Not much more to add here. I guess I could measure it by putting different load resistors on and seeing how much the output dropped, and then figuring out the output resistance (assuming a simple model with a perfect voltage generator and a series resistor, ignoring frequency-related effects). There must be some way to calculate it too though, right? Does anyone have a ready-made formula for this? The output stage is a SE 2A3, cathode biased (and bypassed), through an OPT.
Thanks,
Saurav
Thanks,
Saurav
So the output impedance will get divided by the square of the turns ratio, just like the load does. My OPT is 2500:8, which is a ratio of 312.5, so my 2A3's output impedance will also be divided by 312.5. Is that correct?
Then, how do I calculate the output impedance from the operating point? Let's say I'm at the 'classic' 250V/60mA operating point (I'm not, but I haven't measured it since my latest changes, beyond making sure none of the voltages were too high) - how do I calculate the output impedance from that? Is it just the slope of the plate curve at the operating point?
Then, how do I calculate the output impedance from the operating point? Let's say I'm at the 'classic' 250V/60mA operating point (I'm not, but I haven't measured it since my latest changes, beyond making sure none of the voltages were too high) - how do I calculate the output impedance from that? Is it just the slope of the plate curve at the operating point?
Saurav said:
Calculating it is useful in the design phase. Once you've got an amp on hand, it's easier and better to actually measure it. Your inclination is correct, just switch from an 8 ohm to a 4 ohm load resistor (on the same xfrmer secondary tap) and see the voltage drop.
The frequency variation of source Z is an interesting and useful thing to know, too, and is trivial to get. The simpler formulas won't be accurate for that.
Measuring Zout.
Take a signal generator and add a series resistor to its output so that Z out of generator plus external resistance is 1k ohm. Set the voltage to 1 volts rms ( say 1 KHz) . You now get a current source of 1mA.
Connect the free end to your 8 ohm load which is connected to the OPT. Measure the voltage across the 8 ohm load . It will be in mV. So if you get 0.3mV the output impedance with 8 ohms parallel with it is 0.3mV/1mA = 0.3 ohms.
It is not absolutely correct but pretty close. The actual output impedance will be = (8x0.3)/(8-0.3) = 0.3117 ohm.
You should try this at other frequencies also.
Note : Do not connect any input signal to the amp ! It is best to short the input / plug in a 600 ohm load / turn the volume down to zero etc. but " no input signal".
The signal generator and the mvMeter can be obtained free on the web as software used in conjunction with your sound card.
Try www.dazyweblabs.com and www.rightmark.com amongst others.
Cheers.
Take a signal generator and add a series resistor to its output so that Z out of generator plus external resistance is 1k ohm. Set the voltage to 1 volts rms ( say 1 KHz) . You now get a current source of 1mA.
Connect the free end to your 8 ohm load which is connected to the OPT. Measure the voltage across the 8 ohm load . It will be in mV. So if you get 0.3mV the output impedance with 8 ohms parallel with it is 0.3mV/1mA = 0.3 ohms.
It is not absolutely correct but pretty close. The actual output impedance will be = (8x0.3)/(8-0.3) = 0.3117 ohm.
You should try this at other frequencies also.
Note : Do not connect any input signal to the amp ! It is best to short the input / plug in a 600 ohm load / turn the volume down to zero etc. but " no input signal".
The signal generator and the mvMeter can be obtained free on the web as software used in conjunction with your sound card.
Try www.dazyweblabs.com and www.rightmark.com amongst others.
Cheers.
Actually DCR of primary as well as secondary, plus leakage inductance which will cause a rolloff of damping factor at HF. To figure resistance, figure Rpri in terms of looking through from the secondary, that is, divide it by the impedance ratio, then add it in series with the secondary.
Overall it probably would be easier to measure actual.
Tim
Overall it probably would be easier to measure actual.
Tim
SY's method for measuring output impedance is theoretically correct, but difficult in practice because small resistors possess signficant inductance. A method that suffers less from this problem (and wiring resistance problems) is:
Short-circuit the input of the amplifier.
Connect an oscillator on its 600R (or similar) output resistance directly across the output of the amplifier.
Switch the amplifier on.
Measure how well the amplifier can attenuate the oscillator.
The ratio of the loaded voltage to the unloaded voltage is (very nearly) the same as the ratio of the output resistances. So if the amplifier attenuates a 600R source by a factor of 100, its output resistance is 6R. It helps to set the oscillator to its maximum output voltage...
Short-circuit the input of the amplifier.
Connect an oscillator on its 600R (or similar) output resistance directly across the output of the amplifier.
Switch the amplifier on.
Measure how well the amplifier can attenuate the oscillator.
The ratio of the loaded voltage to the unloaded voltage is (very nearly) the same as the ratio of the output resistances. So if the amplifier attenuates a 600R source by a factor of 100, its output resistance is 6R. It helps to set the oscillator to its maximum output voltage...
"I'm here for an argument."
EC's method suffers from not having the amp loaded with the proper impedance. Not a huge deal for most ss amps, but important to consider when checking out a tube amp.
Inductance in power resistors is not much of an issue at "normal" frequencies. And source Z is usually a small-signal measurement, so you can use 2 watt metal oxide or carbon resistors (or if you want to be exotic, noninductive wirewound resistors), with no worries about inductance.
EC's method suffers from not having the amp loaded with the proper impedance. Not a huge deal for most ss amps, but important to consider when checking out a tube amp.
Inductance in power resistors is not much of an issue at "normal" frequencies. And source Z is usually a small-signal measurement, so you can use 2 watt metal oxide or carbon resistors (or if you want to be exotic, noninductive wirewound resistors), with no worries about inductance.
Re: "I'm here for an argument." (Alright, if you must.)
You can always add the appropriate load resistor across the output of the amplifier. It just makes the calculations a bit harder.
Actually, inductance in low-value wirewound resistors is an issue at audio frequencies, as I recently found to my cost when measuring the frequency response of a single-ended amplifier. (I've since replaced the W/W resistors on my dummy load with 50W metal film resistors). Agreed, the higher resistivity of carbon makes it almost non-inductive, and ordinary metal film aren't bad.
SY said:
You can always add the appropriate load resistor across the output of the amplifier. It just makes the calculations a bit harder.
Actually, inductance in low-value wirewound resistors is an issue at audio frequencies, as I recently found to my cost when measuring the frequency response of a single-ended amplifier. (I've since replaced the W/W resistors on my dummy load with 50W metal film resistors). Agreed, the higher resistivity of carbon makes it almost non-inductive, and ordinary metal film aren't bad.
How about the Radio Shack sand-cast resistors? How inductive are those? That's all I have in 8/4ohm power resistors, though I could easily buy a different type.
All of this came up because of some stuff I was thinking of doing with my speakers. Someone I was talking to said I should know my amp's output impedance to design the enclosure correctly. That's something I hadn't thought of before.
All of this came up because of some stuff I was thinking of doing with my speakers. Someone I was talking to said I should know my amp's output impedance to design the enclosure correctly. That's something I hadn't thought of before.
Free generator for DIY use.
I keep seeing this message that some DIY'ers don't have signal generators or any measurement setup.
Its all available FREE on the Net and I assume you all have your own PC's which have a sound card.
Download software from www.dazyweblabs.com or www.rightmark.com or www.speakerworkshop.com
Both can run the oscillator on the output of the sound card and measure from the input terminal on the sound card at the same time ( on full-duplex cards). The display might be on a spectrum analyser type screen. But if you measure the input signal and the output signal ( from the device under test - DUT in short) you will get a relative reading which you can use to calculate what you want. It's not as hard as it might seem. Do it once and you will see how easy it is.
Cheers.
( Just be cautious of the signal levels that you input to the sound card. You can build a protective input stage to avoid damage if you want to. For all passive device testing this will not be required.)
I keep seeing this message that some DIY'ers don't have signal generators or any measurement setup.
Its all available FREE on the Net and I assume you all have your own PC's which have a sound card.
Download software from www.dazyweblabs.com or www.rightmark.com or www.speakerworkshop.com
Both can run the oscillator on the output of the sound card and measure from the input terminal on the sound card at the same time ( on full-duplex cards). The display might be on a spectrum analyser type screen. But if you measure the input signal and the output signal ( from the device under test - DUT in short) you will get a relative reading which you can use to calculate what you want. It's not as hard as it might seem. Do it once and you will see how easy it is.
Cheers.
( Just be cautious of the signal levels that you input to the sound card. You can build a protective input stage to avoid damage if you want to. For all passive device testing this will not be required.)
My PC is in a different room from my stereo, or I would have had one of these a long time ago. I've used tone generator software to burn CD's of test tones for subwoofer equalization. For testing amps and things like that though, it's not really practical for me to move things into the bedroom every time I need a function generator.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.