the volume control will give you a more useful and less touchy "accelerator".

Something that might both reduce gain and give you more tonal flexibility is to introduce something like a Fender tonestack between the emitter follower and your volume control.  Tonestacks generally are lossy, around 20dB, so this would make good use of the excess gain in the first stage and also give you two or three extra knobs to twiddle (and of course endless fun tweeking tonestack component values, which is where Tone Stack Calc comes in to its own).  Something to ponder.
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

#156
Hmm I'll have to have a think as I don't really have enough room to add an extra three knobs for tone controls as I wanted to keep the amp side of things as far away from the PSU as possible. You'll see what I mean in the photos attached. I could perhaps double up the switch so as I switch between different JFET Source Capacitors it also switches between different RCs for tone......

Roly how did you get the TDA2050 chip into LTSpice? Been trying to figure a way for ages!


Worked out how to include an opamp so I can see the extent of the issue and now it has all become clear why the amp sounds like it does! So to begin with I'm going to change the shunt resistor to 1K and the Feedback resistor to 18K and see where that gets me. I should have about 0.8V before clipping then.

In the end I just replaced the 680ohm shunt resistor with a 1K5 and now the Chipamp can take a nice 1 volt before clipping. The result......a now fantastic sounding amp! It stays pretty much clean with no source bypass capacitor, starts to distort slightly at about 9 on the volume but not much. Then with a bypass capacitor it starts to distort at about midway on the volume. Perfect :).

Roly

Ok-ay!  That's a really nice looking build.   :dbtu:

I think you still have enough front panel room to at least get a two-knob tonestack in there.


Now I'm going to get a bit picky about some details.

I'd like to see some heatshrink over all the exposed terminals on the back of your mains switch.

A plastic cap, coat of paint, or blob of silicone goo on the head of the bolt that holds your toroid transformer would be a good idea too (if that accidentally gets shorted to the chassis it makes a shorted turn and will give you a nice big welding splat).

The knot in your power lead will stop it pulling through if it is tripped over, but it won't stop it turning around and possibly stressing wires internally.  There are several methods of anchoring power leads, clamping grommets, clamp glands, but a simple and effective solution is a small plastic gizmo called a "P-hoop" which bolts down inside and holds the cable firm.  Of course it still needs the protection of a grommet or similar where it passes through the chassis.

Your board mounting nuts are not retained.  You can use star washers for this, 'tho my own trick is to use a bit of cutoff plastic Rawlplug on the screw thread after the nut.



This both holds the nut from coming off and protects your hands from the sharp thread end biting you while working on the amp.

For your next build (and amp building is about as addictive as crack) a few points for the planning stage.

Normally it's a good idea to have only a single point of connection of the amp common to the chassis to avoid forming earth loops.  This means that we either select a point near the power supply common, or at the input socket.  Where the ground point isn't either input or output then it is normal to use sockets that have their common insulated from the chassis.

It appears that both your regulator and chip amp heatsinks are supported by their respective chips.  This means that their leads carry the weight of the heatsink as it twangs around during transport.  There is also the question of where they are getting rid of their heat to - the air trapped inside the case.

When planning your next build you could make a starting point to mount your heatsink(s) on the back panel outside, cutting holes for the chips to mount on them inside, then grow your layout around that.  That way you will have the back panel supporting the heatsink, which in turn will support the chips.


But it is a very nice looking little build.  It has been an interesting journey to get to this point, and pleasing that you are getting to be satisfied with the amp in operation - but I warn you, now homebrew amp or stomp is ever "finished"; there are always new ideas to try.   ;)
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

#158
Thanks Roly

I will get some heatshrink on that as I must admit I do get bit nervous when I have to go near that switch when working on the amp whilst its on :P.

Now do you mean the bolt and the washer so it doesn't conduct with the cover on top of the transformer?

I'll take a look at that as I'll probably have to drill a small hole for the P Hoop.

I can get hold of star washers so thats not a problem :).

With regards to the Amp common it is only connected to the chassis near the PSU, unless I'm missing something?

I did have some issues with heat in this amp with the voltage regulator, it seems to be coping now however I obviously haven't tested it in muggy, sweaty gig conditions. I did try and find a chassis with vents but struggled surprisingly. Will definitely be considered next time though.

My next project is actually going to be my first ever valve amp, although it won't be from scratch it will be a kit from ampmaker.co.uk, the PP-18 to be precise. I want to start off with a kit as I've never built a valve amp before and making mistakes in valve amps can be very expensive. If you make a mistake in an SS build then you might blow a couple of pence of components (Unless of course you cook the mains transformer or somehow destroy the chassis), if you make a mistake in a valve build and blow components there will be tears (Valves, Output transformer etc). Also if I start with a kit then I'll have a good sounding amp as a starting point for many modifications.

Like you said, amp building is as addictive as crack ;)

Roly

Around power mains "nervous" is good.

The amp chassis, torroid mounting bolt, and top plate are electrically connected, and constitute about 3/4's of a turn through the torroid.  If by mischance the circuit were to be completed from the top plate back to the chassis you would have a shorted turn around the tranny core, and one hell of a lot of current would flow.

The P-hoop could possibly use one of the existing board mounting bolts, down against the chassis.

Quote from: Littlewyanunless I'm missing something?

Yes.  The sockets you have used for input and output have their commons connected to the amplifier common, but they also are metal attached to the metal chassis, thus forming two paths back to the common earth point, and a couple of earth loops.  If you replace these with sockets that have bodies insulated from the chassis you may notice another reduction in hum level.

As I said, the heat has to get out of the chassis somehow, or it just keeps getting hotter.

I don't know the exact valve kit but I can tell you a couple of things based on experiences with a large number of Lamington and similar builds.

Apart from a few dropped valves, or crunched-in-transit, nobody has blown anything up worth mentioning.  This is partly because valves tend to be rather more forgiving of mistakes than transistors; so you can forget the bias connection, have the power stage anodes glow red, yet it will come back as if nothing happened when the problem is fixed.  Put a transistor in backwards and it's goodnight at power up before you can even say " :grr ".

I hope you will post a new build thread here with pix so we can keep up with your progress.

Yours is a healthy attitude; build something basic and get it going, then dick around with it.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

Ah right I see I shall look at painting the bolt for the transformer and using a P Hoop.

With regards to the sockets conducting with the chassis that could explain something, as I have found that I sometimes get a bit of hum until the lid is on, would this be connected?

I shall keep you all updated.

Thanks Roly

Roly

Quote from: Littlewyan on September 22, 2013, 02:40:46 PM
With regards to the sockets conducting with the chassis that could explain something, as I have found that I sometimes get a bit of hum until the lid is on, would this be connected?

That's a definite "maybe".

Earth loops are sensitive to magnetic fields, and while torroids generally have very little external field there are plenty of other sources external to the amp.  Assuming the lid is steel it would do a pretty good job of shielding the innards from external fields, but just to be confusing it will also screen electric fields getting in to high impedance circuits.

About the only way you can tell the difference is that magnetic screening will occur just with the lid in the way, while electric field screening is essentially capacitive and requires that the lid be actually electrically connected to the rest of the chassis.  This can give rise to the effect where the hum suddenly stops as you tighten down a retaining screw.

In very loose terms magnetic fields arise from current, while electric fields arise from voltage.
If you say theory and practice don't agree you haven't applied enough theory.