Dummy load

[jeff]

 I was wondering about speaker impedance vs a straight resistance. It's bad to run an amp into a dummy load of 8 ohm resistance instead of a speaker with an 8 ohm impedance because the speakers resistance changes with frequency, right. This got me wondering why. I was thinking of two amps. One is playing a 150 Hz signal and the other is playing a 400 Hz. Now at 150Hz the speaker has a higher resistance and at 400 it's closer to 8. So why wouldn't the load be mismatched at 150Hz.

Is it because the transformers impedance?
 
Does the transformers resistance change with frequency in a similar manner to the speaker?

For example: At 150Hz a speakers resistance is close to 30 ohms. So is the transformers resistance at 150 Hz for a 8K transformer closer to 3K and the ratio between the transformers 3K and the speakers 30 present the tube with the right load(8K:8)? If you were to graph a transformer's resistance VS frequency would it look similar to a speakers?

Is this the reason why using a resistor instead of a speaker is bad?

[sluckey]

It's bad to run an amp into a dummy load of 8 ohm resistance instead of a speaker with an 8 ohm impedance

Where'd you hear that?

[alerich]

Is this the reason why using a resistor instead of a speaker is bad?

I've always believed that a resistive load is actually safer because it dissipates energy evenly across the frequency spectrum.

[Frankenamp]

Is this the reason why using a resistor instead of a speaker is bad?

I've always believed that a resistive load is actually safer because it dissipates energy evenly across the frequency spectrum.

+1  However, tubes are inherently safer than transistors regarding mismatches especially when sorted. Just don't run them with an open circuit. If you are wanting a quick and dirty dummy load, just dunk a 8 ohm 50 watt power resistor (one of those aluminum jobbies) in a quart of veggie oil. (think I saw plans for one around here someplace...)

[jeff]

Where'd you hear that?

Can't find the original page but it directed me to this http://www.aikenamps.com/spkrload.html saying that a pure resistive load is bad for the amp and I was just wondering why. Assuming the transformer had similar spikes and valleys is all I could come up for as to why.

So if the tramsformer has an even resistance throughout as you play a low E 82Hz the speaker is around 8 ohms and the load to the tube is 8K, but what happens as you climb up the E string? When you get to the A# the resistance of the speaker spikes to 30 ohms, so if the transformers even across all frequencies you have a reflected 30K load on the tube. Then as you keep climbing you start to head back towards 8.

So your tube follows an 8K ohm load line when you play an E and a 30K ohm load line when you play a A#?
I never noticed anything but I'll have to crank my amp and listen carefully to what happens as I climb the E string.

I like the sound of my little amp. Originally I wanted to rebuild another but add a line out and just use a 8 ohm 20W resistor instead of a speaker for size, convienence. But I want to make sure this is good to do.

[jojokeo]

A purely resistive load only makes for a poor power attenuator when it comes to a recording or sound reproduction of the amp playing through one. As for primary Z load, much is made of this being "matched" perfectly. It can't be. It can be made as close to optimum as possible but that's it since it's always actively changing.

[HotBluePlates]

Can't find the original page but it directed me to this http://www.aikenamps.com/spkrload.html saying that a pure resistive load is bad for the amp ...

Nowhere does he say that a pure resistive load is "bad for the amp."

But he's also writing an article about a speaker emulator; a resistor would be a poor speaker emulator because it doesn't have all the peaks and dips of a speaker, both in frequency terms and impedance terms.

A resistive dummy load is the way to go.

I was thinking of two amps. One is playing a 150 Hz signal and the other is playing a 400 Hz. Now at 150Hz the speaker has a higher resistance and at 400 it's closer to 8. So why wouldn't the load be mismatched at 150Hz.

Well, it is "mismatched."

For any set of output tubes and operating voltage, there is an ideal load which results in maximum clean power output. If you deviate above of below the ideal, you get more distortion, less output power, or both.

Designers generally pretend that the speaker is a resistor, and choose the transformer as though the peaks at the bass resonant frequency and at highs did not exist. That results in less power and more distortion (in most pentode and beam power tubes) at those points where the impedance is higher than nominal.

But the bass resonant frequency is often below the guitar's range, so the effect is unimportant. If it's close (say a resonant frequency of 75Hz), the lowered power output is offset by the fact the speaker moves more freely at that frequency (so less power = same cone movement).

For highs, you're out of luck. You either use a "corrective filter" (to use RCA's terminology) or "conjunctive filter" (to use Dr. Z's terminology) which offsets the rising impedance at highs, or you use negative feedback from the OT secondary to make the output stage more immune to changing load impedance.

You can find this exact issue discussed in RDH4. Do you now see why some of us feel the whole matching issue has been overemphasized, and why we wouldn't sweat changing OT secondary taps? Speaker impedance might vary 5:1 above nominal, and we're normally just switching taps in a 2:1 ratio.

[PRR]

> the bass resonant frequency is often below the guitar's range

Often not. That PR8 is resonant a half-octave up into the guitar band. 90Hz is common in larger for-guitar speakers. (Yes, ex-PA speakers such as E-120 may be tuned below the guitar band.)

But that's fine, even desired.

At bass resonance the speaker's actual efficiency is higher. But so is its impedance. The higher impedance sucks less true power from the amp. You can balance one against the other so that acoustic output at bass-resonance is equal to output in the midrange, "flat". In guitar amps we often have a cabinet fall-off which is compensated by allowing the bass resonance to run strong, so the result is sorta-flat and bumped-dipped according to our taste in the bottom octave and the speaker's bass-handling ability.

Resistor load is fine for the amp.

Signal tapped from a resistor load does not give the same sound as when driving a speaker. Hence the effort to approximate speaker impedance.

Both the bass resonance and the rise in treble are key speaker design factors. The speakers on the market cover a range of useful "sounds" and tastes (and prices). The usual convention is that the user does NOT worry about the impedance curve. It is the speaker designer's job to make it work good. The user takes the nominal impedance (4, 8, 16) when tapping the amplifier.

> For highs, you're out of luck

That's part of the speaker designer's job, and it isn't a huge problem. "Mis-match" is small within the speaker's useful range, power demand up there is small and the amp does not mind. Most guitar amps are fine. AM radios with naked pentodes often shunt the load with a cap, not really for impedance, but to filter-shape the rising response into a sharp drop past 3KHz-4KHz to trim static.

[alerich]

Can't find the original page but it directed me to this http://www.aikenamps.com/spkrload.html saying that a pure resistive load is bad for the amp and I was just wondering why.

Where on that page does it say that a purely resistive load is bad for an amplifier? I made my earlier comment based mostly on Randy Aiken's comments some time back in a Usenet thread about attenuators.

Edited to add: There's much discussion about whether a purely resistive attenuator sounds as good as one built with reactive elements but that's in the eye of the beholder. Dr. Z has sold a bunch of those purely resistive Air Brake attenuators and they are fairly highly regarded amongst players who use attenuators. Hard to believe how much people will pay for a few resistors, a pot, a switch and a few jacks in a box. I saw one used the other day at a local mom and pop music store (remember those?) for $250. I have a THD Hot Plate and I love it but my home brew low watt amps have simple built in resistive attenuators that don't sound too bad for what they are.

[bluesbear]

I have a large 50 watt wire resistor with a sliding tap for setting ohmage. I have it mounted on a little stand made of a wood base and coat hanger wire. You can measure and set the tap for whatever ohms you need. Of course, if you're dealing with an amp with over 50 watts, you'd need a bigger resistor. I never do. Works for me.
Dave

[jeff]

Can't find the original page but it directed me to this http://www.aikenamps.com/spkrload.html saying that a pure resistive load is bad for the amp and I was just wondering why.

Where on that page does it say that a purely resistive load is bad for an amplifier?

Can't find the original page but it directed me to this http://www.aikenamps.com/spkrload.html saying that a pure resistive load is bad for the amp ...

Nowhere does he say that a pure resistive load is "bad for the amp."

It was on the original page that directed me to that page but I can't find it again. It said it was bad to use a resistor and directed me to that page to build that instead of using the resistor. Maybe I misread it and he was saying bad for sound, not bad for the tubes.

I just wanted to make sure with you guys before I used just the 8 Ohm resistor instead of the speaker. I tried it before until I noticed the tube was red plating. I put the speaker back but it was still red plating. I checked all the voltages and current but couldn't figure it out because the math was right, I was under 12W at idle. It was an old, used tube so I wasn't sure if the resistor caused it or if it was just "it's time". I wanted to check with you guys before I did it again with the new tube.

Thanks
   Jeff

[jeff]

 I'll try the resistor again, probally just a bad, old tube(It was in grandma's TV for 50+ years).

 My question was more out of curiosity. I know the speaker is a coil of wire and I was wondering if this spike was a function of that. Since the transformer is a bunch of windings I was wondering if it too has a similar spike. If this is just the nature of a coil of wire, that it has a spike.

[jeff]

 Come to think of it my combo does have a rattling around the A#. I alway thought it was because it was a combo and the vibrations of the speaker were making vibrations on the tubes or chassis or something. I thought it went away when I took the amp out of the combo but I'll try it again and really listen.

[jojokeo]

Maybe I misread it and he was saying bad for sound, not bad for the tubes.

I just wanted to make sure with you guys before I used just the 8 Ohm resistor instead of the speaker. I tried it before until I noticed the tube was red plating. I put the speaker back but it was still red plating.

I was wondering exactly what you meant by "bad" earlier. Yes a misread - bad for sound not tubes.

Redplating has nothing to do w/ the speaker or resistor load. It comes from exceeding the dissipation of the plate OR the screen. Many people forget about the screen grid

[HotBluePlates]

I know the speaker is a coil of wire and I was wondering if this spike was a function of that. Since the transformer is a bunch of windings I was wondering if it too has a similar spike. If this is just the nature of a coil of wire, that it has a spike.

Pretend the spike in impedance at the bass resonant frequency doesn't exist for a moment.

The remaining rise in impedance gets bigger as the frequency goes up. That is the nature of an inductance; the inductive reactance increases as the frequency increases.

So what about the peak at the resonant frequency? This is the frequency where the cone moves most freely. The movement of the speaker causes an electromotive force (EMF, another term for voltage) that is in opposition to the voltage/current that made the speaker move in the first place. The net effect of the nominal impedance plus this back-EMF is the same as if the impedance were larger, and we always describe it in terms of an impedance rise.

The back-EMF is generated because a coil moving through a magnetic field has a current induced in the coil. The speaker uses this fact to move a coil (mounted at the apex of the speaker cone) back and forth by suspending the coil in a fixed magnetic field. The current fed to the voice coil creates a magnetic field around the coil, and this field reacts to the fixed field provided by the speaker magnet to either pull the cone inward or push it outward.

The effect works the same way whether you're talking about a speaker or a microphone. When I was in the Navy, we used sound-powered microphones for talking over some circuits onboard the ship. A different time, I needed to talk over an internet connection, but didn't have a microphone for my computer. I took a pair of headphones, plugged them in the mic jack, and talked into one of the earpieces. It worked beautifully.

[jeff]

 Cool, Thanks for satisfing my curiosity. If I were a cat I'd be dead by now.

 Speaking of the headphone thing I remember hearing somewhere that to record a bass drum someone mounted a 15" speaker in front of it and used it as a mic.

Redplating has nothing to do w/ the speaker or resistor load. It comes from exceeding the dissipation of the plate OR the screen. Many people forget about the screen grid

  Yeah, that's right, now that you mention it I did have to increase the screen dropping resistor because the screen voltage was higher than the plate. I fixed it, but maybe I damaged the tube without realizing it. The thing is I only noticed the redplating after I used the resistor as a speaker and wanted to double check with you guys if this was coincedence or if the resistor caused it.

[jeff]

Right now I have a speaker jack that disconnects the internal speaker to use an external speaker. If I build this:

INPUT
| V-----------|
|      |           >
|      >           <(100K)
|      <(8 ohm  >
|      > 20W)   |
|      <           >(10K pot)
|      >           <<-------V |  OUTPUT
|____|_______>_________|
             
Plug the input into the speaker jack and the output into the input of an amp?
Is the 100K/10K pot correct?

[jojokeo]

Come to think of it my combo does have a rattling around the A#. I alway thought it was because it was a combo and the vibrations of the speaker were making vibrations on the tubes or chassis or something. I thought it went away when I took the amp out of the combo but I'll try it again and really listen.

Could be a parasitic oscillation at that frequency if you find it is not the other things you mentioned.

BTW, if you have a speaker laying around that's old & worthless, blown, or the frame is warped? Cut out the cone leaving the inner section where the coil wires come through. Cut away the frame leaving only the tag strip part. Instant dummy load in lieu of a resistor. You can also remove the spider and epoxy the coil into the magnet. Be creative depending on your needs.

[jojokeo]

Here's a headphone amp active dummy load (idea from Weber) but could easily be modified for use as an attenuator.

[Willabe]

That looks nice jojokeo, how do you like it?

What is the volume control spec., and where did you get it? It looks like a nice heavy duty one.

What is the little jack on the far left side for?

Is there any unruley vibration from whats left of the speaker with the chassis not being vented to let out the air preasure? I would think it might move around a little, depending on how much volume/signal your trying to dump.

Got a schem. for us?     


              Thanks,        Brad              

[jojokeo]

It works great for what it is and what it's for. I drew it up from using the rheostat's diagram on the package which is the vol control. The rheostat has an impedance & power rating which matches the amp's output which matches the speaker's. The outputs are 1/4" & 1/8" stereo jacks and a 1/4" mono for driving headphones or...? The input jack recieves signal directly from the amp (no speaker) and why the speaker motor assy. It takes the place of the amp's speaker. No vibrations, nothing moves.

[Willabe]

Very nice.

      Thanks,     Brad     

[jeff]

 I just had a thought, maybe this is crazy(90% of my ideas are) but my goal is to basically build a preamp. An amp that doesn't use a speaker, I need a dummy load and a line out. If I'm not going to use a speaker, do I need a transformer? Right now I'm using a 5K transformer at 250V. Could I just use a 5K 25W resistor instead of the transformer and raise the B+ to 490V so the plate is at the same 250V? Then use a cap from the plate to the voltage divider to the line out?

I'm thinking if I use a voltage doubler I can take the plate voltage off the top cap and the screen voltage off the bottom cap. I don't know what I'd use for the output voltage divider resistors(values, wattage). Maybe I could put a cap across the 5K resistor to reduce the high to simulate the speaker?

[jeff]

Something like this?

[kagliostro]

May be this can be of some interest

http://www.duncanamps.com/technical/dummyload.html

Kagliostro

[HotBluePlates]

INPUT
| V-----------|
|      |           >
|      >           <(100K)
|      <(8 ohm  >
|      > 20W)   |
|      <           >(10K pot)
|      >           <<-------V |  OUTPUT
|____|_______>_________|
            
Plug the input into the speaker jack and the output into the input of an amp?
Is the 100K/10K pot correct?

That depends on the power output of the amp.

Let's assume the amp outputs 20w to the 8 ohm speaker. You probably know the equation, Power = Voltage * Current. We care about the amount of voltage feeding the next amp's input, so rearrange that equation using ohm's law to have just voltage and resistance (we'll pretend they're the same thing for simplicity). The equation is now Power = Voltage²/Resistance.

Solve for voltage across the speaker.
20w = Voltage²/8 ohms
160 = Voltage²
Voltage = 12.65v RMS

So our worst case input (with this particular amp) will be in the range of 12.65v. We need to drop that down to around 100mV or less; 100mV/12.65v = 0.0079:1

10k pot and 100k resistor only drops the signal to 10k/110k = 0.0909:1 so we have about 10 times too much signal, so you'd want a 1k pot. 1k/101k = 0.00909:1, which is still a bit too much signal with the pot set to maximum (but maybe that's a good thing).

However, jojokeo's method would probably be better. This gives you an idea of how to figure this stuff for yourself.

[jeff]

Thanks for the formula
The driving amp would be a SE 6V6 ~5W

With jojokeo's method of cutting out the speaker cone, I have a question.
"So what about the peak at the resonant frequency? This is the frequency where the cone moves most freely. The movement of the speaker causes an electromotive force (EMF, another term for voltage) that is in opposition to the voltage/current that made the speaker move in the first place. The net effect of the nominal impedance plus this back-EMF is the same as if the impedance were larger, and we always describe it in terms of an impedance rise."
 If the spike is caused where the cone moves most freely and we cut the cone off, will the coil move more freely and cause spikes/rising resistance at more frequencies? Or will it move the same with/without cone. I assume the stiffness of the cone would add some resistance of the movement of the coil.

Also, I'm just wondering if this schematic would work. A 5K 25W resistor is cheaper than an 8ohm 20W resistor and a transformer. Do we need the transformer if we're not using a speaker?

[jeff]

The output voltage swing would be too big could you split the 5K and tap off that?

[jeff]

Sorry I'm getting carried away here. I'm no designer, I should just cut up a speaker or built a preamp with 12AX7s.

[Jack1962]

I use a Dummy Load resistor all the time , what I do is I have 4 different resistors ,a pair of 8 ohm 300 watt (i think ) in parallel (I couldn't find a 4 ohm) , 1-8 ohm 300 watt and 1-16 ohm 300 watt . I use whichever on is the right impedanceof the amp I am working on , the reason for 300 watt resistors is the dissipation is 3 times that of the rating of the amp.


                                                       Rock On