L-Pad and fixed R bypassing rheostat question

[dai h.]

I've been googling info on L-Pads because I'm going to do a simple experiment essentially just sticking an L-Pad at the output of my Marshall Power Brake to lower the volume further and to see how good (or not good) it sounds that way. One of the bits of info/diagrams I encountered is this (simple L-Pad design for low wattage amp) :

http://blueguitar.org/new/schem/misc_mod/power_sk.jpg

...and for the shunt R, instead of a shunt R only, there is a rheostat (which of course here is to make the output adjustable). My question, is why is there both a shunt R and a rheostat, instead of just a 4 ohm rheostat(4.7 and 25ohms in parallel is 3.96ohms)? Is there some kind of advantage in using a shunt R together with a rheostat? In terms of current handling, my understanding is that a lower resistance value for a given wattage rheostat can handle more current (from looking at the ohmite rheostat datasheet plus if you look at actual rheostats you can see that very low R units have quite a thick ribbon-like physical shape for the resistive element versus the higher R units which have more of a wire), so from that perspective use of a rheostat only would seem to be more advantageous. Perhaps I'm overthinking this and the designer just used what he/she had? Or maybe it's an economical design (meant to be) since it uses a relatively inexpensive 3 watt rheostat?

[cbarrow7625]

You could do that but the power rating of the rheostat would have to be much higher (and larger). In this case, the 4.7 Ohm resistor disappates the most heat from waste power & the 25 Ohm rheostat can be small.

For your scenario, the 4 Ohm rheostat would need to be about a 10-13Watt device. Economics definitely look like a factor to me. I think a 3 watt, wirewound pot is going to be far less expensive than a 10-13 watt rheostat.

Other than that, there is no reason you couldn't do it. Your version would put less resistance in series with the driver and probably sound better.

Chris

[dai h.]

thanks for your response Chris.

I guess that in searching for wirewound rheostats here and finding 25W-50W units not particularly expensive (around $15 or so--seems to be a single dominant manf. here in Japan as far as parts vendors that sell to general consumers)--that was coloring my perspective (just looked at an older Mouser catalog as well as their site for current prices and in fact they are more expensive). I suppose I should look at the diagram as an example to study of utilizing a lower wattage rheostat to get the job done.

Your version would put less resistance in series with the driver and probably sound better.

would this be because the increased R in the path would lead to less L in the signal output path (flatter freq. response, maybe less lively, less responsive)?

(to give a little background--skip if boring! ) I've been wanting to achive a decent tone at very low volume and wondering if it could be done, and in checking out sound clips from a Swedish guy "Ralle"(of the good sounding van halen clips fame), he was (is) apparently using an L-Pad at the output of his Power Brake (for greater than stock attenuation) with great success. (FWIW his PB is also apparently modded, shorting out the low freq. resonance 12.5mH/160uF LC pair. Mine has Rs across the LC and 1mH series L to limit the impedance rise.) And also, I dug far back and found an old post by Randall Aiken where he stated he used a reactive attenator combined with a resistive att. to get really low levels low enough for apartment use (although specifics weren't elaborated, L-Pad after Power Brake seemed to make sense for this configuration).

[cbarrow7625]

Dai,

Yeah, the resistor in series would damp the inductance rise in the speaker. But what would also happen is you would be shifting the frequency response of the speaker somewhat. A series resistor into an inductor to ground is what? Answer: a highpass filter. The version with the 25 Ohm resistor is more likely to affect the frequency spectrum than your 4 ohm version.

The best possible solution is to use multiple L-Pads and a switch. You are not working at high currents here. With 25 Wattsinto 8 Ohms is about 1.75 Amps. A 6 position dual pole rotary with 3 amp contact ratings should be fine for this type of circuit. I'm sure you can find those in mouser.

Use an L-Pad calculator (just google the term, you'll find a bunch of them) to determine 6 different settings you'd like & you're off & running. Each one of them will be optimized without excessive resistance in the signal path and willmaintain the proper 8 ohm load back to the attenuator.

Chris

[dai h.]

(Chris, thanks for editing that--had me confused for a bit, lol... no worries! )

re: the RL HPF, I spent a bit of time googling for info on that and found some literature but still not getting a grip on it mentally (slow brain in effect). I did, however find a calculator which is able to show the effect on the freq. response graphically (low freq. response getting worse) although it's in Japanese:

http://sim.okawa-denshi.jp/LRhikeisan.htm

if the above link doesn't work(direct link seems a bit wonky--seems to bring up the LR one), try

http://sim.okawa-denshi.jp/Fkeisan.htm

and the RL filter calc. is down at the middle to the right.

(under the "R" and "L" box, it says " can use p：pico，n：nano，u：micro，k：kilo，M：mega". Then under that first box, :"freq.analysis", then check boxes for:

[note:I don't understand all the terms so I apologize if I'm in error here.]

Bode plot
Nyquist plot (f=0～∞)
pole, zero-point
phase (allowance?)
vibration analysis

(second box) :
Transient Analysis

(check boxes) :
Step impulse response (G?-inverse LaPlace transform)
overshoot
final step response

(also, not sure if this is a valid measurement but took an L reading across the spk. I'll be using--12in. spk. in M/Boogie .22+ combo cab--and it was about 860uH and tried plugging this in to the calc. with some different values for R.)

as far as the specific values I planned on using, I was going to try an 8 ohm and 1 ohm rheostat in series (spk. off wiper) because in the guy's soundclips above, apparently (though I have not heard the specific one in question), there is one where the volume is actually low enough that the "raw" unamplified electric guitar strings can be heard along with the spk. output signal and (as far as I can gather from what he's posted) his (fixed) L-Pad consists of an 8 ohm and 0.5ohm Rs (obviously with spk. output taken from here and after a Power Brake on the min. position the acoustic volume must be rather low--low enough for an apartment situation). (I already had a couple of 8ohm power WW and found a 1ohm rheostat in surplus.)

My thinking behind using the 8ohm R and 1ohm rheostat (besides being close to the reference values) were that it appeared matching exactly wasn't particularly paramount, and the limited volume range (and the slight difference in R as the spk. "came in" to the circuit as the wiper was raised from ground) wouldn't be a problem either(higher values in parallel could only approach 1ohm and not go higher). (I could also tack a 1ohm across the rheostat easily if needed to lower volume further.)

(Probably there is an advantage there but) I haven't quite been able to mentally work through the advantage of your suggestion of using a multi-position sw. (better with higher R values?--not sure about this yet, sorry for my feeble-mindedness... ).

[cbarrow7625]

Yeah, sorry about the brain fart on the high pass / low pass thing....

8 Ohm series into a 1 ohm rheostat should work OK. That will give you about 20dB of volume reduction at maximum position on the rheostat.

The best calculated series resistor for that setup is actually 7.1 Ohms series (@22 Watts dissapation for 25 Watt input!) with a 1 Ohms shunt resistor (2.5 Watts dissapation). This gives 19dB of reduction.

That will give you a true 8 Ohms impedance back to the powerbrake. A 1 Ohms / 3 Watt rheostat should get the job done but I would make sure to build in some ventilation for it just to be safe!

The multi-position switch option would allow you to use high wattage resistors (which are cheaper than high wattage rheostats) so that you don't have to worry about burning anyhting up. It would also allow you to have a set-up with less than 20dB of attenuation (should you need that). With the 8 Ohm / 1 Ohm combo you are kinda limited to just low volume & no room to jump up in level. You can't simply reduce the value of the 8 Ohm resistor to get louder. If you do, you'll put too much power through the 3 Watt rheostat and burn it up. With a switched setup you could do anything. A combination of the switched setup & your setup may be a real good comprimise. Just have your 8 ohm / 1 ohm rheostat setup as the lowest volume setting on the switch.

Good Luck

Chris

[dai h.]

thanks for the additional input Chris. The 1ohm is a 50W unit (lucked out found one surplus), which I think was good for 7 amps which is total overkill but I figure should give me no worries about burning it up. A thing to note is that the resistive strip is real thick (which seems to contribute to grittiness relative to other ones this size with more resistance and a thinner wire for the strip), so I'm going to try to find a big cheap knob (for better leverage rotating it). I'll report back on how it sounds hopefully in a couple of days. I have a box I can put the rheostat in but need to make some mounting holes (I can solder and assemble decently enough but finally just learning how to make proper holes using center punch, graph paper, etc.etc.). Thx again!

[dai h.]

this is a bit of a followup to the L-Pad business.

Well after reading and learning a bit more about L-Pads I kinda got lazy about trying the rheostat (since it seems to be generally better to try to maintain a constant impedance). Got off my butt yesterday and hooked up a fixed 8 ohm L-Pad (an 8 ohm and 300ohm in parallel as the "series R" and a 0.27ohm and 1ohm in parallel as the "shunt R"--approximately 7.79ohms and 0.247ohms respectively which gives about -30dB in attenuation). Then I connected this fixed L-Pad after the output of my slightly modded Power Brake. Well, this definitely drastically kills the volume (it's so low that I can hear the clock tick and the "raw" gtr. strings) but tone-wise it just wasn't doing it for me. Tried higher settings on the PB, different tone and volume settings on the amp, plus the pot in series with the 470ohm R in the PI, but nothing satisfying. I thought perhaps the Boogie .22+ 1x12 was just simply the wrong speaker(and cab) to use with the Marshall and turned the amp off and called it quits.

Today, I tried reducing the value of the parallel R across the 12.5mH(which puts the low frequency peak in) from about 16ohms down to about 10 ohms. This seemed to help a lot. I suppose there was just too much low frequency response for this low volume setting. (The Ralle guy above seemed to have the right idea in shorting out the low freq. resonance inducing LC pair.) My impression is that at this super low volume it actually better to have the speaker right in front of you to hear it better(I guess not so much bouncing off the walls and getting to your ears?). Still not perfect but improved from previous attempts. Maybe high and low frequency adjustments (from the reactive load) would be useful. I thought about using rheostats in place of the resistors across the two Ls, but it seemed I would need high power ones(not sure exactly how much wattage they would need to be).

[dai h.]

tried a bit more simple experimentation by lowering the R across the 1mH from about 16 ohms down to about 9, then about 5.5ohms. Didn't really notice any difference in particular but maybe if I could use a rheostat and switch faster it'd be easier to notice.

Also tried mic'ing the cab with a cheapo Electro-Voice dynamic. Not quite the best in terms of a mic'ed gtr. amp sound (I would guess in part to some degree from the crappy mic and mic pre--which is the mixer section of an old Tascam 488mkII Portastudio, plus substantially less than stellar mic'ing technique balancing the mic on a coffee can) but not bad and I find I prefer this over my SansAmp and Marshall SE100 spk. emulator. The acoustic volume from the 1x12 cab is super low--so much so that the unamplified gtr. leaks into the microphone. I find the headphones are preferable since I can "turn up" the amp volume (to me only without making the room volume louder) which helps make it "feel" more normal as far as the "playing a gtr. amp" experience goes. A closet to stick the cab into would be more ideal I think (more isolation, could crank the volume a bit more). Also, I think I could go a little bit louder--say a bit loud TV level--without disturbing neighbors. If anyone wants to try this from scratch I think it'd be a good idea to buy an adjustable L-Pad to "fine tune" the volume (but you'd have to be very careful not to overpower the L-Pad). I also tried lowering the AC a bit to around 90V. This seemed to mellow the sound a little bit but wasn't anything super noticable.

re: the reduction of the 12.5mH/160uF LC pair's effect: I did try shorting that before but undid that since it (not unexpectedly) killed too much of the bottom. Maybe the low freq. hump it causes should be in at moderate levels of attenuation (and in "load position") but lessened at greater (extreme?) levels of attenuation. Have parallel R go down as the atten. clicks go down.

I also found this neat (javascript applet??) thing that allows (building and) looking at the voltage and current in the basic reactive load circuit (L, R, then LC).

http://www.walter-fendt.de/ph14e/combrlc.htm

(Actually I found this page before but only figured out what could be done yesterday after seeing it in my bookmarks.)

[Mike Sulzer]

The resistance the speaker looks back into does affect its sound. A guitar speaker is a woofer with a low frequency resonance at about 100 Hz, somewhat modified by the cabinet. The important thing is that the speaker impedance is quite large, maybe 100 ohms, near the resonance. This matters because a tube guitar amp uses pentodes as output devices, with very little, if any, feedback. The amp has a high output impedance in the linear range because the tube characteristics are flat (constant current). It also has a high impedance when overdriven because the tube is current limited. So a guitar amp is kind of like driving the speaker with a current source: the voltage across the speaker depends on the impedance and thus on frequency. If you use an attenuator that presents a low impedance looking back into the speaker, you alter the speaker response. The difference is audible. A high impedance give the sort of full or resonant sound in the lower range that we are used to. A low impedance gives a more analytical and less interesting sound. I have tried this the other way, too. Take a transistor amp, which has a very low output impedance. (Play the amp clean and get the distortion from a pedal or whatever so you do not get bad solid state amp distortion.) Put a resistor in series and it sounds more like a tube amp. The resistor needs to be maybe twice the nominal speaker impedance; making it bigger makes little difference. You lose a lot of power doing this, but if that is the goal, no problem.

The tube amp also should look into a speaker load rather than a resistor. But your power brake should do that for you.

[dai h.]

A high impedance give the sort of full or resonant sound in the lower range that we are used to.

yes the LC (12.5mH and 160uF) is to create that resonance (and my understanding is that the R--an adjustable sliding wirewound R--I put across it gives control over their effect).

But,

A low impedance gives a more analytical and less interesting sound.

this sounds more of a subjective opinion. Some attenuators are pretty much resistors only (Air Brake). The load in the Ultimate Attenuator is a resistor only (followed by a solid state power amp). Some people like them. Plus I can think of an opinion (from what seemed to be a technically knowledgeable person) where he thought a resistor-only attenuator (or resistor only loading of the tube amp output) sounded better than a reactive one even though the reactive one should have been a better electronic equivalent to an actual speaker. So there must be something missing in the calculation esp. at levels of high attenuation (which I think is what most people find problematic from attenuators). And, that (reducing the low freq. resonance) seems to help balance out the sound and make it sound more "correct". Probably there are compromises compared to non-attenuated mic'ed speaker sound, but the benefit here of course is very low apartment/eviction-safe acoustic volumes.

I'd like to go a little further with the experimentation also, lowering the B+ separately from the heaters and lowering the power (section? maybe include PI V3 heater as well) and preamp heaters separately also to help the volume reduction along further and consume less power (maybe get more interesting distortion as well?).

[Mike Sulzer]

Well, when you listen to something and judge the result, it is always subjective. But you have to do it. High attenuation causes an apparent loss of bass; this has to be added back in. But this is different from the resonant effect.

An attenuator can be though of as having three primary functions:
1. Loading the amp;
2. providing the attenuation;
3. getting the impedance looking back from the speaker correct.

#1 affects the sound because when the impedance varies as a function of frequency, it affects how the amp distorts. #3 matters because letting the speaker cone vibrate at the low frequency resonance affects the low frequency transient response (and the high frequency response is affected by the gradual increase of impedance with frequency). I think (subjective) that #3 is more important.

The resistor only attenuator can get #3 correct because the impedance looking back into the amp (due to the plate resistance of the pentodes and their current limiting function) is primarily resistive at the frequencies of interest. This can be simulated by making the resistance looking back into the attenuator high enough. The inductors in the load are intended to make the load on the amp like that of a speaker, less important in my subjective opinion. (You must, of course, load the amp approximately correctly to protect the output transformer.)

[dai h.]

High attenuation causes an apparent loss of bass; this has to be added back in.

well, I thought due to the Fletcher-Munson thing I would need to add bass and treble, but to ear it seems to work better taking bass away. (Or, maybe I should have tried to add treble?) Without the lowering of the LC tank (I think you call it)'s effect the bass resonance was too boomy.

But this is different from the resonant effect.

but the LC tank does effect the bass, right?

(was going to try adjusting the L-Pad a bit to raise the volume a hair, but I feel like kaka now, so it'll have to wait 'til tomorrow... )