Using speaker jack outputs

[dude]

Can more than one output speaker jack be used at the same time. For instances:


A 4 ohm tap with a 4 ohm speaker and an 8 ohms tap with an eight ohm speaker used together at the same time, damage to OT?

[Justa]

An 8 ohm plus a 4 ohm load works out to a total load of 2.66 ohms.  Too low for an amp designed to run on 8 ohms.  Not healthy for the OT either.

[HotBluePlates]

Can more than one output speaker jack be used at the same time. For instances:

A 4 ohm tap with a 4 ohm speaker and an 8 ohms tap with an eight ohm speaker used together at the same time ...

Can you?   Yes.

What's the impact?  The amp sees the same thing as 2x 8Ω loads on the 8Ω tap.  If you had three taps (4, 8, 16) and connected the marked impedance to each tap, it would likewise look the same as 3x 16Ω loads connected to the 16Ω tap (or 3x 8Ω loads connected to the 8Ω tap, etc).  Therefore, the reflected primary impedance is 1/2 (for 2x loads) or 1/3 (for 3x loads) the design value.
If you wanted to reflect the design impedance to the primary, then when connecting 2x loads, use double the marked impedance on each.  For example, you have 4Ω & 8Ω taps, so connect 8Ω or speaker to the 4Ω tap and 16Ω of speaker to the 8Ω tap; the reflected load will be the design value (gotten when connected a single speaker load to the correct tap).

An 8 ohm plus a 4 ohm load works out to a total load of 2.66 ohms. ...

This is correct is both of those speaker loads are on the same tap.

Since the question is about different taps, you should envision the situation as 2x equal loads on a single tap, or half the design loading.

RDH4 in the Library of Information has a section on transformer which explores this exact scenario, and has some math to prove it.  It's easier for you & me to just know the end effect the math shows about the reflected primary impedance.

... damage to OT?

...  Not healthy for the OT either.

I wish people would stop saying this...

Have you ever plugged your Fender combo's internal speaker into the wrong jack?  You get a very quiet, extremely distorted sound out of the speaker.  That's because if nothing is plugged into the main Speaker jack, it shorts hot to ground.  That keeps you from operating the amp with the transformer unloaded, which is arguably the hazardous condition for a tube amp.

I did exactly this way back when I built my first amp... I spent a good half-hour with the speaker plugged into the wrong jack and wondering why the amp had no volume. I was playing through it most of that time, in between checking voltages through the amp.  The OT suffered no ill effect, and I'm still playing through that amp 15 years later.

So, no OT damage but power output will be reduced and/or distortion increased as you shift the reflected primary impedance away from the design value.

As a rule, tube amps don't mind a short-circuit speaker output but can be damaged by an open-circuit speaker output (due to flyback voltage damaging the OT); transistor amps don't mind an open-circuit speaker output (this is no-current drawn from the output transistors) but will be damaged by a short-circuit speaker output (output transistors try to deliver infinite current).

To be fair though, modern transistor amps probably have protection circuitry for the short-circuit condition...

[Justa]

Can more than one output speaker jack be used at the same time. For instances:

A 4 ohm tap with a 4 ohm speaker and an 8 ohms tap with an eight ohm speaker used together at the same time ...

Can you?   Yes.

What's the impact?  The amp sees the same thing as 2x 8Ω loads on the 8Ω tap.  If you had three taps (4, 8, 16) and connected the marked impedance to each tap, it would likewise look the same as 3x 16Ω loads connected to the 16Ω tap (or 3x 8Ω loads connected to the 8Ω tap, etc).  Therefore, the reflected primary impedance is 1/2 (for 2x loads) or 1/3 (for 3x loads) the design value.
If you wanted to reflect the design impedance to the primary, then when connecting 2x loads, use double the marked impedance on each.  For example, you have 4Ω & 8Ω taps, so connect 8Ω or speaker to the 4Ω tap and 16Ω of speaker to the 8Ω tap; the reflected load will be the design value (gotten when connected a single speaker load to the correct tap).

An 8 ohm plus a 4 ohm load works out to a total load of 2.66 ohms. ...

This is correct is both of those speaker loads are on the same tap.

Since the question is about different taps, you should envision the situation as 2x equal loads on a single tap, or half the design loading.

RDH4 in the Library of Information has a section on transformer which explores this exact scenario, and has some math to prove it.  It's easier for you & me to just know the end effect the math shows about the reflected primary impedance.

... damage to OT?

...  Not healthy for the OT either.

I wish people would stop saying this...

Have you ever plugged your Fender combo's internal speaker into the wrong jack?  You get a very quiet, extremely distorted sound out of the speaker.  That's because if nothing is plugged into the main Speaker jack, it shorts hot to ground.  That keeps you from operating the amp with the transformer unloaded, which is arguably the hazardous condition for a tube amp.

I did exactly this way back when I built my first amp... I spent a good half-hour with the speaker plugged into the wrong jack and wondering why the amp had no volume. I was playing through it most of that time, in between checking voltages through the amp.  The OT suffered no ill effect, and I'm still playing through that amp 15 years later.

So, no OT damage but power output will be reduced and/or distortion increased as you shift the reflected primary impedance away from the design value.

As a rule, tube amps don't mind a short-circuit speaker output but can be damaged by an open-circuit speaker output (due to flyback voltage damaging the OT); transistor amps don't mind an open-circuit speaker output (this is no-current drawn from the output transistors) but will be damaged by a short-circuit speaker output (output transistors try to deliver infinite current).

To be fair though, modern transistor amps probably have protection circuitry for the short-circuit condition...

Thanks HotBluePlate for the correction on calculated impedance.

I'm aware that Fender uses shorting jacks for protection in the event you forget to plug in a speaker.  But if you are running the amp on a lower impedance load than it was designed for isn't the current in the OT going to generate a lot more heat than it was designed for.  A momentary short is one thing but playing all night with lower than designed impedance has to have it's limits doesn't it?

[PRR]

> isn't the current in the OT going to generate a lot more heat

No. Or not so you notice.

Audio transformers don't overheat. The low loss needed to preserve precious Audio Power means load heating is very small.

Tubes don't give wild over-current (like transistors can). Current through a vacuum is expensive. If design load is say 100mA peak, we use a tube and screen voltage to give 120mA max before grid-limiting.

[Justa]

Thanks for the explanation PRR!

This contradicts what I had read on the net.  Guess I need to double check my sources!   

[HotBluePlates]

This contradicts what I had read on the net.  Guess I need to double check my sources!   

It's all good, and you're nowhere near the only one to repeat it.

I blame Gerald Weber 1000% for this one, because he put in his late-90's book that it was safe to use a too-high load on an amp, but that a too-low load would damage the OT.  Everyone else is just adding a layer of logic to the false argument saying, "If G.W. is correct, then how does that damage happen?"

I had hoped you see that if a 0Ω speaker load (my shorted jack) didn't damage the OT after a half-hour or more of trying to find the issue, how could some other load substantially-higher than 0Ω hurt the OT?

The real trick here is that the output tubes will only flow a current dictated by their grid voltage, even when the load drops to 0Ω.  That current is not any/much higher than what they flow during normal signal peaks, and the OT handles those fine and also will in the 0Ω load situation.

[MartinB]

If you wanted to reflect the design impedance to the primary, then when connecting 2x loads, use double the marked impedance on each.  For example, you have 4Ω & 8Ω taps, so connect 8Ω or speaker to the 4Ω tap and 16Ω of speaker to the 8Ω tap; the reflected load will be the design value (gotten when connected a single speaker load to the correct tap).

That makes perfect sense when explained that way.  Since I'm currently working on a combo with 4, 8 and 16 ohm taps and an 8 ohm internal speaker, it's useful to know that I could run a 16 ohm extension cab if I needed to, without changing the reflected load. 

[Justa]

This contradicts what I had read on the net.  Guess I need to double check my sources!   

The real trick here is that the output tubes will only flow a current dictated by their grid voltage, even when the load drops to 0Ω.  That current is not any/much higher than what they flow during normal signal peaks, and the OT handles those fine and also will in the 0Ω load situation.

Thanks for that!  I didn't realize that the output tubes were limited by their grid voltage even when the load drops to 0 ohms.

[HotBluePlates]

Thanks for that!  I didn't realize that the output tubes were limited by their grid voltage even when the load drops to 0 ohms.

Check out the attached grid curves for a 6L6.

The tube manufacturer generates a set of grid curves by connecting the tube's cathode to ground, the plate directly to a power supply, and applies a negative voltage to the grid.  The grid is held at the fixed negative voltage while the plate voltage is increased from zero to some upper high voltage.  There is no plate load resistance/impedance, and the plate current is measured at increments along the way.  The process is repeated for a different grid voltage, and individual voltage/current points are connected using a french curve.

The typical limit of driving signal before gross distortion is the 0v gridline, where the grid voltage equals the cathode voltage.  Look at the curves attached:  the Purple line traces the 0v grid line, so even if there is no bias voltage and/or there is a big input signal driving the grid to equal the cathode voltage, this is as high as the tube plate current gets.

The Blue line would be typical of a plate load provided by the OT (here for a supply voltage of 400v, approximating a 4kΩ plate-to-plate load in Class A, giving an effective 2kΩ load to the tube on this side of the push-pull pair).  Because of the voltage drop across the effective 2kΩ plate load, once the plate current reaches the 0v line the tube is passing 170mA of plate current and the plate voltage has dropped to 75v.  You can check this with Ohm's Law: Resistance = Voltage-change/Current-change = (400v-75v)/(170mA-0mA) = ~1900Ω.

What happens if you have no load, or only that provided by the OT's d.c. resistance?  You'll see this case with the Red line... Plate current is still dictated by the 0v gridline, which is now 190mA.  But the plate voltage only dropped by 50v because the load is now only the OT's DCR.  Resistance = (400v-350v)/(190mA-0mA) = ~263Ω.  Even if the resistance was zero (vertical line along 400v up to the 0v gridline), the plate current would only have increased another 3-4mA.

Believe it or not, you already knew grid voltage controlled plate current in this manner, because you already know grid bias sets idle plate current to a fixed value.

You may have intuitively imagined a 0Ω speaker load would be like the OT becoming 0Ω from B+ to ground.  If that happened, current would rise without constraint until a fuse blew.  But the tube is between the OT and ground, and imposes its own limit on current drawn from the power supply and through the OT.

[Justa]

Wow, thanks HotBluePlates for the very comprehensive tutorial that is most helpful!  I won't forget this one for future reference.

[Ritchie200]

This contradicts what I had read on the net.  Guess I need to double check my sources!   

You cant believe anything you read on the net.....

Jim

[Ritchie200]

As usual, thanks PRR and HBP for the great explanation!

Jim

[HotBluePlates]

This contradicts what I had read on the net.  Guess I need to double check my sources!   

You cant believe anything you read on the net.....

I blame Gerald Weber 1000% for this one, because he put in his late-90's book that it was safe to use a too-high load on an amp, but that a too-low load would damage the OT.  Everyone else is just adding a layer of logic to the false argument saying, "If G.W. is correct, then how does that damage happen?"

Or sometimes what you read in print.

Weber's books were a blessing to me when I first picked this stuff up, because there wasn't much of anything on the internet then (or even much of an internet), the old books had long disappeared from libraries, and there was precious-little geared for guitar amp electronics.  Weber also writes in a very easy-to-read style.

When I learned more, his books were sometimes a curse, as there was much I had to unlearn to get deeper into the how & why.

[alerich]

Have you ever plugged your Fender combo's internal speaker into the wrong jack?  You get a very quiet, extremely distorted sound out of the speaker.  That's because if nothing is plugged into the main Speaker jack, it shorts hot to ground.  That keeps you from operating the amp with the transformer unloaded, which is arguably the hazardous condition for a tube amp.

This is obviously a good fail safe for a tube amp. Why doesn't every manufacturer do this? They copied Fender on most everything else (the largely unnecessary standby switch, for example). I know a shorting jack is a little more expensive but nobody was more frugal when it came to building amplifiers than Leo Fender and he still included it. Maybe he had a bad habit of not hooking up a speaker load back in the day when he was testing finished amps and this was his solution to keep him from wasting OT's.

[Justa]

Knowing that tube amps are a lot more tolerable to a low impedance load and can withstand a short for a lot longer than I ever imagined makes me wonder about how much of a load mismatch tube amps can handle.  What I commonly hear is that tube amps can safely handle a mismatch of 2X or 1/2X of the designed load impedance.  After reading this thread this doesn't seem to make much sense as I now expect to see much more amp tolerance of speakers with a lower impedance than the amp is designed.

So how much mismatch is acceptable in both the too high and too low speaker impedance?

Thanks!

[HotBluePlates]

... I ... wonder about how much of a load mismatch tube amps can handle.  What I commonly hear is that tube amps can safely handle a mismatch of 2X or 1/2X of the designed load impedance.  ...

So how much mismatch is acceptable in both the too high and too low speaker impedance?

Look at the attached plot of a Cannabis Rex.  The Peach-colored line is the impedance of the speaker at different frequencies.  You'll see this nominal-8Ω speaker is indeed around 8Ω between 200-500Hz, but this rises with frequency above 500Hz, and also rises at the bass resonant to ~120Ω.

So now you know the OT will handle a short-circuit load, and that real speakers can/do exhibit impedances many times of the nominal impedance marked on the speaker.

The takeaway is, I believe, that most of the recommendations are bogus.  You don't want the OT to run unloaded (∞Ω), at least not with the tubes driven hard enough to cut off in Class AB operation.  This is because the cut-off of tube current is equivalent to a shock in the OT and induces a counteracting flyback voltage spike.  It's that flyback spike which damages the OT insulation.  You're unlikely (in my limited accidental experience) to damage an OT with a few minutes of light playing with no speaker (but I only strummed a few times and asked myself, "why is there no sound?" before I reconnected the speaker).

[EL34]

Justa,
Please note that all your forum notification emails are bouncing back
att.net is blocking forum emails

You need to go see why att.net is blocking your forum notification emails

[Justa]

Justa,
Please note that all your forum notification emails are bouncing back
att.net is blocking forum emails

You need to go see why att.net is blocking your forum notification emails

I have no idea why att.net is blocking my emails so I changed the email address in my profile to a Gmail account.
Sorry for the trouble!

[Justa]

Look at the attached plot of a Cannabis Rex.  The Peach-colored line is the impedance of the speaker at different frequencies.  You'll see this nominal-8Ω speaker is indeed around 8Ω between 200-500Hz, but this rises with frequency above 500Hz, and also rises at the bass resonant to ~120Ω.

So now you know the OT will handle a short-circuit load, and that real speakers can/do exhibit impedances many times of the nominal impedance marked on the speaker.

The takeaway is, I believe, that most of the recommendations are bogus.  You don't want the OT to run unloaded (∞Ω), at least not with the tubes driven hard enough to cut off in Class AB operation.  This is because the cut-off of tube current is equivalent to a shock in the OT and induces a counteracting flyback voltage spike.  It's that flyback spike which damages the OT insulation.  You're unlikely (in my limited accidental experience) to damage an OT with a few minutes of light playing with no speaker (but I only strummed a few times and asked myself, "why is there no sound?" before I reconnected the speaker).

Thanks so much for the explanation HotBluePlates,
I guess bogus is the operative word here.  I didn't realize that the impedances varied to the extent they do with frequency.  So I guess when Fender says 8 ohms total load on the back of an amp that really should be viewed as a guideline and the amp will actually tolerate considerable load range.

[HotBluePlates]

...  So I guess when Fender says 8 ohms total load on the back of an amp that really should be viewed as a guideline and the amp will actually tolerate considerable load range.

Just connect a speaker marked with the impedance indicated on the amp/tap.  If you don't have that speaker load, do the best you can & know clean power will likely be reduced.

The designer likely designed the amp as though there were a resistor of the marked impedance connected to the OT secondary, then possibly added negative feedback or some other technique to compensate for the speaker's varying impedance.

But yes, most amps can tolerate quite a lot of variation.

[PRR]

Cathode bias amps will be happy with any load down to a short and up to several times higher than nominal (not UN-loaded). Power into the tube is relatively unchanged, worst-case is idle which we do a lot so the amp must stand it.

Deep fix-bias amps "can" be strained working in very low-Z loads, though frankly I have never seen it happen even on monster 8417 (cheap 6550) amps.

[Justa]

Thanks PRR,
I didn't realize there would be difference between cathode bias and fixed bias amps.

[PRR]

Cathode bias tries to self-correct the tube current.

Fix-bias allows tube current to increase, doesn't set any limit.

[Justa]

Oh right, now it makes a lot more sense to me.