Passive Low Cut, High Cut, Bandpass Filter

Usually a simple high-pass involves wiring a (bipolar) capacitor in series with one speaker lead, and the simplest low-pass involves wiring an inductor in series.

In your case, though, there is a catch, and it worries me: if, for instance, you wire that cap in series to prevent frequencies below 75 Hz from reaching the speakers, it has the side-effect that you also remove the load from the amp output at frequencies below 75 Hz.

We all know valve (tube) amps don't like operating without a load on the output, so will this lack of a load at low frequencies put your amp in danger? I'm not sure, but I wouldn't want to find out the hard way, with a dead output transformer and/or output valves. 😱

There is a similar issue with the low-pass filter (series inductor) to remove the highs from the speakers. Once again, the load impedance presented to the amp will rise, this time at high frequencies. Once again, I'm not 100% sure how much of a risk this will pose to your amp, but my spidey-sense is tingling on this one - my gut tells me it's a Very Bad Idea (TM).

All is not lost - it principle, at least, it's possible to add additional inductors, resistors, and capacitors in parallel to (mostly) compensate for the impedance changes caused by the bandpass filter. But, in practice, this might turn out to be a bit complicated and expensive.

So what can you do with your project? Well, my first thought is that a normal guitar in standard tuning only puts out frequencies above 83 Hz, so maybe it doesn't matter if the speakers can go lower than that, particularly if the amp design filters out those deep frequencies long before they get to the speaker.

As for the high frequency end, the first idea that pops into my head is to do the filtering acoustically, rather than electronically. In other words, put a wad of foam or pillow-stuffing in front of the speakers, for instance, under the grille cloth, to filter out the high frequencies.

A bit awkward, and Hi-Fi speaker designers will beat their heads against the wall and scream in pain at the very thought, but it should work to muffle excessive high frequencies, and at least it won't hurt your expensive and powerful new amp!

-Gnobuddy

RogerKirchner said:

I'm gonna try the inductor thing... I'll check my ohm readings before I hook it all up

Click to expand...

Roger, good inductors have very low resistance at DC. Ohmmeters measure using DC, so your meter (on the ohms range) will tell you that the inductor is not doing anything, just acting like a little piece of wire.

Now, as you go from DC to AC, and the frequency climbs, the inductor will start to present more and more impedance (more ohms) to the amp. Just like inserting a variable resistor in one speaker line and turning up the ohms, it will turn down the sound from the speaker. But in this case, it will only do this at high frequencies - so you get that lowpass filter you wanted.

The problem is that (at high frequencies) the amp also doesn't get properly loaded by the speaker, because of that high impedance inductor in between. Your ohmmeter won't see this impedance - it only shows up at high frequencies.

I've never destroyed a (tube) guitar amp by running it without a load - but plenty of other people have. You have a high-powered amp here, so it's even more at risk. Worst case, you can burn out the output transformer, burn out the output tubes, burn out their screen grid resistors, and get flashovers and carbon-tracks on the tube sockets, which means they have to be replaced, too.

I wouldn't put that inductor in series, I really wouldn't . I have a very bad feeling about that idea, I think it's very likely that you will end up with a fried amp and a very expensive repair bill.

A couple of dollar-store foam seat cushions in front of the speaker, now - that may be unconventional, and it may not even work, but at least you won't hurt anything expensive by trying it!

-Gnobuddy

RogerKirchner said:

If I hook everything up, band pass filter & speakers, all that will be tied into a 1/4" male jack that will be what I'd plug into the amp output... if i measure the ohms coming off that 1/4" male jack, won't the ohm reading from that be what the amp will see as the load?

Click to expand...

What the ohmmeter shows will be the load that the amp sees at DC.

The trouble is that adding the bandpass filter changes the load at different (AC) frequencies. That will not show up on an ohmmeter, but it is quite possible that it might damage your amp.

RogerKirchner said:

I guess is the 1/4" ohm measurement that i take from that any different than if I have a 2x12 cab loaded with unknown ohm speakers & I don't know how they've been wired?

Click to expand...

With the unknown cab, you also only get the DC resistance. But we've had moving-coil loudspeakers since the 1930s, so we have a pretty good idea how the load varies with frequency - there's a peak at the low frequency end (resonance), and a gradual rise at the high frequency end (due to voice coil inductance). And we know that this doesn't kill (tube) amps.

What you're proposing (inserting a bandpass filter between amp and speaker) will, I think, cause more dramatic load impedance changes with frequency, compared to just bare drivers wired in series or parallel or what-have-you. My worry is that the lack of loading (particularly at the high frequency end of things) will actually cause damage to your amp.

You and I are both on unknown ground here, so I'm not 100% certain that your amp will actually be damaged. I might be completely wrong. But I think there's a significant chance of it happening.

That being the case, considering the potential cost and frustration to you, the least I can do is let you know about my concerns. You take it from there, it's your amp! 😀

RogerKirchner said:

For the I just plug in a speaker cable & if it comes out about 4 ohms it's probably 2 8 ohm Speakers in parallel & if it's 16 ohms then it's probably 2 8 ohm Speakers in series?

Click to expand...

Sure, but the actual impedance can be wildly different from 4 ohms (or 8 ohms, or whatever) at different frequencies. Take a look at the attached image, which is an impedance curve for a 12" guitar speaker I grabbed off the 'Net. You can see that at DC the impedance is, indeed, around 6 ohms or so, but a bit above 100 Hz, the impedance is over 100 ohms. Then, above 500 Hz, the impedance starts to climb again, reaching over 20 ohms by 5 kHz.

That's a lot of variation, but we know that thousands of valve guitar amps have coped successfully with this sort of speaker load for the past eighty or so years. So we know the amp won't self-destruct with this speaker as a load.

But what will the impedance curve look like when you insert a bandpass filter between amp and speaker? The DC load will actually be infinity ohms (because of the series capacitor in the high-pass filter). If you measured with an ohmmeter, it would show you an open-circuit, as though you had a broken speaker lead!

At high frequencies, when the series inductor for the low-pass kicks in, the AC load will also start to climb, due to the inductive reactance. Maybe this won't be any worse than the effect of the speakers own voice coil inductance, as seen in the attached image. I'm really not sure.

As mentioned before, my concern is that the impedance changes due to the (bandpass filter + speaker) will be more severe, more extreme, than normal guitar speakers on their own. And that these more severe impedance changes will potentially damage your guitar amp.

I will try to find time to do an LTSpice circuit simulation of a speaker with a bandpass filter ahead of it. That will tell us more about the actual effect of the filter components. But it still won't answer the $1000 question: "Will the amp be okay with this weird load connected to it?"

To answer that question, you have to find someone who's actually tried it, or you have to venture out into the minefield, and find out for yourself. 😱

I wish Tube_lab (George) would stop by. He knows more about abusing tubes than just about anyone else...he could probably give you a much more certain answer than I can.

-Gnobuddy

I did a little tinkering with LTSpice, inserting a passive, first-order, bandpass filter ahead of a dummy 8 ohm load representing the speaker. Then I plotted frequency response at the speaker, and impedance presented to the amp. The results are in the two attached screenshots.

The frequency response is down 3 dB at about 80 Hz and about 6 kHz, more or less what you wanted, and more or less what we usually expect for an electric guitar speaker.

The impedance changes due to the bandpass filter aren't as bad as I had been imagining. Load impedance climbs to over 60 ohms at 10 Hz (and will climb to infinity as the frequency goes down from 10 Hz to DC). At the high end, it's only about 15 ohms at 10 kHz.

Please note that I didn't model all the variation in a real speaker's own impedance - to keep things simple, I just represented the speaker as an 8 ohm resistor.

That means that with an actual speaker, we can expect more severe and more dramatic impedance changes than the attached plots.

Still, this is a reasonably valid way to find out the effect of the bandpass filter alone. And the answer is that it's not as bad as I thought it might be. It doesn't look a whole heck of a lot worse than the impedance changes of just that single 12-inch guitar speaker by itself (in my previous post), in fact.

So, is it safe to overdrive a 100 watt guitar amp into this sort of load? That, I still don't know!

-Gnobuddy

Thanks, Roger. It's been a rough week, and it's really nice to hear something positive aimed my way. 🙂

-Gnobuddy

RogerKirchner said:

One thing I'm not clear on is the ac/dc thing...

Click to expand...

Basically, there are three different kinds of objects that "resist" the flow of electricity. First there is the plane-jane resistor; it has resistance, and the resistance doesn't change with frequency. electricity.

Now, if you take some copper wire and coil it up neatly on a tube (like a voice coil, or a transformer winding), it certainly does have some (usually low) resistance, due to the copper wire. But the weird thing is, if you apply AC to it instead of DC, it starts to display another kind of "resistance" as well. This second kind of "resistance" is called inductive reactance. Like actual resistance, it's also measured in ohms, and it also opposes the flow of (AC) current in a circuit. But, it changes with frequency, and has some other properties different from true resistance, related to the fact that the coil of wire also creates a magnetic field when current flows through it.

It also turns out that putting a magnetic material (like a slug of iron, or ferrite) inside the coil dramatically increases this inductive reactance.

Two types of resistance down, one more to go!

The third kind of "resistance" comes from capacitors. Two conducting "plates" separated by an insulator - surely the resistance is infinity, then?

It is, if you measure it at DC, with an ohmmeter. But, quite weirdly, when you apply AC voltage, current actually flows right through the insulator, and so the capacitor too has a sort of "resistance". This time, it's called capacitive reactance. It too is measured in ohms, it too changes with frequency.

So far so good, I hope? Okay, onward we go!

A speaker has a voice coil wound with a length of fairly thin copper wire. Because it's thin and long, there are several ohms of resistance. But because it's a coil, and that too one with a chunk of iron in the middle, it's also got inductive reactance, which adds to the actual resistance of the coil.

That explains the DC ohms you measure on a speaker, and also the smooth rise in impedance above 500 Hz or so.

The other noteworthy feature in the guitar speaker impedance plot - the big hump around 100 Hz - comes from something else altogether. It comes from the fact that the voice coil sitting in the speaker magnet also behaves as an AC generator once the speaker starts to vibrate. The coil generates an AC voltage that opposes the voltage trying to drive the speaker in the first place - so you get less current flow - so you get more "resistance". The big hump at 100 Hz is because that's the frequency at which this particular speaker likes to move most freely, i.e., it's the fundamental resonance frequency of that particular speaker.

RogerKirchner said:

so, speakers by themselves ohms are measured in DC ohms , but because of the materials needed to build a passive bandpass filter (caps and inductors) their signal will have AC measurements that basically throws the whole setup into a "random" ohm producing signal that will freak the amp/tubes/transformers into an over worked or under worked roller-coaster based on whatever frequencies I'm producing on the guitar.

Click to expand...

You got it, pretty much! 😀

There are only three kinds of "resistance", or rather, one kind of resistance, and two kinds of reactance. But when you throw a dollop or two of each into a circuit, they all interact and can quickly make a rather complicated mess.

RogerKirchner said:

In stuff like 3 way crossovers in home stereos & car audio stuff is the fluctuations in ohm due to the crossovers not really a big deal because the solid state amps don't really have as hard of time dealing with unstable ohm loads?

Click to expand...

To a large degree, yes. If you completely disconnect the speaker from a solid-state amp, it's usually entirely stable and happy. Do the same with a tube amp, and it will probably fry.

The problem isn't the tubes themselves, really, it's the output transformer in the tube amp. When a transformer is operating, there is a lot of energy stored up in it. Take away the load, and that energy can erupt in unwanted ways, destroying transformer insulation, tube screen grids, etc.

There was a time when early solid-state amps also had output transformers - but transistors very rapidly evolved to a point where it became possible to do away with the output transformers altogether. A lot of problems went away along with the output transformers!

RogerKirchner said:

If I'm correct with that, how do tube driven hifi home stereo amps deal or accept the changing ohms?

Click to expand...

I think part of the answer is that we don't overdrive the heck out of our Hi-Fi amps, so they have an easier time coping with imperfect loads.

And another part of the answer is that Hi-Fi valve amps also can be unhappy with the wrong sort of load - even if they don't die outright, they might not sound right.

I know next to nothing about valve Hi-Fi, though, so I'm not really qualified to give you a good answer.

RogerKirchner said:

...I got a long way to go

Click to expand...

So do we all, it's an awesomely complicated world out there! But clearly you learn really fast, and I love the fact that you're willing to experiment, tinker, and try new things with valve amps. 🙂

-Gnobuddy

RogerKirchner said:

Since full range & 6x9 speakers for guitar are generally a bag full of crap, I really would like to see if trying to roll off the top & bottom to roughly equal the normal frequency range that standard produced guitar speakers can make them sound half way decent at all...

Click to expand...

This is something I've been wondering about, too. IMO, aftermarket "real" guitar speakers are crazy-stupid expensive for the crude and primitive contraptions that they are. So I too want to see if I can persuade cheap and easily available speakers to work "good enough" in a low-power guitar amp for home use.

So far, I found a cheap 8" paging/background music speaker that works really well for jazzy clean guitar tones with a very "vocal" quality - but bass response is so weak as to be truly pathetic.

I also found another cheap 8" dual-cone speaker that isn't as good in the treble, but is much better in the bass, so it's a usable compromise. Bass is still a bit weaker than I'd like, though.

Next up to try is a 2x8 cab with both of those 8" speakers in it. I'm hoping to get the lovely vocal-sounding "cone cry" of the first speaker, mixed with the slightly beefier bass of the second. We'll see how it goes!

On the same topic of using an external filter to reshape guitar speaker frequency response, there was an engineer called Steve Bench who had a very informative website full of all sorts of wonderful tube circuits and info. I stumbled across some of his work a couple of years ago, only to find out that his original website was long gone, his personal life had taken some very tragic and unusual twists (including the loss of his wife, a near-fatal heart attack, and gender-change surgery), and he/she was no longer active in the world of electronics.

One of the things I stumbled across was a guitar amp speaker Steve Bench/ Stefie Bench had made and described. (S)he used four full-range Hi-Fi speakers in it, and added a small series inductor to roll off the excessive high frequency response. It was described as sounding excellent, but this is obviously a subjective thing.

RogerKirchner said:

granted I'll be playing the guitar, so the the best made guitar speakers can't improve my chops,

Click to expand...

I've learned slowly over the years that playing any musical instrument has to become a way of life, otherwise one is pretty much guaranteed to be really bad at it. 🙂

I too struggle with how to spend my (often very limited) hobby time; work on my guitar skills, or build some guitar electronics? Usually, for me, I end up picking up the guitar rather than the soldering iron.

RogerKirchner said:

I like trying things out especially if it's in somewhat uncharted waters...

Click to expand...

I hope you never lose that, the world of electric guitar desperately needs people like you! 😀

I know there are great-sounding guitar amp designs from fifty or sixty years ago, but I cannot believe that those are still the only possibilities today. We have so many components now that were unknown (or too expensive to use) in 1950, from LEDs and cheap power MOSFETS to tiny 10 uF ceramic caps and lightweight, powerful, switching power supplies.

Surely if we combine some of these newer developments with classic old valve circuits, we can come up with new and interesting guitar amp designs?

I really wish "The Hundred Buck Amp Challenge" thread would come back to life; to me, it's far and away the most interesting thread diyAudio has ever produced.

-Gnobuddy

RogerKirchner said:

I left with a pretty sweet '59 Bassman clone, but enough "knowledge" to realize I know nothing about how amps/electronics actually work.

Click to expand...

Yeah, the more you know, the more you realize you still don't know, right? 🙂 I love the fact that you can keep on learning every single day of your life, and never run out of still more things to learn.

Sounds like you're off to a good start with the DIY electronics, though. Thanks to the Internet, there is a lot of good information out there. (There is also even more nonsense out there, unfortunately!)

RogerKirchner said:

I've also put together several pedal kits, but like the amp kit, it helped me get less crappy at soldering, but didn't really increase my understanding of why it does what it does.

Click to expand...

It helps a lot to have one or two actual vintage electronics textbooks, as you do. Piece together the schematic of the pedal with appropriate bits of theory from the book, and it will eventually begin to make sense. (Most pedals are weird, though, in the sense that they are designed to mess up the sound in some way, rather than just amplify or filter of whatever.)

RogerKirchner said:

... had some issues trying to find a way to pull 9v out of somewhere for both typically pedal fx that I planned to build into the setup.

Click to expand...

I love those little wall-wart switching power supplies that are everywhere these days. I can often find them for $2 at thrift stores, and they are often perfect solutions for problems like this. Just add a little bit of filtering in the +9V line to take out any last little switching noise, and mount the power supply far from your preamp input stage to minimize any stray noise pickup.

RogerKirchner said:

I've attached a pic of that as well as 2 other designs I made using a similar speaker configuration.

Click to expand...

Interesting looking amps, quite unique in the usual sea of Fender/Marshall stuff!

-Gnobuddy