Thanks guys
I will have a think what to do with this amp. I have learnt an enormous amount from working on it, but perhaps it's just a noisy amp. I have other SS amps around, and I'm telling you this one has an annoying hiss. It has been a 'project' - with the fantastic help of you guys - to see if we could quieten this amp. I guess if I got it to a technician maybe the source of the hiss could be found, maybe he'd just say it's a hissy amp.

Criticisms of my recording - unfortunately I could only use my phone, so that's not a great start. Yes I put the phone near the amp, but there's already noise and hiss from the phone, so there was no point in having the phone at a distance where you couldn't separate it's own hiss from the hiss of the amp. Believe me from one metre the hiss on this amp is noticeable.

I will put the amp back together, and play it a few more times, and think about what to do.

For the time being perhaps we should close this case - and to quote Spinal Tap - 'best left unsolved'.

I plan to return to this forum soon with my next project - pimping up a Roland Micro Cube.

Thanks
Jim

Hi all
Sorry - another week delay.

I've done the two tests set for me -
Roly - the 0.1uF across pin 4 & 8 of IC3. I noticed that there's + and - rails which links pins 4 and 8 of all the ICs, so the cap could have gone anywhere along that line, but I put the cap very close to IC3 incase that had an impact on noise. Result - no discernible change. I have left the cap on, figuring it can't harm it.

Secondly, I did Phatt's two resistor changes around IC3. A 1k res became 10k, and the 560k outputting the Ch B pre-amp became a 100k. Unfortunately no change. If anything, the mV readings at the speaker changed - they changed to 0.7mV - 2.0mV when the Ch B vol went from 0-10, where as before, and in most other configurations, this was typically 0.7mV - 1.5mV.

I have left the cap in, and I've left the replacement resistors in for the time being, it's a struggle to change components on this board without the pads disintegrating.

This time I have uploaded an audio file - unfortunately recorded with my phone, I simply can't get any recording gear and the amp in the same room conveniently. There is something you can hear there. There is a very slight flanging sound in this audio which isn't in the real thing.

Thanks again for your patience,
Jim

Thanks again Phil

I'll try those resistor changes later this evening - will report back.
Much appreciated.
Jim

Quote from: Roly on February 01, 2015, 02:06:43 AM
Try sticking a 0.1uF (or what have you) directly across the IC power supply pins 4 and 8 on the copper side right under the IC3 chip, nice and close.

{Just a wild idea.}

Elsewise I'd say Phatt's suggestion of lowering the impedances around this stage may help.

Subtle problems can be tricky.

Thanks Roly
OK good idea I'll try the cap under IC3 pin 4 & 8. If this made a change it would certainly prompt an upgrade in the power supply.

But back to Phatt's suggestion about lowering the caps/res around what we now call IC3a, the very last op-amp before output. Bear in mind that I have already done the other experiment on IC3b (AKA the 2nd op-amp in the Clean channel) - where I increased the caps and res 100x - and it made no difference.

After all the confusion of the ICs being mislabelled, and before I start soldering the board again, could I just get you or Phatt to confirm again what this other mod was supposed to be?

Was it about dropping the res's 10x around IC3a? Currently V- or pin2 brings in the outputs of the three circuits - ChA, ChB and Reverb, via 270k, 560k and 470k res respectively. And across pin 1 (Vout) and pin 2 (V-) there's a 1M res, and I put in a 22pF ceramic on blank pads which had no cap (I really don't think this cap helped much if at all).

Thanks
Jim

Hi all

The following is the results of some tests Roly had asked me to do in previous posts...

Test A  - testing mV output with ICs inserted in the order of 3,1,2,4...

"(power off)
Insert IC3 (the output end of the preamp);
(power on)
Measure residual noise at the amp output;
(pwr off)
Insert IC1
(etc)
Then 2 then 4 (reverb section)."

I put the DMM across the speakers, set to 200mVAC... All these tests were done with the 270k Res connecting the OD channel one-leg-unhooked. So just Clean ChB. To save myself having to make a table here, I've attached a spreadsheet below giving the mV results of various permutations of ICs being removed. I hope it's self-explanatory.

Test 2
On Jan 5th, Roly gave me these tests to try...

Quote from: Roly
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

This test came out as 0.7mV - I was getting 0.4mV before - don't know why the change. In this test the pre-amp hiss goes to almost nothing, revealing an underlying hum which I know would improve with a better power supply. I'll think about improving the power supply once I get rid of this f*cking hiss.

Quote from: Roly
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.

I'm inserting IC3 now, because that's what we used to call IC2, the final IC in the chain. Result: 0.7mV (same as if the Return was shorted - but it sounds very different)

Quote from: Roly
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.

Again, IC2 is now IC3 - result - 91.8mV and a terrible moaning hiss. Take the ChB vol pot to zero and it's back to 0.7mV

Quote from: Roly
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.

This test wasn't affected by the IC's being wrongly labelled. Result - with the ChB Vol at 100% this test with the outside leg of the 1M grounded, comes out as 2.1mV. If you take the short off the 1M it's back to 1.8mV (that's with vol 100%).

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

This will give us 0.7mV if the Chb vol is 0%, 1.8mV if the ChB vol is 100%, as per usual.

Are these tests and the spreadsheet showing anything?

Some observations:
* With no IC's at all, the output jumps to 8.1mV and there's a moaning hum.
* Once IC3 is in, there is this hiss floor as usual, even with all pots on 0.
* That once IC3 was in, and no matter if it was in on its own or the other ICs were added it gave the same readings - ChB vol at 0% = 0.7mV, ChB Vol at 100% = 1.8mm. And at 0% the hiss was at it's usual lowest level, and at 100% the hiss rose by the same amount in each instance.

Conclusion?
The hiss is with IC3, and it's there even with all pots at 0. If the problem's in IC3, the final IC in the chain, and IC1 doesn't lift the hiss floor, then that might explain why both ChA and ChB are affected by hiss - because IC1 and IC2 are not at fault, it's the last stage of the pre-amp.

If so - what tests can I start doing on IC3?

Phew this turning into a saga to rival War And Peace.

Thanks again for your patience.
Jim

Many thanks again Roly and Juan

I've learned so much in the past month or two, and the adventure continues...

I'll go back to the actual amp in the next post, this one's about the software oscilloscope...

OK this is brilliant - I certainly have computers and sound cards lying around. How's the Linux version - called xoscope? That'd save me having to get a Windows PC up and running - or is www.zeitnitz.eu the least pain-in-the-a** option?

With your buffer circuit, Juan, are all those resistors 1/4w? And what are the diodes - 4148 or something bigger? I noticed that the xoscope guy had a fancy buffer which cleverly fits into a 5 1/4" disk blank, becoming a front-mounted fixture. http://xoscope.sourceforge.net/hardware/hardware.html

This being yet another thing I have never done, but would like to attempt... so if I did get this 'scope going, what tests would I perform with the probe?

For instance - would I work backwards from the point the pre-amp goes into the power amp, watching the shape of the wave, then keep going til I find the points where the hiss begins?

With any lucky the next test I do, which will be Roly's suggestion of pulling the pre-amp op-amps out, and replacing them in order, that might tell me something. I guess because the hiss happens on both the Clean and OD channels suggests the problem starts early in the chain of op-amps.

Thanks again,
Jim

Thanks Roly
I'll do this test again tomorrow, then do the audio recording when I can over the next few days.

Quote from: RolyGiven that we have now sorted out the IC misnumbering it might be an idea to revisit the progressive re-insertion of the op-amps working from the power amp input back towards the guitar inputs, measuring the hiss at each step.

So this must be the new sequence?...
(power off)
Insert IC3 (the output end of the preamp);
(power on)
Measure residual noise at the amp output;
(pwr off)
Insert IC1
(etc)
Then 2 then 4 (reverb section).

Quote from: RolyPerhaps if you could upload a few seconds of sound sample so we can hear the character of the noise?

Will do in the next few days - will take a little bit of setting up of recording gear.

Roly - if these tests yield nothing conclusive - are we at the limits of what we can diagnose without an oscilloscope? Would this be 15 minutes work running a probe around the circuit? The last tests, isolating these sections of the circuit, showed that both channels are generating hiss, albeit the OD channel being the bigger culprit. 

Jim

Thanks g1

To answer your question:
Yes the clean channel is not as loud as the OD channel - but that's more or less to be expected. The clean channel is reasonably loud, but you get a whole new level of noise (overdriven) if you crank the OD channel. I don't know if that's a problem per se.

Yes there *was* bad 'trailing edge' distortion on a guitar sound, but I think between changing all the pre-amp ICs to LM883N's and putting those 22pF caps across the op-amps which were missing, that problem was reduced by about half - in other words I can still hear some sort of glob of noise just following a note, but it's not quite as bad as it was.

Yes making input 2 (not 1) the main input jack for the instrument seems arse-about doesn't it.

Thanks for clearing up the design questions about this amp - it's always been a bit confusing. I have seen that article before.

This amp is flummoxing me - both pre-amp channels have some hiss/noise, and I just don't know how to track it down. And this hiss can't be 'just how this amp is' because if I bought a new amp and it had this sort of hiss I'd been returning it under warranty.

Jim

Hi there

Quote from: PhattCan I suggest isolating the sections to narrow down the field, divide and conquer.
The mixer is the last stage before poweramp and it mixes 3 signals, those being the 2 channels and the reverb return path.
By lifting one end of those 3 resistors one at a time might give a clue as to which section is causing the most trouble.

The other point
By lifting one end of those mix resistors from ChA and Rev return (points D&R)
Leaves just the clean ChB in circuit and if hiss is still present then just focus on fixing that section first.

I have done this experiment - here's the findings...
There are the three tracks which come into IC3a - from the Clean, OD Channel and Reverb, each with a resistor. I desoldered one leg of these resistors to disconnect these in this sequence...

Just the Reverb circuit removed (so OD and Clean)....
No change - same hiss as usual.

Reverb and OD removed, so just Clean channel...
Quieter, but some some hiss still present.

Rev, OD and Clean all removed...
Quiet, same as if the Return jack was shorted, as you would expect.

Reverb only...
Not totally silent, but acceptable levels of hiss given that it doesn't get noisy till the knob is 6/10.

Reverb and OD channel only...
Significant noise from the OD channel, on the OD and Vol A knobs, once they are past 3/10.

Reverb and Clear only...
Quieter than 'Rev and OD only', but still there's enough hiss to be annoying past 4/10 on the Vol B knob. Not quiet enough to record with. Still more noise than my other little Yamaha JX20 amp at comparible volumes.

So as it stands I have left it with the Clean and Reverb channels still connected, the res on the OD channel still disconnected. The Reverb channel isn't a problem.

The thing is that the Clean channel is just on the wrong side of being acceptable for hiss - just a bit less might be ok, meanwhlie, the OD channel has a worser hiss problem again.

Where do I go from here?

Is there a clue in that switch setting doing something a bit unexpected? Maybe not. What about that other idea mentioned about shielding the caps at the input stages.
Thanks guys for looking at this.
Jim

Hi Phil, Roly, and all riveted onlookers....

Here's a new version of the schematic with corrected op-amp labelling. Note IC2 and IC3 are now completely swapped around.

I will post another thread shortly with a few photos of either side of the PCB - if there's anything to be gleaned from these - eg component placement etc.

I have restored the original cap and res's to IC3b. None of these changes are impacting the hiss either way.

But there's one thing I have to report, and I don't know if it's significant: it's to do with the guitar signal that comes out according to different switch and input settings... bear with me... (I can't say when this problem started, but some time prior to me attempting to fix this amp)

I ran these tests:

Guitar in input 1
Switch on ChA: - Vol A/OD responds ok (while adding hiss), Vol B does nothing (but doesn't add hiss).
Switch on ChB: - same behaviour as above - Vol A/OD works ok (with hiss), Vol B does nothing (but doesn't add hiss).
Switch in middle position: - Vol A/OD responds ok (while adding hiss), but Vol B, when wound in, introduces hiss but no guitar sound. (this is the questionable one I want to highlight... surely the second channel is supposed to be audible on this setting)

Guitar in input 2
Switch on ChA: - Vol A/OD responds ok (while adding hiss), Vol B does nothing (but doesn't add hiss).
Switch on ChB: - Vol A/OD does nothing but adds hiss, Vol B responds normally (while adding hiss).
Switch in middle position: - Vol A/OD responds ok (adding usual hiss), Vol B responds ok (adding usual hiss). (if both channels are responsive when the guitar is in input 2, switch in middle, why is that not the case when the guitar is in input 1, switch middle??)

So - a question mark there: admittedly this amp has a slightly confusing set up with the channels, but surely this is not doing what it's supposed to?

Phil - I will detach legs off those resisters to isolate those pre-amp sections later tonight if I get time.

Jim

Hi Phatt and Roly

I spent some time tonight on the Session amp...

Firstly - I disconnected one leg off each of the 4 input diodes - no change to the hiss - soldered them back.

Following that, I did nearly all the tests and changes as suggested by Phatt - and I will leave what I wrote below so you can see what I did, but now... after again being a bit flummoxed by the PCB and there being a few resistors which were different to spec, and me not being able to find components which should be there, I scratched my head, then did another significant tracing of the PCB, and discovered one more thing:

The schematic got another pair of op-amps swapped in the wrong IC. FFS. IC3a and IC2b - mislabelled each other. The guys who made this board must have had a bloody good laugh because this has certainly confused me. I thought I'd correctly identified all the op-amps the previous time I discovered the other pair had to be swapped. I was correct with them, but I guess I didn't dig far enough into the circuit to see the error with the second pair.

So I'll do another PDF of the schematic, and undo all the changes I've made to the board so far, and get back to you with the new revised schematic, and maybe Phatt, if you don't mind having a look at the revised schematic when I post it tomorrow, and applying your theories to the new layout. We might have to rethink this again.

Below is a record of what I did tonight, though I now know that some of it will be useless because apart from in IC1, I was reading some of the op-amps round the wrong way...

--------------------------
Now - looking at Phatt's suggestions...
Firstly the possibly incorrectly banded resistors...
1 Phatt: "Most obvious candidates, 560k series R at IC2b. if it's 56k instead of 560k then the hiss would be horrendous."

Me before discovering the IC swap: On IC2b - both the 560k and 270k R's on pin 6 of IC2 are 470k, the 1M is 1M. Is this significant that they used 470k for both those? Should I try swapping them?

Me, later after discovering the IC swap: F*ckin glad I didn't: I now know that IC2b and IC3a are swapped - these readings are consistent with what IC3a should be - so ignore what I wrote, other than it confirming the op-amp swap.

2 Phatt: "Same goes for both input stages, that 100k R that goes to ground through a 10nF cap if that is 10k the gain would go way up and hiss would certainly be evident."

Me: On IC1a - pin 2 has 100K and 1M - which is correct, pin 6, the same - no errors on either. (NB: These readings are fine, because IC1 is correct in the schematic).

So I went to IC3b (correctly renamed after the first schematic revision):

3 Phatt: "IC3b pin 6 has 2 R's and 1 Cap and I just tested this for real and it is a big noise issue.
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."

Me: IC3b - I substituted the 10k across pin 6&7 for a 1M, and the 1k to ground became 100k. The 1nF cap across 6&7 became a 22pF. There was no particular change in hiss. I can't test the amp at any volume right now because it's late and people are sleeping, but I can say the hiss characteristics are more or less identical.

Given I've changed the res's and caps on IC3b - and it didn't solve it - should I replace the original parts back in seeing as that wasn't the problem?

So - I'm not sure what's next - I was out of my depth ages ago, and just relying on you guys. This schematic being wrong *twice* has cost me a bit of time - though I guess it's been a lesson is having to trace a board and solve a puzzle.

Thanks
Jim

Thanks Phatt and Roly
Tomorrow night definitely I will get the amp on the bench, and get some readings and results and we'll see where it goes from there.

Thanks guys, this is brilliant, the amp is improving and I am learning an enormous amount along the way.
J

Thanks a lot Phatt and Roly
I will put the amp on the bench and do those experiments this week. I suppose I am a bit reticent to try tests which involve a lot of "solder, connect speaker, power up, power down, disconnect speaker, soldering, power up power down" etc, because the speaker cable is so short, and it's difficult to test the amp without almost fully reassembling it. Plus this circuit board seems weaker than most - one or two solder/desolders on these pads they seem to collapse (and no the iron is not too hot). I guess I could extend the speaker wire - that'd make life easier.

Roly - with the diodes - yes it'd be nice to know which one, but I have an image of the current hiss levels from this fkn amp in my head, and if replacing the 4 diodes made a change, then I might not know which one it was, but at least I'd be able to tell the problem was in that area.

But thanks to both of you - this is getting somewhere. I still think it's a nice little amp, worth persisting with.
J

To Roly, and Phatt.
Firstly, updating this, responding to Roly's previous posts first...
I have put 22pF ceramic caps in those three spots which had missing caps. It seems to have calmed the hiss down a lot - but not completely. I will email Award-Session shortly and see if I can get any info about why those caps were missing - because it clearly needed them - if I'm lucky I might be able to contact the designer Stewart Ward.

But - the hiss not completely gone, so I am interested in continuing to work on this amp, and use Phatt's recent posts.

Also - next time I have the amp on the work table I will also think about replacing those input 4 1N4148 diodes - I have some at home bought for another project. (Roly had previously suggested that could be one possible problem). I could test each one individually, but it seems a lot of work for components worth pennys. It'd be just quicker to replace the four and see how that goes, can't go wrong.


Hi Phatt
Thanks for your posts.

I'll go over what you're saying to clarify your points...

Firstly - you've made two separate suggestions - and I'm taking it that the second one would be the one to try first?

Going over the two posts:
In your first post, you suggest scaling down x10 the resistors around IC2b. You point out that a possible consequence may be an altering of tone - I have to say that I don't want to change the tone. With this amp I am after blues and jazz sounds - not heavy metal. I only ever have it on the edge of overdriving, and there's a sweetspot on Input B/Ch B where I can pull a nice soulful blues sound (Ch A is better if I want to cut through with a rocky lead break).

I don't really want to experiment with dropping various resistors and caps in - though I guess with those resistors being much lower, I could probably get away with leaving the old resistors in and tacking the lower ones across the underside of the board in parallel - that would be a way to test different values without doing too much component replacement (and obviously I'd properly replace the resistors if I liked the change).

Looking now at the second post...
It sounds like you're more convinced about this modification: you are suggesting bringing IC3b into the same cap/resistor combination as IC1a, IC1b, IC2b and IC3a - the 10k res becomes 1M, but the cap goes down from 1nF to 10pF (the others are all on 22pF, though Roly preferred 15pF on IC1a and IC1b).

Roly - have you got an opinion on either of these modifications? I'm more inclined to try the second one first.

And Phatt, you're saying that the mods you are talking about refer to Ch B, and that you didn't get time to look at Ch A. Well as it stands, though the amp still has some hiss, Ch B is better than A. Ch A is more hissy, particularly once the overdrive knob is over 3-4.

I need to get the amp on the work table this week, and try those mods to IC3b and report back. I will also order the parts needed for Roly's power regulation circuitry.

Thanks
J

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