I have a question about what is the standard practice for impedance matching with output transformers. Should the output impedance be the same as the impedance of the speakers you are using or do you use a lower impedance to get a higher damping factor like solid state amps?
Maximum power will be had with a matched impedance, but there will be lower distortion and lower source impedance
with a higher speaker impedance than the tap's rating. Since a speaker's impedance varies quite a lot, experimentation is in order.
Yes .
No.
If you need lower impedance in a tube amp, increase NFB ... good luck with that.
If you need high damping go SS.
That's basically what I am trying to understand. Why do I need damping with SS but tube amps need to have matched impedances?
Sorry but you got it wrong.
1) you don't "need" damping with SS, you just "get" it, for the very good reason that SS amps have very high voltage_NFB which turns them into excellent stable voltage sources.
Tube amps would also be excellent voltage sources, if they could also get high values of voltage NFB .
But they can not, mainly because the output transformer is inside the loop and has important phase shifts all over the place, which complicate things a lot, so with high feedback it's hard to maintain stability.
2) in tube amps , output transformer does NOT match load impedance to internal tube impedance , it's a common misconception.
Transformer presents to tubes the optimal impedance so as to maximize output power, period.
Different taps (4 / 8 / 16) mean that a 40W amp will put out 40W into 4 , 8 or 16 ohms.
That the tube reflected impedance may be equal (damping=1) , larger (damping <1) or smaller (damping >1) is something else.
Notice that said 40W amp will produce 40W into loads matching corresponding output taps, yet depending on feedback applied (even lack of it) damping will change ... without changing transformer taps.
That's what I told you earlier that changing transformer taps was not the way to change damping.
And SS amps, which have no output transformers, can go all the way from high damping to unity to very low damping, just changing a couple NFB resistors, go figure.
In fact, my amps, which are meant for guitar, have damping=1 , by using mixed mode feedback, a combination of voltage and current feedback.
It all depends on what you mean by "output impedance". If you really mean 'output impedance', then definitely not - except for the rare situation where your speakers are deliberately engineered to cope with this DF=1 amplifier (or you enjoy a boomy bass resonance). If you mean 'optimum load impedance' then yes - but the actual output impedance will be significantly lower (just as for SS, but not as low as typical SS).cspirou said:
If you have nominally 8 ohm speakers then your OPT should be designed to provide the optimum load to your output valves when loaded by 8 ohms. 'Optimum load' may mean 'minimum distortion' or 'maximum power' or some compromise between the two. If you use a 4 ohm secondary to drive an 8 ohm speaker then your valves will see twice the correct impedance; this means less power, and possibly higher distortion. It may mean higher damping (higher DF) too, but there are other ways of achieving this if that is what you want.
Ok, I think I am starting to grasp this a bit better. Since tube amps are low powered this means that you want to match the impedance to increase power output. However exclusively focusing on maximizing power could lead to more distortion so ideally you would need to find the right impedance to balance power and distortion. SS amps have power to spare and so don't have as much a need to match impedance to maximize power.
Impedance matching (in the strict sense of actually matching impedances) is almost never used in audio, because it is usually irrelevant. The most common place in electronics to find it is RF engineering, where it is very relevant. Don't think of impedance matching; when most audio people speak of impedance matching they don't really mean matching impedances, but merely presenting the appropriate impedance - which is almost always much higher or lower than a matching impedance would be.
These are two separate things here. Speaker damping is a consequence of output impedance. The Zo of a typical SS amp is quite low because transistors are low voltage, high current (therefore Lo-Z) devices almost universally connected these days as emitter/source followers. Being Lo-Z devices, they can be directly connected to Lo-Z loads with no problems. The Zo is also quite low since transistors have gm's orders of magnitude greater than hollow state devices.
Tubes are high voltage, low current (Hi-Z) devices. The OPT is needed to match the Lo-Z load of the speeks to a much higher load impedance, as determined by the AC loadline you're using. It's transforming hundereds of volts and tenths of an amp to the tens of volts and amps the speeks actually need.
If you were using PP 807s, the Zload= 6K6 (P-2-P) transformed to 8R speeks.
NZ= 6600 : 8= 825 : 1 (PRI : SEC)
The nominal rp= 14K for the 807
With one half of the PP pair cutoff, the impedance seen at the 8R output would be:
Zo= 14E3/(825/4)= 67.88R
For a damping factor of: 8/67.88= 0.117
That's almost no damping at all, and running open loop, expect to hear boomy bass that sounds more like monotone thumping than the actual notes being played.
To fix that, you need either finals with much lower plate resistances, like low-u power triodes, or you need to add NFB to get that effective plate resistance down.
- Status
- Not open for further replies.