Fine tuning/Blueprinting your amp-

['67_Plexi]

putting an R in a ground line sounds like a ground loop is being broken. Maybe the ground point needs to go somewhere else?

It's not a ground loop that is the potential problem, it's oscillation. They are totally different. Ground loops occur due to potential differences between ground points. Ground current flows and cause a potential difference due to the various resistances between the ground points (ohms law). Technically using a resistor in some situations could make an actual ground loop problem worse if there was enough ground current. 10 ohms is high enough not to be a problem in most situations.

[ayan]

Hi Gil,

Firstly I use a steel chassis and a multi-star grounding arrangement.
The amp has dual masters that feed in to a cathode follower via a relay arrangement for the efx loop.
It seems that input impedance of the cathode follower plays a big part in any oscillation issue.

{snip}

Very interesting thing! Just so that I can get my arms around this:

1. Do you have a blocking cap at the input of the CF?
2. If so, have you tried hanging a large resistor to ground from the input terminal of the cap, while really grounding (no 10 ohm resistors) the CF "-" side?

I am 100% babbling here, but I think the very large input impedance of the CF, in the middle of all those leads carryng and radiating lots of signal, is what's doing you in -- I believe I am restating essentially what you already said. If you lowered the input impedance of the CF stage to the maximum level you need it to be without loading your preamp, you would obviously decrease susceptibility to noise and oscillations. If you have 1 Meg masters, typically dialed at maybe 200K to ground or so, I believe hanging a 2.2 Meg resistor (upwind from the input cap) might also kill the problem.

I completely understand that the 10 ohm resistors solve the problem, I am just not sure how the do so!

Cheers,

Gil

[Bob-I]

putting an R in a ground line sounds like a ground loop is being broken. Maybe the ground point needs to go somewhere else?

It's not a ground loop that is the potential problem, it's oscillation. They are totally different. Ground loops occur due to potential differences between ground points. Ground current flows and cause a potential difference due to the various resistances between the ground points (ohms law). Technically using a resistor in some situations could make an actual ground loop problem worse if there was enough ground current. 10 ohms is high enough not to be a problem in most situations.

That's what I was thinking, but what do I know.

I've never really gotten a handle on solving ground loop issues, mostly I just happen to fix them then stick to that scheme.

Many times I've solved a ground loop hum by REMOVING a star ground and seperating the grounds. For example on a 5F6-A I fixed for a friend, he'd connected all controls and the cathodes to one star ground. I split them with cathodes near the PS and controls and input near the input. The hum went away. In this example there was no measurable resistance between points.

Now if I could just solve my Dumble in a Princeton head hum as easily

Great discussion!!

[Luthierwnc]

One handy ground loop chaser I've found is to turn the amp on with no signal. Clip the ground lead of an oscilloscope to your main ground and dial in a very concise line. Then touch the other grounds with the probe. Points with potential will blur the line.

I doubt that is an official way of doing it but it helps.

Skip

[Dai H.]

Some Marshall designs employed a 10 ohm resistor to ground on a similar efx loop circuit, again definately making me feel the problem is with the input to that circuit.

I'm not really sure what is happening in regards to the oscillation, but as far as my logic the bit about a 10 ohm in an efx. loop made me think that the 10 ohm was used to lift the ground a bit to break a loop that would happen when the signal was connected to another AC-powered device via the ground loop (ground loop due to another ground connection through the AC safety ground connection). So that was my thinking there but can't quite picture how things are connected since I'm not sure which *umble amp/schematic is being referred to. Seems to be quite common to use a low value R to do this (far cheaper than using a transformer--which can also break loops--so a huge cost advantage relative to that). For example:

http://sound.westhost.com/earthing.htm

This is in my reiss. Fender Reverb unit (15ohms? plus cap plus diodes). There is also a 10 ohm lifting the entire cct. from chassis ground in my Marshall SE100 Spk. Emulator as well as Power Brake(grd. loop could happen as the SE100 touched other chassis). Sometimes the value can be slightly different. The DOD A/B box also has one (15ohms?). So I thought perhaps part of the circuit's ground was in the right place but another wasn't inducing the problem but putting the 10 ohm fixed it (but that may be completely wrong).

[Tonegeek]

This requires very careful lead dress, but is do-able, especially if you connect the masters to ground via a 10 ohm resistor.

I had a simiar issue. If I turned down the clean master all the way, I got what sounded like a ground loop hum. I think this happended with some other pots at zero as well, but not sure since it was last year when I fixed it. I put a small value (390 ohms) resistor in series to the input of the CF (of the built in loop) and the problem went completely away. I know of one other forumite who applied the same fix. In my case the master being right before the loop was grounding the .05 cap when it was all the way down. I mention that fact, although it may have nothing to do with the hum. I don't think it was a ground loop, although it sounded like one.

[stelligan]

Henry, as far as we know, there is always a 100K there. If Dave's amp had no resistor after the HRM, then it was not "accurate". 100K seems to work well fwiw.

Stelligan's HRM sounded great with no resistor there (using Humbuckers) until he started playing a Strat through that amp - then putting 150K there made all the difference

Hey guys, would it be possible or advisable to put a 150K trimmer there that I could access from the front? Or a switching network that could go from 0-180k?

Thanks in advance,
Dave

[dogears]

I'd use the 100K. Btw, it is in series with the OD master. So, if your OD master measure low, you can raise the series resistor up. Shooting for the 1.1M total is good IMO. Spec a 1M pot and use a 100K resistor. That should work for you.

Henry, as far as we know, there is always a 100K there. If Dave's amp had no resistor after the HRM, then it was not "accurate". 100K seems to work well fwiw.

Stelligan's HRM sounded great with no resistor there (using Humbuckers) until he started playing a Strat through that amp - then putting 150K there made all the difference

Hey guys, would it be possible or advisable to put a 150K trimmer there that I could access from the front? Or a switching network that could go from 0-180k?

Thanks in advance,
Dave

['67_Plexi]

Hi Gil,

Firstly I use a steel chassis and a multi-star grounding arrangement.
The amp has dual masters that feed in to a cathode follower via a relay arrangement for the efx loop.
It seems that input impedance of the cathode follower plays a big part in any oscillation issue.

{snip}

Very interesting thing! Just so that I can get my arms around this:

1. Do you have a blocking cap at the input of the CF?
2. If so, have you tried hanging a large resistor to ground from the input terminal of the cap, while really grounding (no 10 ohm resistors) the CF "-" side?

I am 100% babbling here, but I think the very large input impedance of the CF, in the middle of all those leads carryng and radiating lots of signal, is what's doing you in --
Cheers,

Gil

Absolutely Gil, you are thinking along exactly the same lines as me. There is a blocking cap at the input.....and I think that plays in to this to...... !!

If I used sheilded cable for the masters like Dumble did, the problem is eliminated, but you know...I like how it sounds without them clogging the high end up, plus there's something more Rock 'n' Roll about an amp operasting on the verge of some instability....LOL !!!
Another thing I noticed was the larger the conductor (lower resistance) cable going from the masters to ground, the more susceptible it is to oscillation. The lowest value resistance that eliminates the problem is approx 2 ohms. 10 ohms is 100% safe in all amps.

[ayan]

Absolutely Gil, you are thinking along exactly the same lines as me. There is a blocking cap at the input.....and I think that plays in to this to...... !!

If I used sheilded cable for the masters like Dumble did, the problem is eliminated, but you know...I like how it sounds without them clogging the high end up, plus there's something more Rock 'n' Roll about an amp operasting on the verge of some instability....LOL !!!
Another thing I noticed was the larger the conductor (lower resistance) cable going from the masters to ground, the more susceptible it is to oscillation. The lowest value resistance that eliminates the problem is approx 2 ohms. 10 ohms is 100% safe in all amps.

We're in agreement. And come to think of it, probably the best sounding amp I ever built was a Dumble inside a Fender DR. I didn't use any shielded cable inside, and because of the existing holes on the front panel, my layout was such that a few wires had to go back and forth a bit. I ended up with a very unstable amp. I struggled to get it under control and wrote a few ideas here (almost 10 years ago, yikes! I am getting old!):

http://www.ayanmusic.com/leaddress.htm

Anyway, once I got the amp to behave, it was a thing of beauty. It would take off on me on every note, at any volume, and without much gain. In contrast, all of my Dumble clone builds have been extremely stable amps in terms of being 100% free of oscillations. To get them to sing "clean," they need to be loud.

Cheers,

Gil

[HiGain]

I haven't experimented with the 90s topology too much yet, so this just applies to 80s...

It seems that I've had the greatest success when I focus on a specific frequency I'm targeting. At first altering component values doesn't seem to have too great an effect. However, as you start to hone in on your target sound, smaller and smaller alterations in values become more and more noticeable. Parenthetically, I think this is one of the reasons for the disagreement among internet participants regarding whether you can hear small changes in values or different cap types, etc.

Anyway, I'm guessing each amp must be tuned depending on the transformers and types of components (as many have said). When you get real close to a target sound, changing a single capacitor type can appear to have a huge difference. Something about those "fragile harmonics!!!"

Last, layout has a *HUGE* effect on the apparent frequency spectrum of the amp, especially the midrange. So much is this the case that I wonder if HAD used a second board under the main component board in order to secure the positioning of wires. You never know, maybe he intentionally routed wires in specific ways as a method for TUNING the resonant frequency, i.e., the apparent midrange, of the amp. Hint: Think oscillator.

Warning: this is all from my experience. My electronic know-how is fairly limited!

Jake

[Funkalicousgroove]

There is in-deed a backing board under every circuit board in the amps I've seen, he also uses zip ties to keep things in place.

Lead dress is KEY for the amp to sound right IMHO-

[ToneMerc]

Bump bump

[10thTx]

OK, two "tweaks/mods" that I've done to a D-inspired amp (not a clone) that might be useful to a clone (?) are:

1) OD tone stack which I find very useful to me especially with the
single coil vs. humbucker pickups.

2) Enhance cap across LTPI entrance plate. I have found this to be very
useful on numerous of my builds. Seems to smooth the tone, eliminate
some fizzy/harshness and increase blooming & sustain on my amps. I
can not tell any loss of highs when I keep this cap a low enough value.

With respect, 10thtx

[andyhardy]

Here are some tips from Henry in 2007


Ok say your amp project is completed and now the final tweaking begins.
Here's what I do, but I'm open to better suggestions
Before tweaking make sure you are using an appropriate guitar and speaker combination. i.e. don't tweak with a Jensen if you amp will have a Celeston.
1. Do the voltage changes first - i.e. dropping string adjustments (don't forget to re-bias after each change). Lower = smoother
2. Coupling caps next - if you make a change and like it but think you lost something try changing the bypass value accordingly on the stage ahead of the coupler. i.e. if you gained highs but lost lows increse the bypass cap
3. Bypass caps next - in addition to tone these caps affect gain too, so you can add or take away "character" as the tone changes. i.e. changing to a 10uf (from a 05uf) will add bass and gain, if you like the gain but think it's too bassy try reducing the coupling cap to the next stage.
4. RP and RK values - I've learned not to mess with the ratio of these much, changing the ratio more than a few % hurts the "linearity" for lack of a better word. 220/3k3 180/2k7 150/2k2 120/1k8 100/1k5 pairs can be tried in any location. Sometimes replacing both pairs in the OD section is just what an amps needs i.e. V2 180/2k7 and 120/1k8 changes to 220/3k3 and 150/2k2 or vice versa.
5. Bleeders and filters - you can place a cap and resistor in series to gound anywhere in the signal path to bleed off highs or do a resistor/cap parallel pair to increase highs i.e. a .001 - 330k series from OD2 CC output to ground gets rid of "OD fizz" or i.e.2 lowering the 500p in the 220k/500p filter on the input to CL2 for less mids.
6.PI and output tweaks - I saved the best for last - you can really make or break an amp in this section. For max headroom try and use the smallest PI entrance cap you can (a lot of Fenders used .001!) and larger output tube grid resistors (I like 5k1s better than 3k3s) Screens on 6L6s sound better to me at 500 rather than 1k. If you have a trim pot to balance the plates make sure you can adjust them to within 5 volts of each other if not change one RP until you can, start at 5 volt difference and then very slowly increase as you strum a chord through the amp - you are trying to get maximum overtones - you'll know it when you find it.
.The NFB area is especially fun - max the presence control and cliplead in various values of caps and the FB feed resistor to taylor the gain and NFB cut off frequency.

Now the dilemma - How do you know whether a tweak is good or not? Short of testing it in a live situation I'm not sure. I have learned that in my case the sound on the bench is not to be trusted until I hear it live.