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

March 29, 2024, 07:14:53 AM

Login with username, password and session length

Recent Posts

 

Noob with a soldering iron

Started by Jungle-Jim, December 21, 2014, 09:01:50 PM

Previous topic - Next topic

phatt

Hi JungleJim,
                If the circuit has design flaws no amount of filtering or sexy opamps will rid you of the dreaded hiss. As this rig has a hot channel then likely to be an opamp with a couple of diodes, usually a high value pot is added which gives the OD effect. If this section has no buffer then the hiss can be horrendous.

If my hunch is right then the problem is resistor noise which has little to do with type of opamp.
In the circuit below; if R2 and R3 are large,,,,, expect hiss/fizz.
Meantime Google "boltzmann's noise" or "Johnsons noise".
Roly might like to add comment on that subject. 8)
Phil.


Roly

#16
{Just to explain my first remark above; any component that is connected to or near a socket is somewhat at risk.  Guitar inputs are typically 1 megohm impedance and are therefore somewhat more susceptible to damage from excessive voltage being applied.

ESD - ElectroStatic Discharge is one, particularly in dry air conditioned atmospheres and synthetic carpet (sound like the pub where you play?).  These conditions can lead to people building up very high voltage electrostatic charge of many thousands of volts.  Despite the very high voltages the storage element (e.g. you) has a very small storage capacitance, so the current and therefore power is small, but with bad static conditions the power level can rise high enough to cause subtle damage to components, semiconductor junctions in particular.

By "subtle" I mean the formation of a small crack in the chip lattice that doesn't cause instant failure, but with thermal cycling may get larger and spread, causing odd noises and finally chip failure.  A study done here of chip failures in telephone exchanges showed a significant peak around three days after a passing thunderstorm.}


I expect that simply changing that front end dual to an 833 should give you a noticeable drop in noise/hiss.  The sockets make this A/B test easy.  Just be extra careful when initially aligning the chip pins into the socket - it's dead easy to get one wrong and grinch it - and make sure it's the right way around before switch on.   


Yeah, the '833.  I'm retired at 65 and I've been playing with electronics, mainly audio, since my pre-teens, and in all that time I have never seen anything like the LM833.  I have some in the preamp of my homebrew Twin-50 keyboard stage amp, and they are as close to "a perfect audio op-amp" as I've ever seen.  They are effectively audio silent, distortion-free, and have an extremely wide bandwidth (which was actually a problem), and are excellent op-amps with high input impedance and good output drive capability.  Best of all they cost next to nothing and it's worth the postage overhead to get the ten-up discount and have a few spare kicking around for those "midnight ideas".

A possible exception is in fuzz boxes where the crappy audio characteristics of the LM741 might be significant, but otherwise an LM833 will improve any circuit where it will fit.


Shielding

It should be obvious that all the stray capacitances from a circuit board inside a metal box connected to its own ground, will be to its own ground, and there will be no external signal capacitive coupling or pickup.

What may not be so obvious is that a ground plane mounted close to the circuit board can get most of the same effect by being, by far, the dominant stray capacitance.  All the previous stray coupling of external signals still take place, but the dominant stray capacitance from every point on the board is now its own ground plane, and this may be several orders of magnitude greater than the stray capacitances, so these unwanted signals will be reduced accordingly.

The first element is something conductive; ali roof flashing, kitchen alfoil and glue is good, and something to insulate it from accidental contact with the circuit, a filing card tray, self-stick films of all sorts.  The metal must be connected back to the circuit central ground point, and presto - quietness.  Take these elements and look around you at the various thin conductive things like kitchen alfoil and heavy foil baking trays, and insulating things like cling-film, document protectors, plus handicraft/glues etc.  Graphite paint also works (see modem insides).  It's mainly a matter of exploring what you've got in the home office, or can easily get from the local office supply joint/supermarket.


Improving the power supply

That preamp will have a considerable tolerance to operating voltages.  I would expect it to work just fine from +/-12V to +/-18V, and somewhat better towards the top end (op-amp specs tend to improve with supply voltage, and you obviously also get more signal headroom to overload/clip). 




NOTE CAREFULLY!  THE PINOUTS ARE DIFFERENT!


See attach.




Ultimate noise

Fx pedals and guitar front ends are a doddle when you come to look at quality microphone preamps, say for capacitor-capsule mics.

For a standard "600-ohm" balanced line input the actual load resistor(s) is about 1k3 ohms.

All resistors (or any component having any resistance) generates a tiny amount of electrical noise due to thermal agitation, and the more resistance it has, and the hotter it is, the more electrical noise it produces.

http://en.wikipedia.org/wiki/Noise_figure

When you pass any current through the resistor you generate (yet) another source of noise.  In fact there are several different characteristic noises and sources, some mechanisms generate low end rumble, others high hiss, and others can be wide band.

With a mike like this you are looking for a noise floor perhaps -120dB below full output, but "full output" may be only a few millivolts, so you are starting to get down to the theoretical minimum noise of all your components (unless you want to suspend them in liquid nitrogen).  Thankfully guitar doesn't need ultimate noise.


Bench instrumentation is particularly demanding because you want it to remain good and faithful down to at least one-tenth of what you will be measuring with it.


{ed: typos}
If you say theory and practice don't agree you haven't applied enough theory.

Roly

The attach is a capacitance multipler rather than a voltage regulator as such, since your target is really hum, AC on the DC lines.


Low-Pass Filter time constant;
R1 * C1 = 100 * 470*10^-6 = 0.047s,
1/t = f = 1/0.047 = 21.277Hz -3dB hinge fc (i.e. below 50/60Hz power freq).


Capacitance multiplier;

C1 * hFE = 470 * 20 = 9400uF min at Imax, more at lower currents, e.g.

470 * 70 = 32,900uF

The MJ3055/2955 suggest themselves for ease of heatsinking, but just about any power pair to hand, such as the TIP-series, would do equally or better.

Season with gain or capacitance to taste.


LTSpice sim
There will be a drop of about 7 volts across the transistor(s) for an average current of 1 amp, so some heatsinking will be required for 2 times 7 watts, perhaps a scrounged CPU cooler?

The supply rails are reduced to +/-22V.

After doing a lot of sums I concluded that it will still be "quite loud enough" for studio work.

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

Jungle-Jim

Thanks Roly
I've spent a bit of time today going over your posts to make sure I understand what you're saying...

Firstly - the easy ones - I was searching for an EMI filter amongst computer gear, and can I use one of those inline units you get on video leads - like this? http://www.ebay.co.uk/itm/6-Pcs-Clip-On-EMI-RFI-Noise-Ferrite-Core-Filter-for-5mm-Cable-/331361952793?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item4d26b62419

I will look at making a grounded metallic cover around the amp - but one issue is encasing the heatsinks from the main amp chips. I doubt that's a good idea.

But onto your circuit designs...
Firstly - thanks - this is really interesting, and it forces me onto a learning curve and that's great because it's stuff I always wanted to know.

Out of the two circuits you suggested, did you decided after posting the regulator circuit that you actually thought the capacitance multiplier was a better idea? Which would you suggest I try? I can build either, that's within my capabilities.

With the capacitance multiplier, where would it be inserted (presumably immediately after the power supply?) And is it just for the pre-amp circuit? And is there a negative voltage supply version? (and what's the size for the large caps?)

With the voltage regulator circuit - are you sure 30vdc is ok for the power amp? Why does the designer give it 36v, and why the pre-amp at 16v? These are unusual amounts.

Also - a couple of Q's about that design (excuse them if they're a bit L plate):
* Resistors R1a and R1b in series - both 1 ohm 5w - is that a deliberate doubling up, or because you can't get a 2 ohm R?
* Do the 4 LEDs have a diode function here, or are they cosmetic? They would look pretty but be overkill here.
* I am a bit confused about the fact that your new design starts with a +/-36vdc supply. Does that mean that it's to attached to the output of the original supply (in other words going through the two large caps), or do you mean attaching it straight after the rectifier? In fact what I'm actually confused about here is: across the fuse, prior to the rectifier, it's 53vac, or 27 to ground each side. What would that be straight after the rectifier, as it meets the two big caps  - a crude +/-27vdc? In which case how does it end up 36vdc because of the large caps? I'm missing some basic knowledge here.

As you can see I am doing my best to keep up with this, so please bear with me when I have to ask stupid questions.

Thanks again,
Jim


Roly

Apart from practicing electronics in its many forms I have also spent a fair chunk of my like teaching it, and 'tho it may be an unpopular idea these days, learning anything takes some strain, it has to be worked at, chewed over, until that wonderful light bulb moment when a trainee who has been looking so puzzled suddenly says "oh ... I get it!"


The ferrite EMC filter is similar but a bit different and won't be as effective as a proper line filter, but better than nothing.

This is a repackage I did of an open line filter from a computer monitor but they are also found inside 'puter PSU's, and TV's and other AV gear when they are often built on a corner of the main board that can just be hacksawed off and remounted.  Buy, build or wangle, lots of options (but this is a good idea, not totally essential).

http://www.ozvalveamps.org/mainsfilter.htm#4amp

Quote from: Jungle-JimOut of the two circuits you suggested, did you decided after posting the regulator circuit that you actually thought the capacitance multiplier was a better idea? Which would you suggest I try? I can build either, that's within my capabilities.

Remember, I'm only stumbling along just ahead of you saying "I think it's over this way".

1.  You want to convert the preamp power supply from zeners to 3-pin regs.  There is nothing wrong with doing this, some fun to be had and things to be learned, and you should end up with a superior power supply for the preamp (and therefore some confidence it is problem-free allowing you to concentrate on the shortcomings of the preamp alone).  In technical terms there will be some marginal improvements but I don't expect it to make any audible difference.


2.  Adding the small series resistor and another 2200uF to the main rails should produce some audible improvement in the power amp residual hum, but it's very hard to quantify.

Adding the capacitance multiplier (alone or in conjunction with the improved filtering) should again produce a marked improvement in residual hum levels, albeit for some loss of maximum power output.  In the LTSpice screenshot above you can see that the output hum level (red) is very greatly reduced on the primary supply hum level (green), and this is under full load, worst case.

The next step would be a full-on regulator, but it's also a matter of what is sufficient to the task, and if the capacitance multiplier alone reduces power amp hum to a negligible level under studio conditions, then we can neglect it (unless you want to go full regulation as an educative exercise  ).




You can stop wherever you like, go back, go forward, try something sideways like a big battery bank - when you pick up a screwdriver or soldering iron you are making several major statements; you are going to take control, you aren't going to be a consumer victim, you are going to build your own stuff making heavy use of recycle and personal ingenuity, you are authorised to change things until you are happy with them, break stuff to see how it works and what its limits are.




Quote from: Jungle-JimI will look at making a grounded metallic cover around the amp

...which is why I got into ground planes - they do most of the job without disrupting air flow.  In your case I'd look at some sort of metallic lining on the inside back surface, under the chassis and reverb tank.  Since hum is an issue I'd be looking for a plated steel L-plate to act as a shield between the power transformer and everything else.

These are all shot-in-the-dark ideas because the real way to find out what's needed is to take measurements and try some simple experiments.  Where is the hum getting into the signal path?  Preamp?  Power amp?  Ground loop or common earth path due to poor (PCB) layout?  A common problem is that the high current AC path from the transformer, rectifier, and main smoothing caps, is somewhere shared with a sensitive signal return path (normally somewhere near the main filter caps themselves).  Sometimes cutting tracks and re-routing them using insulated heavy wire to get the PCB layout around the PSU right, is required.  Power and signal ground path problems, where power supply noise gets coupled into the amp signal path, are actually very common even in commercial builds because it is an important topic that gets too little consideration - "meh, it's only ground".  No, it's actually the other half of your power supply flow and layout.


Quote from: Jungle-JimWith the voltage regulator circuit - are you sure 30vdc is ok for the power amp? Why does the designer give it 36v, and why the pre-amp at 16v? These are unusual amounts.

No.  With the 36V you measured 35V caps aren't good practice, 50V or 64V would be more the go, but at least with a margin of 20% over the highest switch-on, off-load voltage they will ever see.

If you fit 3-pin regulators the preamp rails should be ironing-board flat, a "blameless" power supply as Doug Self would say.  It may not be perfect under a microscope, but for our needs it may as well be, particularly as op-amps are built to have a very high power supply signal rejection, typically -120dB.



The negative supply version of the capacitance multiplier is simply the inversion, the cap reversed (+ve to ground) and the PNP MJ2955 used in place of the NPN MJ3055 of the +ve version (see attached).

Power amps are pretty tolerant to their supplies (which is why they are normally so basic, as here).  As the supply rails drop the power amp circuitry does a pretty good job of struggling on, but if anything may need to be reset for best sound at reduce supply it will be the idle bias current, but that may prove trivial too - see what it sounds like before you go solving a problem you may not have.


R1a and R1b are two one ohm 5 watt.  The "right" component would be 2.2r/10W but 2x 1r/5W gives us the same result with the waste heat spread over more area, the possibility of trying only one, or the two in parallel if the sag turns out to be excessive (trading supply sag for hum reduction here).  And if they turn out to be left overs later, are more generally useful than a single 2.2r/10W.  Getting the job in hand done is only part of the considerations of a designer; will this component still be available when the amp breaks down?  How hard/expensive will it be to get/replace/substitute?  Here I am also factoring in that you are a tinkerer and will have uses for floating parts somewhere downstream.



Quote from: Jungle-JimDo the 4 LEDs have a diode function here, or are they cosmetic? They would look pretty but be overkill here.

Okay, a bit of glitz, but they do have a value, particularly if your amp is playing up and you are trying to fix it in the gloom of backstage half and hour before you go on.  An extreme for built-in instant diagnostics was a broadcast compressor/limiter I worked on that had tell-tail LED's a glimmin' and a flashin' all over the board, like a New York ant city after dark.



Why the supply voltages?

The preamp rails in op-amp preamps are a pretty arbitrary choice (which may have more to do with having thousands of 16V zeners in the store that need to be turned over).  Component supplies may also level their stocks by making a special offer on something similar but different, e.g. you order 35V electros for a 30V supply and they offer you 64V at the same price because they are overstocked, and you accept  for the extra safety margin.  All sorts of non-technical reasons.

Some op-amps are comfortable on +/-4.5V (split 9V battery), most need +/-12V, and most have a maximum of +/-18V where they typically give their best, and where you also get maximum headroom to clip/overload.  While op-amp supplies are often amongst the best, cleanest and most stable, they are really overkill for op-amps which are the most tolerant of rough or "moody" supplies.


The rails in the power amp are driven by the need for a given peak voltage across the nominal speaker impedance to develop the required watts.

Starting assumptions;
R = 8 ohm speaker load (treat as 8 ohm resistor)
P = 30 watts

From these we can calculate the voltage for that power in that load (and the current), and find the peak voltage required, the minimum possible supply voltage for the power amp.

P = E2/R

P R = E2

E = sqroot(P R)

(30 * 8.0)^0.5 = 15.5VRMS

Epk = root(2) + VRMS

1.414 * 15.5 = 21.9Vpk

So at least +/-22V rails are needed for 30 watts into an 8 ohm speaker.

The devices in the output amp are not ideal, so some voltage will be lost across those, so the actual rail voltages have to be increased to compensate or the amp won't make rated power when the output stage is clipping.  This will normally only be a few volts due to the drops across the output transistor (due to its own Veb and the Veb of the driving transistor, the saturation voltage of the Voltage Amplifier Stage before that) and the drop across the obligatory 0.22r/5W emitter resistor, roughly around 4-5 volts total.  So the supply rails might be nominally 28V on load rising to 32V off load.


Quote from: Jungle-Jim53vac, or 27 to ground each side

Ah!  Somewhere above I thought I picked up that you had measured 36VDC.  +/-27VDC is much more reasonable on 35V caps.  Cool.  Sorry.


53/2 = 26.5 * 1.414 = 37.5Vpk
37.5 - (2 * 0.6) = 36.3V off load (corrected for rectifier diode drops)

26.5 * 1.1 = 29.2V on-load estimate


The overview remains the same, just scale the voltages down a bit.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.

phatt

This might help with rectification.
Phil.

phatt

I just re-read most of this and correct me if I've missed something but back at "Reply 10" Jim noted that things had improved after replacing caps and other stuff but still the hiss remained.
If this is all that is wrong then surely the problem is more likely to be circuit design not shielding.
Though by the look of the pics some extra shielding would be helpful. :cheesy:

I'd hate to see Jim throw hours of work into this only to find the hiss still remains. :grr

Worth note;
A lot of bedroom guitar players tend to turn the gain way up and master volume low and often complain about the annoying background hiss/fizz.
But in a live gig you would never hear the background noise until the band stopped playing.
Of course if it's really bad then it needs looking into.
Phil.

Roly

Yeah, we're just discussing this while the op-amps are cooking (but there was some residual hum, and its a recording amp, so as a soundie I'd like him to present with a nice studio-friendly amp).

As you say about the minimalist "chassis".  But not too hard to fix on the kitchen table.   :dbtu:  It's more tedious than tricky or expensive, a bit of alfoil and some sticky schoolbook covering.

And as for the hiss, "whatever it takes baby" as Prime Minister Credlin would say.   :trouble

My money is on IC1.    8|
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Roly
Thanks again - this is great. I am following this with great interest and making sure I get what you're saying. And phatt - Roly isn't getting me to do work that needn't be done. Nobody is wasting anybody's time here. As Roly says -

"...when you pick up a screwdriver or soldering iron you are making several major statements; you are going to take control, you aren't going to be a consumer victim, you are going to build your own stuff making heavy use of recycle and personal ingenuity, you are authorised to change things until you are happy with them, break stuff to see how it works and what its limits are."

That pretty well sums up my attitude to most things: learning to fix the things you use is a way to take control, and save money of course. I have the same approach to my car - I keep an old Volvo 240 on the road (here in UK) - and it's old-school enough that I can fix anything, and I know its systems. People with new cars can't even find the dipstick.

The work on this Session amp is in many ways going beyond simply trying to fix the amp itself: it's a good choice of amp for this purpose - it's simple, and the schematics were published. These amps were well regarded - their reputed strength being a good tone, and being loud for their wattage - so they are probably best suited to live work, maximum bang for the bucks and not too many frills. So it's a good project to look at the power supply and noise issues.

It's new year but the LM833N's will arrive in the post soon, so that'll tell us something once they're in about noise levels. However, the project has turned into an exercise in improving this amp, and possibly making it better than when it was new.

Back to responding to your last post Roly:


I get this bit:

E = sqroot(P R)

(30 * 8.0)^0.5 = 15.5VRMS

Then you write...

Epk = root(2) + VRMS

1.414 * 15.5 = 21.9Vpk

(L Plates out)
- are you multiplying the Vrms by 1.414 because the power rating was AC, and it needs to be recalculated into DC?
- and what does 'Epk = root(2) + Vrms' refer to? because it doesn't appear to correspond to the equation on the following line.

Also, thanks very much for doing the revised schmetics for the modified circuits - they look great. I will look inside the box and work out where to insert an additional circuit board, or whether there's room to hack the existing board to insert new parts.

But a question is - you were saying that the board needs at least 21.9V to give out 30W, but yet your revised power circuit has +/-21Vdc for the power amp. At the moment the board is getting +/-35Vdc - that's quite a drop down to 21.


Onto the next topic - the EMI filter pre-power supply:
I have computers PSU's lying around. Here's one I disassembled, with a photo of the choke/filter section. Can I cut that off, and insert it at its AC power inputs, and outputting into the amp transformer from the two points where the 4 rectifying diodes begin? And I notice that this part of the circuit includes a big 2-lead thing which looks like a transformer but is in fact some sort of inductor - so I guess that has to go in as well. Am I on the right track here?



And lastly - 'ground plane'.
Thinking about some sort of grounded cover for the circuit, you have mentioned 'ground planes' - by this do you mean at least giving it a grounded metallic shield for one side, in this case beneath the circuit board, which could soak up most of the interference? If so I could try that first.

Thanks again, and have a great new year (it's probably already happened in Australia).

Thanks
Jim

Roly

E = sqroot(P R)

...gives us the RMS (or DC equivalent) voltage for the power in the load.  This is presumed to be a sine wave, and the peak value of a sine wave, Epk, is;

Epk = root(2) + VRMS

root(2) is 1.414

Therefore the peak voltage (minimum supply voltage) for the given power in the load is;

1.414 * VRMS


Quote from: Jungle-JimAt the moment the board is getting +/-35Vdc

Now I'm confused.  This is what I initially understood you to say, then you revised it down to 27V (just above), now we are back to 36V - I don't get it.  What are the supply rails under idle conditions?

The first two stabs, the resistors, and the capacitance multiplier are not voltage regulators so the output voltages will wander up and down with the supply as the load varies with playing; it only subtracts a few volts from the incoming supply for its own headroom.  If the incoming is 36V then you will get around 32V out, if it's 27V in then around 22V out.  The drop can be reduced but this also reduces your hum immunity.


Quote from: Jungle-JimI will look inside the box and work out where to insert an additional circuit board

Actually the parts are minimal, it's the heatsink for the regulators and capacitance multiplier transistors you will have to find room/mounting for - most of the mods mount on this.  Mount a heasink (e.g. CPU cooler) between the PCB and the reverb line, and mount a small bit of board on that to carry the few components that need it.


I think that is a low voltage transformer that provides the standby and USB power when the 'puter is "off".

The line filter section is attached.  If you are not clear about the mains side of 'puter PSU's then for safety sake I suggest that you hunt around for one of the enclosed IEC type (pic posted above) with spade terminals on the end.


Quote from: Jungle-Jimby this do you mean at least giving it a grounded metallic shield for one side, in this case beneath the circuit board, which could soak up most of the interference? If so I could try that first.

I do.  This is pretty easy to busk up out of minimal stuff such as alfoil and glue, or if you want to get fancy a sheet of thin ali from the hardware store.  I've lined many sensitive things with alfoil, built some radio station mic preamps in sardine cans, chopped up other cans to make solderable tinplate shields on boards, and it tends to be high profit for the effort.
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Thanks Roly,
I'll reply to this now, and go out later to celebrate NYE.

Quote from: RolyEpk = root(2) + VRMS

Do you mean Epk = root(2) X VRMS? I think I get what root(2) is now - thanks.

Quote from: RolyNow I'm confused.  This is what I initially understood you to say, then you revised it down to 27V (just above), now we are back to 36V - I don't get it.  What are the supply rails under idle conditions?

Sorry for the confusion.
I just measured again to be sure:
Coming out of the transformer, across the fuses, before the rectifier - it's 26.9VAC each side.
After the rectifier and big caps, in other words at the end of the power supply stage, it's + and - 37.5VDC for the power amp supplies (and after the big resistors etc the pre-amp supplies are +/-16.3VDC). And I don't know if it's significant, but those readings are with the speaker disconnected (because it's hard to get in there with the speaker connected).

Regarding the Noise Filter - earlier I did a graphic of this part of the circuit board, based on a reversed photo of the back of the circuit board, to clarify what's there, and I am assuming this is the noise filter. It's the first bit of the PSU to get power, and it's before the rectifying diodes. What else can it be? One curve-ball is that it's got that large inductor - is it assumed that this should also go back on? The unit you showed was tiny compared with all those components, let alone that large inductor. Am I on the right track here? Mains stuff isn't my forte, but I'm sure I can handle cutting this part off and re-using it. That power supply hasn't been used for months or longer. Shall I short all the electro caps on it?

Thanks again,
J

Roly

Disconnected speaker is okay for idle measurement.


I'd cut it fairly square, much as my cropped pic above, between the rectifier and the torroid, up past the mains connections, then across, leave some room for drilling mounting holes.  Forget the large inductor.

Mount it in a plastic box.

Profit.

{incidentally that four pin flat pack sitting over a slot in the board just near the mains connection is an opto-coupler.}


We hardly ever use the "average" value, it's mainly RMS or peak.
N.B. - 0.707 = 1 / sqroot(2) = 1 / 1.414


Yeah, you can short the two big caps next to the rectifier diodes, but if you've been handling the board and haven't been bitten, they're discharged.
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Hi all - and Roly when you see this.
Hope you all had a good new year.

There is an update on this Rockette-30 amp...

Firstly - I cut the noise filter out of the computer PSU, cleaned it up and mounted inside the cabinet on a plastic base with plastic spacers and screws. I can make a plastic lid for it, but it's secure. Did it make a difference - I really can't say - before and after comparisons are only from memory. At the least I am glad that it went in fine, the amp still works, and I learned a bit and got some confidence from attempting this.

Then I put a ground plane behind the circuit board - again - that was a bit of fun, and when I was out getting some takeaway Indian food, I got them to give me some ally food containers, and used those which was a good bit of improvising - flattened and glued down with contact cement. Again - it didn't solve all the noise problems, but it can't have harmed it.

Then today the LM883N's arrived, they went in, and they really didn't solve the noise/hiss issues, in fact I can't even say whether they improved it at all. 

What next? The hum is not shocking, but is still worse than the little Yamaha JX20 I use as a practise amp for my keyboard, which is very quiet. And the hiss is still hardly there with the instrument in the FX Return, but appears when you wind the pots up a bit.

I will look at Roly's schematics for the power supply modifications now, and chase up ordering the bits.

Thanks
J

Roly

I don't know if I specifically mentioned that your new Indian ground plane ( :dbtu:) has to be securely connected to the amplifier common/earth.  I didn't expect the filter or ground plane to make an obvious difference, that's more a realisation one day that you haven't heard fridge pops or drill whine from your amp in a while (normally triggered by somebody else's doing it and yours not).

Like your Line Filter, but I would like to see an insulating cover, less exposed mains the better.

The LM833's not improving the hiss is not what I expected at all.

I would try to localise the source (if this is even possible) using the following steps;

(since they are happily in sockets) pull all the preamps IC's;

Shorting plug into Fx Return and measure the residual noise with a millivoltmeter*;

Remove short and confirm the noise level stays the same;

(power off)
Insert IC2 (the output end of the preamp);
(power on)
Measure residual noise at the amp output;
(pwr off)
Insert IC1
(etc)
Then 3 then 4 (reverb section).

Your hiss will either suddenly reappear at one of these steps (good) or it will just get gradually louder as the IC's go back (accumulated noise, not good).

I still think that IC1 is your prime suspect, but I'm starting to think of second order causes, and the only one I can think of ATM is that one of the protection diodes on the inputs has taken a belting at some point and gone leaky (noise generator).


{* don't got no millivoltmeter?  Beg, borrow, or steal something like a cassette deck with VU meters on the front, anything with VU meters on the front.  It won't be calibrated but it should still give you a good relative idea of level.  Fx Send to Rec In.}
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Hi Roly
Thanks for being patient here...

I took these readings from the speaker cable, with the speaker removed, with the DMM at the 200 mV AC range:

The 4 LM883N are IC1-4.

With FX Return grounded, and no ICs - 0.4mV
With FX Return grounded, all 4 ICs - 0.4mV

The following are with no shorted cable in FX Return...
No IC's         8.6mV
IC 1 only      8.6mV
IC 2 only      8.6mV
IC 1 & 2       8.6mV
IC 3 only      0.7-0.8mV
IC 3 only - using a TL072  0.7-0.8mV
IC 2 & 3       1.6mV
IC 3 & 4 - with Reverb at 0 -   0.6-0.7mV
IC 3 & 4 - with Reverb at 10 - 0.9-1.0mV
IC 3 & 4 - with Reverb at 10, while tapping the reverb box - fluctuating up to 10mV
IC 1, 3 & 4   0.7-0.8mV
IC 1,2,3 & 4  1.0-1.1mv

With the speaker connected - with all ICs and FX Return not grounded - 1.0-1.1mV (appears to be the same if the speaker was connected or not.)

Is this useful, or did I need to test those ICs in and out, with the speaker connected?

Also, playing a guitar through the amp today, at a rehearsal volume, putting it through its paces, can I describe the sound: amongst the hiss, which I still think is too much, I noticed that when you hit a note or chord, that trailing behind the note, slightly delayed, is a slightly mushy clump of hissy distortion. Obviously if you wind the overdrive up, this is even more noticeable, as is the noise. Is this a sign of anything?

I am inclined to try to solve this pre-amp noise first, rather than modify the power supply - because unless I can solve this, and know the amp can sound good, tackling the remaining underlying PS hum is a secondary issue which I know I can deal with later.

(Oh and yes the ground plane is earthed).

Thanks
J