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

March 28, 2024, 07:40:45 AM

Login with username, password and session length

Recent Posts

 

18W Amp Build - Few Minor Issues

Started by Littlewyan, November 11, 2013, 06:55:11 PM

Previous topic - Next topic

Littlewyan

#15
 <3)I've checked the Input Jacks and they are fine, it does this fizz noise on Input 1 as well. Now I've tried the conjunctive filter and it does seem to help a bit. Also tried a Zobel network but that didn't seem to do much.

I might try lowering the Plate Voltage from 350V to 280V just to see if i still get the issue. Other than that looking at other schematics my amp isn't much different from others so cant see why it has this issue. Unless its something like the EL84s or the OT

Roly

Just reviewing this thread.

Quote from: LittlewyanI have taken photos throughout the entire build process but obviously cannot post them all here

Get a free account on an imageserver, put all your pix up there in a build-thread folder, and post the link here.


Quote from: LittlewyanAlso tried a Zobel network but that didn't seem to do much.

It won't on a valve amp.  These are fitted to s.s. amps to prevent HF/VHF instability.


You seem to be focusing on the OP transformer, and I seriously doubt that is where your frizz trouble is coming from.


Looking at your underchassis pic I notice that while the jacks are insulated from the chassis there are two earthing bolts with lugs to the chassis, some distance apart - what's that about?  You suspect parasitic instability, here is a place to start looking.

I assume that this grounding is as per the kit instructions, but it doesn't look right to me; "single point" means just that.  Have you tried searching to see if anybody has had similar problems with this kit?
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

#17
Good to hear from you Roly!

Off the top of my head I believe there are three grounding lugs in this amp which are per the kit instructions. The Grounding Lug near the Input Jacks grounds the ground line on one half of the turret board and there is a ground cable going from the input jacks to that side of the turret board. I did check all of the connections as I built the amp and all were good, also checked the grounding again on the input jacks and they are grounded when nothing is plugged in.

I did read a few topics over at 18watt.com where a few people had similar issues but they'd done different things to fix it. Barry at Ampmaker believes that it is a characteristic of the amp but it can't be, I've heard Amps that are of very similar design and they sound nothing like this amp. The only thing that does sound similar to this amp is one of the sound clips on ampmaker.com, you can hear the fizzy noise in it.

I've been comparing my amp with the Marshall 1974x which I've heard clips of and has no fizzyness.

http://www.webphix.com/schematic%20heaven/www.schematicheaven.com/marshallamps/marshall_18watt_schem.pdf

Only real differences that may make a difference are the HT Voltage, Tone Stack, Capacitor between Tone Stack and PI and the capacitor on the otherside from the Tremolo Channel (to ground in my case).

Littlewyan

Just realised this topic is in the wrong place! Could a moderator move it please?

Roly

I see the close similarity with the Marshall.

The circuit is never the full story, there are always unmentioned details of technique that can make all the difference.  The circuit shows you what they did, but it doesn't show you how they did it.  For that you need other information like layout diagrammes.

The layout diagramme is only as good as the care that went into the basic investigation; it is one thing to build one and document it, it is quite another to manufacture something in bulk (or as kits), and there is a constant need to feed experience from builds back into the design.

You do not want to underestimate the importance of the ground circuit, as it might appear the designer of this amp has.

Combining the suggestion that this amp may have a generic stability problem (I can't see the 18watt.com discussions) with an observed multi-point grounding scheme, is pretty suspicious.

Quote from: LittlewyanBarry at Ampmaker believes that it is a characteristic of the amp

I think he's right.  I think it's marginally unstable.  Stability is not just the circuit, but includes other variables such as component spread and lead dress.

There is a lot in the Hi-Fi world that is rubbish, but radial single-point grounding is tried, tested and true.

The whole point about power supply interstage decoupling R's and C's is to remove impedances in common between stages; if you don't then stages can impress voltages on other stages via these common impedances.  When it's anti-phase you get some loss of gain, but when it's in-phase you get overshoot, transient and parasitic instability, and outright oscillation, at very low, mid, or supersonic frequencies.

{When one of the preamp bypass caps fails in a Fender Twin you get the very odd effect that the volume goes to zero at "1", then comes back up a little down to "0".  Figure that one out.  ;) }

The nodes are the main filter cap, and the sub-nodes are the individual bypass caps for each section.  Each of the sub-node grounds should be radially wired back to the main filter cap node, which can also be one of the bolts on the chassis.


Another point is that there is no specific HF rolloff in this amp, and something to try first might be a small C, say 100pF to 220pF to taste (full HT voltage rating) between the phase inverter anodes.  There may be an optimum value that minimises the frizz without making the amp muffled or dull.

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

Littlewyan

#20
With regards to the layout you're definitely right, I managed to find a drawing of the Marshall Layout and whereas they do have two ground lugs in the chassis, there is clearly one main ground lug that most of the ground connections go to.

I did think about a Cap on the PI tails, however I was finding it a bit difficult to find a cap with a voltage value that high, going to have a look on ebay again. 100pF and 200pF with 1KV rating on order :)

In the meantime I tried out the Paul Ruby Zener Fix on the amp, but unfortunately.............I only had one 8.1V and one 9.1V Zener Diode, whereas really I need two 13.1V (minimum) Zeners. So as you can imagine this added a bit of distortion (which didn't sound too bad tbh!) making it hard to hear if I'd fixed the issue or if the extra distortion has just masked it somewhat.

Going to get hold of the correct Zeners on Monday to try it again. With regards to the grounding however, basically I've just got to rewire the amp so it uses only one ground lug, is that right?

Also Toby I forgot to mention, I know you used my audio clip in Cool Edit to get that WaveForm, however, I did record that audio clip with my iPhone and additional buzzing could be caused by the microphone on the iPhone clipping as I did have it a bit close to the amp. Although I must say I was quite impressed by it!

phatt

Hi Littlewyan,
Sadly comparisons between utube clips won't help much as there are many unknowns. The voltage swing entering the amp alone can vary widely with different pu's.
That 1 meg grid resistor has been around since the dawn of Guitar Valve Amplifiers and ~50 years back it was indeed needed as pu output was rather limited so a very sensitive input was needed. Not so now, Heck even the most basic model guitar will likely have high output PU's.

What may not be obvious to some is that may give you more pu power but means the signal distorts earlier on the volume dial which gives the illusion of more power but may not equate to better tone.
Much like hotting up a Car,, you might improve acceleration times but find the top speed is still much the same as before but drive-ability is lost and fuel economy may suffer. :'(

Murphy's law;
You solve one problem only to find you have unwittingly created several others.  :duh

Assuming all else is ok then research the speaker you have, try out darker speakers.
/Use a guitar with lower output pu's.
/If you have to resort to conjunctive filters then you may have a hifi rated OT which can't be overcome short of replacement.
/Add series resistance after first stage which will help pull down hi-freq response as well as keep the signal from going silly.

My guess is that you have powerful pu's with increased bandwidth.
Add the best cables ever made. (more hifreq)
The OT is over engineered delivering even (more bandwidth).
The speaker has a fancy name (more Bwidth)
a lot of speakers are way too bright now days)
Yarda yarda,,,
All this stuff is cumulative and it adds up to crap tone. :'( :'( :'(
Unless I've been reading the wrong stuff wide bandwidth and distortion is a recipe for ugly fizz. So if you have fizz on your fuzz then a series resistor will help somewhat.
Assuming you have exhausted all the obvious stuff mentioned and no major issues with layout/grounding, I would Turn up the Amp and find the sweet spot now power down and measure the resistance from top of volume to centre wiper lug. That is you series R value.
If the Amp still has ice pick fizz then I'd look at OT's and the speaker.

Re Output Transformers;
I've messed around with conjunctive filters and many hot rod tweaks but the real break through for Valve Amps I've built came from lower speced OT's.
Others here are far better qualified to explain the hidden detail but in my experience I found that lowering the primary winding from the HiFi ideal will pull far more mojo than circuit tweaks.

HiFi  OT's (Over Engineered for guitar) are far more likely to give you the problem you noted.
That wonderful Man *Roly* has a short article I wrote some years back which might help you come to terms with why the icepick treble happens.

http://www.ozvalveamps.org/6dq6substitution.htm

My observations are down the page where I swapped OT's from 6k6 down to 2k6 (no other mods) and noted a dramatic improvement. <3)
Even with Roly's prodding I've still not fully understood the reasons but all I know is it made one hell of a difference.

In the end you will come to realize that many factors effect the final result all I've done is to try to find the easiest/cheapest way for a bloke with limited knowledge and $$$ to produce a good tone as I don't have another 3 life times to learn the teck as well as learn the guitar.  :lmao:

I'm just a guitar player who got so frustrated with all the BS in the industry I decided to go back and build stuff from the ground up,,, sadly that takes a lot of time but in the end it is liberating to have at least some understanding of Why.

I strongly urge you to find some cad sim programs and learn to use them as it will speed up the process of trouble shooting amp designs.
Sure scoping things is very helpful but you can make multiple mod tweaks in seconds on screen that would require hours to setup on real amps.

I'm still betting that a series resistor will help a lot as I'd say the signal is too big going into the PI.

Oh yes,, Roly has a good point about the grounding issue.
I've read more than once where some teck guys use a clip lead to find the best ground point.

Oh Oh!! yes go find Valve Wizard site,, some very well written grounding lead dressings there. How to wire up each stage with nice piks of layout.
Phil.

Littlewyan

Phil, quite an essay there, must admit I read just after I woke up and couldn't take it all in so had to reread it after breakfast :P.

With regards to the guitar pickups I did actually install Tonerider Rocksongs in my guitar not too long ago and immediately found that my Marshall distorted more at lower volumes. As for the grid stoppers those are on my list of things to try, but first i'll finish trying the Paul Ruby fix as i'm halfway through that.

I'm using a Marshall 2x12 cab for this and when I scoped the amp I used one of my Grandad's Hi Fi speakers and found the fizz to be a LOT worse. Only issue with trying a different OT is I'm not quite sure what I'm looking at yet with OTs, haven't got that far yet!

One thing I did find that sounded nice last night was the conjunctive filter in place with the Paul Ruby Fix. However I've disconnected the filter for now as I just want to try the Paul Ruby Fix without it. I'm getting the correct Zeners tomorrow so tomorrow night I'll be trying all of these mods.


Roly

Quote from: Littlewyana Cap on the PI tails

Actually it's a "Long-Tailed Pair", LTP, so the bottom part is the tail-in-common (a.k.a. a Differential or diff pair) - I don't think we call the anodes anything special - just "anodes".   :loco

Quote from: Littlewyanbasically I've just got to rewire the amp so it uses only one ground lug, is that right?

AND make sure that all the earths radiate from the single earth point, are not daisy-chained from stage to stage with sundry stage HT return paths sharing the same wire.

The circuit simulator of the moment is LTSpice, Spice compatible (mostly), and best of all, free!  I've used a number of sims over the years, some of which were like having teeth pulled (so eventually wrote my own), but Spice has become a de facto standard because it is pretty quick and easy to sketch up a circuit and give it a run.  Nothing beats the simulation using real components in a prototype, but Spice gets you close quickly and the prototype may need little or no variation.  There's also a major Yahoo support group and component library.  If you are into tinkering with amps it really is worthwhile getting and learning how to drive.  Kinda like Tone Stack Calc except that you can enter any circuit you like.

@phabb - repeats a couple of common observations about output transformers; that cheap mid-fi trannies sound better for guitar than expensive Hi-Fi ones; and that sometimes tonal improvements can be had by reducing the anode load below the "optimum" match.  The first is fairly easy to explain by the distortion products produced by cheaper iron laminations, bad for Hi-Fi but cream for guitarists.  The second observation is a bit harder to explain, but I'm sure someone can if they think about it long enough - but until somebody does it's an idea that is worth tying to find out if the results are agreeable to you.

Quote from: Littlewyanwhen I scoped the amp I used one of my Grandad's Hi Fi speakers and found the fizz to be a LOT worse.

Any Hi-Fi speaker system worthy of the name (and a few that aren't) will go up to the limit of hearing, 10+kHz, while a guitar speaker is typically a 12-inch that rolls off at maybe 5kHz, and a guitar doesn't have a lot of content even at that frequency anyway.

The rule of thumb is that a speaker rolls off at around 80 times it's free air resonance.  With a typical resonance around 50Hz that's 50 * 80 = 4000Hz.  Get a sig gen or download a multi-frequency test file and sweep your 2x12 - you may be astonished at just how "mid" mid-Fi can be.  Even better if you can A/B with grandpa's speaker.

If you're shopping for output transformers have a read of this, then check out what is locally available.

HTH

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

Littlewyan

Thanks for that Roly I'll give that a read as I do plan on building another amp (Like you said addictive as crack) next year from SCRATCH! So as you can imagine I've been looking at plenty of different circuits working out what does what.

Now an update on the amp I think the Paul Ruby Fix helped lower the buzz although its still there, I then added a 470K Grid Stopper and I think this improved the amp sound overall. Unfortunately its difficult as once I've finished testing the amp I have to wait for the EL84s to cool down so I can take them out otherwise I can't turn it upside down to work on it, by which time I've almost forgotten what it sounded like! Anyway the amp sounds better to my ears, probably because the 470K should stop a lot of the higher frequencies from getting through so there is less going on if you know what I mean. Once the capacitors that I've ordered arrive then I will be able to cancel out the higher frequencies in the power amp even more so hopefully this will improve the sound further.

Now one thing that puzzles me is looking at Marshall's 1987xl (one I own myself) they have the 470K Grid Stopper in the Pre Amp bypassed by a 470pF capacitor. Now according to LTSpice this basically allows everything through that the 470K Grid Stopper is put in to stop. Either way it shouldn't really need bypassing, sure the 470K does cut into the 5Khz range a bit but in my case the amp sounded better. Any ideas?

Roly

470k is quite a high value for a grid stopper, somewhere between 1-10k is more common.


Quote from: Littlewyan470K Grid Stopper in the Pre Amp bypassed by a 470pF capacitor

If we stick a 470k in series with the signal path, then parallel it with a 470pF cap we have created a high-pass network.

The hinge frequency will be when the cap has the same reactance as the resistor value.

So at what frequency does a 470pF cap have a reactance of 470k?

Xc = 1/2 Pi f C
  where:
    Xc = the reactance in ohms
    f = the frequency in hertz
    C = the capacity in farads

transpose

f = 1/2 Pi Xc C

substitute

1/(2 * Pi * 470*10^3 * 470*10^-12) = 720Hz (roughly the F an octave and a half above Middle-C)

So starting around 720Hz the response will rise at 6dB/octave, fairly serious treble boost.

Quote from: LittlewyanAny ideas?

Yeah, do what sounds best to your ears.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.

Littlewyan

#26
This one is just before the 22nf cap going to the PI, would it not be called a Grid Stopper there?

Also i put the resistor there to stop high frequencies above 5Khz being as present as they were, so with the 470pF it wouldnt do that. So i'm just wondering why you'd bother having it there in the first place? As it wouldnt be doing it's job anymore surely

Roly

Quote from: LittlewyanThis one is just before the 22nf cap going to the PI, would it not be called a Grid Stopper there?

Not really.  A "grid stopper" is a smallish (1-10k) resistance in series with the control grid and normally placed very close to the grid connection, i.e. on the socket VHF-style.  It's primary function is to prevent high frequency/ultrasonic/RF oscillations, hence oscillation "stopper".  On the first stage the reason for the classic 68k in series is as an RF stopper to prevent radio station breakthrough.  These days a ferrite bead or ring on the input lead behind the socket should do the same job.

A secondary function in guitar amps (where stages are often overdriven) is to limit grid current when the grid is driven positive.  If this is not limited then a charge will rapidly accumulate on the input coupling cap and shift the bias point of the stage until it leaks away through the grid resistor.  Because of the direction of the Grid-Cathode "diode" this will result in the stage being over biased towards cutoff.  Generally this is more significant in the output stage than in earlier stages.

How much this series resistor alone attenuates higher frequencies depends on the effective capacitance to ground of the following stage, and with a triode this will mainly be due to Miller capacitance, but it will be pretty small, depending on layout, and therefore a bit hard to predict.

In this case, given its position before the coupling cap can have no influence on grid current, its role here seems to be just as a treble boost.

HTH

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

Littlewyan

Well definitely improved the sound anyway, the capacitors i ordered still haven't arrived yet so can't do anything with the PI anodes yet. Was thinking about getting extra distortion out of this amp by cascading the first preamp valve but due to the layout it wouldnt be an easy task.

Also need to lower the bass a bit, always got the tone control on 10 and on input 1 (single triode) its still too bassy! Going to try lowering the .1uF in the PI to .01uF and maybe the .022uF to .01uF and then may change the .0005uF on the Tone Control to 10nF. I think the first two cap changes will just tighten up the bass a bit and the tone control cap change will make the sweet spot a bit more central on the knob

Roly

If you don't already have Duncan's Tone Stack Calculator then download it and have a play.
If you say theory and practice don't agree you haven't applied enough theory.