| Author |
Topic: A bit more, re: Odd amp input circuit mod question... |
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 9 Oct 2011 11:39 pm |
|
My Matchless Lightning has only a high level input, pushing the first gain stage pretty hard, resulting in not a lot of clean headroom...I am playing a pedal steel now, with a super high output pickup, and needed more headroom....and the amp also had a lot of hum, probably due to the hot single coil and high gain...the steel worked fine (good headroom, no hum) on the low level input of my HC-30 head (which has both low and high level inputs), essentially a double power Lightning (but set at half power), but not on the Lightning. So it wasn't something outside the amp....
The amp has an input divider circuit with a 1 meg to ground and a 68K to the input tube...I spoke with Phil Jamison at Matchless and he suggested two things, neither of which worked very well...
1. change the 68K to 68 ohms. Strangely, this resulted in no output at all!
2. change the 1 meg to 68K - this cut and cleaned up the input just a tiny bit...should have made it just like the low level input on the HC-30, but didn't..
I ended up replacing the 1 meg to ground with a 33K to ground....so there is now a divider from the input with a 33K to ground and a 68K to the first stage input tube - this worked great, lots more headroom, hum imperceptible except when all the way up on volume and master, and still a lot of overdrive available from the volume knob....
Anyone know why this works? And why Phil's changes didn't? Or what I should have done instead?
thanks!
Last edited by Steve Lipsey on 15 Oct 2011 1:22 pm; edited 1 time in total |
|
|
 |
Tim Marcus
From:
San Francisco, CA
|
Posted 10 Oct 2011 6:40 am
|
|
68k and 1m are fender values, so they should be fine. That's pretty typical, and gives the tube a 1M input impedance which should be plenty high for a steel guitar.
Chances are something else in the circuit is tweaked to make a guitar break up. Remember: 99.9% of tube amps are tweaked for guitar and yet the steel has a very different dynamic to it with different harmonics. Sometimes this results in an extremely low headroom amp when a steel is plugged in to a guitar amp.
_________________
Milkmansound.com |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 10 Oct 2011 10:15 am
|
|
Tim-
The lightning has an extra gain stage that takes it a bit higher than some other Matchless amps (e.g., the Spitfire is similar, but one less gain stage) - it is designed to give you breakup that is very controllable from your guitar volume knob - but it is a lot to ask from only 15 watts (even 15 Matchless watts, which seem louder than many others) to keep up in a full band unmic'ed situation....so I wanted to take the gain down.
Also, there was a HUGE hum, which went away with my mod...
Everyone-
The 68k/1meg divider is totally standard everywhere....and on two input amps the standard switching input jacks switch that to 68K/68K when something is plugged into the high level input...but converting my one-input amp to that didn't work... |
|
|
|
Tim Marcus
From:
San Francisco, CA
|
Posted 10 Oct 2011 6:57 pm
|
|
oh - you can always bypass that extra gain stage! That will most likely bring a ton of headroom back to the amp, and also reduce noise from the stages before.
When it comes to tubes and pedal steel - I usually defer to Fender designs. Anything even remotely marshall will simply not have the headroom.
That said, Sunn and Ampeg designs are also good - just less common for some reason.
_________________
Milkmansound.com |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 10 Oct 2011 8:10 pm
|
|
Tim-
Didn't know about your amp! Do you post on the Steel Guitar Forum? I switched to Matchless just before I decided to go to pedal steel....I wasn't sure that the Vox-based amp would work out, but the tone is pretty great for any instrument and kind of music, so I'm pretty happy....but always looking for the next better thing...the Lightning was too high-gain until the mod, but the HC-30 head worked just fine....the interactive tone controls provide a lot of latitude, and the Class A tone is pretty nice, the chime actually works well for steel
But as I said, I'll be checking out your site more now! |
|
|
|
Tim Marcus
From:
San Francisco, CA
|
Posted 10 Oct 2011 8:40 pm
|
|
Eric Heywood used a Vox AC50 (I think - maybe a 30) on the first Sun Volt record, so its definitely been done before, but its not an ideal circuit for steel, in my opinion.
I emailed you back 
_________________
Milkmansound.com |
|
|
|
Brad Sarno
From:
St. Louis, MO USA
|
Posted 11 Oct 2011 7:30 am
|
|
Try removing the cathode bypass capacitor on the first (input) tube.
You'll reduce the gain and increase the clean input headroom. One of the best and simplest steel guitar mods you can do on a tube guitar amp.
Brad |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 11 Oct 2011 9:42 am
|
|
Brad- Which cap is that? The 25mfd?
Here is the schematic from an online archive, probably still valid - except my Lightning is the Reverb model, so it only has input 2, not the low input (input 1?), and changing the resistors (1 meg ->68K) to match the low input didn't do the job, as described above....
 |
|
|
|
J Fletcher
From:
London,Ont,Canada
|
Posted 11 Oct 2011 10:09 am
|
|
Hi Steve
Most likely your amp has the #1 input. Changing the 68k resistor to 33k will have little effect on the input signal strength, as the two resistors,1 meg and 68k, are not set up as a voltage divider. Changing the 1meg to a low value like 33k doesn't seem like a good idea, it lowers the input impedance too much. Brad's idea is a good one, remove the 25u/25volt cap across the 1k5 cathode resistor...Jerry |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 11 Oct 2011 10:52 am
|
|
Jerry-
Can my change of the 1 meg to 33K actually hurt anything? Because it sounds fine now...(actually I just added a 33K across the 1 meg)....
Someday I will definitely open it up again and try some of this stuff....but right now it sounds good, input is definitely lower with the added resistor and hum is gone, and amp sounds good... |
|
|
|
J Fletcher
From:
London,Ont,Canada
|
Posted 11 Oct 2011 11:05 am
|
|
| Can't hurt anything, if it sounds good, it is good! 33k is a pretty low input impedance though. 1 meg is typical. I'm wondering if the original 1 meg resistor is open, or not connected properly. |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 11 Oct 2011 1:07 pm
|
|
Nope, it measured out at 1 meg and connection was solid...
What is the impact of a low input impedance? FYI, the pickup in the pedal steel is a Wallace Trutone single coil, measuring out at 18.5K ohms... |
|
|
|
Brad Sarno
From:
St. Louis, MO USA
|
Posted 11 Oct 2011 1:27 pm
|
|
Modifying the input resistor network is one way to pad things down, but really, you can leave that stuff alone. For pedal steel guitar in particular, I really like the mod of simply removing that first 25uF/25v cap. Just remove or snip it.
It's actually a fairly complex thing that happens because without that cap, the bias voltage is no longer held perfectly stable. But that's a good thing here because this condition sets up a "local negative feedback" loop that significantly increases your clean input headroom, lowers distortion, and also lowers noise. It's a cleaner, more steel-friendly sound, and perfectly compensates for the high voltage that comes from a steel pickup. It simultaneously provides the clean headroom AND pads the signal down about 6dB so the rest of the preamp circuit can operate as normal and the whole thing will be much more steel guitar friendly.
On real Fender amps, I like to ditch the bright cap and use the bright switch as a toggle between guitar/steel. The steel position merely disengages that first stage cathode bypass cap's connection to ground.
Matchless does that cool thing with the paralleled triodes. I really like that about the handful of Matchless amps that do that. It's part of the Matchless mojo. I use that same parallel thing on the SMS Classic's 3rd (final) gain stage.
Brad |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 11 Oct 2011 2:54 pm
|
|
Nothing new in this post....just a huge "Thank you" to all those who took the time to help me out here.
-Steve |
|
|
|
Jay Fagerlie
From:
Lotus, California, USA
|
Posted 11 Oct 2011 3:15 pm
|
|
Hey Steve-
OOPS! I guess that input is grounded when no jack is plugged in....
Try this- remove the cap that connects from pins 3/8 to ground on that first stage-that will decrease the gain by quite a bit. The 2.5 uF @25V electrolytic.
Try replacing that first tube with a U7 or T7 instead of a X7
Replace the phase inverter with a T7.
Let me know how it goes
You can PM me for more help if you need it
Jay |
|
|
|
Ken Fox
From:
Nashville GA USA
|
Posted 11 Oct 2011 4:01 pm
|
|
The effect that I see on a lower input impedance is more of the signal is dropped across the internal impedance of the source. Typically source signal to input is not matched (1 to 1 impedance ratio) but in fact bridged (in Hi=Z inputs. A bridged source to input is typically 10 times or more lower than the input impedance, hence the typical the 1 meg input. That allows maximun voltage transfer to teh input (n\not max power transfer, as in a matched system) Don't be confused by the DC resistace (R) of a pickup, it is far lower than its true impedance (Z). Good example of that on this spec sheet:
My own tube reverb unit uses a variable input impedance of 27K to 1Meg + 27K. The lower the input impedance the more the high end seems to be affected (slightly rolled off). We also use a 12AU7 in the input circuit (gain of 20). Brad's Black box has a very similar variable input impedance control.
Last edited by Ken Fox on 11 Oct 2011 6:05 pm; edited 1 time in total |
|
|
|
Ken Fox
From:
Nashville GA USA
|
Posted 11 Oct 2011 4:04 pm
|
|
| Also notice the typical two input amp with two 68K resistors present a voltage divider to the source signal of 34K when the #1 jack one is used (the two 68K are in parallel) |
|
|
|
Ken Fox
From:
Nashville GA USA
|
Posted 11 Oct 2011 4:05 pm
|
|
Some good info from the net:
Grid Resistors - Why Are They Used?
General
If you look at the schematic of a typical guitar amplifier, you will notice that there is a resistor in series with the grid of the first tube, usually around 68K or so, and there is also a resistor in series with the grid of the power tubes, usually 1.5K or 5.6K, and you may occasionally see very large value resistors, such as 470K or greater, in series with tube grids in high-gain preamps. Some amplifiers have no such grid resistors, and occasionally people will recommend removing them to supposedly "increase the gain" of that stage. What is the purpose of these resistors, and should you remove them?
Reasons for the resistors
These resistors, which are commonly called "grid stoppers", are not put on the control grid of the tube for signal level attenuation purposes; rather, they act as a very high frequency low-pass filter in conjunction with the input capacitance of the triode (which is the sum of the grid-to-cathode capacitance and the Miller capacitance, and can get as high as 100pF or more). In the normal operating mode of a vacuum tube, the grid is biased negatively with respect to the cathode. Because of this, there is no current flow into the grid element, and it looks like a very high impedance circuit node. This means that there can be little or no midband attenuation of the input signal, because the voltage divider formed by the series resistor and the high input impedance of the tube is very small. For all practical purposes, the attenuation is negligible at midband, so there is no "increase in gain" by removing these resistors. Attenuation only occurs at the higher frequencies, above the frequency breakpoint caused by the series resistance and the input capacitance.
The grid resistor accomplishes the following things:
It helps prevent high frequency parasitic oscillation in the tube itself
It helps prevent radio frequencies from getting into the input stage, where they can be rectified and lowpass filtered (AM detection) and become audible at the amplifier output
It can limit grid current when the tube is driven into the positive grid region, which helps in preventing "blocking" distortion
Where to put the resistors
In order to take advantage of the parasitic suppression benefits of these grid resistors, they must be placed as close as possible to the socket pin of the tube, preferably soldered directly to the pin with a very short lead. The resistor should be placed after the grid-to-ground resistor (usually 1 Meg or so), to avoid attenuation and to keep the signal path short. If the resistor is connected in series with the input jack and before the 1 Meg grid resistor, there is a small loss of the input signal, although, in most cases the attenuation is not enough to be concerned with (0.94 times for a 68K grid stopper and a 1 Meg grid resistor), and in amplifiers with a high and low level input, the grid stoppers also serve as attenuators. When designing an amplifier, it is better to use separate resistors for input attenuation purposes in order to be able to locate the grid stoppers as close to the input grid pins as possible, rather than mounting them on the input jacks.
How large should they be?
The grid resistor value typically varies from as low as 1.5K to as high as 470K.
Most output stages use relatively small grid resistors, such as the 1.5K seen on the grids of 6L6 tubes in most Fenders, and the 5.6K seen on the grids of EL34 tubes in most Marshalls. In general, the grid resistor at the grid of the power tubes can be as high as 56K to 100K before any noticeable loss of high frequencies occurs. Higher values can help in reducing "blocking" distortion as noted above, and can also take some of the "edge" off of an overly brittle sounding output stage. If the resistor value is made too low, it may not be enough to prevent parasitic oscillations, and the amplifier may exhibit instability in the higher frequency range. This may or may not be audible. Symptoms of oscillations include: high-pitched "squeal", glowing plates at "safe" bias currents, harsh treble response, lack of power, undesirable overtones, and unusual frequency response which makes the amp sound funny. Note that power tubes have a specification for maximum resistance that can be in series with the grid terminal before the tube becomes unstable due to grid current. The maximum allowable resistance is larger in cathode biased circuits than it is in fixed bias circuits because the cathode bias provides some "self-limiting" protection against bias runaway. The total resistance is the sum of the series grid resistor and the bias feed or "grid-to-ground" resistances, so if the max spec is 300K, for example, and there is a 220K bias feed resistor, the largest grid resistance that can safely be used is 80K. Of course, in practice, tubes should not be run that close to the edge of their specifications, to insure reliability.
The grid resistor on the preamp stages typically ranges from 0 to 68K, although very large values, such as 470K, are sometimes used in high-gain preamps to shape the frequency response and prevent "blocking" distortion in the preamp section under heavy overdrive conditions. The Miller capacitance of a typical 12AX7 is around 151pF, so the upper frequency response -3dB cutoff point of a stage using a 68K grid resistor is around 15.5kHz. The frequency response drops to around 2.2kHz if a 470k grid resistor is used. This "free" response rolloff can be used to tame the "buzziness" of high-gain preamp stages without having to add additional rolloff capacitors. Perhaps the most important grid resistor is the one that goes to the grid of the very first stage, right after the input jack. This resistor is the one that prevents oscillations and pickup of radio stations and other noise due to long or poorly-shielded cables. It is not usually a good idea to eliminate this resistor. Ideally, it should be soldered directly to the grid pins of the socket, with very short leads. |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 11 Oct 2011 9:48 pm
|
|
Ah, thanks, Ken
...so the 68K grid input resistor seems like a good thing to keep.
...and I guess the impedance mismatch (high to lower) from cutting the 1 meg-to-ground resistor to 33K might change the frequency response in some fashion (any ideas on exactly what this might be?)....I wasn't really listening for that when I had it on the bench, was looking for headroom and mainly, getting rid of the huge hum....
It sounded fine at the gig (nicely balanced frequency response, as far as I could tell, without being able to compare "before-and-after"), and still has lots of drive when the gain volume knob is turned up, even with lower output sources like guitars....so I'm tempted to leave it alone...
Matchless emphasizes the "no negative feedback" aspect of the amps....I wonder what clipping the cathode cap (thus introducing a negative feedback loop) would have relative to this aspect?
This thread is really helping me understand a lot more of what is going on in this part of the amp! |
|
|
|
J Fletcher
From:
London,Ont,Canada
|
Posted 12 Oct 2011 5:18 am
|
|
| I think the "no negative feedback" refers to the output section. There isn't a feedback loop from the speaker to the driver tube, as far as I know. After all, the cathode follower that preceeds the tone control has lots of neg feedback...Jerry |
|
|
|
Michael Brebes
From:
Northridge CA
|
Posted 12 Oct 2011 7:51 am
|
|
If you don't want to mod the amp, then make a little box with two 1/4 mono jacks. 68K resistor from hot to hot. 68K resistor from hot to ground on the second connector (output). Plug steel into input and have a short little cable from out to input of amp. There's your Input 2 with the lower gain.
_________________
Michael Brebes
Instrument/amp/ pickup repair
MSA D10 Classic/Rickenbacher B6/
Dickerson MOTS/Dobro D32 Hawaiian/
Goldtone Paul Beard Reso
Mesa Boogie Studio Pre/Hafler 3000
RP1/MPX100 |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 12 Oct 2011 10:22 am
|
|
Michael-
good idea in concept - but there is already an internal 68K to the input tube in the amp, this would be another in series....and I tried 68K to ground and didn't get enough change, so I used 33K (just tacked on in parallel with the 1 meg that was already there), which worked perfectly...
So an even simpler solution would be a 33K to ground inside the plug in the cord from the steel! (or use a two-jack box, as you suggested, for flexibility later in changing cords).
Lots of options to try here... |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 12 Oct 2011 10:24 am Last Remaining Question Is....
|
|
| anybody have any idea how my current solution - changing the 1 meg to ground into a 33K to ground, leaving the 68K to the input tube alone - would change the frequency response of the amp? |
|
|
|
J Fletcher
From:
London,Ont,Canada
|
Posted 12 Oct 2011 11:09 am
|
|
| I would say you are going to have some noticeable high frequency loss with a 33k resistor. Your pickup will be loaded down, unless there is an active effect, delay, distortion, etc, between your steel and the amp...Jerry |
|
|
|
Steve Lipsey
From:
Portland, Oregon, USA
|
Posted 12 Oct 2011 11:15 am
|
|
| Ah - yes, I do have a pedal board in between...(Chi Wah Wah, Tuner, Keeley Comp, Hotcake OD)....and my old passive Ernie Ball volume pedal works OK at the end of that line, without killing the tone (I use it for practice at home, leave the Hilton in the seat), so I expect that the steel freq loss will be minimized also..the amp doesn't see the steel directly, ever... |
|
|
|
