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

March 28, 2024, 09:23:08 PM

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

Roly

Quote from: Roly(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).

I should have make it clear that the order of trying the IC's, 2, 1, 3, 4, was specific.  There is no point in trying IC1 alone since there is no pathway to the output if IC2 isn't in place.  The idea being to progressively add stages from the power amp back towards the preamp input.

Quote from: Jungle-Jim
With FX Return grounded, and no ICs - 0.4mV

No IC's         8.6mV

IC 1,2,3 & 4  1.0-1.1mv

This is a bit confounding.  With the power amp input shorted you have about half a millivolt of residual (which looks quite reasonable), but even with all the IC's in this rises to only about a millivolt.  I would think that this is also reasonable, and am surprised that you can even hear it - a millivolt into a speaker should be inaudible.

Just a check on terminology to be certain we are on the same page; "hiss" to a tech is a high pitched sound like steam escaping, this is what you are hearing?

I should have asked before, but specifically with the clean channel, is the hiss effected by any of the channel controls, volume, treble and bass?  I'd try to ignore the overdrive and reverb paths for the moment, try and get a handle on the simplest case, the clean channel.

{there is a side point here about needing a DC blocking cap in series with your meter when measuring AC where there might also be DC (which confuses the meter), but this normally results in the meter over-reading, so we can at least take your readings as "no worse than".}

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

Jungle-Jim

Roly
Listening to the amp, without taking it apart to do voltage readings, and not plugging in an instrument because it's too late at night...

My definition here of hiss - it's probably closer to white noise or a snow blizzard rather than steam escaping, but it's not unlike that. And it's more of a constant, smooth stream rather than a coarse or jagged noise. And when all knobs are on zero, the hiss is more-or-less not there. It's only when you start bringing the knobs up from zero...

With the switch on Channel B, the clean channel, it does the same whether it's got a shorted jack (in the input 2 jack in this case) or nothing: from about 30% upwards (on the vol knob) the hiss becomes noticeable, very much so upwards of 60%. This channel on its own is fairly low-volume, so you'd have the vol over 50% to get anything out it, by which stage the hiss is undoubtedly present. The bass and treble knobs don't seem to introduce any new hisses of their own, but particularly the treble knob obviously emphasises the hiss in the channel.

The hiss in Ch B is not completely appalling on its own, but the usual way to play this amp is to combine both channels to get any volume or tone out of it (it's a funny set up with the two channels combining or being separable) so that means having both channels up and running. The thing is as soon as you've got a bit of volume on both channels, and a bit on the overdrive knob - to even just get to 'bedroom' volume - you've got a very noticeable hiss - you wouldn't want to record with this.

With Channel A on its own via the switch, with jack in input 1, again, on the vol knob there's not a lot of hiss introduced (but then there's virtually no instrument volume either unless you introduce the overdrive knob), but as soon as you give the overdrive knob anything, the hiss starts, and is noticeable over about 10%, annoying upwards of that. Also - stop press - I've just noticed that the overdrive knob is a bit crackly if you wind it up full - I hadn't noticed that before.

Also - I said in the previous post about another thing it does - I'll re-paste it as I described it:
"...playing a guitar through the amp today, at a rehearsal volume ...amongst the hiss... I noticed that when you hit a note or chord, that trailing behind the note, slightly delayed, is a mushy clump of hissy distortion, like an echo. Obviously if you wind the volumes up, this is even more noticeable, as is the noise."

Another observation I made was that the Ch B has a nicer tone - warm - while it's really hard to get a nice tone out of Ch A - sure it's overdriven, but it's always a bit scratchy, no depth about it.

It's clearly in the pre-amp, because playing the guitar through the FX Return is admittedly a fairly dry sound, but still it's a decent tone without hiss.

What about other possible causes then - grounding? shielding? poor power supply, input diodes, dirty pots? Sorry if this is turning into a wild goose chase.

Cheers again, much appreciated for looking at this,
J

phatt

Hi Jim,
I'd swap the chip Roly mentioned then if that does not help consider tracing the circuit,, not hard as it's single side board and the issue will be in the preamp section so you only need to work out that part.

The symptoms you describe are found in many small combo amps, some are as you say unusable at any high volume setting.

Also be prepaired to find build mistakes, It is not uncommon to find a 100k resistor on a board with 10k printed clearly on pcb.  :duh
It can be that simple but insanely hard to find if you do not have a good idea of what should be in place.

If it helps,,,I wus once a knob with a soldering iron and long before internet I drew up many circuits by simply copying the PCB back to a schematic.
Working from the component side hold it up to strong light and you will be able to follow the copper tracks on the other side with ease.
You will waste quite a few pages but you will get there and then you will have a real schematic and learn heaps in the process. :tu:

Phil.

Roly

Quote from: Jungle-JimWhat about other possible causes then - grounding? shielding? poor power supply, input diodes, dirty pots? Sorry if this is turning into a wild goose chase.

Nah, it's just getting interesting.    8|


...um... it's a bit of a shot in the dark, but hiss, odd lingering distortion ... I'm starting to wonder ... could there be a supersonic instability issue emergent here?  Aging/broken 100nF supply bypass caps perhaps? {and it hasn't been done right, each rail to ground when it should be bypassed rail-to-rail.}


Good advice from Phil, teaching yourself how to trace a PCB becomes mightily useful on the service bench.


On the circuit we've got;
Clean Ch
x10 (HF rolloff) -> tonestack -> x10 (HF rolloff) -> volume (top coupled) -> x2 (HF rolloff) -> power amp

This does not compute; my Twin-50 LM833 preamp has almost this configuration and is utterly silent (so I'm puzzled...)


The values of the HF rolloff caps isn't given on the circuit, but I'm beginning to wonder about off the circuit, if this amp has been hacked/modded at some point perhaps to get extra treble, and that the HF rolloff caps have been removed ... or something.

So are there any signs of vacant holes, desoldered pads?

I'd also try a test with the protection diodes at the inputs disconnected (lift one leg of each).
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Quote from: RolyNah, it's just getting interesting.

I'm glad about that Roly - and again thanks so much for your help.

Quote from: Roly...um... it's a bit of a shot in the dark, but hiss, odd lingering distortion ... I'm starting to wonder ... could there be a supersonic instability issue emergent here?  Aging/broken 100nF supply bypass caps perhaps? {and it hasn't been done right, each rail to ground when it should be bypassed rail-to-rail.}

Is it worth a shot to replace those 2 100nF (0.1uF) 250v MKT caps? (for reference they are orange MKT metallised polyester film caps)
These amps were designed by Stewart Ward - who presumably is something of a SS amp guru - why would he get that wrong?

Quote from: RolyGood advice from Phil, teaching yourself how to trace a PCB becomes mightily useful on the service bench.

OK - this is a good point. Clearly there is a limit to how much someone can diagnose a circuit like an SS amp without learning to follow a circuit in a meaningful way and knowing what readings to take from components. If so - can somebody teach themselves to do this? Should I get an oscilloscope? And probably a signal generator? I have done quite a bit of backyard electronics over the years - fixing things by eye or hunch or basic knowledge - but my knowledge falls short fairly quickly. However I seem to be able to follow this SS amp circuit using the schematics - it's just that I can't really troubleshoot.

So anyway back to the amp...

Quote from: RolyI'd also try a test with the protection diodes at the inputs disconnected (lift one leg of each).
OK - I've located those 4 IN4148 diodes on schematic and board. That's doable - but test them one by one? and get microvolt readings or just listening by ear for changes? I'll see if I've got any spares of those - they're so common.

Quote from: RolyThe values of the HF rolloff caps isn't given on the circuit, but I'm beginning to wonder about off the circuit, if this amp has been hacked/modded at some point perhaps to get extra treble, and that the HF rolloff caps have been removed ... or something.

Well done - good bit of detective work there - I don't know your reasoning, but you're onto something here - does this help...
Looking at the schematic, I could see 3 caps which didn't have values. They were across pins 1&2, 6&7 of IC1, and pins 6&7 of IC2. All those caps are parallel with a 1M res. Guess what, looking at the actual circuitboard: While IC2 pin6&7 has a 1M res and a 22pF green ceramic cap, on IC1, pins 1&2, 6&7, the 1M res's are there alright, but where the caps were supposed to be is unused pads - there's never been caps on them. What does that mean?

Quote from: RolySo are there any signs of vacant holes, desoldered pads?
Nothing that's been removed, leaving desoldered pads - but there's unused spots like those two caps from IC1 and also blank (never went in) are D9 & D10 - the pair of opposing parallel IN4148 diodes between earth and the Ch A/B switch.

Any of that useful info?
J

Roly

#35
Quote from: Jungle-JimIs it worth a shot to replace those 2 100nF (0.1uF) 250v MKT caps? (for reference they are orange MKT metallised polyester film caps)
These amps were designed by Stewart Ward - who presumably is something of a SS amp guru - why would he get that wrong?

You could, but I personally wouldn't bother unless they show signs of damage; these are normally pretty damn reliable class of components.

I'm a friggin' Guru too (in fact lotsa Gurus on this site  8| ), and I make mistakes, and have bad days, and misunderstand, and forget...  but I wasn't meaning to imply that there was some instability as an initial design fault, but rather from a aging/damaged/failed component, crazy mod, joint, PCB trace...


Quote from: Jungle-Jimcan somebody teach themselves to do this? Should I get an oscilloscope? And probably a signal generator?

Yes, yes*, buy, or more interesting, build.  {you can make various test signal recordings in Audacity (freeware) and play them back on an MP3 player, or similar with cassettes.  A really basic spot frequency test oscillator can be built in a sardine can around a single transistor, and there are a wide range of circuits ranging in complexity/quality up from there.}

*There are a number of free PC 'scopes and many unloved desktop machines that they will run on.  Between these and Audacity there is a lot you can wangle.


The front of my bench setup has evolved to be the CRO and the homebrew Noise and Distortion/audio millivoltmeter sitting on top.

Normally the bench signal probe is connected to the input of the mVmeter and all range setting is done with its range switch.  It has a normalised output which is distributed to the CRO channel A and the bench frequency counter behind (CRO-B Ch also has a probe and is under direct control).  The input range of the mVmeter is 1mVAC to 200VAC in 10dB steps.  (This NaD/mVmeter also has high purity sine spots available for 100Hz, 1kHz, and 10kHz distortion measurements).  The normal bench signal source is a modified ZA-1202 kit from (old) Dick Smith using an LM13600 OTA, a good design obviously bought in.


Quote from: Jungle-Jimit's just that I can't really troubleshoot.

Da kneebode connected to da thigh bone,
and the thigh bone connected to the ...


Up the river from New Orleans, step by step methodical.  Get a very firm grip on Ohms Law, the relationship  between voltage, resistance and current, and circuit theory.

Your best test instrument are your eyes.  Observation.  Be suspicious.

Designers (like me) are lazy, we reuse and recycle everything we know works.  Look in most valve guitar amps and you will find a 12AX7 first stage with a 100k anode load, 1k5 and 25uF in the cathode - it's a Lego(tm) brick, and most of the gear we deal with is, it's highly generic.

Solid state preamps tend to be a lot more "imaginative", but mainly we are still dealing with the same set of Lego(tm) bricks, just that there has been a whole lot of channel switching and control introduced, but that's pretty generic in itself too.

Work out your flows of power and signals, power flows downhill from the supply to the most insignificant LED (always check the supply(s) first), and signals from inputs through components and devices, switching, to the power stage and speaker.

A lot of faultfinding is binary division, splitting in half and finding which half the fault is still in, then splitting that in half, and so on until you are at the stage, then the component.  Which is cause, and which is effect?




Input protection diodes - lift one end of each of the four diodes, and see if the hiss stops/reduces, is one of these diode generating the hiss? {reasoning: there is no series resistance to protect these protection diodes themselves, so it's possible some excessive voltage has been applied, say from another amp, or electrostatic discharge, that has caused one or more to go leaky and thus noisy?}


Quote from: Jungle-JimI don't know your reasoning

Funny noises, tinfoil in the piano, "squrglies" (that sometimes also appear in digital systems), here makes me think that a possible cause is ultrasonic instability, either constant, or parasitic (depending on a signal to drive it over the edge).

Possible causes are;
loss of supply line bypassing, allowing unwanted coupling between stages
loss of a deliberate stability rolloff.  All the caps I mentioned (and some others) are used to reduce or limit the upper frequency bandpass, locally on each stage.  These are mostly small caps, a few tens to a few hundred pF, most likely tiny disk ceramics with obscure markings, maybe just a short number.

These are in parallel with the feedback resistors of each stage, so as the frequency rises so their capacitive reactance,
Xc = 1/2 Pi f C,
(or "AC resistance") falls to equal the resistor it is in parallel with, and at that frequency (the hinge, corner, or fc) the response starts dropping at -6dB per octave (-20dB per decade).

Had some prune decided to create a "treble monster" by cutting out all those naughty frequency limiting caps - then I can play like Hendrix?  Wow!  Well... maybe not, but in old gear anything is possible (including Pythons  :o ).




Okay, here's what I would do now to try and get a handle on it.

1. Select channel B, Ch A volume, overdrive, filter to zero.

Starting with the Fx Return shorted, speaker connected, and AC millivoltmeter clipped across the speaker.
Confirm you are still getting your 0.4mV reference noise level


2. Remove short from Fx. (oops!)
Insert IC2, Ch B volume at zero.
This will give you the noise contribution of the IC2b stage only, the preamp output buffer (which has a gain of about x2).  Record.


3. Center the Treble and Bass controls, short the wiper of the treble pot to ground with cliplead o.n.o, Ch B volume to max.
This will give you the cascaded noise contributions of IC2a and b (~x10 x2).  Record.


4. Remove the short from the treble pot and place it across the 1M resistor connected to pin 3 of IC1a.  Insert IC1.  Set treble and bass to 12 o'clock/50%.
This will now include the noise contribution of IC1a (~x10 x10 x2).  Record.


5. Remove the short on the 1M resistor, and record open input noise level.



{ed to correct #1}
If you say theory and practice don't agree you haven't applied enough theory.

g1

Roly, you omitted at which step he should remove the short from FX return.  I'm guessing at the end of step #1 ?

Jungle-Jim

Roly
Thanks for that last post - there's several different things you're talking about which are all of great interest.

If I get time tonight I'll do those tests - but - in case you see this before I post my results - you didn't respond to what seems like a significant find here, as I described in the previous post:

You're talking about caps being missing - well - there's supposed to be caps (in parallel with 1M res's) coming off pins 1&2, 6&7 of IC1 of the pre-amp - but - well these caps aren't there, and never were - the pads are unused. What does that mean? There are also some diodes also missing - see the last bit of my previous post for more about this.

I'll send this now and go back to digesting all the stuff in your previous post.

Thanks a lot Roly - you're very generous and patient. There was no insinuation that you weren't 'guru' status - far from it - your ability to predict that those caps were missing was amazing. I'd buy you a beer but I'm on the other side of the planet!

Cheers
J

Roly

Quote from: g1Roly, you omitted at which step he should remove the short from FX return.  I'm guessing at the end of step #1 ?

Yep.   :-[   Corrected, thx.




Quote from: Jungle-JimYou're talking about caps being missing - well - there's supposed to be caps (in parallel with 1M res's) coming off pins 1&2, 6&7 of IC1 of the pre-amp - but - well these caps aren't there, and never were - the pads are unused. What does that mean? There are also some diodes also missing - see the last bit of my previous post for more about this.

Well let's put the missing diodes on the back burner; they are limiters for the signal coming from the Overdrive channel, which we currently have no interest in.  Later Mon.


The missing caps means that this preamp is wide open, no "dominant First Pole", i.e. a frequency selective network(s) that starts to roll the high frequency response just about our highest needs (20kHz for "Hi-Fi", maybe 10kHz for guitar).

The original TL07n series is no bandwidth slouch, but you have now fitted LM833's and these have a bandwidth out to 3.5MHz!  Yes, radio frequencies.  So these are particularly likely to find stray capacitance feedback paths and oscillate.  (the first incarnation of my Twin-50 LM833 preamp happily oscillated at 455kHz until I forcefully cut the bandwidth back to 31kHz.

A bit of doodling in LTSpice produces the values for an overall 10kHz top cutoff.

At IC1(a), a really tiny 15pF.

At IC2(a), 1nF.

At IC2(b), another 15pF.
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

Hi Roly (and whoever else is following this)

I'm just writing a note to say - no I haven't disappeared, but progress slowed because I hit a problem which took me a while to figure out.

The problem started because I couldn't find one of the caps Roly referred to, and I think one of tests he suggested in the previous post didn't appear to correspond with the PCB of the amp. To clear this matter up, I began following all the traces from the pre-amp IC's, but ended up wasting a bit of time thinking I was too idiotic to read a schematic, because it wasn't matching up, and I kept getting weird readings.

Then I ended up realising that the schematic has this error: two of the op-amps are listed as being in the wrong IC: IC2(a) and IC3(b) are in fact the wrong way around. I have taken the schematic into Indesign (a graphics program) and correctly relabelled those two op-amps.

One question: are the two op-amps in each IC working in tandem? - presumably not or else they wouldn't be able to be swapped like this.

So when I get a minute (I've been a bit busy this week), I need to print out the revised schematic, and do all of Roly's tests, then probably order a few ceramic caps and take it from there. I really question why those caps were missing, and I am keen to put in caps of the right spec and see what happens.

I'll get back in a few days when I've had more progress. 

J

Roly

Life is what gets in the way when you are trying to do things.   ;)

Quote from: Jungle-Jimtwo of the op-amps are listed as being in the wrong IC: IC2(a) and IC3(b) are in fact the wrong way around.

Oh great, that helps a lot (but at least we have a circuit).  This sort of thing (the board overprint not matching the circuit) is a not an uncommon error, but as you have found it can have you running in circles until you work it out.


Quote from: Jungle-Jimare the two op-amps in each IC working in tandem?

With few exceptions the op-amps in duals and quads are quite independent of each other apart from sharing power supply pins.  You can ignore the fact that there are two or four in the package and treat them as if they were singles.

Quote from: Jungle-JimI really question why those caps were missing

This is the sort of thing that is modified during manufacture, a production line mod, based on experience with the amp as production proceeds.  Sometimes you will find unused component holes, other times you will find extra components tacked on to the copper side of the board.  The boards may only cost a dollar each but if you have ten thousand in the store then cutting or linking tracks, omitting components or tacking extra ones in is better than junking ten grands worth of boards for such minor changes.  We can only guess that the manufactured amps turned out to be a bit duller than the production prototype, and an easy way to brighten it up a bit was to reduce or omit these top end rolloff caps.  Now that you have wider bandwidth LM833's in there it looks like these caps are again needed to keep the preamp stable.
If you say theory and practice don't agree you haven't applied enough theory.

Jungle-Jim

#41
Roly

I have attached a revised version of the schematic. I have done a number of tracing tests to confirm that IC2(a) and IC3(b) need to be renamed one another, and info about the caps on several of the op-amps.

Today I was in the local Maplins electronics shop looking for some 15pF Ceramics, and they didn't have them so I bought a pack which had a wide selection of ceramic caps - I figured it would come in handy anyway - but nearest to 15pF, it's got 10pF or 22pF. You specified 15pF across both IC1(a) and IC1(b) - but which out of 10 or 22pF would you suggest trying instead, or do I need to specifically use 15pF and if so order them in? Also - IC3(a) specifies 22pF on the schematic, but there's nothing installed on the board. Should I just put a 22pF in there?

But before that, I am conscious that in a previous post you gave a set of tests for me to do - involving putting the DMM across the speaker at 200mVAC, taking the ICs out and doing a series of tests reintroducing them. Should I still do these tests, or can I just try the caps in and see how it sounds? The thing is I was now hoping to avoid taking the LM833N ICs out again because during the last tests they went in and out several times, and it was hard not to bend pin legs, and one LM833 actually went bad during the swapping, luckily I had a few extras.

So - do those tests again - or just put caps in see how it sounds? (And if I put the caps in, is there an order to try them in, or just stick all three in?)
Thanks again,
J

Roly

Nah, you can just whack in whatever you feel like and see how it flies.  I'd start with the higher option, 22pF, and if you think it's too muffled you can change it down in value.  Don't forget that you can obtain other values by placing caps (or resistors) in series or in parallel.  For example two 27pF in series will give you 27/2 = 13.5pF, two 33pF 33/2 = 16.5pF, and so on.

The important point is to try and get the amp clean and stable, no funny sound effects.  If you overdo it the amp will sound muffled and lacking in tops, but better to overdo it and ensure the amp is stable and not supersonically oscillating (which is what I think was going on), then you can try reducing the cap values if you need more tops, but suck it and see.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.

phatt

IF?? this was in front of me I'd be changing resistor values around IC2b (mixer chip,Top right) as that in my umble experience will be the main source of the hiss. :-X
I've just finished testing a circuit I designed which was plagued with the dreaded hiss at high levels,, so tiss all fresh in my mind.

Can't hurt to change 4 resistors,, suck it and see. ;)

You have 4 resistors connected to pin 6 of IC2b these can all be scaled down x10,, so 560k becomes 56k and so forth,, 27k, 47k. 1Meg is then 100k.

(Roly can correct me if I'm wrong here) but that should give very close to the same output but at a much reduced impedance.
The only potential gotcha is that it will likely alter the tone which could actually be a good thing depending on your musical preference.

Re those caps; My test circuit is similar in that it has 2 stages before IC2b and the cap in my circuit is 220pF. (my fb res is 100k) Now if your music is all heavy metal then you may want more top end but if you want a bluesy sound you won't like heaps of top edge,, as Roly says,, try different values.

If I get time I'll test it out as my Breadboard is running hot from much testing in the last few weeks.  :lmao:
cheers,, Phil.

phatt

Just been fiddling with some ideas and here is what I found;
Channel B,
IC3b pin 6 has 2 R's and 1 Cap and I just tested this for real and it is a big noise issue. :grr

Raising the value of both R's x 100 will deliver same gain and noise is down a lot, So 100k to ground and 1Meg between pins 6&7. You will need to change the 1nF Cap to say 10pF for the same freq response. Left at 1nF you will have no treble.

Ran out of time to workout the A Ch and it's a little more complex,, but give me time. :)
Phil.