Impedance selection

[Andy]

I am currently involved in a debate on another forum on whether a tube amp with a 3-way impedance selector switch (4-8-16) can be run safely in any of the three positions with an 8-ohm load. I have heard guys on here repeatedly say that it will work, and will not damage the amp in most cases. everyone on the other forum seems to think I'm smoking crack or something when I tell them that. they go 'by the book', and by 'book' I mean user manual. the user manual for almost all amps says to run the amp into a load matching the amp's settings. failure to do so would result in damage to the tubes, amplifier, and/or speaker.

that being said, could someone (preferably enzo or one of the other highly experienced individuals) please give me a complete technical description of what would happen to said amplifier (power level, frequency response, etc) in all three positions, with an 8 ohm load. I will probably repost it on the other forum to help its members understand the truth. I have personally mismatched speaker loads in the attempt to alter the tone of the amp, and I can certainly say that there was a difference, and no damage occurred, but I would like to hear it from someone with more than my few years of electronics experience.

[Enzo]

Oh, I am the wrong guy to ask what the exact science of the mismatch will bring. For freq response changes and stuff, you need kg or RG or someone else with a G in their name.

When I worked the shop at a music store, we used to sell Mesa Mark 2 amps. The amp came with the 4 ohm spealer plugged into the 8 ohm jack - or was it the 8 ohm speaker plugged into the 4 ohm jack, I forget - anyway, Mesa said to run it that way because they felt it sounded better.

There is a "supposed to" factor in most anything. SOme stuff like bias has taken on a life of its own. Bias ABSOLUTELY MUST be set to 70% of max. Not 60% and not 71%, but 70%. otherwise life on earth will end. I am sure you have heard such things. It is baloney. The bias can be way "off" and still the amp will work. Sure you can go too far, and sure the bias can affect the tone, but it is not written in stone. Same with speaker connections. Connected to the proper winding, the speaker will get the most efficient transfer of energy. Use the "wrong" tap, and you might slide off the optimal part of the power tube curves. But you also might like the sound better.

Manufacturers like to tell you to only use official approved parts and so on. What they are doing is telling you a way to operate their product that they know will be successful. They want to discourage people from experimenting, because some of those people will screw up their amps. SO only use official Marshall fuses. Failure to follow the guide lines COULD cause damage under certain circumstances, but probably won't.

I knew a technician at a music store who thought this was the law of the land. He bought 12AX7s for Fender amps from Fender, and he ordered 12AX7s for Marshall amps from Marshall. he'd buy 4558 op amps for Peaevy mixers from Peavey, and the same 4558s for Crate mixers from SLM. he had some rationale that the factories tested each part and had carefully selected each and every one for optimal service in each amp. Baloney.

If you are off by one notch on the impedance selector, you are not likely to damage anything. You will never go wrong by going by the book. But that is not the same thing as not going by the book always screwing things up.

Remember the bias? People hear that bias MUST be adjusted every time you change tubes. Then they are baffled when they can't find an adjustment on their amp. But how can I...?

If you smoke crack, well, you're stupid, but your amp will still run even if you switch to the "wrong" impedance. Tube amps are remakably flexible, remarkably forgiving, and remarkably versatile.

[Andy]

Thanks for the response, Enzo. you have confirmed what I believed to be true all along.

maybe one of the G's will chime in and give me a mathematical/scientific explanation.

[Ray Ivers]

Andy,

None of my names contain a 'g', so I won't venture into the math/science thing - it's a pretty simple overvoltage/overcurrent situation - but to add to the excellent points Enzo made; it could be that the posters on the other boards had seen recommendations they made 'edited' while making the rounds on the Internet, and rather than take the chance of someone possibly destroying their OT after reading that "so-and-so says it's OK to mismatch... as much as you... want..." (when what was actually posted was totally different), they might want to be a little more conservative and err on the side of caution.

BTW, guitar speaker impedances are only nominal values, and can vary both below and way above their ratings, especially in ported cabs near resonance.

FWIW, I second Enzo's recommendation about mismatching one step max.

Ray

[SlidePicker]

Andy, I don't have a G in my name, either. But I can tell you that the standard references on electronics engineering from the tube era, such as the RDH4, contain graphs showing pentode and beam power tubes' power output and content of various harmonics over a wide range of load impedances. If you look up the original data for these tubes, like the 6L6 and 6BQ5 here, you'll find such graphs for all of them. What you'll see is that the power varies little over a wide range of impedances, but the levels of various harmonics vary quite a bit. The impedance selected as optimal is an arbitrary one chosen for the lowest distortion level.

I've got a 1956 text at home on vacuum tubes and electronic devices that explains it best. (My ancient brain can't think of the title and author at the moment.) Triodes transfer the maximum power to the load when the load impedance matches the plate resistance (which is the internal impedance of the equivalent "generator"), as predicted by standard theory from the 1800s. But they produce more second harmonic distortion than engineers generally thought acceptable in those days, so normally they're loaded with about twice the plate resistance for maximum undistorted power. Beam power and pentode tubes have such high plate resistances that it's impractical to match the plate resistance to the load impedance, so the load is arbitrary.

Some people may be thick enough to think that power transformers actually have some sort of fixed impedances, such as 8 ohms on the secondary and 6k6 on the primary. That's not true. All transformers have is an impedance ratio, so the primary impedance is a simple multiple of whatever impedance is hooked to the secondary.

Also, it should be remembered that speaker impedance varies quite widely over the frequency spectrum applied to the speaker.

You're definitely not going to fry anything with a 2:1 or 4:1 mismatch. It's common practice to load output transformers with a power resistor of about 20x the nominal output impedance to kill the high-voltage spikes that can result from playing withouta speaker load. What you're doing is adding enough resistance to lower the Q of the coil to a level where it won't build up the high voltage. It's like a car spark plug that won't fire because a little carbon on the insulator trickles off the charge and keeps the voltage from building high enough to spark across the gap.

[Joe L]

Be a bit hesitant when missmatching EL34s

I found out thru hard knocks (3 pairs - ouch!) that the EH EL34 will go up like a firecracker almost instantly if you have the amp set to 4 ohms and the speaker is wired at 16 and you push them hard. Didn't have any problems with other brands and types over the few months that I had forgotten that I had rewired the cab to 16 ohms.

[Ray Ivers]

Joe,

You're lucky it wasn't your OT! In most 400V B+ guitar amps, it's reasonable to expect @ 800V peak swings on the OT primaries WRT ground with impedances matched 1:1, and maybe 1200 - 1600V with a 2:1 mismatch; AFAIK most OT's use 2kV-rated insulation, so you're still OK (at 400V B+). With a 4:1 mismatch, you'll easily exceed 2kV with the amp cranked up, and while this may not result in failures in this or that particular amp, it is still an unsafe practice (if you care about toasting your OT, arcing tubes/sockets, etc., that is). Connecting a loading resistor across the output will definitely buy you some time to figure out the the speaker cab is unplugged/fried or the cable's NG, but unless it's sized to draw an appreciable percentage of the amp's output you'll still be slamming the OT primary with really huge voltage swings, and failure will occur eventually.

In most Music Man, SVT, Hiwatt 400, Fender 400PS, etc. amplifiers, where really high B+ voltages are present, even a 2:1 mismatch is unsafe IMO, and 4:1... well, when you do the math you'll realize that's just asking for trouble.

There's a big difference in my mind between trying something oneself - where the consequences are solely on one's own possessions - and recommending that others do it as well, based on a possibly lucky experience - that's why I'm not a big fan of impedance mismatching, overloading PT windings, running tubes glowing hot, etc. - IME it's quite possible to get an excellent sound without doing any of these things, but YMMV.

Ray

[Joe L]

Hello Ray!

You're lucky it wasn't your OT!

What saved me was I was using an OEI 50 watt on one amp (a gutted Bassman as a testbed) and a OEI 100 watt on my Soldano 100 clone. Joe (OEI) appears to have overengineered the OTs and beside a few cheap fried EL34s on the testbed chassis and some 2 amp fuses, no harm has been done.

And no way would I try to gig regularly with a missmatch and I make a living for "The Man" so none of my amps leave the house.

[Steve Conner]

Way back on the original Ampage, I did a bunch of experiments on mismatching tube amps and even running them with no load whatsoever to see what happened and help resolve these debates. Here's what I found:

1) The only really bad thing you can do is dime the amp while using a speaker load impedance that's too high compared to what you set the amp's selector to. This causes the screen grids in your power tubes to melt. (The "Going up like a firecracker" of EL34s that a previous poster mentioned.)

This failure mode is well known in radio engineering, and RF amps that use expensive power tubes tend to have screen protection circuits. I built a screen protection circuit into one of my experimental amps, and proved that it could save tubes from destruction even with a complete open circuit speaker load, but it tends to nuisance trip out on loud bass notes with some speakers.

2) On the other hand, using a speaker of too low an impedance for the amp selection doesn't really seem to hurt things. The power tubes run less efficiently and can red-plate when the amp is cranked. The output power falls somewhat compared to the proper impedance match. If the OT was saturating before, the extra load will probably stop it saturating, so bass notes won't sound as "fat".

I used to use a 50 watt tube amp that only had a 16 ohm output, with a 4x10" bass cab wired for 8 ohms, and really beat on it to get a dirty funk bass sound. I also tried a 2x10" that could be switched between 4 and 16 ohms, and I found the proper 16 ohm match gave the loudest (and to my ears the best) result. I once tried it with a SVT 8x10" (4 ohms) and it sounded weedy and pathetic.

steve
www.scopeboy.com

[dai h.]

re: the Mesa, my .22+'s instructions say (with the stock 8ohm spk.) you can go to 4 or 16ohms. Supposedly Marshalls are more sensitive to a mismatch for some reason. *I think* I caused damage to an MV by setting to 8ohms output with a 16ohm load(the stock spk), but it was so long ago that perhaps I did something really stupid like switch with the power on. There definitely was damage I found years later when I looked inside since the grid stoppers and a 1k screen R were replaced, plus there was arcing on a tube socket.